Migraine disability and its recognition and assessment

Herkennen en boordelen van beperkingen door migraine

Thesis

to obtain the degree of Doctor from the Erasmus University Rotterdam

by command of the
rector magnificus

Prof.dr. S.W.J. Lamberts

and in accordance with the decision of the Doctorate Board

The public defence shall be held on

Wednesday 9th November 2005 at 09.45 hrs

by

Andrew John Dowson

born in Middlesbrough, United Kingdom

1
Doctoral Committee

Promotor: Prof.dr. J. Passchier

Other members: Prof.dr. W.F.M. Arts

Prof.dr. J. Klein
Prof.dr. M.J. Trappenburg

Copromotors: Dr. N.P. van Duijn

Dr. M. Peters

2
Acknowledgements

I am very pleased to acknowledge the co-authors of the papers that have been edited for
inclusion in this thesis: Drs Shaun Kilminster, Michele Peters, Susan Lipscombe, Trevor
Rees, Frances Carter, Stan Darling, Stewart Tepper, Carl Dahlöf, Matthew Clark, Michael
Bundy, Vicente Baos, Francis Baudet, and Domenico D’Amico, and Ms Rebecca Salt, Ann
Turner, and Christine Glover. All gave their permission freely for the editing and publication
of the papers in this format. I am also indebted to the members of the Migraine in Primary
Care Advisors (MIPCA) and the Migraine Action Association (MAA), for their input to several
of the projects contained here, particularly Drs Susan Lipscombe, Trevor Rees, Frances
Carter, Stan Darling, David Watson and Jerome Sender (MIPCA), and Ms Ann Turner
(MAA).

The copyright to the cover illustration is owned jointly by the Migraine Action Association and
Boehringer Ingelheim, and is used with their permission.

3
Contents
1. Introduction 5
1.1 General introduction 5
1.2 Headache Classification 6
1.3 Epidemiology of headache 8
1.4 Burden of headache 11
1.5 Current status of managing headache in the clinic 16
1.5.1 The prevalence and diagnosis of migraine in a primary care setting: 18
insights from the Landmark Study
1.5.2 Analysis of the patients attending a specialist UK headache clinic over 23
a 3-year period
1.6 Shortcomings in available knowledge (rationale for thesis) 32
1.7 Objectives of the thesis 34
1.8 Research questions 35

2. Methods 41

3. Investigating headache-related disability 48

3.1 Understanding the evidence: evaluating the efficacy of migraine 49
medications in clinical practice
3.2 Patients’ preference for triptans and other medications as a tool for 67
assessing the efficacy of acute treatments for migraine
3.3 A general practice study on the prevalence of headache, depression and 80
bodily pain, and the disability associated with headaches occurring inside
and outside the menstrual period in migraine sufferers
3.4 Emotional function with tension-type headache, migraine and chronic 94
daily headache

4. The clinical utility of assessing disability in primary care clinical 110

practice
4.1 Assessing the impact of migraine and other headaches 111
4.2 The outcome of headache management following nurse intervention: 127
assessment in clinical practice using the Migraine Disability Assessment
(MIDAS) Questionnaire
4.3 Outcome measures compared: Headache Impact Test (HIT) and Short 137
Pain Inventory© (SPI)
4.4 Development and validation of the Headache Diagnostic Screening 145
Questionnaire (DSQ): a new questionnaire for the differential diagnosis of
headache for use in primary care
4.5 Identifying patients who require a change in their current acute migraine 157
treatment: the Migraine Assessment of Current Therapy (Migraine-ACT)
questionnaire
4.6 The Migraine Assessment of Current Therapy (Migraine-ACT) 171
questionnaire: investigation of reliability, validity and clinical utility in a
multinational study

5. General discussion 183

5.1 Main results 183
5.2 Implications: methodological, research and clinical 191
5.3 Methodological issues and suggestions for future research 192
5.4 Final remarks 196

6. Summary 201

4
1

Introduction
Summary
General introduction: This section reviews the classification, epidemiology and burden of
illness associated with headache disorders. The discussion concentrates on the common
benign disorders of tension-type headache (TTH), migraine and chronic daily headache
(CDH), these being the main types of headache seen in everyday clinical practice. The
methodology of assessing headache burden is discussed, concentrating on disability and
impact tools. Migraine and CDH are both disabling conditions, leading to significant
impairment of daily activities in sufferers.
Primary care: Population-based studies show that migraine is under-estimated, under-
diagnosed and under-treated in the clinic. Many patients never consult for headache, and
drop out rates are high even among those who do. A study conducted by the author in
primary care showed that patients who consulted with episodic headaches almost always
had migraine. Patients with TTH hardly ever consulted. The conclusion was that migraine
should be the default diagnosis for patients presenting to primary care with episodic
headaches.
Specialist care: A second study conducted by the author showed that the headache pattern
is very different for patients who consult for specialist care. Sixty percent of patients
presenting to a UK specialist clinic had CDH, while only 33% had migraine. Other types of
headache were seen infrequently.
Commentary: New methods are required to deal with the variable pattern of patients seen
with headache in clinical practice. Innovative diagnostic and evaluation tools have been
developed recently that utilise assessments of disability.
Shortcomings in available knowledge: Headache remains a generally unrecognised
disorder, with its epidemiology and burden still not fully characterised. It would be particularly
useful to identify the factors that predict medical need in patients. Assessments of disability
have the potential to aid the diagnosis, evaluation of medical need and assess outcome in
studies.
Objectives of the thesis: The thesis is based on a series of studies conducted in primary
care to evaluate the evidence base of disability assessments and their potential in aiding
diagnosis and management of headache.

1.1 General introduction

Headache is the most common neurological condition, and due to this is the neurological
condition most likely to be seen in primary care. However, UK medical training typically
provides less than one day of education on headache management. Primary care physicians
may therefore be unprepared for the everyday management of headache, and many patients
end up being referred to secondary care, with the inevitable delays this entails. The paradox
is that most types of headache are eminently suitable for treatment in primary care. In
addition, there has been an explosion of research on headache over the past 15 years or so,
and guidelines for managing common headache subtypes have been published in several
countries.

Historically, research on headache was conducted on an ad hoc basis, with no common

diagnostic criteria used. This makes it difficult to apply studies conducted before 1990 to the
present day. Year zero for headache was 1988, the year of publication of the International

5
Headache Society (IHS) classification criteria for headache subtypes.1 Since its inception, it
has become possible to diagnose patients accurately and consistently. The criteria have
since been revised, and new classification criteria were published in 2004.2 Guidelines on
how to conduct studies with acute medications were first put forward by Glaxo in 1991,3
followed by IHS-recommended guidelines a few years later.4 These initiatives have
transformed research in headache, providing consistent methodology for diagnosing
patients, designing studies and defining endpoints. We are now able to use evidence-based
criteria to evaluate headache research, which has greatly expanded our understanding of the
subject. Due to commercial considerations and clinical importance, most recent headache
research has concentrated on migraine, but increasing amounts of data are now emerging
on other headache subtypes.

Most recent research on headache has been conducted in secondary care centres, resulting
in guidance being produced for the specialist physician. Only in the past few years have
evidence-based guidelines been produced for the primary care physician. In addition, little
research has been conducted in the primary care setting. The studies included in this thesis
are an attempt to rectify this situation, involving naturalistic clinical studies on the types of
patients seen in everyday clinical practice and the development of new, simple to use
questionnaires that are applicable to primary care. In this section, I review data on the
epidemiology, illness burden and clinical management of the common headaches, and
analyse the strengths and weaknesses of the methodology used.

1.2 Headache classification

Headaches can be classified into three major types: primary (benign) headaches, secondary
(possibly sinister or worrisome) headaches, and the cranial neuralgias and facial pain
syndromes (Table 1).2

6
Table 1. Classification of the common headache disorders.2

Primary headaches Episodic migraine (with or without aura)

Chronic migraine
Rare migraine variants (e.g. familial
hemiplegic migraine, basilar migraine,
retinal migraine)
Episodic tension-type headache (TTH:
infrequent or frequent)
Chronic TTH
Cluster headache (episodic or chronic)
Other trigeminal autonomic neuralgias
Other primary headaches (stabbing,
cough, exertional, sexual activity, hypnic,
thunderclap, hemicrania continua and new
daily-persistent headache (NDPH)
Secondary headaches Related to:
Head/neck trauma (e.g. whiplash, post-
traumatic disorder)
Cranial/cervical vascular disorder (e.g.
stroke)
Non-vascular intracranial disorder (e.g.
brain tumour)
A substance or its withdrawal (e.g.
medication overuse headache [MOH])
Infection (e.g. meningitis, systemic
bacterial or viral infection)
Disorder of homeostasis (e.g.
hypertension)
Disorder of cranial structures (e.g. neck,
eye, nose, ear, jaw disorders)
Psychiatric disorder (e.g. somatisation,
psychotic disorders)
Cranial neuralgias and facial pain Trigeminal neuralgia and other neuralgias
syndromes Compression headaches
Cold-stimulus headache
Eye-related headaches
Headaches related to Herpes zoster
Central causes of facial pain.

The most frequently reported headaches are the benign primary headaches of episodic
tension-type headache (TTH), episodic migraine and chronic migraine/ chronic TTH. Several
types of headache are commonly and arbitrarily grouped into the term ‘chronic daily
headache’ (CDH). CDH comprises daily or near-daily headaches that last for more than 4
hours on average and are often thought to be linked to medication overuse. They usually
arise from a primary, episodic headache disorder. These headaches are included in the
primary disorders of chronic migraine, chronic TTH, hemicrania continua and new persistent-
daily headache (NPDH) and the secondary disorder of medication overuse headache (MOH)
in the new IHS criteria.2 Table 2 illustrates the differences in presentation in CDH disorders.5

7
Table 2. Presentation of CDH disorders.5

Headache Frequency Presentation

subtype
Chronic migraine ≥ 15 d/mo for > Primary headache is migraine
3 mo May present with migraine or TTH-like features
Chronic TTH ≥ 15 d/mo for > Primary headache is episodic TTH
3 mo Presents typically with TTH-like features
(subject to revision)
Medication > 15 d/mo Chronic migraine, CTTH or mixed migraine and
overuse headache TTH-like features
(MOH) Overuse (≥ 10 d/mo for ≥ 3 mo) of ergots,
triptans, opioids, other pain medications and
combinations (≥ 15 d/mo)
Hemicrania Daily and Unilateral, moderate intensity, with
continua continuous exacerbations of severe pain
Conjunctival injection / lacrimation, nasal
congestion / rhinorrhoea / ptosis / miosis
Responds to indomethacin
New daily- Daily and Presents typically with TTH-like symptoms
persistent unremitting
headache from < 3 d from
onset

1.3 Epidemiology of headache

Studies of epidemiology need to be population-based to avoid the risk of biases. This is

frequently achieved by administering questionnaires in the community by interview or via the
telephone.6 The important outcomes of epidemiology studies are incidence and prevalence
(see Box 1). In practice, prevalence is assessed far more often than is incidence. To obtain
further information, prevalence data is usually analysed for gender, age, race and other
social and epidemiological factors.6

Box 1
Incidence: the number of new cases of a disease to be reported in a 1-year period.
Prevalence: the total number of cases of a disease, calculated as lifetime prevalence (the
total number who have ever experienced the disease) or 1-year prevalence (the number who
experienced the disease over the previous 1-year period). Comparing prevalence data is
often hindered by the differences in time periods used in studies.

Overall, almost everyone gets headache: the lifetime prevalence is estimated as 93%, with a
point prevalence of 11% in men and 22% in women.7 Prevalence studies, with diagnoses
confirmed using the IHS criteria,1 indicate that the only headache subtypes with a prevalence
> 1% in the general population are TTH, migraine and CDH. These headaches are
considered separately below.

Tension-type headache
The pioneering studies of Rasmussen and colleagues showed that TTH is very common,
indeed almost ubiquitous. The 1-year prevalence of TTH in Denmark was 63% in men and
86% in women, with a male: female ratio of 4:5.7 Similar results were reported in a
population-based study conducted in the USA, although with a lower 1-year prevalence of
38%.8 The prevalence decreased with increasing age.7 In Denmark, socio-demographic
variables of marital status, cohabitation, educational level, occupational category or

8
employment status were not significantly associated with TTH. However, TTH was positively
associated with neuroticism, fatigue in both sexes, time-pressure at work in women and
exposure to fumes in men.9 In the USA, TTH prevalence was associated with increasing
educational levels and prevalence was higher in Caucasians than in African Americans.8

Migraine
There are relatively few population-based data on the incidence of migraine. Studies show
that incidence peaks in childhood and adolescence, then declines over time.10,11 The overall
incidence per year is about 0.2% for boys and 0.6% for girls. The mean age of incidence is
lower for boys than for girls (13.7 versus 17.6 years).12

In contrast, there are large amounts of information on migraine prevalence. 1-year

prevalence rates from population-based studies conducted around the world are summarised
in Table 3.

Table 3. 1-year prevalence of migraine from population-based studies.

Prevalence (%)
Country, year n Age range (y) Men Women Total
USA, 199213 20,648 12–85 5.7 17.6 10.5
Canada, 199414 2,992 18+ 7.4 21.9 15.3
Denmark, 19917 1,000 25–64 5.9 15.3 10.4
Netherlands, 199515 1,008 12+ 5.0 12.0 9.0
UK, 200316 4,007 16–65 7.6 18.3 ND
France, 200217 1,486 15+ 4.0 11.2 7.9
Japan, 199718 1,597 15+ 3.6 12.9 8.4
Malaysia, 199619 595 6+ 6.7 11.3 9.0
Latin America, 200520 2,637 15+ 2.9–7.8 6.1–17.4 ND
ND = No data

The overall prevalence of migraine worldwide seems to be about 10% overall, 6% in men
and 15% in women. One large, international prevalence study showed that migraine without
aura was much more common than migraine with aura; less than one-third of the migraine
sufferers experienced aura symptoms.21 The prevalence of migraine with aura has been
estimated at 3% for men and 5% for women.11

Migraine is highly gender- and age-dependent (Figure 1).13

9
Figure 1. Gender- and age-specific prevalence of migraine.13

Stewart WF et al. JAMA 1992;267:64-9.

Overall, approximately 2–3 times as many women as men report migraine, most likely due to
the influence of female sex hormones. Migraine is about twice as likely to occur at the time of
the menstrual period as at other times in the menstrual cycle.22 Up to the age of 12, more
boys than girls have migraine, but the female preponderance starts at age 13, when
menarche has usually occurred.23 Prevalence in women increases up to about 40 years, then
declines. A similar pattern is seen in men, with peak prevalence occurring at about 35 years.
Prevalence declines thereafter for both genders, and < 5% of the population is affected after
age 70 years.13

Some studies show a low prevalence of migraine in certain Asian and African countries,
which have been ascribed to cultural and environmental differences between races.6 A study
in Americans showed that migraine prevalence was higher in Caucasians (20.4%) than in
African Americans (16.2%) and Asian Americans (9.2%).24 There therefore do appear to be
real racial differences in migraine epidemiology, although the condition is common in all
races. There is some evidence for variations in migraine prevalence in terms of social forces.
The prevalence seems to increase as the levels of education6,13 and income13 decrease.

Chronic daily headache

Very few population-based studies have been conducted on CDH. The available evidence
suggests that CDH is a relatively common condition, affecting about 4–5% of the general
adult population.25 CDH can occur at all ages, from 5 to > 80 years, and without treatment,
the condition can persist for years or decades.26 Population-based studies have provided 1-
year prevalence rates of 2.2% for chronic TTH8,25 and 2.4% for chronic migraine.25
Hemicrania continua and NDPH were reported rarely.25,27 There are difficulties in assessing
the prevalence of chronic migraine (often known as ‘transformed migraine’), as studies
published to date did not use the criteria recently published by the IHS.2 Further studies are
therefore necessary to investigate the epidemiology of chronic migraine.

As with migraine, CDH was much more common in women than in men.25 A population-
based study showed that CDH was also associated with being Caucasian, lower education

10
status, and in those with a history of marriage.28 Population- and clinic-based studies
identified co-morbid conditions of obesity, diabetes, arthritis, allergy, asthma, hypothyroidism
and hypertension, and associated with daily caffeine consumption and habitual snoring.28–30
New-onset CDH was associated with the baseline headache frequency and obesity.28 In the
clinic, the presentation of CDH is typically of daily TTH-like headaches, with exacerbations of
migraine-like headaches.31

Medication overuse headache (MOH) is present in a substantial proportion of people with

CDH. The overall population prevalence of MOH in a Spanish study was 1.4%, much higher
in women (2.6%) than in men (0.19%).32 Prevalence increased with age, and the mean age
of sufferers was 56 years. Most patients with MOH have a longstanding history of primary
headaches, with a decade or more of overuse of symptomatic medications.33 A clinic-based
study indicted that MOH was associated with asthma, hypertension and daily caffeine
consumption.30

Other chronic headaches are much less prevalent than CDH. The best known, cluster
headache, has an overall prevalence of about 0.4%; 0.3% in men and 0.1% in women.34

Conclusions
In conclusion, the primary, benign headaches are common conditions that affect both men
and women (but particularly women) during their adult lives.

1.4 Burden of headache

The burden exerted by headache on sufferers is typically reported in terms of

symptomatology, quality of life, disability (effect on sufferers’ daily activities) and resultant
social and economic consequences. This section reviews the first three of these items below,
for TTH, migraine and CDH because these reflect the patient’s perspective.

Methodology
As with epidemiological studies, population-based studies can be conducted that compare
discrete headache populations with each other and the general population. Clinic-based
studies are also important, as they contain specific populations of more severely-affected
patients that are of clinical interest. However, the data do not coincide with the population
distribution of patients. Headache symptomatology has been studied in many of the
epidemiological studies described above.

Quality of life (QOL) is a measure of illness severity that combines subjective perceptions of
the sufferer’s life situation and objective assessments of health factors. QOL is assessed
using questionnaires that patients complete at typically 1-month intervals. Questionnaires
can be generic (to be used for any condition) or migraine-specific. Generic questionnaires
are suitable for comparisons between study populations and different diseases. Disease-
specific questionnaires focus on specific problems associated with a disease and are
generally more sensitive. They may be better suited to assess changes in QOL following
clinical interventions.

The best known of the generic questionnaires is the Medical Outcomes Study Short-Form 36
(SF-36) questionnaire. This contains items on physical, social and emotional functioning,
pain, general medical and mental health.35 Scores of patients are related to those of the
general healthy population. The SF-36 has proven its methodological robustness and clinical
utility in numerous studies in different disease areas over many years. Other generic QOL
questionnaires include the EQ-5D and the WHO-QOL/WHO-QOL bref.

11
Several migraine-specific QOL questionnaires (e.g. MSQOL, MSQ, 24-hour QOLQ and
MIGSEV) are now available, which have also been shown to have good reliability and validity
(for definitions see Box 2).36–38 However, their complexity of assessments and scoring limit
general use in clinical practice.

In contrast to QOL measures, disability is a considerably less complex and more accessible
concept. Disability is a consequence of illness and an important indicator of unmet treatment
need. In part, disability is the product of poor QOL that is attributable to illness. The World
Health Organization (WHO) defines disability as the inability to work and function in other
roles.39 Questionnaires assessing headache-related disability quantify the effect on sufferers’
daily activities. They are more specific than the more general health-focused QOL and
impact questionnaires. Impact questionnaires include disability measures, in concert with
other factors influencing the life of patients (e.g. symptoms, QOL and patient suffering). Two
disability-based questionnaires have been developed, the disability-specific Migraine
Disability Assessment (MIDAS) questionnaire and the more wide-ranging Headache Impact
Test (HIT). Both questionnaires cover all headache subtypes, not just migraine.

MIDAS is a paper questionnaire aimed to be accessible at physicians’ surgeries and

pharmacies. Headache sufferers answer five questions in three activity domains covering the
previous 3-month period.40 They score the number of lost days due to headache in
employment, household work and family and social activities. Sufferers also report the
number of additional days with significant limitations to activity (defined as at least 50%
reduced productivity) in the employment and household work domains. The total MIDAS
score is obtained by summing the answers to the five questions as lost days due to
headache. This can be higher than the actual number of lost headache days due to any one
day being counted in more than one domain. The score is categorised into four severity
grades:
• Grade I = 0–5 (minimal or infrequent disability)
• Grade II = 6–10 (mild or infrequent disability)
• Grade III = 11–20 (moderate disability)
• Grade IV = 21 and over (severe disability)

Two other questions (A and B) are not scored, but were designed to provide the physician
with clinically relevant information on headache frequency and pain intensity. On testing,
MIDAS exhibited face validity, reliability, validity, ease of use and scoring and intuitive
meaning to physicians.40–43 Some studies have also shown it is sensitive to change following
clinical intervention (see Box 2).44,45 However, MIDAS also has some weaknesses. It only
assesses about 35% of the range of headache severity between moderate and severe
intensity, and does not cover milder or the most severe headaches.46 Therefore, the MIDAS
grade score may not always indicate the true medical need of patients due to floor effects
(patients scoring zero on the questionnaire),47 and the score may be weighted towards the
measurement of headache frequency over disability. Nevertheless, MIDAS has proved to be
a useful tool in specialist headache clinics.48

HIT was first developed as a web-based test, designed to be accessible via the Internet. It is
a dynamic questionnaire, with items derived from validated headache questionnaires
sampling all areas of headache impact.49 It is described as dynamic because the answer to
one question drives the choice of the next question, and so on. Patients are questioned until
clinical standards of score precision are met, although five questions are sufficient to grade
the majority of headache sufferers.49 It takes only 1–2 minutes to complete. On testing,
Internet-HIT was shown to be reliable and valid, and covered the whole spectrum of
headache,46,50,51 much more than did MIDAS.46 Internet-HIT differentiated sufferers on the
basis of diagnosis and characteristics such as headache severity and frequency.50

12
HIT-6 is a paper-based, short-form questionnaire based on the Internet-HIT question pool.
Six questions cover pain severity, loss of work and recreational activities, tiredness, mood
alterations and cognition. Each question is scored on a five-point scale, with the scores being
added to produce the final score.52 HIT-6 scores are categorised into four grades,
representing minimal, mild, moderate and severe impact due to headache. Studies have
shown that HIT-6 is reliable and valid,52 and in the clinic can be used to help diagnose
headaches (particularly) and also to select management strategies according to headache
severity.53,54 Also, HIT-6 scores were related to workplace productivity loss due to
headache.55 Internet-HIT and HIT-6 scores compared well to each other when the two forms
of the questionnaire were tested on a group of headache sufferers.56 The main disadvantage
of HIT is that the grade scores are not intuitively meaningful, comprised as they are from a
constellation of items.49,52

Assessments of disability and impact are clearly useful clinically, and have the advantage of
greater simplicity of use and understanding over QOL questionnaires. They can also be used
to assess all headaches, and not just migraine, as with some of the QOL questionnaires. In
fact, MIDAS items57 and HIT-6 scores53 both correlated with QOL and other clinical
assessments of headache severity. The question of which questionnaire is the best to use is
probably best left to the individual investigator. Compared to each other, MIDAS has the
advantage of clinically-meaningful units, while HIT covers a larger area of the headache
spectrum from mild to very severe intensity.

13
Box 2. Research and clinical practice criteria for evaluating clinical measures for headache
severity58

Overview: To be useful for clinical practice, an illness severity instrument should at minimum
meet a number of key criteria. Briefly, the instrument needs to be reliable, valid and internally
consistent and, if it is to be used to follow patients, it must be sensitive to change. For use in
general clinical practice, the tool also needs to be easy to use and score and the score
should be intuitively meaningful (i.e. face validity) to patients and physicians.

Criteria Analysis
Research criteria

Internal consistency The extent to which the items comprising

the measure are related to each other
Test-retest reliability Stability and reproducibility of results
when the instrument is administered
twice to the same person (test–retest)
Content validity Correlation between the instrument-
based measure and a ‘gold standard’
measure
Correlation between expert (researcher,
physician) and patient input
Construct validity The extent to which an instrument
measures what it purports to measure
and fits into a theoretical scheme about
the variable of interest
Discriminant validity The extent to which an instrument
distinguishes two conditions or states
known to be different
External validity The extent to which the instrument-based
measure is related to other measures
considered to be relevant
Sensitivity to change Instrument detects real change over
time, such as improvement in outcome in
response to effective therapy
Clinical practice criteria

Face validity Judgement that the measure

corresponds to an individuals’
perception, e.g. by selecting items
deemed to be important to the disease
sufferer or physician.
Ease of use The instrument should be simple to use,
score and interpret.
Intuitively meaningful The instrument should correlate with
physicians’ judgements of illness severity
and treatment need

14
Burden of episodic tension-type headache
TTH is typified by mild-to-moderate headache2,8 that is bilateral, pressing or tightening in
quality. It may be accompanied by photophobia or phonophobia, but nausea is absent.1,2
There seems to be little effect on the patient’s quality of life, or impact on daily activities, but
interestingly, sufferers are frequently fatigued,9 and suffer from neurotic and psychotic
conditions.59 One population-based study has shown that TTH can lead to lost work days
and reduced efficiency at work.8 However, HIT-6 scores for TTH are significantly lower than
those for migraine and CDH.53 TTH appears to have a complex aetiology that is affected by
several psychosocial factors.11 However, relatively little research has been conducted on
TTH, and the condition is not fully understood.

Burden of migraine
Migraine symptoms reported in population-based studies include moderate to severe
headaches, which are usually throbbing, one-sided and aggravated by physical activity. The
headache is usually accompanied by photophobia and/or phonophobia and nausea, and less
frequently by vomiting and aura symptoms. Attacks occur on average about once or twice a
month (range < 6 to > 50 attacks per year), last for about 24 hours (range < 4 to 72 hours),
and are separated by symptom-free intervals.21,60 However, migraine is an unpredictable,
heterogeneous disorder and attacks vary widely in frequency, duration, severity and reported
symptoms.61

A study using the SF-36 generic questionnaire showed that migraine sufferers taking part in
a clinical study had significantly lower QOL than the US general population.62 Increasing
severity of migraine was associated with worsening QOL. The main aspects of QOL affected
by the migraine were pain, and interference with daily activities and social lives. When
compared with other serious illnesses, migraine sufferers had lower QOL, in at least some
items, than patients with hypertension, depression, osteoarthritis and type II diabetes.62 A
further study indicated that migraine sufferers have impaired QOL even when they are
supposedly free from symptoms between attacks. Migraine sufferers had more physical and
emotional problems compared with healthy controls.63 Similar results are seen with the
migraine-specific QOL questionnaires.36–38 Migraine is closely linked to psychiatric disorders,
particularly depression and anxiety.6 For depression, the link is bi-directional, indicating that
each disorder increases the risk of onset of the other.64

Migraine is associated with high levels of disability. Population-based studies in the USA,61
Canada65 and Japan18 indicated that about three-quarters of sufferers have a limited ability to
function during their attacks. Population-based studies with the MIDAS and HIT
questionnaires show that migraine is associated with severe impact to sufferers’ daily
activities. Two studies, one in Europe and one in the USA, indicated that the mean MIDAS
score was in the Grade III range (moderate disability) with about 50% of sufferers scoring
Grade III or IV (moderate or severe disability).40,41 The MIDAS score for migraine was
significantly higher than that for non-migraine headaches.40 Numerous studies indicate that
migraine leads to lost time at work16,66,67 and in family and social activities.66,68,69 Studies with
the HIT-6 questionnaire in clinical practice also indicate that migraine is associated with
severe impact.53,54

Burden of chronic daily headache

There are relatively few population-based studies of CDH, and most studies are clinic based.
CDH presents in a variety of forms. Sufferers typically start with one of the episodic primary
headaches, usually migraine or TTH, which changes over time to a chronic headache
presentation. In its chronic form, the balance between the TTH and migraine vary from
patient to patient:
• At one extreme, some patients experience mild-to-moderate severity TTH only
(chronic TTH).

15
 • At the other extreme, patients may have daily, or near-daily, moderate-to-severe
severity migraine attacks without significant TTH (chronic migraine). A study of
patients with chronic migraine showed that over 80% suffered from fatigue, with two-
thirds meeting the criteria for chronic fatigue syndrome.70
• Patients with a previous history of episodic migraine may develop daily headaches
which may symptomatically resemble TTH. Periodic, more severe exacerbations that
meet criteria for migraine usually occur up to or more than 15 days per month. Such
patients should probably be classified as having chronic migraine in the new IHS
criteria,31 although this is not currently the case.2 Using the McGill Pain
Questionnaire, chronic migraine was shown to be more severe than chronic TTH, and
the two conditions had differing patterns of descriptive characteristics.71
All forms of CDH are associated with personality disorders, particularly neuroticism and
psychoticism, but these symptoms are most common in MOH sufferers.59

The key feature of the CDH patient with MOH is the daily or near daily overuse of headache
medications, such as simple analgesics (e.g. aspirin or paracetamol),72 combination
analgesics containing caffeine, codeine or barbiturates, opioids, ergotamine73 or triptans.74
Many patients are taking multiple medications.75 There tends to be a clinical syndrome of the
patient obtaining transient relief from their headache medication. This is then followed by a
return of the headache (withdrawal phenomenon) that necessitates further medication, and
so on until a vicious cycle is established. If the headache medication is stopped, withdrawal
symptoms are commonly reported. Patients with migraine as their primary headache develop
migraine-like daily headaches or an increase in their migraine frequency and TTH-like daily
headaches. Patients with initial TTH or combined headaches nearly always developed TTH-
like daily headaches.76

Three population-based studies have investigated the QOL of CDH sufferers. Two studies
showed that CDH without MOH77 and MOH32 patients had lower QOL than healthy controls
on all domains of the SF-36 questionnaire. In addition, CDH patients had poorer QOL than
those with episodic migraine (especially in general and mental health and vitality), and MOH
patients had poorer QOL than those with CDH without MOH (especially in physical
functioning and pain).77 Similar results were reported when using a migraine-specific QOL
questionnaire.78

Clinic-based studies show that headache-related disability is high in patients with CDH, with
and without MOH. The mean MIDAS score in these patients was in the high Grade IV range
(i.e. severe disability).45,79–81 The mean MIDAS score was significantly higher in patients with
chronic migraine compared to those with episodic migraine (34.9 versus 19.3, p < 0.001).81

Conclusions
In conclusion, the common primary headache subtypes of migraine and CDH are both
associated with severe symptoms, reduced QOL and significant impact on the sufferers’ daily
activities. The available data indicates that episodic TTH may be a more severe condition
than is generally thought, even though the symptoms are relatively mild.

1.5 Current status of managing headache in the clinic

Numerous population-based studies have investigated the proportion of IHS-defined

migraine sufferers who actually consult a physician for care (Table 4).18,21,65,82–88 These
studies consistently show that only about 50% or fewer migraine sufferers are currently
consulting a physician. Although the majority of sufferers consult at some time, many
subsequently drop out from care. Interestingly, these data have not changed in studies
conducted over the past decade, despite the number of educational initiatives conducted in
this time. Factors that are related to consultation include being female and married, being
older than 50 years, and the overall attack frequency, duration, severity and associated

16
disability.82 Factors that are related to non-consultation and dropping out from consultation
include a lack of perception of the severity of headache, good efficacy of OTC medications
and feeling that the physician could not help them, or had failed to do so in the past.21,65,88

Table 4. Consultation patterns for migraine. Proportion of migraineurs consulting physicians

in population-based epidemiological studies. * Study conducted in Belgium, Canada, Italy,
Sweden and the UK; **Study conducted in France, Germany, Italy, UK and USA

Study Ever Current Drop out Never

consulted consulter consulters consulted
(%) (%) (%) (%)
Canada, 199365 81 36 45 19
International*, 199321 66 31 35 34
USA, 199882 67 47 21 32
Denmark, 199283 56 ND ND 44
UK, 199684 58 ND ND 42
199985 67 47 21 32
Japan, 199718 31 15 16 69
USA, 200286 69 48 21 31
UK, 200387 86 65 21 14
International**, 200388 41–63 37–59
ND = No data

The profile of patients in the population with headache, and those consulting with primary
and secondary care physicians, differ markedly. Sections 1.5.1 and 1.5.2 below describe
concisely clinic-based studies that the author has conducted on the consultation pattern in
primary and secondary care.

17
1.5.1 The prevalence and diagnosis of migraine in a primary care setting: insights
from the Landmark Study*

Andrew Dowson1, Carl Dahlöf2, Stewart Tepper3, Lawrence Newman4 and Landmark Study
Group5
1
University of London, Kings College Hospital, London; 2Gothenburg Migraine Clinic,
Gothenburg, Sweden; 3New England Headache Center, Stamford, CT; 4The Headache
Institute, St. Luke’s Roosevelt Hospital Center, New York, 5GlaxoSmithKline; London,
Toronto, and North Carolina.

Background
Migraine affects about 3% to 8% of males and 11% to 18% of females in developed
countries.1 It is recognised that most migraine sufferers do not seek medical consultation for
their headaches,2 and nearly half of migraine sufferers are under-diagnosed or
misdiagnosed.3–5 The prevalence of migraine is well-established in population-based studies.
However the prevalence of migraine in patients presenting to primary care with headache is
limited.

Objective
To determine the prevalence and diagnosis of migraine in subjects presenting to primary
care physicians (PCPs) with a complaint of headache.

Design and methods

The Landmark Study was a prospective, international, open-label study designed to examine
the association of headache impact assessed by HIT-6 and migraine diagnosis in subjects
presenting to primary care physicians with a complaint of headache. All subjects who
presented with headache completed a paper-based HIT-6 questionnaire and a headache
survey. The investigator categorised patients as having migraine or non-migraine headaches
by using his/her customary diagnostic practice. Subjects diagnosed with a secondary
headache disorder or chronic daily headache were withdrawn from the study.

Patients newly-diagnosed as migraineurs and non-migraineurs completed diary cards for

their first 6 headaches or up to 3 months, whichever came first. At the end of the study, an
expert panel of headache specialists reviewed the diary cards utilising IHS criteria to provide
a final headache diagnosis.
– IHS 1.1 Migraine without aura
– IHS 1.2 Migraine with aura
– IHS 1.7 Migrainous headache
– IHS 2.1 Episodic tension-type headache (TTH)
– IHS 13 Headache not classifiable

*The full Landmark paper has now been published (Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and
diagnosis of migraine in patients consulting their physician with a complaint of headache: data from the Landmark
Study. Headache 2004;44:856–64). This summary of the full article is edited from another published article
(Dowson A, Dahlöf C, Tepper S et al. The prevalence and diagnosis of migraine in a primary care setting: insights
from the Landmark Study. Headache Care 2004;1:137–9).

Results
Full details of the study results are published in the full Landmark paper.6 A total of 1217
subjects were enrolled and 1204 completed the screening phase. 377 subjects were
classified by the Expert Panel according to their completed headache diaries.

18
Initial PCP Diagnoses
A total of 1017 subjects (85%) presenting to primary care with a complaint of headache
received a migraine diagnosis: 306 subjects were newly diagnosed migraineurs (Figure 1)
and 711 subjects had been previously diagnosed with migraine. A total of 142 subjects (12%)
were diagnosed with non-migraine primary headache (Figure 2). Forty five subjects (4%)
were initially diagnosed with secondary and/or other headaches.

19
Figure 1: Expert panel diary review of newly diagnosed migraine.

Newly diagnosed migraine

n=306

Expert panel review

n=272

IHS 1.1 IHS 1.2 IHS 1.7 IHS 2.1 IHS 13

n=138 n=99 n=31 n=1 n=3
51% 36% 11 % 0.4 % 1%

87% (237/272) of subjects had migraine diagnosis confirmed

Figure 2: Expert panel diary review of non-migraine primary headache.

Non-migraine primary headache

n=142

Expert panel review

n=105

IHS 1.1 IHS 1.2 IHS 1.7 IHS 2.1 IHS 13

n=38 n=13 n=36 n=10 n=8
36 % 12 % 34 % 9% 8%

49% (51/105) of subjects previously diagnosed as non-

migraine received a migraine diagnosis

Figure 3: Prevalence of headache based upon expert panel diary review (combined newly
diagnosed migraine and non-migraine)

20
 3% 3%

Migraine (n = 288)
Migrainous (n = 67)
ETTH (n = 11)
18% Unclassifiable (n = 11)

76%

Limitations of the study

Study limitations include the use of a diary study which may be subject to recall bias.
Diaries were to be completed at the time of the events. However, people do not always
complete diaries when they should. Biases by reviewers were limited by the
establishment of guidelines for review a priori. The inter-reviewer reliability coefficient
showed substantial agreement among panellists (ICC=0.71).

Conclusions
85% of the patients presenting to primary care with a complaint of headache suffered
from migraine. Few patients consulted for TTH.

An initial PCP diagnosis of migraine (IHS 1.1/1.2) was found to be correct 87% of the
time. However, almost 49% of PCP non-migraine diagnoses were reclassified as a
migraine (IHS 1.1/1.2) disorder by the expert panel. In the absence of contradictory
evidence, migraine should be the default diagnosis for patients presenting to primary care
with episodic headache.

Landmark Study Group

Members of the Landmark Study Group are Carl Dahlöf (Gothenburg Migraine Clinic, Gothenburg Sweden),
Andrew Dowson (University of London, Kings College Hospital, London, UK), Susan Jarvis (GlaxoSmithKline,
Harlow, UK), Martin Jones (GlaxoSmithKline, Greenford, UK), Frances Kinrade (GlaxoSmithKline, Greenford,
UK), Jackie Kwong (GlaxoSmithKline, Research Triangle Park, USA), Laura Lisk (GlaxoSmithKline,
Greenford, UK), Jane Loftus (GlaxoSmithKline, Greenford, UK), Hank Mansbach (GlaxoSmithKline,
Research Triangle Park, USA), Lawrence Newman (St. Luke’s Roosevelt Hospital Center, New York, USA),
Ba Pham (GlaxoSmithKline, Mississauga, Ontario, Canada), Mary S. Richardson (GlaxoSmithKline,
Research Triangle Park, USA), Andrew Scott (GlaxoSmithKline, Research Triangle Park, USA), Stewart
Tepper (New England Headache Center, Stamford, CT) and Christopher J. Webster (GlaxoSmithKline,
Research Triangle Park, USA).

References
1. Stewart WF, Shechter A, Rasmussen BK. Migraine prevalence. A review of
population-based studies. Neurology 1994;44 (Suppl 4):S17–23.
2. Lipton RB, Stewart WF, Simon D. Medical consultation for migraine: results from the
American Migraine Study. Headache 1998;38:87–96.

21
3. Diamond, ML. The role of concomitant headache types and non-headache co-
morbidities in the underdiagnosis of migraine.Neurology 2002;58 (9 Suppl 6):S3–9.
4. Lipton RB, Diamond S, Diamond M et al. Prevalence of migraine headache in the
United States: data from the American Migraine Study II. Headache 2001;41:646–57.
5. Lipton RB, Diamond S, Reed M et al. Migraine diagnosis and treatment: Results from
the American Migraine Study II. Headache 2001;41:638–45.
6. Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and diagnosis of migraine in
patients consulting their physician with a complaint of headache: data from the
Landmark Study. Headache 2004;44:856–64.

22
1.5.2 Analysis of the patients attending a specialist UK headache clinic over a
3-year period*

Andrew J Dowson
Director of the King’s Headache Service, King’s College Hospital, London, UK.
.
Abstract
Objective: This study analysed the profile of patients who attended a specialist UK
headache clinic over a 3-year period.
Methods: An audit was conducted of the clinical records of patients attending the specialist
headache clinic at King’s College, London, between January 1997 and January 2000. Data
were collected for diagnoses given, current medications taken, medications prescribed and
recommended and investigations conducted. Results were calculated as numbers and
proportions of patients for the 3-year period and for the three separate 12-month periods.
Results: A total of 458 patients were included in the audit. Most patients were diagnosed as
having chronic daily headache (CDH, 60%) or migraine (33%). Prior to the clinic visit, most
CDH and migraine patients treated their headaches with analgesics, and there was little use
of prophylactic medications. In the clinic, 74% of CDH patients and 85% of migraine patients
were prescribed or recommended prophylactic medications and 81% of migraine patients
were given triptans for acute treatment. Investigations (imaging procedures) were conducted
for 12% of patients, all proving negative.
Conclusions: CDH and migraine were the most common headache types encountered in
this UK secondary-care clinic. Withdrawal of analgesics and use of prophylactic medications
were used to manage CDH, while prophylaxis and triptans were used to manage migraine.
The results indicate that management of CDH and migraine in UK primary care is sub-
optimal, and educational initiatives are needed there to improve headache management.

*Summarised from the published version of this article (Dowson AJ. Analysis of the patients attending a specialist
UK headache clinic over a 3-year period. Headache 2003;43:14–18).

23
Background
Headaches are often not managed effectively in primary care, where migraine in particular is
under-estimated, under-diagnosed and under-treated.1 The field of headache research has
expanded dramatically over the past decade, with new therapies (especially the triptans)
introduced for migraine2 and increased understanding of the aetiology and treatment goals
for CDH.3 Clinical practice has changed to reflect these developments, especially in
specialist clinics, although primary care has been slower to respond.

Objectives
This study analyzed the diagnoses and management of patients attending one specialist
headache clinic in the UK over a 3-year period. The study aims were to examine the profile
of headache patients attending specialist clinics, to monitor changes in clinical practice and
to gauge referral practices from primary care.

Patients and methods

This was a retrospective audit of the clinical records of patients attending the specialist
headache clinic at King’s College Hospital, London, UK. Records from January 1997 to
January 2000 were analysed separately for three time periods: 1997–8 (January 1997 to
January 1998); 1998–9 (January 1998 to January 1999); and 1999–2000 (January 1999 to
January 2000). For 1997–8 and 1998–9, all patients who attended the clinic were analyzed.
For 1999–2000, only the first 77 consecutive patients who attended the hospital clinic were
analysed.

Data were analysed for the diagnosis given, current (pre-clinic visit) medications taken,
medications prescribed and recommended at the clinic visit and investigations conducted.
Results were recorded as numbers and percentages of patients.

Results
Patients: 458 patients attended the clinic and were analysed over the 3-year audit period,
215 in 1997–8, 166 in 1998–9 and the first 77 attending in 1999–2000. Patients came from
the large, urban, London area. Most patients were women. Demographic data were collected
for 1999–2000, where 51 patients (66.2%) were women and 26 (33.8%) were men.

Diagnoses (Table 1): The total number of diagnoses was greater than the number of patients
because some patients were diagnosed with more than one headache type. Overall, the
most frequent diagnosis given was for CDH (60%). The proportion of patients with this
diagnosis remained relatively constant over the 3-year period. About twice as many women
as men consulted with CDH in 1999–2000 (30 women and 16 men, 65.2% versus 34.8%).
The second most common diagnosis was for migraine (33%). The proportion of patients with
migraine fell slightly over the 3-year period, from 37% to 29%. The vast majority of patients
who consulted for migraine in 1999–2000 were women (18 women and four men, 81.8%
versus 18.2%). Short, sharp headaches (4%) and cluster headache (5%) were diagnosed
infrequently, cluster headache being only reported by men in 1999–2000. Headaches were
rarely diagnosed due to partial seizures (one patient), intermittent tension-type headache
(one patient), epilepsy (one patient), possible demyelination (one patient), possible
movement disorder (one patient), orgasm headache (three patients), sneeze-related
headache (one patient), cranial arteritis (one patient), inner ear infection (one patient) and
mobile phone-related headache (one patient).

24
Table 1. Diagnoses given at the headache clinic
Number of patients (%)*
Diagnosis 1997–8 1998–9 1999–2000 3-year total
n=215 n=166 n=77 n=458
Chronic daily 118 (55) 112 (64) 46 (60) 276 (60)
headache
(CDH)
Migraine 80 (37) 51 (31) 22 (29) 153 (33)
Short, sharp 11 (5) 5 (3) 3 (4) 19 (4)
headaches
Cluster 7 (3) 13 (8) 4 (5) 24 (5)
headache

*Some patients were diagnosed with more than one headache

Treatments
CDH (Tables 2 and 3): Baseline: Overall, most patients were using regular (≥1 dose/day)
analgesia (66%) and this proportion did not change markedly during the 3-year period.
Additionally, 14% of patients were using triptans as acute medications. Relatively few
patients (15%) were taking prophylactic medications.
Prescribed: Patients were advised to avoid analgesics and none were prescribed. Overall,
74% of patients were prescribed prophylaxis, the proportion nearly doubling from 1997–8 to
1998–9. The most commonly prescribed prophylactic drugs were dothiepin, sodium valproate
and paroxetine (Table 3). Triptans were prescribed rarely (4%), to treat ‘peak’ symptoms
only.

Table 2. Medications used to treat chronic daily headache (CDH): medications used prior to
the clinic visit
Number of patients (%)
Medication 1997–8 1998–9 1999–2000 3-year total
n=118 n=112 n=46 n=276
Regular 76 (64) 76 (68) 31 (67) 183 (66)
analgesia
Prophylaxis 20 (17) 14 (13) 8 (17) 42 (15)
Triptans 17 (14) 12 (1) 9 (20) 38 (14)
Diazepam 3 (3) 0 0 3 (1)
Ergots 0 1 (1) 0 1 (<1)

Table 3. Medications used to treat chronic daily headache (CDH): medications prescribed at
the clinic visit
Number of patients (%)
Medication 1997–8 1998–9 1999–2000 3-year total
n=118 n=112 n=46 n=276
Analgesia 0 0 0 0
Prophylaxis 60 (51) 104 (93) 40 (86) 204 (74)*
Triptans 3 (3) 7 (6%) 0 10 (4)
Diazepam 0 0 0 0

*Prophylactic drugs includes dothiepin (176 patients), sodium valproate (154 patients),
paroxetine (115 patients), gabapentin (seven patients) and carbamazepine (two patients)

Migraine (Tables 4 and 5): Baseline: Overall, most patients were using regular (≥1 dose/day)
analgesia (58%) and this proportion did not change markedly during the 3-year period.

25
Twenty nine percent of patients had been prescribed triptans as acute medication by their
primary care physician. This proportion increased only slightly, from 26% to 32%, over the 3-
year period. Use of ergotamine was low and declined from 9% of patients in 1997 to none in
1999. Relatively few patients (12%) had been prescribed prophylactic medications.

Prescribed: For acute medications overall, analgesics were prescribed or recommended for
9% of patients and triptans for 81%. Ergotamine was not used for any patient. Eighty five
percent of patients were prescribed or recommended prophylaxis for their migraine, with 8%
provided with hormonal manipulation for menstrual migraine. By far the most commonly
prescribed or recommended prophylactic drug was propranolol, with only a few patients
given pizotifen, cyproheptadine, dothiepin and verapamil. These figures remained relatively
constant over the 3-year study period.

Table 4. Medications used to treat migraine: medications used prior to the clinic visit

Number of patients (%)

Medication 1997–8 1998–9 1999–2000 3-year total
n=80 n=51 n=22 n=153
Analgesia 47 (59) 30 (59) 12 (55) 89 (58)
Triptan 21 (26) 16 (31) 7 (32) 44 (29)
Prophylaxis 15 (19) 1 (2) 3 (14) 19 (12)
Ergotamine 7 (9) 4 (8) 0 11 (7)

Table 5. Medications used to treat migraine: medications prescribed and recommended at

the clinic visit
Number of patients (%)
Medication 1997–8 1998–9 1999–2000 3-year total
n=80 n=51 n=22 n=153
Analgesia
Prescribed 1 (1) 1 (2) 0 2 (1)
Recommended 8 (10) 1 (2) 3 (14) 12 (8)
Triptan
Prescribed 23 (29) 25 (49) 2 (9) 50 (33)
Recommended 39 (49) 18 (35) 17 (77) 74 (48)
Prophylaxis
Prescribed 2 (3) 1 (2) 0 3 (2)
Recommended 64 (80) 43 (84) 20 (91) 127 (83)
Ergotamine
Prescribed 0 0 0 0
Recommended 0 0 0 0
Hormonal
manipulation
Prescribed 0 0 0 0
Recommended 10 (13) 2 (4) 0 12 (8)

26
Investigations
Investigations were conducted for 19 patients (9%) in 1997–8, 17 (10%) in 1998–9 and 17
(22%) in 1999–2000. Most investigations were for patients with CDH (58%) and migraine
(11%). CT and MRI scans were mainly used, with ESR, FBC and INR, EEG and
demyelination investigations being used occasionally. All investigations proved to be
negative.

Discussion
Over half the patients attending the two headache clinics were suffering from CDH, with a
third having migraine. These proportions of patients remained stable over the period 1997 to
2000. It appears, therefore, that the main headache currently managed in secondary care is
CDH, rather than migraine. Epidemiological studies show that CDH has a prevalence of
about 5%,3 whilst migraine affects 10% or more of the general population.4 This indicates
that migraine may be managed better than CDH in primary care. Cluster headache was
diagnosed infrequently and primary care physicians are likely to encounter cases only rarely.
To optimise resources, primary care educational initiatives for headache are probably best
concentrated on CDH and migraine.

CDH disorders include chronic TTH, chronic migraine (sometimes, but not synonymously,
known as transformed migraine), new daily-persistent headache and hemicrania continua.3
The natural history of CDH usually starts with a primary intermittent headache disorder such
as tension-type headache or migraine. Over a period of years, these conditions transform
into daily or near-daily headaches. The cause of this transformation is often due to patients’
overuse of acute headache medications (especially analgesics, ergots and triptans), leading
eventually to medication overuse headache (MOH) linked to CDH.3 Results from the present
study showed that patients attending the clinic had this pattern of CDH, relying on regular
analgesia (which could be prescribed or obtained over the counter without a prescription) for
therapy, with a substantial minority also using prescribed triptans for relief. Disappointingly
few were using headache prophylaxis before attending the clinic. The general principle of
management at the clinic was to treat the chronic headache by withdrawing analgesia and
prescribing prophylaxis for most patients, with triptans restricted for the acute treatment of
‘peak symptoms’. The most frequently prescribed prophylactic drugs were antidepressants
(dothiepin and paroxetine) and/or the anticonvulsant sodium valproate. When the chronic
headache was controlled, prophylaxis could be withdrawn and acute medications prescribed
for the underlying primary headache disorder. Currently recommended treatment strategies
for CDH incorporate withdrawal of analgesics, use of prophylactic medications and limitations
in the weekly use of acute agents.5 This strategy is similar to that used in the present study.

Migraine patients referred to the specialist clinic in this study relied on analgesia and, to a
lesser extent, triptans for acute therapy. Relatively few patients had been prescribed
prophylaxis. Management at the clinic increased the use of triptans and reduced the reliance
on analgesia for acute therapy, while providing most patients with prophylaxis (in the vast
majority of cases the beta-blocker propranolol). This management pattern is close to those of
the evidence-based guidelines of the US Headache Consortium6 and the Migraine in Primary
Care Advisors’ (MIPCA)7 in the UK, although the present study was completed before
publication of these two documents.

Investigational procedures were conducted on a small minority of patients (12%) in this

study. All procedures proved to be negative, and were conducted primarily for the
reassurance of the patients at their request.

In conclusion, CDH and migraine were the most common headache types encountered in
this UK secondary-care clinic. Withdrawal of analgesics and use of prophylactic medications
were used to manage CDH, while prophylaxis and triptans were used to manage migraine.

27
The results indicate that management of CDH and migraine in UK primary care is
suboptimal, and educational initiatives are needed there to improve headache management.

References
1. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
2. Goadsby PJ, Olesen J. Diagnosis and management of migraine. BMJ 1996;312:1279–
83.
3. Silberstein SD, Lipton RB. Chronic daily headache. Curr Opin Neurol 2000;13:277–83.
4. Stewart WF, Lipton RB, Celentano DD et al. Prevalence of migraine headache in the
United States: Relation to age, income, race and other sociodemographic factors. JAMA
1992;267:64–9.
5. Redillas C, Solomon S. Prophylactic pharmacologic treatment of chronic daily headache.
Headache 2000;40:83–102.
6. Matchar DB, Young WB, Rosenberg JH et al. Multispecialty consensus on diagnosis and
treatment of headache: pharmacological management of acute attacks. Neurology
2000;54:www.aan.com/public/practiceguidelines/03.pdf
7. Dowson AJ, Gruffydd-Jones K, Hackett G et al. Migraine Management Guidelines.
London: Synergy Medical Education, 2000.

28
Commentary on the differences between primary care patients (1.5.1) and secondary
care patients (1.5.2)
Taken together with Rasmussen et al’s paper on the epidemiology of headache in the
general population,7 these two papers illustrate the different profile of headaches seen in the
population and in primary and secondary care clinics (Figure 2).

Figure 2. Proportion of patients presenting with episodic TTH, migraine and CDH in the
general population,7 and in primary89.90 and secondary care headache clinics.91

100 94
90
80 78

70
Patients (%)

60
60 Population
50 Primary care
40 33 Secondary care
30
20 16
10 3 4
0 ND
0
TTH Migraine CDH

• In the general population, TTH was reported by the majority of people, while migraine
affected about one in six people and CDH about one in 20.7
• In patients presenting to primary care with episodic headaches, virtually all had migraine,
irrespective of the physician and patient diagnoses.89,90 Very few people consulted for
TTH. Patients with chronic headaches were excluded from this study.
• In secondary care, the majority of patients had CDH, while about one third had migraine.
No patients were referred for TTH.91 One study of patients with CDH in a secondary care
clinic showed that 78% had transformed migraine, 15% chronic TTH and 7% other
headache disorders.92

These studies also have important implications for headache diagnosis, which are discussed
below.

Diagnosis
Population-based studies demonstrate that a high proportion of migraine sufferers who
consult physicians do not receive a correct diagnosis.82,86,87,93,94 Over the past decade, the
diagnostic rate has only increased from about 45% to 60%, again despite the many initiatives
designed to improve this. Features associated with an accurate diagnosis of migraine
included female gender, the presence of an aura, the absence of co-morbid depression, and
the presence of headache-related disability.65,82,95

Table 5. Proportion of patients with migraine correctly diagnosed by their physicians in

population-based studies.

29
Study Proportion of patients correctly
diagnosed with migraine (%)
USA, 199482 45
USA, 199893 44
France, 199994 42
USA, 200286 56
UK, 200387 67

The IHS classification criteria have provided guidelines for the diagnosis of migraine and
TTH for well over a decade now,1,2 although CDH (chronic migraine, chronic TTH,
hemicrania continua and NDPH) was only defined in the 2004 document.2 While these
criteria are comprehensive, they are also lengthy, the physician being asked to potentially
read through 160 pages to make a correct diagnosis. However, the diagnosis of CDH is not
so simple, even using the IHS criteria. Several studies have demonstrated that many patients
cannot be classified using the IHS criteria,31,92,96 with many patients being classified with
migraine. Multiple and cumbersome diagnostic procedures were necessary to diagnose
patients with chronic headaches.31,96

Clearly, such complex criteria are not applicable to primary care. What are needed are
simple inclusive questionnaires and/or checklists to screen for multiple headache subtypes in
conjunction with exclusive questionnaires to specify the individual headache. From the
Landmark study described above,89,90 it was deduced that migraine should be the default
diagnosis for patients presenting to primary care with episodic disabling headache, in the
absence of contradictory evidence. Recently, several new questionnaires have been
developed to aid in headache diagnosis, with most being exclusive questionnaires to
diagnose migraine:
• Maizels and Burchette developed the Brief Headache Screen to diagnose different
headache subtypes.97 Migraine, daily headache syndromes and medication overuse
were distinguished using three questions: How often do you get severe headaches?;
How often do you get mild or less severe headaches?; and How often do you take
pain relievers, or any medication to relieve headache symptoms?
• Lipton et al developed a 3-item screener that sensitively and specifically diagnosed
migraine: Has a headache limited your activities for a day or more in the last three
months?; Are you nauseated or sick to your stomach when you have a headache?;
Does light bother you when you have a headache?98 Two or three ‘yes’ answers
coincided with a diagnosis of migraine.
• Cady et al developed a 3-item screener that sensitively diagnosed migraine: Do you
have recurrent headaches that interfere with work, family, or social functions?; Do
your headaches last at least 4 hours?; and Have you had new or different headaches
in the past 6 months?99 Migraine was indicted by a ‘yes’ answer to the first two
questions and a ‘no’ answer to the third.
• Pryse-Phillips et al developed a 3-item screener: Do you have a headache every
day?; Is your headache on one side of your head only?; Does your headache stop
you from doing things? The questionnaire distinguished between pure migraine and
other headache diagnoses with high reliability, validity and sensitivity.100
The clinical utility of these questionnaires remains to be elucidated.

Treatment
Population-based studies indicate that the majority of migraine sufferers rely on over-the-
counter (OTC) medications for treatment (Table 6).65,86,101,102 This is the case even for those
who also have medications prescribed by their physician.65,86.88 Again, this situation seems to
have changed little over the past decade.

30
Table 6. Proportion of migraine sufferers using prescription and over-the-counter
medications.
% Sufferers
Country Prescription OTC No
medications medications medications
USA, 1992101 37 59 4
Canada, 199365 44* 91* ND
102
Canada, 1998 12 88 0
USA, 200286 46* 72* 5
ND = No data
*Some sufferers took both prescription and OTC medications

Two studies in migraine patients showed that little more than a quarter of them (27%88 and
29%103 respectively) were satisfied with their acute medications. The main reasons for
dissatisfaction related to efficacy rather than to the incidence of side effects (Table 7).103

Table 7. Unmet treatment needs of migraine sufferers. Proportion of sufferers reporting the
reason for dissatisfaction with their usual acute migraine treatment to be subjectively
‘important’ or ‘very important’.103

Reason for dissatisfaction Sufferers (%)

Pain relief takes too long 87
Does not relieve all the pain 87
Does not always work 5
Headache comes back 71
Too many side effects 35

When asked, patients stated that they primarily wanted acute medications to provide rapid
and complete relief, with no recurrence. They were also interested in medications that were
convenient to take at any time and place.88,103

The methods used in assessing efficacy of acute migraine medications follows those
developed for double-blind, placebo-controlled studies with the triptans. A four-point scale is
used: severe, moderate, mild and no headache, respectively Grades 3, 2, 1 and 0.
Headache relief, defined as an improvement from severe or moderate to mild or none at 2
hours after treatment, was developed by Glaxo in the late 1980s to use in their clinical
studies with sumatriptan.3 Improvement to pain-free (Grade 3/2 to 0) has since been
advocated by the IHS as a more clinically-relevant endpoint.4 More recently still, these
endpoints have been further refined. An improvement of any headache to pain-free is now
advocated by many researchers, especially as recent recommendations state that triptans
should be given as soon as possible after the headache starts, so long as the patient is sure
that it is a migraine.104 All these endpoints are illustrated in Figure 3.

31
 Figure 3. Assessments of the efficacy of acute migraine treatments in clinical
studies3,4,104.

Pain-free (‘IHS’)

Pain relief (‘Glaxo’)

3 2 1 0
Severe Moderate Mild None

Pain-free (population)

Other researchers propose the concept of ‘sustained pain-free’, in which the patient must be
without pain for 24 or 48 hours after treatment, with no headache recurrence or use of rescue
medications.105 Such strict endpoints result in quite low efficacy rates, that can be criticised in
terms of their clinical relevance.

Endpoints used for preventive treatments (used for frequent migraine and CDH to reduce the
frequency of attacks) relate to reductions in the frequency, duration and severity of attacks.
Perhaps the most stringent endpoint (and the gold standard) is a reduction of over 50% in
the number of attacks in a defined time period. The latest studies with the neuromodulator
topiramate in migraine illustrate the optimal treatment endpoints (change from baseline in
mean monthly frequency, ≥ 50% reduction in monthly migraine frequency, number of monthly
migraine days, severity, duration, days per month requiring rescue medication and adverse
events).106–108 However, preventive treatment of migraine and CDH is notoriously difficult,
and high drop-out rates from treatment are reported due to wrong diagnoses, incorrect use of
the available drugs (using inappropriate doses for inadequate time periods), patient
expectations of a ‘cure’ rather than the usual reduction in frequency and the incidence of
bothersome side effects.109,110

A survey discovered the following reasons for treatment failure that led to patients being
referred to secondary care headache clinics:
• The diagnosis was incorrect or incomplete
• Important exacerbating factors were missed
• Prescribed and non-prescribed treatments were inadequate
• Patients had unrealistic expectations or co-morbidities.111
Improvements in the diagnosis and treatment of headache are therefore required to improve
the management of refractory patients.

1.6 Shortcomings in available knowledge (rationale for thesis)

Population-based studies provide information on the pattern of illness in the general

population, while clinic-based studies show the profile of patients who actually see a
physician for their condition. Comparing data from these two types of studies allows the

32
tentative analysis of factors that drive consultation and aids in the identification of sufferers
whose needs are not being met.

An enormous amount of material is available from population- and clinic-based studies on

the epidemiology, burden of illness and management of migraine, much of it published in the
last decade or so. Despite this, migraine is generally under-recognised, under-diagnosed and
under-treated in the clinic.58 The situation is even worse for TTH and CDH. The Landmark
Study showed that very few TTH sufferers consult with a physician,89,90 and while most
sufferers are likely to be only moderately affected by the disorder, the pattern of co-
morbidities suggest that there is a subgroup of sufferers who may be quite severely affected,
but who are invisible to medical services at present. We have no information as yet from
population-based studies on the proportion of CDH sufferers who consult, but we can deduce
likely trends from other data. Chronic migraine sufferers are seen in clinics much more often
than chronic TTH sufferers,92 although they are equally prevalent in the population.8 This
suggests that chronic TTH is particularly unrecognised and managed poorly at the moment.
The fact that consultation rates for migraine persist at about 50%18,21,65,82–88 also suggests
that chronic migraine sufferers are also probably under-recognised, under-diagnosed and
under-treated. There is therefore an enormous unmet clinical need for the common benign,
primary headaches, which has not been addressed significantly in recent years. Research
needs include the following:
• Population-based epidemiological studies in TTH and CDH, particularly with respect
to the definition of what constitutes chronic migraine
• Identifying factors that predict medical need in TTH sufferers (episodic and chronic)
• Population-based studies on recognition, diagnosis and management of CDH
• Clinic-based studies on the management of TTH, migraine and CDH
• In addition, educational initiatives are required to encourage sufferers with unmet
medical needs to consult a physician or other healthcare professional.

The need to improve knowledge of headache severity and headache management practices
has driven researchers and physicians to devise new ways of assessing the burden of
illness. Methods to assess QOL are well known and generic and headache-specific QOL
questionnaires have been developed that are methodologically robust and have proved
valuable in research.35–38 They are perhaps not so valuable in the clinic, where the lack of
intuitively-understandable questions and units make them difficult to use with patients.
Researchers have instead turned to assessments of impact (synonymous with disability) to
overcome these barriers.

Assessments of disability and impact describe the consequences of illness; the effects on the
everyday lives of sufferers. They are therefore potentially understandable to both the
physician and the patient. MIDAS and HIT are the most investigated of these questionnaires.
Both are generally robust methodologically, but are not as well categorised in terms of
clinical utility. The strengths of MIDAS are that it is reliable and valid, is easy to use and has
units that are easily understandable to the patient and physician.40–43,48 Its weaknesses are
that patients find it difficult to complete without supervision and that it may only show utility in
more disabled patients. In secondary care clinics, many patients have high MIDAS scores, in
the Grade IV range.45 In this setting, MIDAS has proved useful as a tool to demonstrate the
efficacy of treatments. In primary care and below (e.g. in the pharmacy), however, many
headache sufferers score 0 on MIDAS, making it impossible to use as an outcome tool.
Some of these patients may have unmet medical needs that are not detectable with MIDAS.
The strengths of HIT are that it is reliable and valid, and covers the full range of headache
severity, thus overcoming the floor effects seen with MIDAS.46,49,50,53 However, it has not
demonstrated the clinical utility that MIDAS has, and may be more suitable as a diagnostic
tool than as an impact tool.53,54 Research needs with impact questionnaires therefore include
the following:

33
 • Further investigation of the clinical utility of MIDAS and HIT
• Investigating other ways and tools to investigate impact
• Investigating impact as a tool to help with diagnosis.

Migraine is generally poorly diagnosed in primary care,82,86,87,93,94 and an inaccurate diagnosis

drives inadequate management to a large extent.111 By extrapolation, diagnosis of TTH and
CDH are also likely to be poor. While several new questionnaires have been developed in
recent years that diagnose migraine specifically,98–100 questionnaires to screen for different
headaches are thin on the ground and not fully categorised in terms of methodology and
clinical utility.97 Research needs include the following:
• Developing new questionnaires to screen for headache subtype for research and the
clinic (for patients presenting to GPs and other healthcare professionals)
• Deducing the factors that predict diagnosis, to develop a unified headache diagnostic
questionnaire.

Treatments for migraine remain clearly inadequate, with most sufferers still relying on OTC
medications,65,86,101,102 even though effective acute and preventive medications are available
for prescription by the physician. The situation for disabled TTH and CDH sufferers is likely to
be even worse, due to the lack of knowledge about these conditions. One possible way of
potentially improving treatments is to have a clear and simple way for the physician to assess
the efficacy of therapies. They can then change (or not) the patient’s treatment based on
objective criteria. Currently-used endpoints are those developed for clinical trials conducted
for regulatory purposes.3,4 These may not be applicable to everyday clinical practice.
Research needs include the following:
• Evaluation of the currently-used endpoints for their accuracy and clinical utility
• Development of new valid but simple endpoints for use in everyday clinical practice.

1.7 Objectives of the thesis

This thesis is based on research conducted over the past decade in primary care clinical
practice. The major aim was to conduct clinic-based naturalistic studies designed to generate
data that are applicable to everyday clinical practice. All of the studies were instigated and
designed by the author, in conjunction with colleagues. Some of the studies were conducted
in the author’s own clinical practice and others in the area of Surrey, UK. Others were
conducted nationally in the UK under the auspices of the Migraine in Primary Care Advisors,
a UK-based charity dedicated to the improvement of headache management in primary care
(www.mipca.org.uk), of which the author is the chairman.

The main objective of the thesis was to investigate headache-related disability, how to
recognise and assess disabled patients and the clinical utility of assessing disability in
primary care clinical practice. Individual objectives within this area were as follows:
1. To critically evaluate the evidence base for the quality of clinical evidence in acute
migraine studies.
2. To investigate whether alternative clinical study endpoints may be useful in evaluating
patients’ response to acute migraine therapies.
3. To critically evaluate the evidence base for the use of disability and impact tools in
migraine.
4. To investigate the disability experienced by different groups of headache patients in
the clinical setting.
5. To investigate the efficacy of the MIDAS questionnaire when used by nurses in
headache screening.
6. To investigate factors associated with headache diagnosis and develop a new
screening questionnaire for different headache subtypes.

34
 7. To investigate factors associated with headache management and develop a new
questionnaire to help physicians decide on the need to change acute migraine
medications.

1.8 Research questions

1. What is the quality of the evidence that underpins the use of acute migraine
medications and how can it be used to provide guidance to physicians on how to
evaluate clinical evidence rationally and in an evidence-based way?
2. Is patient preference a useful endpoint in clinical studies of acute migraine
medications?
3. Does the disability associated with the headaches experienced by female migraine
sufferers differ inside and outside the menstrual period?
4. What are the emotional factors implicated in how patients with TTH, migraine and
CDH cope with their headaches on a chronic basis?
5. What is the methodological robustness and the clinical utility of the impact
questionnaires MIDAS and HIT used to assess the severity of headaches?
6. What is the clinical utility of the MIDAS questionnaire when used by nurses in the
initial management of headache patients, before they see a physician?
7. What is the clinical utility of the HIT questionnaire in terms of its use as a diagnostic
or impact tool?
8. Can we develop a new diagnostic screening questionnaire for headache for use in
primary care? Does such a questionnaire exhibit methodological robustness and
potential clinical utility for patients and healthcare professionals?
9. Can we develop a new questionnaire to be used as screener for the need to change
patients’ acute migraine medications? Does such a questionnaire exhibit
methodological robustness and potential clinical utility for primary care physicians?

References

1. Headache Classification Committee of the International Headache Society.

Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8(Suppl 7):1–92.
2. Headache Classification Committee of the International Headache Society. The
international classification of headache disorders; 2nd Edition. Cephalalgia
2004;24(Suppl 1):1–160.
3. Pilgrim AJ. Methodology of clinical trials of sumatriptan in migraine and cluster
headache. Eur Neurol 1991;31:295–9.
4. International Headache Society Clinical Trials Subcommittee. Guidelines for
controlled trials of drugs in migraine: second edition. Cephalalgia 2000;20:765–86.
5. Dowson AJ, Bradford S, Lipscombe S et al. Managing chronic headaches in the
clinic. Int J Clin Pract 2004;58:1142–51.
6. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and
co-morbidities. Neurology 2001;56 (Suppl 1):4–12.
7. Rasmussen BK, Jensen R, Schroll M et al. Epidemiology of headache in a general
population: a prevalence study. J Clin Epidemiol. 1991;44:1147–57.
8. Schwartz, BS, Stewart WF, Simon D et al. Epidemiology of tension-type headache.
JAMA 1998;279:381–3.
9. Rasmussen BK. Migraine and tension-type headache in a general population:
psychosocial factors. Int J Epidemiol. 1992;21:1138–43.
10. Stewart WF, Linet MS, Celentano DD et al. Age- and sex-specific incidence rates of
migraine with and without aura. Am J Epidemiol 1991;134:1111–20.
11. Rasmussen BK. Epidemiology of headache. Cephalalgia 1995;15:45–68.
12. Breslau N, Davis GC, Andreski P. Migraine, psychiatric disorders, and suicide
attempts: an epidemiologic study of young adults. Psychiatry Res 1991;37:11–23.

35
13. Stewart WF, Lipton RB, Celentano DD et al. Prevalence of migraine headache in the
United States. Relation to age, income, race and other sociodemographic factors.
JAMA 1992;267:64–9.
14. O’Brien B, Goeree R, Streiner D. Prevalence of migraine headache in Canada: a
population-based survey. Int J Epidemiol 1994;23:1020–6.
15. van Roijen L, Essink-Bot M-L, Koopmanschap MA et al. Societal perspective on the
burden of migraine in the Netherlands. PharmcaoEconomics 1995;7:170–9.
16. Steiner TJ, Scher AI, Stewart WF et al. The prevalence and disability burden of adult
migraine in England and their relationships to age, gender and ethnicity. Cephalalgia.
2003;23:519–27.
17. Henry P, Auray JP. Gaudin AF et al. Prevalence and clinical characteristics of
migraine in France. Neurology 2002;59:232–7.
18. Sakai F, Igarashi H. Epidemiology of migraine in Japan. Cephalalgia 1997;17:15–22.
19. Alders EEA, Hentzen A, Tan CT. A community-based prevalence study on headache
in Malaysia. Headache 1996;36:379–84.
20. Morillo LE, Alarcon F, Aranaga N et al. Prevalence of migraine in Latin America.
Headache 2005;45:106–17.
21. Micieli G. Suffering in silence. In Edmeads J, ed. Migraine: A Brighter Future.
Worthing: Cambridge Medical Publications 1993;p1–7.
22. MacGregor EA, Hackshaw A. Prevalence of migraine on each day of the natural
menstrual cycle. Neurology 2004;63:351–3.
23. Abu-Arefeh I, Russell G. Prevalence of headache and migraine in schoolchildren.
BMJ 1994;309:765–69.
24. Stewart WF, Lipton RB, Liberman J. Variation in migraine prevalence by race.
Neurology 1996;47:52–9.
25. Castillo J, Muñoz P, Guitera V et al. Epidemiology of chronic daily headache in the
general population. Headache 1999;39:190–6.
26. Silberstein SD, Lipton RB. Chronic daily headache. Curr Opin Neurol 2000;13:277–
83.
27. Tepper SJ, Rapoport AM, Sheftell FD et al. Chronic daily headache – an update.
Headache Care 2004;1:233–45.
28. Scher AI, Stewart WF, Ricci A et al. Factors associated with the onset and remission
of chronic daily headache in a population-based study. Pain 2003;106:81–9.
29. Scher AI, Lipton RB, Stewart WF. Habitual snoring as a risk factor for chronic daily
headache. Neurology 2003;60:1366–8.
30. Bigal ME, Sheftell FD, Rapoport AM et al. Chronic daily headache: identification of
factors associated with induction and remission. Headache 2002;42:575–81.
31. Bigal ME, Tepper SJ, Sheftell FD et al. Chronic daily headache: correlation between
the 2004 and the 1988 International Headache Society diagnostic criteria. Headache
2004;44:684–91.
32. Colás R, Muñoz P, Temprano R et al. Chronic daily headache with analgesic
overuse : epidemiology and impact on quality of life. Neurology 2004;62:1138–42.
33. Diener H-C, Katsarava Z. Medication overuse headache. Curr Med Res Opin 2001;17
(Suppl 1):17–21.
34. Sjaastad O, Bakketeig LS. Cluster headache prevalence: Vågå study of headache
epidemiology. Cephalalgia 2003;23:528–33.
35. Anon. Norm-based scoring of SF-36 scales. In: SF-36 Summary Measures Manual.
Lincoln, RI, USA: QualityMetric Inc. 2003; p 22–35.
36. Santanello NC, Davies G, Allen C et al. Determinants of migraine-specific quality of
life. Cephalalgia 2002;22:680–5.
37. Martin BC, Pathak DS, Sharfman MI et al. Validity and reliability of the migraine-
specific quality of life questionnaire (MSQ Version 2.1). Headache 2000;40:204–15.
38. Wagner TH, Patrick DL, Galer BS et al. A new instrument to assess the long-term
quality of life effects from migraine: development and psychometric testing of the
MSQOL. Headache 1996;36:484–92.

36
39. National Academy of Sciences/Institute of Medicine (NAS/IOM). Disability in America:
toward a national agenda for prevention. Washington, DC: NAS Press, 1991.
40. Stewart WF, Lipton RB, Kolodner K et al. Reliability of the migraine disability
assessment score in a population-based sample of headache sufferers. Cephalalgia
1999;19:107–14.
41. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of
the Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.
42. Stewart WF, Lipton RB, Kolodner KB et al. Validity of the Migraine Disability
Assessment (MIDAS) score in comparison to a diary-based measure in a population
sample of migraine sufferers. Pain 2000;88:41–52.
43. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56(Suppl1):S20–8.
44. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
45. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine,
and cluster headache with topiramate. Headache 2002;42:796–803.
46. Ware JE, Bjorner JB, Kosinski M. Practical implications of item response theory and
computerized adaptive testing. Medical Care 2000;38:S73–82.
47. Lipton RB, Sheftell F, Stewart WF et al. The MIDAS perception study: headache
characteristics and treatment at low MIDAS grades. Cephalalgia 2001;21:326–7.
Abstract.
48. Edmeads J, Láinez JM, Brandes JL et al. Potential of the Migraine Disability
Assessment (MIDAS) Questionnaire as a public health initiative and in clinical
practice. Neurology 2001;56(Suppl 1):S29–34.
49. Bjorner JB, Kosinski M, Ware JE Jr. Calibration of an item pool for assessing the
burden of headaches: an application of item response theory to the headache impact
test (HIT). Qual Life Res 2003;12:913–33.
50. Bayliss MS, Dewey JE, Dunlap I et al. A study of the feasibility of Internet
administration of a computerized health survey: the headache impact test (HIT). Qual
Life Res 2003;12:953–61.
51. Bjorner JB, Kosinski M, Ware JE Jr. Using item response theory to calibrate the
Headache Impact Test (HIT) to the metric of traditional headache scales. Qual Life
Res 2003;12:981–1002.
52. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
53. Nachit-Ouinekh F, Dartigues J-F, Henry P et al. Use of the headache impact test
(HIT-6) in general practice: relationship with quality of life and severity. Eur J Neurol
2005;12:189–93.
54. Vuillaume De Diego E, Lanteri-Minet M. Recognition and management of migraine in
primary care: influence of functional impact measured by the headache impact test
(HIT). Cephalalgia 2005;25:184–90.
55. Bayliss MS, Kosinski M, Diamond M et al. HIT-6 scores discriminate among
headache sufferers differing in headache-associated workplace productivity loss.
Cephalalgia 2001;21:333–4. Abstract.
56. Dahlof C, Tepper S, Pryse-Phillips W et al. Comparability of scores on the Headache
Impact Test™ (HIT) and HIT-6™. Cephalalgia 2001;21:334–5. Abstract.
57. Lipton RB, Liberman JN, Kolodner KB et al. Migraine headache disability and health-
related quality-of-life: a population-based case-control study from England.
Cephalalgia 2003;23:441–50.
58. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
59. Guitera V, Gutierrez E, Muñoz P et al. Personality changes in chronic daily headache:
a study in the general population. Neurologia 2001;16:11–16. (in Spanish)

37
60. Henry P, Michel P, Brochet B et al. A nationwide survey of migraine in France:
prevalence and clinical features in adults. Cephalalgia 1992;12:229–37.
61. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity. Disability, pain intensity,
and attack frequency and duration. Neurology 1994;44(Suppl 4):S24–39.
62. Osterhaus JT, Townsend RJ, Gandek B et al. Measuring the functional status and
well-being of patients with migraine headache. Headache 1994;34:337–43.
63. Dahlöf CGH, Dimenäs E. Migraine patients experience poorer subjective well-
being/quality of life even between attacks. Cephalalgia 1995;15:31–6.
64. Breslau N, Davis GC, Schultz LR et al. Migraine and major depression: a longitudinal
study. Headache 1994;34:387–93.
65. Edmeads J, Findlay H, Tugwell P et al. Impact of migraine and tension-type
headache on life-style, consulting behaviour, and medication use: a Canadian
population survey. Can J Neurol Sci 1993;20:131–7.
66. Clarke CE, MacMillan L, Sondhi S et al. Economic and social impact of migraine. Q J
Med 1996;89:77–84.
67. Von Korff M, Stewart WF, Simon DJ et al. Migraine and reduced work performance: a
population-based diary study. Neurology 1998;50:1741–5.
68. Smith R. Impact of migraine on the family. Headache 1998;38:423–6.
69. Lipton RB, Bigal ME, Kolodner K et al. The family impact of migraine: population-
based studies in the USA and the UK. Cephalalgia 2003;23:429–40.
70. Peres MF, Zukerman E, Young WB et al. Fatigue in chronic migraine patients.
Cephalalgia 2002;22:720–4.
71. Mongini F, Deregibus A, Raviola F et al. Confirmation of the distinction between
chronic migraine and chronic tension-type headache by the McGill Pain
Questionnaire. Headache 2003;43:867–77.
72. Zwart J-A, Dyb G, Hagen K et al. Analgesic use: A predictor of chronic pain and
medication overuse headache: The Head-HUNT Study. Neurology 2003;61:160–4.
73. Meyler WJ. Side effects of ergotamine. Cephalalgia 1996;16:5–10.
74. Katsarava Z, Fritsche G, Muessig M et al. Clinical features of withdrawal headache
following overuse of triptans and other headache drugs. Neurology. 2001;57:1694–8.
75. Mathew NT, Kurman R, Perez F. Drug induced refractory headache – clinical features
and management. Headache 1990;30:634–8.
76. Limmroth V, Kazarawa S, Fritsche G et al. Features of medication overuse headache
following overuse of different acute headache drugs. Neurology 2002;59:1011–4.
77. Guitera V, Muñoz P, Castillo J et al. Quality of life in chronic daily headache.
Neurology 2002;58:1062–5.
78. Duru G, Auray J-P, Gaudin A-F et al. Impact of headache on quality of life in a
general population survey in France (GRIM2000 Study). Headache 2004;44:571–80.
79. Harpole LH, Samsa GP, Jurgelski AE et al. Headache management program
improves outcome for chronic headache. Headache 2003;43:715–24.
80. Grazzi L, Andrasik F, D’Amico D et al. Disability in chronic migraine patients with
medication overuse: treatment effects at 1-year follow-up. Headache 2004;44:678–
83.
81. Bigal ME, Rapoport AM, Lipton RB et al. Assessment of migraine disability using the
migraine disability assessment (MIDAS) questionnaire: a comparison of chronic
migraine with episodic migraine. Headache 2003;43:336–42.
82. Lipton RB, Stewart WF, Simon D. Medical consultation for migraine: Results from the
American Migraine Study. Headache 1998;38:87–96.
83. Rasmussen BK, Jensen R, Olesen J. Impact of headache on sickness absence and
utilisation of medical services: A Danish population study. J Epidemiol Community
Health 1992;46:443–6.
84. Mounstephen AH, Harrison RK. A study of migraine and its effects in a working
population. Occup Med 1995;45:311–7.
85. Lipton RB, Stewart WF, Liberman J et al. Patterns of healthcare utilization for
migraine in England. Cephalalgia 1999;19:305 (Abstract).

38
86. Lipton RB, Scher AI, Kolodner K et al. Migraine in the United States: epidemiology
and patterns of health care use. Neurology 2002;58:885–94.
87. Lipton RB, Scher AI, Steiner TJ et al. Patterns of health care utilization for migraine in
England and in the United States. Neurology 2003;60:441–8.
88. MacGregor EA, Brandes J, Eikermann A. Migraine prevalence and treatment
patterns: the global Migraine And Zolmitriptan Evaluation survey. Headache
2003;43:19–26.
89. Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and diagnosis of migraine in
patients consulting their physician with a complaint of headache: data from the
Landmark Study. Headache 2004;44:856–64.
90. Dowson A, Dahlöf C, Tepper S et al. The prevalence and diagnosis of migraine in a
primary care setting: insights from the Landmark Study. Headache Care 2004;1:137–
9.
91. Dowson AJ. Analysis of the patients attending a specialist UK headache clinic over a
3-year period. Headache 2003;43:14–18.
92. Silberstein SD, Lipton RB, Sliwinski M. Classification of daily and near-daily
headaches: field trial of revised IHS criteria. Neurology 1996;47:638–9.
93. Stang PE, Von Korff M. The diagnosis of headache in primary care: factors in the
agreement of clinical and standardized diagnoses. Headache 1994;34:138–42.
94. Richard A, Massiou H, Herrmann MA. Frequency and profile of migraine patients
seen by general practitioners. Sem Hop Paris 1999;75:440–6.
95. Lipton RB, Stewart WF, Celentano DD et al. Undiagnosed migraine headaches. A
comparison of symptom-based and reported physician diagnosis. Arch Intern Med
1992;152:1273–8.
96. Bigal ME, Sheftell FD, Rapoport AM et al. Chronic daily headache in a tertiary care
population: correlation between the International Headache Society diagnostic criteria
and proposed revisions of criteria for chronic daily headache. Cephalalgia
2002;22:432–8.
97. Maizels M, Burchette R. Rapid and sensitive paradigm for screening patients with
headache in primary care settings. Headache 2003;43:441–50.
98. Lipton RB, Dodick D, Sadovsky R et al. A self-administered screener for migraine in
primary care: The ID Migraine validation study. Neurology 2003;61:375–82.
99. Cady RK, Borchert LD, Spalding W et al. Simple and efficient recognition of migraine
with 3-question headache screen. Headache 2004;44:323–7.
100. Pryse-Phillips W, Aube M, Gawel M et al. A headache diagnosis project.
Headache 2002;42:728–37.
101. Celentano DD, Stewart WF, Lipton RB et al. Medication use and disability
among migraineurs: a national probability sample survey. Headache 1992;32:223–8.
102. Forward SP, McGrath PJ, MacKinnon D et al. Medication patterns of recurrent
headache sufferers: a community study. Cephalalgia 1998;18:146–51.
103. Lipton RB, Stewart WF. Acute migraine therapy: do doctors understand what
migraine patients want from therapy? Headache 1999;39 (Suppl 2):20–26.
104. Ferrari MD. Should we advise patients to treat migraine attacks early?
Cephalalgia 2004;24:915–17.
105. Adelman JU, Lipton RB, Ferrari MD et al. Comparison of rizatriptan and other
triptans on stringent measures of efficacy. Neurology 2001;57:1377–83.
106. Brandes JL, Saper JR, Diamond M et al. Topiramate for migraine prevention:
a randomized controlled trial. JAMA 2004;291:965–73.
107. Silberstein SD, Neto W, Schmitt J et al. Topiramate in migraine prevention:
results of a large controlled trial. Arch Neurol 2004;61:490–5.
108. Diener HC, Tfelt-Hansen P, Dahlöf C et al. Topiramate in migraine prophylaxis
– results from a placebo-controlled trial with propranolol as an active control. J Neurol
2004;251:943–50.
109. Diener H-C, Limmroth V. The management of migraine. Rev Contemp
Pharmacother 1994;5:271–84.

39
110. Dowson AJ, Dodick DW, Limmroth V. Medication overuse headache in
patients with primary headache disorders: epidemiology, management and
pathogenesis. CNS Drugs 2005;19:483–97.
111. Lipton RB, Silberstein SD, Saper JR et al. Why headache treatment fails.
Neurology 2003;60:1064–70.

40
2

Methods
Introduction: Outline of thesis
This thesis poses several questions relating to the clinical importance of headache-related
disability, how to recognise it and how to assess it.

Clinical importance of assessing disability

1. Chapter 2.1. What is the quality of the evidence that underpins the use of acute
migraine medications and how can it be used to provide guidance to physicians on
how to evaluate clinical evidence rationally and in an evidence-based way?
2. Chapter 2.2. Is patient preference a useful endpoint in clinical studies of acute
migraine medications?
These items are analysed by searches and reviews of the relevant clinical literature on the
subjects.

Recognising headache-related disability

3. Chapter 2.3. Does the disability associated with the headaches experienced by
female migraine sufferers differ inside and outside the menstrual period?
4. Chapter 2.4. What are the emotional factors implicated in how patients with TTH,
migraine and CDH cope with their headaches on a chronic basis?
These items are analysed by separate naturalistic studies conducted in primary care clinical
practice.

Assessing headache-related disability

5. Chapter 3.1. What is the methodological robustness and the clinical utility of the
disability-based questionnaires MIDAS and HIT used to assess the severity of
headaches?
This item is analysed by searches and reviews of the relevant clinical literature on the
subject.

6. Chapter 3.2. What is the clinical utility of the MIDAS questionnaire when used by
nurses in the initial management of headache patients, before they see a physician?
7. Chapter 3.3. What is the clinical utility of the HIT questionnaire in terms of its use as a
diagnostic or impact tool?
8. Chapter 3.4. Can we develop a new diagnostic screening questionnaire for headache
for use in primary care? Does such a questionnaire exhibit methodological
robustness and potential clinical utility for patients and healthcare professionals?
9. Chapters 3.5 and 3.6. Can we develop a new questionnaire to be used as screener
for the need to change patients’ acute migraine medications? Does such a
questionnaire exhibit methodological robustness and potential clinical utility for
primary care physicians?
These four items are analysed by separate naturalistic studies conducted in primary care
clinical practice.

Literature searches and reviews

Literature searches were conducted by a combination of three different ways:

1. Through searches of MedLine and journal websites.
2. By attending and monitoring presentations at international congresses on headache
and neurology: International Headache Congress, European Headache Society,

41
 American Headache Society, Migraine Trust International Symposium, Headache
Care for Practising Clinicians, American Academy of Neurology, European
Federation of Neurological Societies.
3. By collating discussions at meetings of the UK Migraine in Primary Care Advisors
(MIPCA) and the UK Migraine Action Association (MAA).

Studies
Design
The studies included in this thesis were all open, naturalistic studies conducted in primary
care clinical practice. They were designed to mimic the situation in everyday clinical practice
in terms of patients selected, drugs and doses administered and outcomes assessed.

Participants
The patients taking part in the studies were people being treated for TTH, migraine or CDH
at primary care clinics in the UK. Many of the patients came from the author’s own General
Practice-based clinics. Other patients were referred to the author from the MAA, the UK
patient support group. Finally, multicentre studies also included patients from the practices of
MIPCA members, all of whom are primary care physicians. Patients were diagnosed
according to the criteria of the International Headache Society (IHS) pertaining at the time,1
in conjunction with the author’s own clinical experience. Patients had to be between 18 and
65 years of age, not be abusing recreational drugs and not suffering from conditions that
could interfere with the study results or put the patient at risk. Few restrictions were placed
on the patients taking part in the studies deliberately to ensure, as far as possible, that the
patients were representative of those in clinical practice.

All patients gave their written informed consent to take part in the studies and, where
appropriate, institutional review body approval was obtained for the studies from local ethics
committees.

Procedures
Demographic and baseline characteristics were recorded for all patients at baseline. In
addition, patients were assessed with other questionnaires to assess their level of headache-
related disability. Patients were then typically followed up using these questionnaires over a
period of time when they were allowed to treat their headaches with their usual medications.
Details of these assessment procedures are shown in Table 1.

42
Table 1. Details of the studies: investigations, assessments and timelines.

Chapter Study investigation Baseline Time Follow-up

assessments of assessments
follow
up
2.32 Disability of headaches inside Headache, 2 Disability
and outside the menstrual depression and bodily months questionnaire
period pain questionnaire
2.43 Emotional factors associated Short Pain Inventory 7 days Short Pain
with TTH, migraine and CDH (SPI©) Inventory (SPI©)
3.24 Utility of the MIDAS Questionnaire on 6 Questionnaire
questionnaire when used by headache features months on headache
nurses and consultations; features and
MIDAS questionnaire consultations;
MIDAS
questionnaire
3.35 Validity and clinical utility of Internet HIT; 7 days Internet HIT;
Internet HIT and SPI SPI SPI
3.46 Development of a new Diagnostic Screening ND ND
diagnostic screening Questionnaire (DSQ)
questionnaire for headache
3.57 and Development of a new SF-36 QOL 1 week SF-36 QOL
3.68 questionnaire to screen for the questionnaire; questionnaire;
need to change patients’ acute MIDAS questionnaire; MIDAS
medications Migraine Therapy questionnaire;
Assessment MTAQ
Questionnaire
(MTAQ)
ND = no data

Instruments
Headache, depression and bodily pain questionnaire
A questionnaire was developed by the study investigators to assess the headaches,
depression and bodily pain severity of patients attending a UK primary care practice (see
Chapter 2.3).2 Three questions assessed headache, five assessed depression and four
assessed bodily pain severity. The answers to the questions were used to define the
diagnosis and illness severity. The questionnaire was devised to be simple to use and
potentially applicable to primary care clinical practice.

Disability questionnaire
A disability questionnaire was developed by the study investigators to assess headaches that
took place inside and outside the menstrual period over a 2-month period (see Chapter 2.3).2
Criteria used to develop the questionnaire included the IHS diagnostic criteria pertaining at
the time.1 Questions assessing the time lost from normal activities and the further time spent
at less than 50% of normal capabilities are similar to those used in the MIDAS
questionnaire,9 and therefore have potential validity.

Short Pain Inventory (SPI©)

The SPI was used to assess the emotional factors associated with TTH, migraine and CDH3
and to assess the validity and clinical utility of the Internet HIT and SPI questionnaires.5

The SPI is a 17-item self-rating questionnaire dedicated to measuring any pain, not only
headache. The questions assess pain intensity, social interaction skills and components of
emotional function. Patients rate each item on a five-point Likert scale: not at all, a little,
moderately, very much so and extremely. Five subscales measure sedation, social
interaction, anxiety, sadness and anger, and have been fine-tuned by maximising items that

43
are specific to pain. The SPI takes about one minute to complete,10 and can be accessed via
the internet with automatic scoring.11 A scale of Total Mood Disturbance (TMD, 14 items) can
be summated from the SPI, which measures the subtle mood changes that specifically
covary with mild to severe pain (including sadness, anger, anxiety and sedation scores
[social interaction scores are removed because this is a cognitive-behavioural parameter]).
All 17 items can be summated to form the Total Pain Disturbance (TPD) score.

As the physical severity of pain worsens, so generally does each of the emotional
components.12–14 If a patient’s emotional score is exaggerated or flattened relative to the
normal values, then a coping score can also be given as to how well a patient is coping with
their pain, relative to others at the same physical level of pain. In this way an emotional
‘footprint’ of the headache can be created. This can be very useful in the individual
management of patients. It has been found that around 60% of the variance of physical pain
is common to the overall mood of the patient.12,13 The SPI has been shown to be far more
powerful at discriminating between varying levels of physical pain severity than the McGill
Pain Questionnaire (used as the gold standard measure), with the majority of the pain
variance being captured by the SPI.12

Other studies have demonstrated that the SPI subscales could differentiate the footprint of
pain resulting from headache and other pain states such as dental pain, osteoarthritis and
chronic pain states.5,13,14 In general, headache patients had the greatest level of sedation,
while dental patients had the greatest levels of anger, sadness and anxiety. Patients with
osteoarthritis had the lowest level of mood disturbance of all the patient groups. Overall,
headache patients had a level of mood disturbance at least as severe as patients who were
attending a secondary care chronic pain clinic.

MIDAS questionnaire
The sensitivity of the MIDAS questionnaire to change was investigated in a study where a
nurse provided headache advice to patients attending a primary care practice.4 It was also
used to validate the Migraine-ACT questionnaire, a new questionnaire to screen for the need
to change patients’ acute medications.7,8

The development, testing and features of MIDAS are discussed in detail in Chapter 1. In
brief, it assesses headache-related disability as the time lost from daily activities
(employment, household work and family and leisure activities) and is scored as the number
of days lost in a 3-month period.9

HIT questionnaire
The validity and clinical utility of the Internet HIT questionnaire were assessed in studies
comparing it to the SPI questionnaire,5 and it was also used in conjunction with the SPI to
assess the emotional factors associated with TTH, migraine and CDH.3 It was also used to
validate the Migraine-ACT questionnaire, a new questionnaire to screen for the need to
change patients’ acute medications.7,8

The development, testing and features of Internet HIT and HIT-6 are discussed in detail in
Chapter 1. In brief, it assesses headache impact as a constellation of pain, disability (activity
limitations), tiredness and mood changes, and is scored numerically using arbitrary units.15

SF-36 QOL questionnaire

The SF-36 questionnaire was used to validate the Migraine-ACT questionnaire, a new
questionnaire to screen for the need to change patients’ acute medications.7,8

The SF-36 is a relatively brief generic QOL questionnaire, used in many chronic diseases,
which contains 36 items, analysed as eight scales (physical functioning, role-physical, bodily
pain, general health, vitality, social functioning, role-emotional and mental health) and a one-

44
item measure of change in health. It is scored between 0 and 100, with a score of 50
coinciding with the score for the healthy US general population.16 Investigations in migraine
sufferers have shown that the SF-36 score correlates with migraine severity and
discriminates between migraine and other serious medical conditions.17

Migraine Therapy Assessment Questionnaire (MTAQ)

The MTAQ was used to validate the Migraine-ACT questionnaire, a new questionnaire to
screen for the need to change patients’ acute medications.7,8

MTAQ is a migraine-specific disease management questionnaire, which contains nine

questions on migraine frequency and triggers, use of medications and their effectiveness,
healthcare utilisation and time lost from everyday activities.18 Each question is scored as
‘yes’ or ‘no’, and the score ranges from 0–6 (three questions are not scored). A score of ≥ 2
is used as an indicator that the patient requires follow up. It is proposed to be used in primary
care to identify patients who require additional care.18

Psychometric properties of the questionnaires

The instruments used in the studies were either developed for use or testing by the author or
were established questionnaires with well categorised psychometric properties. Table 2
shows the psychometric properties of the established questionnaires.

Table 2. Psychometric properties of the established questionnaires used in the studies.

Questionnaire Internal Test-retest Validity Validation measure

consistency reliability
(Cronbach
Alpha)
Short Pain 0.93–0.9412,13 0.9312 0.775 Headache severity5
Inventory (SPI©) 0.7912 McGill Pain
Questionnaire12
MIDAS questionnaire 0.769 0.80–0.839 0.639 Headache diary data9
Internet HIT ND 0.7919 0.8919 Headache severity19
SF-36 QOL 0.9720 0.8520 0.9221 Psychometric and clinical
questionnaire tests21
MTAQ ND18 0.7118 0.6718 SF-36 QOL Questionnaire
MIDAS Questionnaire
Treatment satisfaction
Healthcare resource use18
NA = no data

Overall, the SPI, MIDAS, HIT, SF-36 and MTAQ questionnaires are all highly reliable and
valid, and therefore suitable for use in clinical research. In addition, the MIDAS22 and SF-3623
questionnaires have also been shown to be sensitive to change and are therefore suitable for
use in clinical practice.

Statistical analyses
Statistical analyses were performed using the Statistica V4 (Tulsa USA) statistical package
for MacIntosh. All data were double- or triple-checked for accuracy, and all analyses checked
for normality. Parametric analyses were used for normally-distributed data and non-
parametric analyses for non-normally distributed data. Analyses were two-sided and the level
of significance was set at p-values less than 0.05. Patient demography was analysed using
descriptive statistics only.

Parametric analyses included discriminant function analysis, simple t-tests and test-retest
Pearson product moment correlation coefficients,3,5 t-tests for dependent samples4 and test-
retest Pearson’s product moment and Spearman rank measures, discriminatory t-tests and

45
factor analysis.7,8 Non-parametric analyses conducted included the Mann-Whitney U Test,2
Wilcoxon rank tests,4 and the Wilcoxon matched-pairs test.6

References
1. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8 (Suppl 7):1–96.
2. Dowson AJ, Kilminster SG, Salt R et al. Disability associated with headaches
occurring inside and outside the menstrual period in those with migraine: A general
practice study. Headache 2005;45:274–82.
3. Dowson AJ, Bundy M, Kilminster SG. Assessing patients: coping with headache,
emotional function and healthcare utilisation associated with episodic tension-type
headache, migraine and chronic daily headache. Headache Care 2005; submitted for
publication.
4. Dowson AJ, Salt R, Kilminster S. The outcome of headache management following
nurse intervention: assessment in clinical practice using the Migraine Disability
Assessment (MIDAS) Questionnaire. Headache Care 2004;1:177–81.
5. Kilminster SG, Dowson A, Bundy M. The Headache Impact Test® and the Short Pain
Inventory©: Outcome measures compared. Int J Pharm Med 2003;17:23–32.
6. Dowson AJ, Turner A, Kilminster S et al. Development and validation of the headache
Diagnostic Screening Questionnaire (DSQ): a new questionnaire for the differential
diagnosis of headache for use in primary care. Headache Care 2005;2:111–18.
7. Dowson AJ, Tepper SJ, Baos V et al. Identifying patients who require a change in
their current acute migraine treatment: the Migraine Assessment of Current Therapy
(Migraine-ACT) questionnaire. Curr Med Res Opin 2004;20:1125–35.
8. Kilminster S, Dowson AJ, Tepper SJ et al. The Migraine Assessment of Current
Therapy (Migraine-ACT) questionnaire: investigation of reliability, validity and clinical
utility in a multinational study. Headache 2005; submitted for publication.
9. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56 (Suppl 1):S20–28.
10. Manual. The Short Pain Inventory. Selsey, West Sussex: Lawrencian Clinical Ltd.
England.
11. The SPI Online Test for headache pain. www.headachetest.co.uk
12. Kilminster SG, Mould GP. Comparison of the internal reliability and validity of the
McGill Pain Questionnaire and Short Pain Inventory. Int J Pharm Med 2002;16:87–
95.
13. Kilminster SG, Power M, Fozard JR. Survey of pain in two medical and dental clinics,
with non patient controls using the Short Pain Inventory. Int J Pharm Med
2000;14:137–47.
14. Kilminster SG, Mould GP. Comparison of diclofenac spray and gel on knee joints of
patients with osteoarthritic pain. Clin Drug Invest 1999;18:345–54.
15. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
16. Ware JE, Snow KK, Kosinski M et al. SF-36 Health Survey manual and interpretation
guide. Boston: New England Medical Center, 1993.
17. Osterhaus JT, Townsend RJ, Gandek B et al. Measuring the functional status and
well-being of patients with migraine headache. Headache 1994;34:337–43.
18. Chatterton ML, Lofland JH, Shechter A et al. Reliability and validity of the Migraine
Therapy Assessment Questionnaire. Headache 2002;42:1006–15.
19. Ware JE Jr, Kosinski M, Bjorner JB et al. Applications of computerized adaptive
testing (CAT) to the assessment of headache impact. Qual Life Res 2003;12:935–52.
20. McHorney CA, Ware JE Jr, Lu JF et al. The MOS 36-item Short-Form Health Survey
(SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse
patient groups. Med Care 1994;32:40–66.

46
21. McHorney CA, Ware JE Jr, Raczek AE. The MOS 36-Item Short-Form Health Survey
(SF-36): II. Psychometric and clinical tests of validity in measuring physical and
mental health constructs. Med Care 1993;31:247–63.
22. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
23. Solomon GD, Skobieranda FG, Genzen JR. Quality of life assessment among
migraine patients treated with sumatriptan. Headache 1995;35:449–54.

47
3

Investigating headache-related disability

48
3.1

Understanding the evidence: evaluating the efficacy of migraine

medications in clinical practice*

Abstract

Large, randomised, double-blind, controlled clinical trials form the gold standard of clinical
evidence for evaluating new medications. Meta-analyses and post hoc analyses of existing
trials are increasingly used to compare therapies, in part to spare the expense and time
required to conduct direct comparator studies. However, these analyses have been criticised
methodologically, and should not be used on their own to evaluate therapies. Controlled
clinical trials evaluating triptans in the acute treatment of migraine demonstrate clear
superiority of the drugs over placebo, but only small differences are reported in studies
comparing individual triptans and triptans with non-triptan medications. In contrast, patients
can clearly distinguish between triptans and non-triptans and between different triptans in
clinical practice. This incongruity may be due to the relative insensitivity of the standard
clinical trial endpoint, relief of headache. New and more sensitive clinical trial endpoints are
required for use in clinical studies that reflect everyday general practice (naturalistic studies).
Among the potential endpoints are assessments of patient preference, impact on the
patient’s daily life and quality of life. However, novel endpoints should be developed, in
collaboration with patients, which can summarise the whole migraine experience.

* Edited from the published manuscript: Dowson AJ, Kilminster S, Peters M, Lipscombe S, Rees T,
Carter F, Darling S. Understanding the evidence: evaluating the efficacy of migraine medications in
clinical practice. Headache Care 2005;2:133–43.

49
Introduction

Migraine therapies are generally divided into acute (treatment of individual attacks as they
occur) and prophylactic (pharmacological and non-pharmacological treatments taken daily to
prevent attacks occurring). Most migraine treatments are relatively mature and the design
and methodology of investigational studies varied quite widely.

The introduction of the triptans to the armamentarium of acute treatments for migraine has
revolutionised the management of the condition over the past decade or so. Seven triptans
are now available to the physician in many countries: sumatriptan, naratriptan, zolmitriptan,
rizatriptan, almotriptan, eletriptan and frovatriptan. Triptans can be prescribed in several
different formulations: conventional tablets, melt-in-the mouth or orally disintegrating tablets
(ODTs), nasal sprays, subcutaneous injections and, in some countries, suppositories are all
available. In general, the primary studies with the triptans were of good quality and were
conducted to similar principles of design and methodology.1 However, the resultant
competition between the different manufacturers has led to a plethora of supporting clinical
studies, and meta-analyses and post hoc analyses of clinical studies, all trying to establish
superiority of one triptan over another. The result is a confusion of different study designs,
endpoints and analyses, which can make it difficult for the physician to evaluate the evidence
and to select the most appropriate drug for the individual patient.

This article aims to provide an objective overview of the clinical methodology used in, and
data resulting from, studies conducted with the triptans and other acute migraine treatments.
We also analyse the misunderstandings and pitfalls, which have biased the results from
studies, and provide guidance on what to look for when evaluating the evidence from
migraine studies. This general guidance is intended to be applicable to the evaluation of
other migraine treatments (acute and preventive) and to treatments for other therapy areas.
This article is based on discussions held at a meeting of the Migraine in Primary Care
Advisors (MIPCA), which took place on 1 May 2003.

Clinical trial methodology

Historically, the design and methodology of clinical studies of acute migraine treatments were
decided on an ad hoc basis, as no agreed criteria were available. Indeed, diagnostic criteria
were not established by the International Headache Society (IHS) until 1988.2 However,
there was increasing pressure from regulatory authorities during the 1970s and 1980s to
demonstrate a strong evidence base of efficacy and safety for new drugs. Scientists and
clinicians took up this challenge in the late 1980s and early 1990s and developed new
guidelines for study design, selection of patients, study methodology, endpoints and
analyses of study data that are mostly still in use today in short-term controlled studies in
migraine:1
• Studies are multicentre, randomised, double-blind, placebo- or comparator-controlled
and generally parallel group in design. A crossover design is used only infrequently,
due to the possibility of carryover effects.
• Eligible patients are those diagnosed with migraine according to the IHS criteria.1
Patients usually have to be adults with established migraine who experience between
one and four migraine attacks per month and without significant co-morbidities.
• Patients treat either one or three migraine attacks of moderate-to-severe intensity
with study medication and evaluate the effect of treatment on record cards.
• Patients self-rate the intensity of their headache and other symptoms. The primary
endpoint chosen was the relief of headache, defined as an improvement from severe
or moderate to mild or no pain at 2 hours after treatment (Figure 1). Secondary
endpoints included: headache relief at other time points, pain-free status at 2 hours
and other time points, time to self-rated ‘meaningful relief’, time to return of normal

50
 daily activities, presence of non-headache symptoms (nausea, vomiting, photophobia
and phonophobia), and the incidence of headache recurrence and adverse events.
• Large numbers of patients are required and must be recruited to achieve sufficient
power for the statistical analyses. Data are collected as the proportion of patients
reporting the respective endpoints and are analysed using non-parametric statistical
tests.

In contrast, long-term studies may be stand-alone, or open-label extensions to the controlled

clinical studies. Patients treat all their migraine attacks with the triptan for a 6- or 12-month
period. The primary endpoint is usually the incidence and nature of reported adverse events.
Efficacy data are also collected in terms of the percentage of patients or attacks where
headache relief is reported.

Figure 1. Primary endpoints from migraine clinical studies: Standard1 and IHS4
recommended outcomes.

Pain relief (Standard)

3 2 1 0
Severe Moderate Mild None

Pain-free (IHS)

Perspective on clinical trial methodology

When reviewing data from separate clinical studies, it is important to review the methodology
used. Evidence from clinical studies can be graded in terms of quality:3
• Grade A encompasses consistent evidence from one or more well-designed and
randomised clinical studies, as well as that from systematic meta-analyses.
• Grade B evidence is derived from well designed but non-randomised studies, and
from non-systematic meta-analyses and post hoc analyses of controlled studies.
• Grade C evidence is derived from opinion and practices used in the absence of
objective evidence.
Grade A evidence is essential in order to make clear recommendations for clinical practice.
Grade B and C evidence should only be used to guide the objectives and design of future
clinical studies.3 In practice, evidence from Grade B and Grade C studies is used to guide
clinical decisions in the absence of Grade A evidence.

Clinical studies need to be powered appropriately to detect expected differences between

study drugs. Increasing the numbers of patients directly increases the power and clinical

51
relevance of the results. It is also important that study groups (e.g. active treatment and
placebo) are of equivalent size. In migraine studies, about 100–200 patients per group are
usually sufficient to distinguish between active treatment and placebo, but several hundred
patients or more may be needed in each group to power differences between two active
treatments of similar efficacy.

It is important that patient groups are similar in different treatment groups. For example,
hospital-based patients may respond differently to outpatients or primary care patients. The
physician should be aware that it is possible to manipulate patient populations to obtain the
result desired. Patients may not always be from a homogeneous group. For example, the
typical recruitment of patients with 1–4 headache attacks per month in triptan studies may
include sufferers of both migraine and chronic headache. The latter may be difficult to treat
and it is unlikely that a triptan would be effective on its own for patients with chronic
headaches. Furthermore, large differences in placebo response may point to large
differences in the case mix.

The design and endpoints should also be consistent across studies. Most randomised,
controlled studies with the triptans follow the standard criteria1 but studies comparing triptans
to non-triptan medications frequently use different endpoints that may lead to atypical results.
The original standard primary endpoint of headache relief1 has come under challenge
recently. The IHS now recommends that patients be pain-free at the 2-hour endpoint (Figure
1).4 An even more stringent endpoint of ‘sustained pain-free’ is also advocated in recent
publications.5 This specifies that patients be pain-free at 2 hours after treatment, and that
they experience no headache recurrence and use no rescue medication in the 24 or 48
hours after treatment. One problem with these newer endpoints is the low proportion of
patients who respond, and that these definitions of efficacy may not correlate with a clinically
relevant response. Interestingly, the results from many secondary endpoints (e.g.
‘meaningful relief’, return to normal functioning and relief of non-headache symptoms)
usually parallel headache relief rates in clinical studies. Headache relief may, therefore, be
the most clinically-meaningful endpoint that we have for migraine studies. However, the 4-
point ‘Glaxo’ scale may not be optimal, as 5- or 7-point Likert scales are favoured in most
therapy areas.6 In addition, tolerability forms an important endpoint in all studies, assessed
as the incidence, nature and intensity of reported adverse events. It should be noted that
tolerability may not be related to safety, and that it is the side effects that are clinically
significant. Drug-related adverse events, rather than the total incidence of all events, may be
the clinically important endpoint.

Clinical study data on the triptans from well-designed, randomised studies

Due to the similarities in study design, methodology and endpoints used in most clinical
studies with the triptans, the data are directly comparable. We review data from placebo-
controlled and active comparator-controlled studies below.

Placebo-controlled studies
A summary of the efficacy of the different triptan drugs from placebo-controlled studies is
shown in Table 1.7–11

52
Table 1. Summary of the efficacy of the different triptan drugs from randomised, placebo-
controlled clinical trials: proportion of patients reporting headache relief (improvement from
severe or moderate to mild or no pain).7–11 This table is adapted and updated from Reference
7.

Proportion of patients (%)

Triptan Dose and route of Headache relief Headache relief from
administration from triptan (%) placebo (%)
Sumatriptan 6 mg subcutaneous 81–82 31–39
Sumatriptan 100 mg film-coated 56–62 17–26
tablet
Sumatriptan 100 mg fast- 72 42
disintegrating tablets86
Sumatriptan 50 mg film-coated tablet 50–62 17–32
Sumatriptan 50 mg fast-disintegrating 67 42
tablets86
Sumatriptan 25 mg tablet 52 17–27
Sumatriptan 20 mg nasal spray 55–64 25–36
Sumatriptan 12.5 mg suppository 43–69 21–48
Sumatriptan 25 mg suppository 64–74 21–48
Naratriptan 2.5 mg tablet 43–50 18–27
Zolmitriptan 2.5 mg tablet 62–65 34–36
Zolmitriptan 2.5 mg ODT 63 22
Zolmitriptan 5 mg nasal spray 70 30
Rizatriptan 10 mg tablet 67–77 35–40
Rizatriptan 10 mg ODT 74 28
Almotriptan 12.5 mg tablet 57–70 38–42
Eletriptan 40 mg tablet 62–67 19–26
Frovatriptan 2.5 mg tablet 38–40 22–35
ODT = orally disintegrating tablet

Reviewing the primary endpoint (headache relief at 2 hours) data shows that:
1. The placebo response rate is highly variable, ranging from <20–48%.
2. All triptan modalities and formulations are significantly superior to placebo in terms of
headache relief.
3. Most of the oral triptan formulations (including the conventional and orally
disintegrating tablets [ODTs]) provide broadly similar levels of headache relief,
although efficacy may vary somewhat in different studies with the same drug.
However, naratriptan and frovatriptan have lower reported 2-hour efficacy rates
compared to the other oral triptans.
4. Subcutaneous sumatriptan is more effective than any other triptan formulation. Nasal
spay formulations may be superior to oral forms, and certainly have a faster onset of
action (within 15 minutes as compared to 30 minutes or more for tablets).

Active comparator studies

Relatively few well-designed, head-to-head, Grade A comparator studies between the oral
triptans have been published and we have only found 12 to date conducted with marketed
formulations (Table 2).12–23

53
Table 2. Summary of the efficacy of the marketed oral triptans from randomised, controlled,
direct comparator clinical studies: proportion of patients reporting headache relief
(improvement from severe or moderate to mild or no headache pain) at 2 hours after
treatment.12–23

Sumatriptan comparator studies

Proportion of patients (%)
Study Comparator Sumatriptan Placebo p-value
drug (active
comparators)
Zolmitriptan 5 mg vs. 59 61 44 NS
Sumatriptan 100 mg12
(n=1,058)
Zolmitriptan 2.5 mg vs. 63 67 None NS
Sumatriptan 50 mg13
(n=1,522)
Rizatriptan 10 mg vs. 67 62 40 NS
Sumatriptan 100 mg14
(n=1,268)
Rizatriptan 10 mg vs. 72 68 38 NS
Sumatriptan 50 mg15
(n=1,329)
Eletriptan 40 mg vs. 65 55 24 NS
Sumatriptan 100 mg16
(n=692)
Eletriptan 40 mg vs. 67 59 26 <0.001
Sumatriptan 100 mg17
(n=2,113)
Naratriptan 1, 2.5, 5, 7.5, 10 52–69 60 31 NS
mg vs. Sumatriptan 100
mg18 (n=643)
Almotriptan 12.5 mg vs. 58 57 None NS
Sumatriptan 50 mg19
(n=1,173)
Almotriptan 12.5 mg vs. 57 64 42 NS
Sumatriptan 100 mg20
(n=668)

54
Eletriptan comparator studies

Proportion of patients (%)

Study Comparator Naratriptan Placebo p-value
drug (active
comparators)
Eletriptan 40 mg vs. 60 64 22 NS
Zolmitriptan 2.5 mg21
(n=1,312)

Naratriptan comparator studies

Proportion of patients (%)

Study Comparator Naratriptan Placebo p-value
drug (active
comparators)
Rizatriptan 10 mg vs. 45* 21* None <0.001
Naratriptan 2.5 mg22 (n=522)
Eletriptan 40 mg vs. 56 42 31 <0.01
Naratriptan 2.5 mg23 (n=548)
NS = not significant
* Pain-free at 2 hours
In this context, ‘comparator’ refers to the triptan being compared with sumatriptan, eletriptan
and naratriptan in the three sub-tables.

Overall, headache relief rates at 2 hours did not differ significantly in studies comparing
sumatriptan (50 mg and 100 mg) with zolmitriptan 2.5 mg and 5 mg,12,13 rizatriptan 10 mg,14,15
eletriptan 40 mg,16 naratriptan18 and almotriptan 12.5 mg.19,20 One study showed that
eletriptan 40 mg was significantly superior to sumatriptan 100 mg,17 but the numerical
advantage was small and it required a study population of over 2,000 patients to achieve
statistical significance. A further large study showed similar headache relief rates with
eletriptan 40 mg and zolmitriptan 2.5 mg.21 One other factor confounds these results.
Sumatriptan (but not eletriptan) was encapsulated in all comparator studies with eletriptan,
and studies have shown that the absorption of sumatriptan is reduced by this procedure.24
These results should, therefore, be approached with caution. Eletriptan 80 mg was shown to
be significantly superior to sumatriptan 100 mg17 and zolmitriptan 2.5 mg,21 but this high dose
is not recommended for use in clinical practice. However, two studies have shown that
rizatriptan and eletriptan are significantly superior to naratriptan in terms of headache relief
rates 2 hours after treatment.22,23

Rizatriptan and almotriptan do have some evidence of superiority over sumatriptan in these
studies. Rizatriptan appears to be more rapidly acting than sumatriptan,14,15 with a
significantly higher headache relief rate reported at 1 hour in one study.14 Almotriptan was
reported to be as effective as sumatriptan, but was associated with a significantly lower
incidence of adverse events,19,20 with significantly fewer chest symptoms reported in one
study.19 In these studies, both almotriptan and sumatriptan were encapsulated.

Comparison of triptans with non-triptan acute treatments for migraine

Well-designed Grade A clinical studies comparing triptans and non-triptan medications using
the endpoint of headache relief are also relatively scarce, and we have located nine studies
comparing oral triptans with ergotamine and analgesic-based preparations (Table 3).25–33

55
Table 3. Summary of the efficacy of the marketed oral triptans and non-triptan medications
from randomised, controlled, direct comparator clinical studies: proportion of patients
reporting headache relief (improvement from severe or moderate to mild or no headache
pain) at 2 hours after treatment.25–33

Proportion of patients (%)

Study Triptan Non-triptan Placebo p-value
drug (active
comparators)
Sumatriptan 100 mg vs. 66 48 None <0.001
ergotamine 2 mg + caffeine
200 mg25 (n=580)
Eletriptan 40 mg vs. 54 33 None <0.001
ergotamine 2 mg + caffeine
200 mg26 (n=733)
Sumatriptan 100 mg vs. 56 45 None NS
aspirin 900 mg +
metoclopramide 10 mg27
(n=358)
Sumatriptan 100 mg vs. 53 57 24 NS
aspirin 900 mg +
metoclopramide 10 mg28
(n=421)
Zolmitriptan 2.5 mg vs. 33 33 None NS
aspirin 900 mg +
metoclopramide 10 mg29
(n=666)
Sumatriptan 50 mg vs. 56 53* 31 NS
aspirin 1,000 mg vs. 60**
ibuprofen 400 mg30 (n=312)
Sumatriptan 100 mg vs. 79 77 29 NS
tolfenamic acid 200 mg31
(n=141)
Sumatriptan 50 mg vs. 53 53 29 NS
naproxen 500 mg +
metoclopramide 16 mg32
(n=546)
Sumatriptan 50 mg vs. 33 36 None NS
paracetamol 500 mg
+domperidone 10 mg33
(n=120)
* Aspirin 1,000 mg; ** Ibuprofen 400 mg; NS = not significant

Two studies compared triptans with Cafergot, an oral preparation of ergotamine 2 mg plus
caffeine 200 mg. Both sumatriptan 100 mg and eletriptan 40 mg resulted in significantly
superior headache relief, compared with Cafergot.25,26 However, the drug preparations may
not have been equivalent in these studies. Oral ergotamine formulations have poor
bioavailability and suboptimal efficacy. Parenteral ergotamine formulations, delivered rectally
or by injection, are superior to oral formulations and a better study design would have been
to use a double-dummy procedure to compare the different formulations.3

Several studies have compared oral triptans to analgesic-based drugs: aspirin plus
metoclopramide,27–29 aspirin alone or non-steroidal anti-inflammatory drugs (NSAIDs),30–32
and paracetamol plus domperidone.33 In all of these studies, headache relief rates were

56
similar in the triptan and analgesic groups, although the triptans had superior pain-free data
in some.

One multicentre, randomised study has compared sumatriptan suppositories with a fixed
combination of indomethacin, prochlorperazine and caffeine in 112 patients.34 Again,
headache relief at 2 hours was similar in the two groups (sumatriptan = 65% versus 71% for
the combination), although 2-hour pain-free data were significantly superior for the
indomethacin combination (34% versus 49%, p<0.01).

Perspective on clinical study data

Evidence from randomised and well-designed Grade A clinical studies shows that, compared
with placebo, all triptans are effective and well-tolerated in the acute treatment of migraine.
Of the oral triptans, sumatriptan, zolmitriptan, rizatriptan, almotriptan and eletriptan have
similar efficacy profiles. Direct comparator studies indicate that rizatriptan and eletriptan may
be slightly superior to sumatriptan, although the differences are numerically small and,
therefore, of uncertain clinical significance. The data for eletriptan are also confounded by
problems with the formulation of sumatriptan used in the comparisons. Almotriptan may be
better tolerated than sumatriptan, especially in terms of chest-related symptoms, but again
the clinical significance of these results is questionable. Overall, naratriptan and frovatriptan
appear to be less effective than the other oral triptans. Subcutaneous and nasal spray triptan
formulations are more effective and faster-acting than the oral formulations.

Oral sumatriptan and eletriptan are both significantly more effective than oral Cafergot.
However, since oral formulations of ergotamine do not provide optimal efficacy, the clinical
significance of these data is questionable. Several controlled studies using the conventional
design and endpoints showed that oral triptans were not superior to oral analgesic-based
therapies. However, the triptan response rate in these studies tended to be less than that
reported in placebo-controlled studies, in some cases markedly so.29,33 This may (but not
necessarily) indicate that these studies had design faults, which could limit their reliability.
Numerous studies have shown that patients consider triptans to be much more effective than
non-triptans when used in everyday clinical practice.35–38

Meta-analyses and post hoc analyses with the triptans

Meta-analyses and post hoc analyses can be used to combine and compare data from
multiple clinical studies. Formal meta-analyses combine data from published and
unpublished studies, and can be considered Grade A evidence if they are comprehensive
and reproducible. Other analyses for combining data include the calculation of therapeutic
gain (TG), number needed to treat (NNT) and number needed to harm (NNH, Table 4),39 and
form Grade B evidence. These approaches have been used frequently with the triptans, in
part probably to compensate for the relative lack of controlled comparator studies between
them.

57
Table 4. Definitions of post hoc clinical endpoints: therapeutic gain, number needed to treat
and number needed to harm39

Therapeutic gain (TG): the proportion of patients benefiting from treatment, adjusted to the
placebo rate

= Proportion of patients _ Proportion of patients

benefiting on treatment benefiting on placebo

Number needed to treat (NNT): the number of patients that have to be treated in order for
one patient to be treated successfully, adjusted for placebo

= 100 / Therapeutic gain (% of patients)

Number needed to harm (NNH): the number of patients that have to be treated in order for
one patient to report an adverse event, adjusted for placebo

= 100 / Proportion of patients reporting adverse events on treatment – Proportion of

patients with adverse events on placebo

Meta-analyses
Two meta-analyses have been published comparing the oral triptans, using sumatriptan 100
mg as the baseline comparator. Ferrari and co-workers analysed 53 published and
unpublished clinical studies involving 24,089 patients.40 Their analyses indicated that,
compared with sumatriptan, rizatriptan 10 mg exhibited better efficacy and consistency of
response, and similar tolerability. Almotriptan 12.5 mg exhibited better sustained pain-free,
consistency of response and tolerability data, while naratriptan showed superior tolerability.
All other triptans (except frovatriptan, which was not included in the analyses) had similar
clinical profiles to sumatriptan, and all were effective and well tolerated. The authors
concluded that, of the marketed doses, rizatriptan 10 mg and almotriptan 12.5 mg offered the
best chance of overall success. Belsey has published a further meta-analysis of 28 studies
involving 13,204 patients. He analysed TG, NNT and NNH data and concluded that, of the
oral triptans, only rizatriptan 10 mg was markedly superior to sumatriptan 100 mg in terms of
overall efficacy and tolerability.41

Post hoc analyses

Several post hoc analyses of clinical data have been conducted with the oral triptans,
investigating a variety of endpoints. One study analysed TG data, which showed clearly that
subcutaneous sumatriptan 6 mg provided superior efficacy to any of the oral triptans.42 All
the oral triptans were effective, although naratriptan and frovatriptan were the least effective,
and almotriptan also seemed to be relatively less effective than oral sumatriptan,
zolmitriptan, rizatriptan and eletriptan.42,43 An analysis combining data from three head-to-
head comparator studies between eletriptan 40 mg and sumatriptan 100 mg indicated that
eletriptan was the more effective drug.44 Significantly more patients on eletriptan (67%
versus 57%, p < 0.0001) reported headache relief at 2 hours after treatment, the primary
endpoint, and secondary endpoints gave similar results. In a different type of post hoc
analysis, almotriptan 12.5 mg proved to be more effective than rizatriptan 10 mg and
sumatriptan 50 mg and 100 mg, in terms of a composite endpoint of cost-effectiveness ratio
for sustained pain relief and no adverse events.45,46

The manufacturers of rizatriptan have conducted several post hoc analyses comparing their
drug to the other oral triptans. Combining data from several clinical studies, more patients
taking rizatriptan were satisfied with their therapy,47 were pain-free at 2 hours and 24 hours
after treatment,5 and were free of nausea after 2 hours than with other triptans.48

58
Perspective on meta-analyses and post hoc analyses
Overall, the data from the meta-analyses and post hoc analyses with the triptans are
consistent with those from controlled clinical studies. All the oral triptans were effective and
well tolerated, and there were more similarities than differences between the drugs.
Rizatriptan 10 mg and eletriptan 40 mg exhibited the best efficacy profiles, while naratriptan
2.5 mg and almotriptan 12.5 mg exhibited the best tolerability. However, these analyses have
come under criticism from several sources.

Rigorous meta-analyses do provide Grade A clinical evidence.3 However, there is evidence

that the meta-analyses conducted with the triptans may not fulfil the necessary criteria for
this. The two meta-analyses did not provide consistent data, and the selection of studies,
patient populations and endpoints differed. The meta-analysis by Ferrari et al40 is the most
robust of the two, involving a substantially larger number of studies and patients than that of
Belsey.41 However, even this large analysis has been criticised in terms of its methodology,
statistical analyses and the clinical significance of its results.49–53 Using analyses of TG and
NNT to compare triptans has also been criticised. Placebo response rates are crucial to the
calculation of these values, and are highly variable in migraine studies54 while the response
to active treatments tends to be less variable (Table 1). Several pieces of research have
shown that NNT values are correlated more strongly to the placebo response than to the
active treatment response.55 A systematic analysis of placebo-controlled studies with the
triptans showed that the triptan response rate followed a normal distribution, while that of
placebo did not.56 Both TG and NNT analyses failed the test of normality. The authors
concluded that these transformations of data are a potential source of bias in meta-analyses
of acute migraine treatments. Another group of authors has concluded that analysing TG and
NNT adds no new information to the evaluation of placebo-controlled trials.55 There is also
some argument as to whether the placebo response should be so disregarded, as placebo
may activate the same pain modulating brain structures as pain-relieving drugs.57,58
Estimations of TG and NNT may, therefore, significantly underestimate the efficacy of active
treatments. It has been concluded that migraine therapies can only be effectively compared
using well-designed head-to-head studies, and not by meta-analyses.55

Triptans in clinical practice

One of the characteristic features of the triptans is that their reported efficacy is greater when
used in clinical practice than when used in controlled clinical trials. In studies which reflected
everyday clinical practice (naturalistic studies), reported headache relief rates following
treatment with sumatriptan and zolmitriptan were typically close to 80%.59,60 There are
several possible reasons that may explain these findings:
• The populations of patients may differ. Young and middle-aged adults predominate in
clinical trials, while the general population of migraine sufferers contains all age
groups, from children and adolescents, to older people.61
• The severity of the headache may differ. Migraine sufferers in the general population
experience, on average, about one migraine attack per month.61 In clinical studies,
patients may have up to four attacks per month.1 It is likely that some of these
patients have chronic daily headache, a condition that is more severe and more
difficult to treat than migraine.62
• Clinical studies are mostly conducted in Western Europe and North America, in
Caucasian patients. Other countries and races may respond differently to the triptans.
• Patients in clinical practice may not use the strict doses that are given in clinical
studies. Dose-titration studies have shown that, given the choice, the majority of
patients end up using higher doses of the triptans, e.g. sumatriptan 100 mg and
zolmitriptan 5 mg.59,60 For some triptans, e.g. naratriptan and frovatriptan, the
recommended dose may not be at the top of the dose-response curve. The use of

59
 relatively high doses may, therefore, result in a markedly greater efficacy than that
reported in clinical studies.
• In clinical practice, patients may not wait until the headache is moderate or severe
before treating with a triptan. In fact, recent evidence from randomised and controlled
studies indicates that triptans are most effective when taken early in the migraine
attack when the headache is mild. With sumatriptan 100 mg, pain-free rates of 71%
were achieved 2 hours after the treatment of mild headache, compared to 54% when
the initial headache was moderate or severe (p<0.05).63 Similar results have been
reported with zolmitriptan64 and almotriptan,65 and it seems likely that future treatment
guidelines will recommend that triptans be taken early in the attack when the
headache is mild. Suggestions have recently been published on how to conduct
randomised, controlled ‘early treatment’ studies with the triptans.66

The search for new study endpoints

The primary endpoints used in clinical trials with the triptans were designed to demonstrate
significant differences between active treatments and placebo.1 This was done to comply
with the need to produce documents suitable for submission to regulatory authorities
worldwide, rather than to devise endpoints suitable for use in everyday clinical practice. In
fact, the main endpoint of headache relief proved unable to distinguish between triptans and
other active treatments. More sensitive endpoints may be required for use in clinical practice.
In a survey, migraine patients reported the following attributes of acute treatments to be most
important to them: complete pain relief, no headache recurrence, rapid onset of action, no
side effects and relief of non-headache associated symptoms.67 A global measure of efficacy
and tolerability is probably the best option to cover all these areas. Such global measures
that have been used in migraine studies include assessments of patient preference and
satisfaction, the impact on patients’ daily lives and quality of life.

Patient preference and satisfaction

Patient preference is a subjective global measure of efficacy and tolerability that can be
simply assessed by the patient. Numerous migraine studies have used these assessments
as secondary endpoints. In these studies, patients consistently preferred oral triptans over
non-triptan acute medications.68 Additionally, patients were able to distinguish between
different oral triptans, and expressed clear preferences between them. However, their
responses were idiosyncratic. The main reasons patients gave for preferring one triptan over
the other was a faster onset of action, a longer duration of effect and fewer adverse events.
However, these reasons were all given for sumatriptan, zolmitriptan and rizatriptan in
different studies, and the response could not be predicted in advance.69–71 One caveat of
assessing patient preference is that it can only relate the performance of one drug to a
second one. This endpoint does not provide absolute information about a drug, and
precludes meta-analysis.

Impact assessments
Assessing the impact of migraine on patients’ daily lives has become an increasingly used
measure of migraine severity over the past few years. Impact questionnaires have been
developed, validated, and tested as outcome measures. The Migraine Disability Assessment
(MIDAS) Questionnaire has shown sensitivity to change in studies of migraine and chronic
daily headache,72,73 and further research may prove it to be a useful clinical study endpoint.
Two other impact questionnaires, the Headache Impact Test (HIT)74 and the Short Pain
Inventory (SPI),75 require development as outcome measures.

Quality of life measures

Migraine sufferers are known to have significantly poorer quality of life than the general
population during their attacks, and also between attacks when they are regarded as
symptom-free.76,77 Generic and migraine-specific QOL questionnaires have been developed

60
for migraine, and consistently show that effective migraine treatment results in improved
patient QOL.78 Using one of the migraine-specific questionnaires as an outcome measure
(the 24-hour Migraine-Specific QOL Questionnaire), it was demonstrated that rizatriptan 10
mg was significantly superior to non-triptan acute medications in terms of work and social
functioning, energy, and patients’ feelings and symptoms.79

New initiatives
All the above questionnaires have the potential to summarise the whole migraine experience
and act as effective outcome measures. However, more work is required to define patients’
responses and validate these assessment tools against the existing gold standard measure
of headache relief. A new questionnaire, the Migraine Assessment of Current Therapy
(Migraine-ACT) assesses the need for reviewing acute medications in terms of four
questions in clinically-relevant domains: impact, global assessment of relief, consistency of
response and emotional response.80 Patients answer the questions as ‘yes’ or ‘no’ and an
increasing number of ‘no’ answers indicates increasing medical needs. It therefore combines
elements of patient satisfaction, impact and QOL. A large prospective study showed that
Migraine-ACT exhibited high reliability and validity, and had considerable potential clinical
utility.80 What is now needed is to investigate the sensitivity to change of Migraine-ACT and
to develop a similar questionnaire for preventive medications.

Another approach is to examine the principles underlying patient response to therapy, using
qualitative research. Qualitative research aims to increase the understanding of patients and
physicians and could help to provide a more realistic assessment of headache treatments
and to develop relevant patient-centred endpoints.81 Open ended-questions allow patients
and physicians to speak freely about their perspectives, including beliefs about interventions,
and the benefits and preferences of therapies. Asking a different type of questions,
qualitative research can provide information not available from conventional quantitative
methods. Qualitative research can be combined with quantitative methods, by conducting a
qualitative study to generate hypotheses, which can be tested using quantitative methods.82
Qualitative research is now recommended by the UK Medical Research Council for use in
the development of clinical trials programmes, to gain better understanding of patients’ needs
and to develop new clinical endpoints.83 Qualitative methods can also follow quantitative
methods, to help explain quantitative findings, e.g. explaining anomalous results such as the
unexpected similarity in efficacy of triptan and non-triptan medications and differences in
placebo response in separate patient populations.84 Combining qualitative and quantitative
methods in randomised controlled trials is believed by some researchers to be the best
means of assessing whether (and why) interventions work (efficacy) and whether (and how)
they work in clinical practice (effectiveness). Such initiatives can increase the generalisability
of results and generate information to help successfully implement and refine effective
interventions in natural settings.85

Conclusions – understanding the evidence

In evaluating clinical evidence, the randomised, double-blind, controlled clinical trial remains
the gold standard for evaluating the clinical profile of migraine drugs. However, even these
studies may have deficiencies, and there are several items to look out for when evaluating
good quality headache studies:
• The number of patients in the study needs to be appropriate for the analyses
planned. In general, there should be at least 50 patients per treatment group, with
equal numbers of patients in the active and placebo (or active competitor) groups. It
is interesting that large numbers of patients are required to show significance in
evaluating migraine therapies, up to 200 per group in placebo-controlled studies2,7
and up to 1,000 or more in triptan comparator studies.17

61
 • Patient withdrawals from the study should be low, and not exceed 10% of the total.
The proportion of patients withdrawing should be similar in the separate treatment
groups.
• Outcome measures should be quantifiable, discriminate well and have intuitively
meaningful units. Categorical scales (e.g. five- or seven-point Likert scales) are
preferable to visual analogue scales, although 10-point numerical scales work well in
the Netherlands.
• The data should be clinically relevant. For treating migraine, speed of onset, duration
of response, and assessments of functional status or well-being are important in
addition to headache relief at defined time points.3,8

On the whole, the results from meta-analyses and post hoc analyses were congruent with
those from controlled clinical studies. However, they have also generated controversy, and
their chief utility may be to suggest avenues for future clinical studies, rather than being used
on their own as a means of selecting acute treatments for migraine.

However, results from clinical trials may not reflect the situation in clinical practice for
migraine. For example, clinical studies show only small clinical differences between the oral
triptans,12–21 whereas patients exhibit clear, if individual, preferences between them.69–71 The
challenge is to develop studies that can effectively distinguish between the triptans, or to
distinguish between patient-dependent sensitivity to individual triptans. This may be achieved
by conducting naturalistic studies in everyday clinical practice with the full range of patients
who suffer from migraine, allowing them to use the drugs according to normal prescription
recommendations. The development of new, more sensitive and valid endpoints to assess
migraine relief in collaboration with patients would facilitate this process.

Acknowledgements

The authors are pleased to acknowledge the input of the other delegates at the MIPCA
meeting held on 1 May 2003, and additional MIPCA members who reviewed this manuscript:
Dr Jerry Sender, Dr Ahmet Fuat, Dr David Watson, Dr Gregory Parkin-Smith (physicians),
Ms Heather MacBean (nurse practitioner) and Ms Ann Turner (Director of the Migraine
Action Association, the UK patient support group). We are also grateful to AstraZeneca, who
sponsored the meeting with an unrestricted educational grant.

References

1. Pilgrim AJ. Methodology of clinical trials of sumatriptan in migraine and cluster

headache. Eur Neurol 1991;31:295–9.
2. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8(Suppl 7):1–92.
3. Silberstein SD, for the US Headache Consortium. Practice parameter: evidence-
based guidelines for migraine headache (an evidence-based review). Report of the
Quality Standards Subcommittee of the American Academy of Neurology. Neurology
2000; 55:754–62.
4. International Headache Society Clinical Trials Subcommittee. Guidelines for
controlled trials of drugs in migraine: second edition. Cephalalgia 2000;20:765–86.
5. Adelman JU, Lipton RB, Ferrari MD et al. Comparison of rizatriptan and other triptans
on stringent measures of efficacy. Neurology 2001;57:1377–83.
6. Atkinson MJ, Sinha A, Hass SL et al. Validation of a general measure of treatment
satisfaction, the Treatment Satisfaction Questionnaire for Medication (TSQM) using a
national panel study of clinical disease. Health Qual Life Outcomes 2004;26:12.

62
7. Sheftell FD, Fox AW. Acute migraine treatment outcome measures: a clinician’s view.
Cephalalgia 2000;20 (Suppl 2):14–24.
8. Dowson AJ, Lipscombe S, Sender J et al. New guidelines for the management of
migraine in primary care. Curr Med Res Opin 2002;18:414–39.
9. Savani N, Brautaset NJ, Reunanen M et al. A double-blind placebo-controlled study
assessing the efficacy and tolerability of 50 mg sumatriptan tablets in the acute
treatment of migraine. Int J Clin Pract 1999; 105 Suppl:7–15.
10. Mathew NT, Schoenen J, Winner P et al. Comparative efficacy of eletriptan 40 mg
versus sumatriptan100 mg. Headache 2003;43:214–22.
11. Sheftell F, Ryan R, Pitman V et al. Efficacy, safety, and tolerability of oral eletriptan
for treatment of acute migraine: a multicenter, double-blind, placebo-controlled study
conducted in the United States. Headache 2003;43:202–13.
12. Geraud G, Olesen J, Pfaffenrath V et al. Comparison of the efficacy of zolmitriptan
and sumatriptan: issues in migraine trial design. Cephalalgia 2000;20:30–8.
13. Gruffydd-Jones K, Kies B, Middleton A et al. Zolmitriptan versus sumatriptan for the
acute oral treatment of migraine: a randomised, double-blind, international study. Eur
J Neurol 2001;8:237–45.
14. Tfelt Hansen P, Teall J, Rodriguez F et al. Oral rizatriptan versus oral sumatriptan: a
direct comparative study in the acute treatment of migraine. Headache 1998;38:748–
55.
15. Goldstein J, Ryan R, Jiang K et al. Crossover comparison of rizatriptan 5 mg and 10
mg versus sumatriptan 25 mg and 50 mg in migraine. Headache 1998;38:737–47.
16. Goadsby PJ, Ferrari MD, Olesen J et al. Eletriptan in acute migraine: a double-blind,
placebo-controlled comparison to sumatriptan. Neurology 2000;54:156–63.
17. Mathew NT, Schoenen J, Winner P et al. Comparative efficacy of eletriptan 40 mg
versus sumatriptan 100 mg. Headache 2003;43:214–22.
18. Havanka H, Dahlöf C, Pop PH et al. Efficacy of naratriptan tablets in the acute
treatment of migraine: a dose-ranging study. Clin Ther 2000;22:970–80.
19. Spierings EL, Gomez-Mancilla B, Grosz DE et al. Oral almotriptan vs. oral
sumatriptan in the abortive treatment of migraine: a double-blind, randomized,
parallel-group, optimum-dose comparison. Arch Neurol 2001;58:944–50.
20. Dowson AJ, Massiou H, Lainez JM et al. Almotriptan is effective and well-tolerated for
migraine pain: results of a randomized, double-blind, placebo-controlled clinical trial.
Cepahalalgia 2002;22:453–61.
21. Steiner TJ, Diener H-C, MacGregor EA et al. Comparative efficacy of eletriptan and
zolmitriptan in the acute treatment of migraine. Cephalalgia 2003;23:942–52.
22. Bomhof M, Paz J, Legg N et al. Comparison of rizatriptan 10 mg vs. naratriptan 2.5
mg in migraine. Eur Neurol 1999;42:173–9.
23. Garcia-Ramos G, MacGregor EA, Hilliard B et al. Comparative efficacy of eletriptan
vs. naratriptan in the acute treatment of migraine. Cephalalgia 2003;23:869–76.
24. Fuseau E, Petricoul O, Sabin A et al. Effect of encapsulation on absorption of
sumatriptan tablets: data from healthy volunteers and patients during a migraine. Clin
Ther 2001;23:242–51.
25. The Multinational Oral Sumatriptan and Cafergot Comparative Study Group. A
randomized, double-blind comparison of sumatriptan and cafergot in the acute
treatment of migraine. Eur Neurol 1991;31:314–22.
26. Diener HC, Jansen JP, Reches A et al. Efficacy, tolerability and safety of oral
eletriptan and ergotamine plus caffeine (Cafergot) in the acute treatment of migraine:
a multicentre, randomised, double-blind, placebo-controlled comparison. Eur Neurol
2002;47:99–107.
27. The Oral Sumatriptan and Aspirin plus Metoclopramide Comparative Study Group. A
study to compare oral sumatriptan with oral aspirin plus oral metoclopramide in the
acute treatment of migraine. Eur Neurol 1992;32:177–84.

63
28. Tfelt-Hansen P, Henry P, Mulder LJ et al. The effectiveness of combined oral lysine
acetylsalicylate and metoclopramide compared with oral sumatriptan for migraine.
Lancet 1995;346:923–6.
29. Geraud G, Compagnon A, Rossi A. Zolmitriptan versus a combination of
acetylsalicylic acid and metoclopramide in the acute oral treatment of migraine: a
double-blind, randomised, three-attack study. Eur Neurol 2002;47:88–98.
30. Diener H-C, Bussone G, de Liano H et al. Placebo-controlled comparison of
effervescent acetylsalicylic acid, sumatriptan and ibuprofen in the treatment of
migraine attacks. Cephalalgia 2004;24:947–54.
31. Myllylä VV, Havanka H, Herrala L et al. Tolfenamic acid rapid release versus
sumatriptan in the acute treatment of migraine: comparable effect in a double-blind,
randomized, controlled, parallel-group study. Headache 1998;38:201–7.
32. Pellegrino R. Comparative efficacy of MT 100 and sumatriptan 50 mg in treating
acute migraine subjects with severe pain or nausea at baseline. Cephalalgia
2003;23:701 (Abstract).
33. Dowson A, Ball K, Haworth D. Comparison of a fixed combination of domperidone
and paracetamol (Domperamol®) with sumatriptan 50 mg in moderate to severe
migraine: a randomised UK primary care study. Curr Med Res Opin 2000;16:190–7.
34. Di Monda V, Nicolodi M, Aloisio A et al. Efficacy of a fixed combination of
indomethacin, prochlorperazine and caffeine versus sumatriptan in acute treatment of
multiple migraine attacks: a multicenter, randomized, crossover trial. Headache
2003;43:835–44.
35. Kwong WJ, Chalal R, Putnam G et al. Migraine patients prefer Imitrex tablets 50 mg
to their usual non-triptan prescription or over-the-counter therapy. Cephalalgia
2001;21:411 (Abstract).
36. Block GA, Goldstein J, Polis A et al. Efficacy and safety of rizatriptan versus standard
care during long-term treatment for migraine. Headache 1998;38:764–71.
37. Powers C, Szeto S, Pangtay D et al. Evaluation of migraineurs’ preferences for
naratriptan over conventional first-line agents. Arch Fam Med 2000;9:753–8.
38. Robbins L. Triptans vs analgesics: patient preference. Cephalalgia 2001;21:406
(Abstract).
39. Goadsby PJ. The scientific basis of medication choice in symptomatic migraine
treatment. Can J Neurol Sci. 1999;26 (Suppl 3):S20–6.
40. Ferrari MD, Roon KI, Lipton RB et al. Oral triptans (serotonin 5-HT1B/1D agonists) in
acute migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358:1668–75.
41. Belsey JD. Reconciling effectiveness and tolerability in oral triptan therapy: a
quantitative approach to decision making in migraine management. J Clin Res
2001;4:105–27.
42. Tfelt-Hansen P, De Vries P, Saxena PR. Triptans in migraine: a comparative review
of pharmacology, pharmacokinetics and efficacy. Drugs 2000;60:1259–87.
43. Dahlöf CGH, Dodick D, Dowson AJ, Pascual J. How does almotriptan compare with
other triptans? A review of data from placebo-controlled clinical trials. Headache
2002;42:99–113.
44. Diener HC, Ryan R, Sun W et al. The 40-mg dose of eletriptan: comparative efficacy
and tolerability versus sumatriptan 100 mg. Eur J Neurol. 2004;11:125–34.
45. Williams P, Reeder CE. A cost effectiveness comparison of almotriptan and
rizatriptan in the acute treatment of migraine. Am J Man Care 2003; in press.
46. Williams P, Reeder CE. A comparison of the cost-effectiveness of almotriptan and
sumatriptan in the treatment of acute migraine using a composite efficacy/tolerability
end point. J Manag Care Pharm. 2004;10:259–65.
47. Gerth WC, McCarroll KA, Santanello NC et al. Patient satisfaction with rizatriptan
versus other triptans: direct head-to-head comparisons. Int J Clin Pract 2001;55:552–
6.
48. Lipton RB, Pascual J, Goadsby PJ et al. Effect of rizatriptan and other triptans on the
nausea symptom of migraine: a post hoc analysis. Headache 2001;41:754–63.

64
49. Palmer JBD, Salonen R. Triptan medications to treat acute migraine. Lancet
2002;359:1151 (Letter).
50. Dowson A, Kilminster S. Triptan medications to treat acute migraine. Lancet
2002;359:1151–2 (Letter).
51. Tfelt-Hansen P. Triptan medications to treat acute migraine. Lancet 2002;359:1152
(Letter).
52. Kernick DP. Triptan medications to treat acute migraine. Lancet 2002;359:1152
(Letter).
53. Cates C. Triptan medications to treat acute migraine. Lancet 2002;359:1152 (Letter).
54. Bendtsen I, Mattson P, Zwart J-A et al. Placebo response in clinical randomized trials
of analgesics in migraine. Cephalalgia 2003;23:487–90.
55. Sheftell FD, Fox AW, Weeks RE et al. Differentiating the efficacy of 5-HT(1B/1D)
agonists. Headache 2001;41:257–63.
56. Fox AW, Keywood C, Sheftell FD et al. Consequences of transforming measures of
efficacy for acute therapies: 5-HT1B/1D as a worked example. Headache 2004;44:48–
52.
57. Diener H-C. Placebo in headache trials. Cephalalgia 2003;23:485–6.
58. Petrovic P, Kalso E, Petersson KM et al. Placebo and opioid analgesia – imaging a
shared neuronal network. Science 2002;295:1737–40.
59. Dowson AJ, Ashford EA, Prendergast S et al. Patient-selected dosing in a six-month
open-label study evaluating oral sumatriptan in the acute treatment of migraine. Int J
Clin Pract 1999;105 Suppl:25–33.
60. Tepper SJ, Donnan GA, Dowson AJ et al. A long-term study to maximise migraine
relief with zolmitriptan. Curr Med Res Opin. 1999;15:254–71.
61. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and
co-morbidities. Neurology 2001;56 (Suppl 1):4–12.
62. Silberstein SD, Lipton RB. Chronic daily headache Curr Opin Neurol 2000;13:277–83.
63. Scholpp J, Schellenberg R, Moeckesch B et al. Early treatment of a migraine attack
while pain is still mild increases the efficacy of sumatriptan. Cephalalgia
2004;24:925–33.
64. Klapper J, Lucas C, Røsjø Ø et al. Benefits of treating highly disabled migraine
patients with zolmitriptan while pain is mild. Cephalalgia 2004;24:918–24.
65. Pascual J, Cabarrocas X. Within-patient early versus delayed treatment of migraine
attacks with almotriptan: the sooner the better. Headache 2002;42:28–31.
66. Ferrari MD. Should we advise patients to treat migraine attacks early? Cephalalgia
2004;24:915–7.
67. Lipton RB, Stewart WF. Acute migraine therapy: do doctors understand what patients
want from therapy. Headache 1999;39 (Suppl 2):20–6.
68. Dowson AJ, Tepper SJ, Dahlöf C. Patients’ preference for triptans and other
medications as a tool for assessing the efficacy of acute treatments for migraine. J
Headache Pain 2005; 6:112–20.
69. Loder E, Brandes JL, Silberstein S et al. Preference comparison of rizatriptan ODT 10
mg and sumatriptan 50 mg tablet in migraine. Headache 2001;41:745–53.
70. Pascual J, Munoz R, Leira R. An open preference study with sumatriptan 50 mg and
zolmitriptan 2.5 mg in 100 migraine patients. Cephalalgia 2001;21:680–4.
71. Pascual J, Bussone G, Hernandez JF et al. Comparison of preference for rizatriptan
10 mg wafer versus sumatriptan 50 mg tablet in migraine. Eur Neurol 2001;45:275–
83.
72. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
73. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine,
and cluster headache with topiramate. Headache 2002;42:796–803.
74. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.

65
75. Kilminster SG, Dowson A, Bundy M. The Headache Impact Test® and the Short Pain
Inventory©: outcome measures compared. Int J Pharm Med 2003;17:23–32.
76. Osterhaus JT, Townsend RJ, Gandek B et al. Measuring the functional status and
well-being of patients with migraine headache. Headache 1994;34:337–43.
77. Dahlöf CGH, Dimenas E. Migraine patients experience poorer subjective well-
being/quality of life even between attacks. Cephalalgia 1995;15:31–6.
78. Solomon GD, Skobieranda FG, Genzen JR. Quality of life among migraine patients
treated with sumatriptan. Headache 1995;35:449–54.
79. Gerth WC, Ruggles KH, Stark SR et al. Improvement in health-related quality of life
with rizatriptan 10 mg compared with standard migraine therapy. Clin Drug Invest
2001;21:853–60.
80. Dowson AJ, Tepper SJ, Baos V et al. Identifying patients who require a change in
their acute migraine treatment: The Migraine Assessment of Current Therapy
(Migraine-ACT) Questionnaire. Curr Med Res Opin 2004;20:1125–35.
81. Peters M, Huijer Abu-Saad H, Vydelingum V et al. Research into headache: the
contribution of qualitative methods. Headache 2002;42:1051–9.
82. Sofaer S. Qualitative methods: what are they and why use them? Hlth Serv Res
1999;34 (5 Part II):1101–18.
83. Medical Research Council. A framework for development and evaluation of RCTs for
complex interventions to improve health. London: MRC, April 2000.
84. Lewis DW, Kellstein D, Dahl G et al. Children’s ibuprofen suspension for the acute
treatment of pediatric migraine. Headache 2002;42:780–6.
85. Weaver T, Renton A, Tyrer P et al. Combining qualitative studies with randomised
controlled trials is often useful. BMJ 1996;313:629.
86. Sheftell FD, Dahlöf CGH, Brandes JL et al. Two replicate randomized, double-blind,
placebo-controlled trials of the time to onset of pain relief in the acute treatment of
migraine with a fast-disintegrating/rapid-release formulation of sumatriptan tablets.
Clin Ther 2005;27:407–17.

66
3.2

Patients’ preference for triptans and other medications as a tool for

assessing the efficacy of acute treatments for migraine*

Abstract

Oral triptans are effective and well tolerated acute treatments for migraine, but clinical
differences between them are small and difficult to measure in conventional clinical trials.
Patient preference assesses a global measure of efficacy and tolerability, and may be a
more sensitive means of distinguishing between these drugs. In a series of studies, patients
consistently expressed a clear preference for triptans over their usual non-triptan acute
medications, e.g. analgesics and ergotamine. Direct comparator studies of patient preference
with oral triptans showed that patients could distinguish between different triptans, and
between different formulations of the same triptan. Patients could even distinguish between
the three oral doses of sumatriptan. The most frequently provided reasons for preference
were speed of response and overall effectiveness. Patient preference is a sensitive and valid
clinical trial endpoint and physicians should consider using it when reviewing the efficacy of
acute migraine medications.

* Edited from the full published article: Dowson AJ, Tepper SJ, Dahlöf C. Patients’ preference for
triptans and other medications as a tool for assessing the efficacy of acute treatments for migraine. J
Headache Pain 2005;6:112–20.

67
Introduction

Oral triptans (5-hydroxytryptamine-1B/1D receptor agonists) are effective and well tolerated
acute treatments for migraine. [1, 2] However, direct comparator trials and systematic
reviews indicate that differences between oral triptans are relatively small. [3–5] The
standard clinical trial endpoint of headache relief [6] may be relatively insensitive and not
relevant to everyday clinical practice. [7] There are also issues of study design and
encapsulation of certain formulations that may reduce the clinical applicability of some study
results. [8] Meta-analyses may be rather blunt measures of efficacy, better reflecting placebo
response rather than active response [9]. Therefore, other measures of efficacy need much
greater study. [10]

Since patients are treated on an individual basis, the more important question is not which
triptan is best relative to another, but whether the chosen triptan provides the outcome
desired by the patient and healthcare provider. A measure that evaluates the patient’s
subjective judgement of the efficacy and tolerability of therapy may be a more sensitive and
valid measure of efficacy than the standard endpoints. Patient preference evaluates a global
measure of the clinical profile, encompassing efficacy, speed of onset of action, tolerability,
consistency of response, ease of use and feelings of well being on an individual basis. [5]
Using patient preference and satisfaction data may be one approach to comparing the
triptans that provides a more real-life perspective. [11, 12] The International Headache
Society (IHS) guidelines for controlled trials of migraine treatments state that the global
evaluation of migraine medications by patients is a clinically relevant measure. [13]

Many triptan studies have reported patients’ overall evaluation of their migraine treatments.
This article reviews these data, from three types of patient preference study: comparisons of
triptans versus patients’ usual non-triptan treatments, direct comparisons of different triptans
and comparisons of different formulations or doses of the same triptan. We also review the
methodological robustness of the studies for study design and symmetry in the groups
compared, and in blinding techniques.

Comparing triptans with non-triptan medications

These comparisons were conducted in two ways. Firstly, a large meta-analysis was
conducted to capture data on patients’ satisfaction with their usual acute medications before
they entered clinical studies. Secondly, a series of patient preference studies compared
triptans with patients’ usual non-triptan medications.

Meta-analysis of sumatriptan clinical studies

A meta-analysis was conducted to investigate patients’ satisfaction with their usual acute
migraine medications before entering 10 UK clinical studies with sumatriptan. [14] In all
studies there were assessments of acute medications taken for the migraine attacks and
their efficacy. The proportions of patients rating each of the drug categories as
ineffective/poor/reasonable and good/excellent were calculated, and 95% confidence
intervals constructed for the proportions using the normal approximation (Table 1).

68
Table 1. Patients’ (n = 3,378) assessment of acute migraine treatments in the meta-analysis
of sumatriptan versus usual acute treatments [14]

Medication Patient rating %

a
(95% confidence interval)

Ineffective/poor Good/excellent Number of

/reasonable patients

Sumatriptan 12% 88% 688

(6% to 17%) (83% to 94%)
Ergotamine 62% 38% 249
(53% to 71%) (29% to 47%)
Paracetamol/codeine/ 75% 25% 530
buclizine (71% to 79%) (21% to 29%)
Aspirin/metoclopramide 77% 23% 110
(69% to 85%) (15% to 31%)
Paracetamol/metoclopramide 81% 19% 307
(77% to 85%) (15% to 23%)
Ibuprofen 83% 17% 233
(78% to 88%) (12% to 22%)
Paracetamol/codeine 88% 12% 355
(85% to 92%) (8% to 15%)
Aspirin 90% 10% 210
(86% to 94%) (6% to 14%)
Co-proxamol 91% 9% 166
(87% to 95%) (5% to 13%)
Paracetamol 97% 3% 530
(96% to 98%) (2% to 4%)
a
Approximate 95% confidence intervals are given, the exact confidence coefficients may be
lower.

Overall, the majority of patients (88%) who used sumatriptan as their usual migraine
medication rated it as good or excellent. In contrast, only 38% of patients who usually used
ergotamine, 25% of those who normally used paracetamol/codeine/buclizine, 23% of those
who used aspirin/metoclopramide and 19% of those who used paracetamol/metoclopramide
rated these medications as good or excellent.

69
Patient preference studies (Table 2)

Table 2. Summary of studies comparing patients’ preference for triptans versus their usual
non-triptan acute treatments for migraine.

% patients
Triptan and dose Preference for Preference for No preference
triptan usual non-triptan
therapy
Sumatriptan 85 10 5
subcutaneous 6 mg
[15] (n = 217)
Sumatriptan oral 50 73 18 9
mg [16] (n = 402)
Sumatriptan oral 50 69 16 14
mg [18] (n = 29)
Naratriptan oral 2.5 63 27 10
mg [19] (n = 115)
Triptans [22] (n = 52* 21* 9
663)

* 18% of patients preferred to use both a triptan and an analgesic to treat individual migraine
attacks

Sumatriptan
A prospective, multicentre, open-label, 2-month crossover study compared patients’
preference for subcutaneous sumatriptan 6 mg with their usual acute migraine treatments.
[15] Single and combination analgesics were used by 49% of patients, ergotamine by 24%,
non-steroidal anti-inflammatory drugs (NSAIDs) by 19% and dihydroergotamine (DHE) by
7%. At the end of the study, 85% of patients expressed a preference for subcutaneous
sumatriptan, 10% preferred their usual treatments and 5% had no preference (p<0.001).
Some of the comparisons in this study necessitated asymmetric comparison of formulations,
i.e., injection versus tablet.

A large, open-label, 4-attack observational study compared patients’ preference for and
satisfaction with oral sumatriptan 50 mg with those for their usual non-triptan prescription or
over-the-counter therapy (mostly non-narcotic analgesics and NSAIDs). [16, 17] At the end of
the study, 73% of patients expressed a preference for sumatriptan, and only 18% preferred
their usual therapy. The most common reasons given for preferring sumatriptan were
effective pain relief (98% of patients), restored ability to function (93%), requirement for fewer
doses (93%), relief of migraine-associated symptoms (89%), rapid onset of efficacy (86%),
no tired feelings (85%) and fewer side effects (81%). Significantly more patients were
satisfied with sumatriptan as compared with their usual therapies (p<0.001) and with the
overall quality of their medical care when it included sumatriptan (p<0.001).

An open-label, observational, multi-attack preference study in US primary care clinical

practice allowed patients not using triptans to switch to oral sumatriptan 50 mg to treat their
migraine attacks. [18] At baseline, patients were mostly using NSAIDs and other simple
analgesics (69%), OTC or prescription combination therapies (28%) and narcotics (10%),
with the majority (76%) being dissatisfied with their nontriptan therapies. At the end of the
study 69% of patients expressed a preference for sumatriptan, 16% for their previous therapy
and 14% had no preference. The main reasons given for preferring sumatriptan were speed
of relief and overall effectiveness (69% and 30% of patients, respectively). Use of
sumatriptan correlated with a reduction in unscheduled physician visits, emergency room
visits and hospitalisations for migraine.

70
Naratriptan
An open-label study conducted in US primary care assessed migraine patients’ satisfaction
with and preference for oral naratriptan 2.5 mg compared with their previous non-triptan
therapies (simple analgesics [59%], combination products [46%] and narcotics [13%]). [19]
After three treated attacks, more patients were satisfied with naratriptan than with their
previous therapies (75% versus 47%), and 63% preferred naratriptan, 27% their non-triptan
therapy and 10% expressed no preference. The main reasons for preferring naratriptan were
effective pain relief (86% of patients) and restoration of ability to function (81%).

Zolmitriptan
An open-label, multicentre study of 112 patients treating 281 migraine attacks assessed
efficacy, safety and patient acceptance of oral zolmitriptan 2.5 mg. [20] At the end of the
study, 78% of patients stated that zolmitriptan was superior to their previously used abortive
treatments (analgesics and NSAIDs).

Rizatriptan
An open-label, single-attack crossover study compared migraine patients’ (n = 216)
satisfaction with two formulations of oral rizatriptan 10 mg (conventional tablet and orally
disintegrating tablet [ODT]) over their previous non-triptan medications. [21] The study
reflected normal clinical practice, with all patients being triptan naïve. At the end of the study,
more than twice as many patients taking rizatriptan reported that they were ‘very’ or
‘somewhat satisfied’ with the medication compared with their previous non-triptan
medications (p<0.05). In all studies in which the ODT preparations are compared with
conventional tablets, there is an asymmetry in comparison groups with respect to the
formulations assessed.

Overall preference for triptans versus analgesics

A study conducted in US secondary care assessed the choice of acute migraine medications
in patients who had been prescribed triptans in the past. [22] Patients were asked whether
they currently preferred to use triptans or analgesics (OTC, prescription simple and
combination analgesics, and prescription narcotics). Fifty two percent of the patients
preferred using a triptan alone, 21% analgesics alone, 18% triptans plus analgesics for the
same attack and 9% had no preference. The main reasons for preferring triptans over
analgesics were efficacy (62% of patients), reduced side effects (8%) and a combination of
the two (30%).

71
Studies comparing the different triptans (Table 3)

Table 3. Summary of studies comparing patient preference for different triptans.

% patients
Comparison Preference for Preference for No preference
sumatriptan comparator triptan
Sumatriptan oral 50 29 44 27
mg versus
zolmitriptan oral 2.5
mg [23]
(n = 94)
Sumatriptan oral 36 45 19
versus zolmitriptan
oral 2.5 mg [20]
(n = 112)
Sumatriptan oral 50 43 57** 0
mg versus rizatriptan
ODT 10 mg [12]
(n = 374)
Sumatriptan oral 50 36 64*** 0
mg versus rizatriptan
ODT 10 mg [27]
(n = 481)

** p<0.01; *** p≤0.001; ND = no data

Almost all preference studies that compare the triptans use sumatriptan as the comparator
drug. There is a relative lack of comparative clinical data between the newer triptans, both for
conventional efficacy measures and for patient preference and satisfaction measures.

Sumatriptan versus zolmitriptan

An open-label, non-randomised, crossover study investigated patients’ preferences for oral
sumatriptan 50 mg versus oral zolmitriptan 2.5 mg tablets. [23] Patients treated three attacks
with each triptan (in any order) then completed a preference questionnaire. At the end of the
study, 42 patients (44%, CI 34–58%) preferred zolmitriptan, 27 patients (29%, CI 20–38)
preferred sumatriptan and 25 patients (27%, CI 18-36%) reported no preference. The
reasons given in the 69 patients who expressed a preference between the triptans were:
faster onset of action (73%), longer duration of effect (39%), fewer adverse events (35%) and
lower price (13%). Only one-quarter of the patients reported that sumatriptan and zolmitriptan
were equivalent. These results are similar to those from the open-label, multicentre study
conducted described earlier, [20] in which 45% of patients assessed zolmitriptan as superior
to sumatriptan and 36% assessed sumatriptan as superior to zolmitriptan.

Sumatriptan versus rizatriptan

A randomised, double-blind, triple-dummy, parallel group study compared rizatriptan tablets
5 mg and 10 mg, sumatriptan 100 mg and placebo in 1,268 patients treating a single
migraine attack. [24] Headache relief rates after rizatriptan 10 mg were reported to be
somewhat higher than those after sumatriptan. However, patient satisfaction data were also
collected, and showed no significant differences between the rizatriptan and sumatriptan
groups. [25]

Two studies have compared patient preference for sumatriptan conventional tablets with the
ODT formulation of rizatriptan. A multicentre, randomised, open-label, two-period crossover
study compared the proportion of patients who preferred rizatriptan ODT 10 mg to

72
sumatriptan 50 mg tablet. [12] Patients treated two migraine attacks, one each with
rizatriptan and sumatriptan. Significantly more patients preferred rizatriptan to sumatriptan at
the end of the study (57% versus 43%, p<0.01). A post hoc analysis of the data indicated
that patients tended to prefer the triptan that supplied the most rapid pain relief. [26] A
second randomised, open-label, crossover study assessed patient preference for rizatriptan
ODT 10 mg versus sumatriptan 50 mg conventional tablet to treat a single migraine attack.
[27] At the end of the study, significantly more patients preferred rizatriptan to sumatriptan
(64.3% versus 35.7%, p≤0.001). Faster headache relief was the most important reason given
for preference of both drugs (46.9% and 43.4% of patients preferring rizatriptan and
sumatriptan, respectively). Two hours after treatment of the attacks, significantly more
patients receiving rizatriptan than sumatriptan (73.3% versus 59.0%, p≤0.001) reported
satisfaction (completely, very or somewhat satisfied) with therapy and found the drug
convenient (very convenient, convenient or somewhat convenient) to take (87.2% versus
76.3%, p≤0.001). The crossover design of these two studies helps mitigate the asymmetry of
the comparison groups.

Sumatriptan versus eletriptan

A randomised, double-blind, parallel-group study compared the efficacy, safety and
tolerability of oral eletriptan 20 mg, 40 mg and 80 mg versus oral sumatriptan 100 mg and
placebo for a single migraine attack (n = 692). [28] Patients were asked at a follow-up visit to
rate the acceptability of their study medication compared to medications used previously.
Patients rated sumatriptan (64%) and all doses of eletriptan (64%, 74% and 84% for the 20
mg, 40 mg and 80 mg doses, respectively) more acceptable than placebo (32%), with the
highest acceptability rate reported for eletriptan 80 mg. Sumatriptan, but not eletriptan was
encapsulated for blinding purposes in the study, making the comparative groups asymmetric,
a potential bias that was maximized by the parallel group design. Encapsulation of
sumatriptan has been shown to negatively affect its pharmacokinetics and absorption. [8]

Sumatriptan versus almotriptan

A double-blind, multicentre, randomised, parallel-group study (n = 1,173) compared
treatment satisfaction, functional status and health-related quality of life (HRQOL) of patients
treated with oral almotriptan 12.5 mg or oral sumatriptan 50 mg for one migraine attack. [29]
The patients reported similar satisfaction with pain relief associated with the two drugs, but
were significantly less bothered with side effects from almotriptan than sumatriptan
(p=0.016). Improvements in functional status and HRQOL were similar in the two treatment
groups. Both almotriptan and sumatriptan were encapsulated for blinding purposes in the
study, thus no bias based on blinding was present.

Multiple comparisons between the triptans

A post hoc comparison was made of patients’ overall satisfaction with treatment from five
double-blind, placebo-controlled studies in which rizatriptan 10 mg conventional tablets were
compared with other oral triptans. [30] Three studies compared rizatriptan with sumatriptan
(rizatriptan 10 mg versus sumatriptan 100 mg in a parallel-group study, n = 916; rizatriptan
10 mg versus sumatriptan 50 mg in two crossover studies, n = 1,599). One study compared
rizatriptan 10 mg with naratriptan 2.5 mg (n = 502) and another compared rizatriptan 10 mg
with zolmitriptan 2.5 mg (n = 701), both being parallel-group studies. Patients reported their
satisfaction with treatment on a seven-point scale at 2 hours after treatment. Significantly
more patients receiving rizatriptan 10 mg than all the other triptans reported that they were
‘completely’ or ‘very’ satisfied: rizatriptan versus sumatriptan 100 mg (33% versus 26%,
p<0.05); rizatriptan versus sumatriptan 50 mg (40% versus 35%, p<0.05); rizatriptan versus
naratriptan 2.5 mg (33% versus 19%, p<0.01); and rizatriptan versus zolmitriptan 2.5 mg
(38% versus 30%, p<0.05).

A randomised, multicentre, open-label, five-way crossover study assessed patient preference

for sumatriptan 50 mg and 100 mg, naratriptan 2.5 mg, zolmitriptan 2.5 mg and rizatriptan 10

73
mg (n = 372). [11, 31] Patients were randomised to treat one migraine attack with each of the
five triptans in sequence, in a total of 119 possible treatment sequences. Patients assessed
which triptan they preferred at the end of the study. The results showed that sumatriptan 100
mg was preferred by 33% of patients, significantly higher than the random preference rate of
20% (p<0.001). Preference rates for sumatriptan 50 mg, naratriptan, rizatriptan, and
zolmitriptan were not significantly higher than the random preference rate. The patients’
primary reason for preferring a medication was ‘best relief of migraine pain’, and the
treatment that patients preferred corresponded to the medication that was most likely to
confer for them a pain-free response 2 hours postdose.

A second small study (n = 28) conducted in clinical practice compared patient preference to
sumatriptan 50 mg or 100 mg, naratriptan 2.5 mg or 5 mg and zolmitriptan 2.5 mg or 5 mg.
[32] Patients were randomised to treat two attacks with each of the triptans. At the end of the
study, 50% of patients preferred sumatriptan, 32% naratriptan and 18% zolmitriptan.

A retrospective audit of patient data from a secondary care headache clinic (n = 176)
investigated the pattern of preference for and switching between sumatriptan, naratriptan and
zolmitriptan in clinical practice. [33] Most patients (68%) had switched between triptans at
least once in the previous 2 years. No triptan showed a significantly higher level of
preference, although there were some gender differences. Women tended to prefer
zolmitriptan over the other two triptans and switched between triptans more often than men.
Most patients reporting migraine with aura used sumatriptan to treat their attacks.

In a retrospective review of 386 patients who used subcutaneous sumatriptan and were
switched to a different triptan or formulation, 19.5% returned to subcutaneous sumatriptan.
[34] For the other triptans/formulations, the percentages for returning were: sumatriptan 25
mg, 7.8%; sumatriptan 50 or 100 mg, 42.3%; sumatriptan nasal spray, 17.7%; zolmitriptan,
17.6%; rizatriptan, 16.5%; naratriptan: 9.4%. Of those who used more than three triptans or
formulations, the last triptan used was: sumatriptan, 29.5%; zolmitriptan, 31.8%; rizatriptan,
25.0%; naratriptan, 12.5%. Different formulations of sumatriptan were used by 129 subjects
(33.4%). Of the patients who used sumatriptan as the first triptan and switched to other
triptans, sumatriptan was also the last triptan used by 53.8% of them. This study involved
asymmetries of formulations in assessing patient preferences and reasons for switching
behaviours.

A Swedish study has investigated migraine patients’ preference for zolmitriptan 5 mg nasal
spray compared with that to oral triptans in a realistic clinical practice setting (n = 83). [35]
Patients, 96% of whom were currently using a triptan (usually oral), were invited to try
zolmitriptan nasal spray 5 mg for up to six migraine attacks, to see if efficacy could be
improved. Initial data indicated that 76% of patients wanted to continue to use zolmitriptan
nasal spray. The main reasons for this preference were a fast onset of action, a lack of
adverse events and only needing to take a single dose. The first reason may be intrinsic to a
nasal spray compared to a tablet; the other two reasons should not have been impacted by
asymmetry of compared formulations.

Studies comparing different formulations of the same triptan

Sumatriptan, zolmitriptan and rizatriptan are available in different formulations, and a small
number of studies have compared patient preference for different formulations or doses of
these drugs.

Sumatriptan
An open, multicentre, randomised, crossover study with an optional open, parallel-group
extension (n = 385) investigated the efficacy, safety and patient preference for oral
sumatriptan 100 mg and subcutaneous sumatriptan 6 mg formulations. [36] Patient

74
preference for the subcutaneous formulation more than doubled from the pre-treatment
phase to the end of the crossover period in those patients previously naïve to sumatriptan.
During the optional parallel-group phase of the study, 38% of patients chose to use both
sumatriptan formulations, treating some attacks with subcutaneous sumatriptan and some
with oral sumatriptan. The main reason for choosing subcutaneous sumatriptan was speed of
relief, while convenience was the major reason for choosing the tablet.

An open, randomised, three-attack crossover study compared patient opinions of oral

sumatriptan 100 mg with subcutaneous sumatriptan 6 mg (n = 124). [37] At the end of the
study, patient opinion was more often positive after subcutaneous sumatriptan than after oral
sumatriptan. Subcutaneous sumatriptan was significantly more effective than oral
sumatriptan, but more adverse events were reported following the subcutaneous formulation.

A telephone survey was conducted in 707 patients who had used sumatriptan tablets and/or
injection long-term for migraine in clinical practice. [38] Results showed that more patients
preferred the tablets over the injection, but that more patients reported that the injection was
the most effective formulation. The most frequently given reasons for the injection being
superior were efficacy and speed of action. The most frequently given reasons for the tablet
being superior were fewer side effects and lack of experience with other formulations. Most
patients (94%) reported that sumatriptan was superior to their previous non-triptan therapies.

An open, randomised, crossover study compared patient preference for sumatriptan 50 mg

tablets and sumatriptan 20 mg nasal spray. [39] Patients, who were naïve to both
formulations, preferred both formulations approximately equally (47% for tablets and 53% for
nasal spray). Patients preferred the nasal spray for its fast onset of action and the tablets for
their convenience.

Rizatriptan
Patients (n = 367) taking part in a clinical study of rizatriptan were allowed to continue open-
label treatment with both the film-coated tablet and ODT formulations for a 6-month period.
[40] At the end of the study, 51.2% preferred the ODT and 48.8% the film-coated tablet.
Although individual patients had strong reasons for preferring one formulation over the other,
no group preferences were detected for the individual formulations.

Comparing the different doses of oral sumatriptan

A multinational, randomised, double-blind, crossover, 8-week study was conducted to assess
patient dose preference, efficacy and tolerability for oral sumatriptan 25 mg, 50 mg and 100
mg in the acute treatment of migraine. [41] Patients (n = 257) were randomised to treat three
migraine attacks, using a different dose for each. At the end of the study, 34.6% of patients
preferred the 100 mg dose, 30.4% the 50 mg, 20.6% the 25 mg dose and 12.8% expressed
no preference. Efficacy and speed of action were the two main reasons given for preferring
the higher doses. However, adverse events were rarely given as a reason for preferring the
lower doses of sumatriptan. Although the 50 mg dose has been shown to have the optimal
benefit: risk ratio of the formulations, [42] some patients clearly preferred a higher dose.

Discussion

Patients’ assessments of their preference for, and satisfaction with, their migraine treatments
may be measures of clinical efficacy relevant to real-life clinical practice, taking into account
both efficacy and tolerability. We now have considerable clinical data on patient preference,
allowing the evaluation of the triptans and other acute migraine treatments.

All studies of preference [14–19, 22] and satisfaction [17, 19, 20, 21] for triptans compared
with patients’ usual non-triptan medications have demonstrated the superiority of the triptans.
Patients’ most commonly given reasons for preferring triptans were effective relief, speed of

75
relief, restored ability to function and fewer side effects. [16, 18, 19, 22] These are significant
results, as some controlled clinical trials have shown that sumatriptan was not superior to
rapid-release tolfenamic acid, [43] paracetamol/domperidone, [44] aspirin/metoclopramide,
[45, 46] isometheptene/paracetamol/dichlorphenazone, [47] and
paracetamol/aspirin/caffeine. [48] In contrast, a controlled clinical trial showed that oral
sumatriptan 100 mg was significantly superior to oral ergotamine plus caffeine. [49] These
results indicate that patient preference may be a more sensitive and valid measure of
efficacy and clinical utility than conventional clinical trial endpoints.

Relatively few clinical trials have directly compared patients’ preference for individual
triptans, and all included sumatriptan. [12, 20, 23, 27] Data from these four studies were
broadly similar, (Table 3) some patients preferring one triptan and some the other, even
though fewer patients preferred sumatriptan to the comparator triptan in all cases. The main
reason for preference was a faster onset of action. Other reasons given included a longer
duration of effect and fewer adverse events. Each of these reasons was given for all the
triptans. The data in these preference studies were broadly similar to those from randomised,
double-blind comparator studies between these triptans. [50, 51]

Five further studies compared patient preference between multiple triptans. Two studies
showed that more patients preferred oral sumatriptan 50 mg or 100 mg than other oral
triptans. [11, 31, 32] Two studies showed few differences between several oral triptans, [33,
34] and a further study showed that patients preferred zolmitriptan nasal spray over oral
triptans. [35] Different study designs and the patients’ initial triptan may have biased the
results from these studies.

Patient satisfaction data from double-blind, controlled clinical trials showed similar trends to
those reported above for patient preference. Patients were equally satisfied with the efficacy
of sumatriptan 100 mg and rizatriptan 10 mg, [24, 25] sumatriptan 100 mg and eletriptan 20
mg and 40 mg [28] and sumatriptan 50 mg and almotriptan 12.5 mg. [29] In a review of five
clinical trials, the proportions of patients preferring sumatriptan, zolmitriptan and rizatriptan
were not markedly different from each other. [30] Patients were also able to express
preferences between different formulations of the triptans, [36–38, 40] although these data
are biased due to the asymmetry of the treatment groups. Patients could even distinguish
between different doses of oral sumatriptan. [41]

There are several limitations to the available analyses of patient preference and satisfaction.
Many of the studies are published only as abstracts, and much of the desired data is not
recorded. As secondary endpoints, assessments of patient preference were not subjected to
rigorous statistical testing, as were the study primary endpoints. The preference and
satisfaction studies were open-label and bias could therefore occur. Patients may have had
previous access to one or more of the drugs being investigated. It is interesting that in all
cases, the triptan preferred most was that of the company sponsoring the study. In the
double-blind, controlled studies analysed, the patient preference/satisfaction endpoint was a
secondary or post hoc endpoint, and may therefore not have been powered appropriately.
Studies involving different formulations are biased by the asymmetries previously discussed
in this article. The solution is to conduct double-blind, controlled studies with patient
preference as the primary endpoint, and to investigate which set of triptans in which order,
scheme, dosage and formulation leads to the fastest effectiveness in a particular patient. In
such a design, non-responders are switched to another triptan, dosage or formulation to
reach (in theory) 100% effectiveness. Such a naturalistic design would best serve the needs
of the clinician who is interested in the best fit of a patient with a triptan.

Clinical implications
When given the opportunity, most migraine patients are able to distinguish between triptans
and non-triptan acute therapies for migraine, and between the individual oral triptans. This

76
sophisticated individual preference is not usually seen in controlled parallel design clinical
trials, where differences between the oral triptans, when statistically present, are small. [4]
Unfortunately for the physician, patients have very individual preferences for triptans that are
not predictable in advance. The problem therefore, is to investigate the fastest route to detect
the right match between patient and triptan (in terms of dose, order of switching and
formulation). Patients clearly prefer triptans to simple and combination analgesics and
ergotamine, and thus triptans should be a first-line alternative in most migraine patients. [52]
In assessing individual triptans in clinical practice, patients are looking for a therapy that
provides rapid and effective relief of the migraine. [18, 19, 22, 38] and are willing to switch
between triptans to achieve this goal. [33]

Patient preference is clearly a sensitive and valid overall measure of the clinical profile of
triptans, encompassing both efficacy and tolerability. In reviewing migraine patients, the
physician should elicit their preference for, and satisfaction with, their current medication
before making further treatment decisions. There is no need to change the patients’
medication if they are satisfied with their current medications and prefer them to those used
previously. When the patient’s medication is changed, at review the physician should ask
about the patient’s preference for the new medication. In addition, physicians should also
take into consideration patients’ preference for a specific delivery system. For patients with
attacks of varying severity and/or lifestyle needs, more than one formulation may be
appropriate.

In conclusion, patient preference is a sensitive and valid clinical trial endpoint and physicians
should consider using it when reviewing the efficacy of acute migraine medications.

Acknowledgements
The statistical analysis for the meta-analysis of sumatriptan versus normal acute medications
[14] was carried out by Leanne Rice of Statistics and Data Management Department, Glaxo
Laboratories UK Ltd, Stockley Park West, Uxbridge, UB11 1BT, UK.

References
1. Tfelt-Hansen P, De Vries P, Saxena PR. Triptans in migraine: a comparative review of
pharmacology, pharmacokinetics and efficacy. Drugs 2000;60:1259–87.
2. Fox AW. Comparative tolerability of oral 5-HT1B/1D agonists. Headache 2000;40:521–7.
3. Gawel MJ, Worthington I, Maggisano A. A systematic review of the use of triptans in
acute migraine. Can J Neurol Sci 2001;28:30–41.
4. Ferrari MD, Roon KI, Lipton RB et al. Oral triptans (serotonin 5-HT1B/1D agonists) in acute
migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358:1668–75.
5. Salonen R. Drug comparisons: why are they so difficult? Cephalalgia 2000;20 (Suppl
2):25–32.
6. Pilgrim AJ. Methodology of clinical trials of sumatriptan in migraine and cluster headache.
Eur Neurol 1991;31:295–9.
7. Cady RK, Sheftell F, Lipton RB et al. Early treatment with sumatriptan enhances pain-
free response: retrospective analysis from three clinical trials. Clin Ther 2000;22:1035–
48.
8. Fuseau E, Petricoul O, Sabin A et al. Effect of encapsulation on absorption of
sumatriptan tablets: data from healthy volunteers and patients during a migraine. Clin
Ther 2001;23:242–51.
9. Sheftell FD, Fox AW, Weeks RE et al. Differentiating the efficacy of 5HT 1B/D agonists.
Headache 2001;41:257–63.
10. Goadsby PJ. The scientific basis of medication choice in symptomatic migraine
treatment. Can J Neurol Sci 1999;26 (Suppl 3):20–6.
11. Dahlöf C. Assessing patient preference in migraine treatment. Cephalalgia 2001;21:791–
5.

77
12. Loder E, Brandes JL, Silberstein S et al. Preference comparison of rizatriptan ODT 10 mg
and sumatriptan 50 mg tablet in migraine. Headache 2001;41:745–53.
13. Tfelt-Hansen P, Block G, Dahlöf C et al. International Headache Society Clinical Trials
Subcommittee. Guidelines for controlled trials of drugs in migraine: second edition.
Cephalalgia 2000;20:765–86.
14. Rothwell K, Dowson AJ. Patients’ evaluation of acute migraine treatments. Poster
presentation at the 6th International Headache Research Seminar, Copenhagen,
Denmark, 1995.
15. Boureau F, Chazot G, Emile J et al. Comparison of subcutaneous sumatriptan with usual
acute treatments for migraine. French Sumatriptan Study Group. Eur Neurol
1995;35:264–9.
16. Kwong WJ, Chalal R, Putman G et al. Migraine patients prefer Imitrex tablets 50 mg to
their usual non-triptan prescription or over-the-counter therapy. Cephalalgia 2001;21:411
(Abstract).
17. Chalal R, Kwong WJ, Putnam G et al. Patient satisfaction with Imitrex tablets 50 mg
compared with their usual non-triptan prescription or over-the-counter therapy.
Cephalalgia 2001;21:411 (Abstract).
18. Rederich GJ, Newkirk TA, Killilea L et al. Migraineurs’ preference for sumatriptan over
non-triptan therapy in a managed care population. Cephalalgia 2001;21:412 (Abstract).
19. Powers C, Szeto S, Pangtay D et al. Evaluation of migraineurs’ preferences for
naratriptan over conventional first-line agents. Arch Fam Med 2000;9:753–8.
20. Sturzenegger M, Daems K, Billeter M. Zolmitriptan in treatment of migraine attack: the
ARES study. Schweiz Med Wochenschr 2000;130:1984–93 (in German).
21. Solomon S, Frishberg B, Hu XH et al. Migraine treatment outcomes with rizatriptan in
triptan-naïve patients: a naturalistic study. Clin Ther 2001;23:886–90.
22. Robbins L. Triptans vs analgesics: patient preference. Cephalalgia 2001;21:406
(Abstract).
23. Pascual J, Munoz R, Leira R. An open preference study with sumatriptan 50 mg and
zolmitriptan 2.5 mg in 100 migraine patients. Cephalalgia 2001;21:680–4.
24. Tfelt-Hansen P, Teall J, Rodriguez F et al. Oral rizatriptan versus oral sumatriptan: a
direct comparative study in the acute treatment of migraine. Rizatriptan 030 Study Group.
Headache 1998;38:748–55.
25. Tfelt-Hansen P, Goldstein J, Malbecq W et al. Comparison of rizatriptan and sumatriptan.
Headache 1999;39:340–1 (Letter).
26. Bohidar N, Loder E, Guerra F et al. Relationship between patient preference for either
rizatriptan orally disintegrating tablet (ODT) 10 mg or sumatriptan tablet 50 mg and speed
of pain relief. Cephalalgia 2001;21:422 (Abstract).
27. Pascual J, Bussone G, Hernandez JF et al. Comparison of preference for rizatriptan 10
mg wafer versus sumatriptan 50 mg tablet in migraine. Eur Neurol 2001;45:275–83.
28. Goadsby PJ, Ferrari MD, Olesen J et al. Eletriptan in acute migraine: a double-blind,
placebo-controlled comparison to sumatriptan. Eletriptan Steering Committee. Neurology
2000;54:156–63.
29. Colman SS, Brod MI, Krishnamurthy A et al. Treatment satisfaction, functional status,
and health-related quality of life of migraine patients treated with almotriptan or
sumatriptan. Clin Ther 2001;23:127–45.
30. Gerth WC, McCarroll KA, Santanello NC et al. Patient satisfaction with rizatriptan versus
other triptans: direct head-to-head comparisons. Int J Clin Pract 2001;55:552–6.
31. Dahlöf C, Jones M, Davis K et al. A comparison of preference for and efficacy of tablet
formulations of sumatriptan (50 mg and 100 mg), naratriptan (2.5 mg), rizatriptan (10 mg)
and zolmitriptan 2.5 mg) in the acute treatment of migraine. J Headache Pain
2004;5:115–22.
32. Dahlöf C. How to assess patient preference of migraine treatments. Cephalalgia 1999;19
(Suppl 24):2–5.
33. Sandor P, Agosti R. Which triptan would you like? A retrospective triptan preference
study. Cephalalgia 2001;21:408 (Abstract).

78
34. Sheftell FD, Feleppa M, Tepper SJ et al. Patterns of use of triptans and reasons for
switching them in a tertiary care migraine population. Headache 2004;44:661–8.
35. Dahlöf C. Clinical applications of new therapeutic deliveries in migraine. Neurology
2003;61 (8 Suppl 4):S31–4.
36. Gruffydd-Jones K, Hood CA, Price DB. A within-patient comparison of subcutaneous and
oral sumatriptan in the acute treatment of migraine. Cephalalgia 1997;17:31–6.
37. Carpay HA, Matthijsse P, Steinbuch M et al. Oral and subcutaneous sumatriptan in the
acute treatment of migraine: an open randomized cross-over study. Cephalalgia
1997;17:591–5.
38. Dahlöf CG, Saiers J. Sumatriptan injection and tablets in clinical practice: results of a
survey of 707 migraineurs. Headache 1998;38:756–63.
39. Dahlöf CGH. Sumatriptan : pharmacological basis and clinical results. Curr Med Res
Opin 2001;17 (Suppl 1):s35–45.
40. Adelman JU, Mannix LK, Von Seggern RL. Rizatriptan tablet versus wafer: patient
preference. Headache 2000;40: 371–2.
41. Salonen R, Ashford EA, Gibbs M et al. Patient preference for oral sumatriptan 25 mg, 50
mg, or 100 mg in the acute treatment of migraine: a double-blind, randomized, crossover
study. Sumatriptan Tablets S2CM11 Study Group. Int J Clin Pract Suppl 1999;105:16–
24.
42. Pfaffenrath V, Cunin G, Sjonell G et al. Efficacy and safety of sumatriptan tablets (25 mg,
50 mg, 100 mg) in the acute treatment of migraine: defining the optimum doses of oral
sumatriptan. Headache 1998;38:184–90.
43. Myllylä VV, Havanka H, Herrala L et al. Tolfenamic acid rapid release versus sumatriptan
in the acute treatment of migraine: comparable effect in a double-blind, randomized,
controlled, parallel-group study. Headache 1998;38:201–7.
44. Dowson A, Ball K, Haworth D. Comparison of a fixed combination of domperidone and
paracetamol (Domperamol) with sumatriptan 50 mg in moderate to severe migraine: a
randomised UK primary care study. Curr Med Res Opin 2000;16:190–7.
45. Oral sumatriptan and Aspirin-plus-Metoclopramide Comparative Study Group. A study to
compare oral sumatriptan with oral aspirin plus oral metoclopramide in the acute
treatment of migraine. Eur Neurol 1992;32:177–84.
46. Tfelt-Hansen P, Henry P, Mulder LJ et al. The effectiveness of combined oral lysine
acetylsalicylate and metoclopramide compared with oral sumatriptan for migraine. Lancet
1995;346:923–6.
47. Freitag FG, Cady R, DiSerio F et al. Comparative study of a combination of
isometheptene mucate, dichloralphenazone with acetaminophen and sumatriptan
succinate in the treatment of migraine. Headache 2001;41:391–8.
48. Goldstein J, Silberstein SD, Elkind AH et al. A placebo-controlled comparison of the
combination of acetaminophen, aspirin, and caffeine with sumatriptan succinate in the
early treatment of migraine: Results from the Asset trial. Neurology 2003;60 (Suppl 1):
A171. (Abstract)
49. Multinational Oral Sumatriptan and Cafergot Comparative Study Group. A randomized,
double-blind comparison of sumatriptan and Cafergot in the acute treatment of migraine.
Eur Neurol 1991;31:314–22.
50. Gruffydd-Jones K, Kies B, Middleton A et al. Zolmitriptan versus sumatriptan for the acute
oral treatment of migraine: a randomized, double-blind, international study. Eur J Neurol
2001;8:237–45.
51. Adelman JU, Lipton RB, Ferrari MD et al. Comparison of rizatriptan and other triptans on
stringent measures of efficacy. Neurology 2001;57:1377–83.
52. Silberstein SD, for the US Headache Consortium. Practice parameter: evidence-based
guidelines for migraine headache (an evidence-based review). Report of the Quality
Standards Subcommittee of the American Academy of Neurology. Neurology
2000;55:754–62.

79
3.3

A general practice study on the prevalence of headache,

depression and bodily pain, and the disability associated with
headaches occurring inside and outside the menstrual period in
migraine sufferers*

Abstract

Objectives: This study investigated the prevalence of headache, depression and bodily pain
in women attending a UK general practice. The disability associated with migraine and other
headache attacks occurring during and outside the menstrual period was assessed in those
women with migraine.
Methods: 1,434 of 3,470 female patients (41.3%) aged 14–50 years registered at a UK
general practice completed two questionnaires. The first questionnaire assessed the point
prevalence of headache and bodily pain, and the 5-year prevalence of depression in the total
population. The second questionnaire assessed the disability of all headaches over a 2-
month period (to capture a complete menstrual cycle) for patients reporting migraine who
were still menstruating. Disability was assessed as the time lost and time spent at less than
50% productivity in normal activities due to headache, and analyzed as rank sums using the
Mann-Whitney U Test.
Results: The first part of the study showed that the prevalence of headache (66.1%),
depression (55.4%) and bodily pain (40.6%) were high in this population of women. In the
second part of the study, 30 migraine patients who were still menstruating (11.1% of those
eligible) reported 89 migraine and 114 non-migraine headache episodes. For migraine, the
rank order of time at less than 50% productivity was significantly greater for attacks taking
place inside the menstrual period than for those occurring outside the menstrual period
(p=0.01). For non-migraine headaches, the time lost appeared to be numerically greater for
attacks taking place outside the menstrual period than for those occurring inside the
menstrual period, and the comparison approached significance (p=0.06).
Conclusions: The patients reported a high prevalence of headache, depression and bodily
pain. For migraine sufferers, migraine attacks that took place during the menstrual period
tended to be more disabling than those taking place outside the menstrual period, but the
opposite was true for non-migraine headache.

* Edited from the full published paper: Dowson AJ, Kilminster SG, Salt R, Clark M, Bundy
MJ. Disability associated with headaches occurring inside and outside the menstrual period
in those with migraine: A general practice study. Headache 2005;45:274–82. Additional
statistical analyses were conducted by Professor Jan Passchier.

80
Introduction

Headache and depression are two of the commonest conditions that the primary care
physician has to deal with. Most people suffer from headaches, usually tension-type
headaches (TTH), migraine or chronic daily headache (CDH). TTH affects over 50% of the
general population,1 migraine about 12%2 and CDH about 5%.3 Major depression is reported
to affect up to 12% of the general population.4 Migraine,5 CDH6 and depression4 are all
disabling conditions to the patient and cause major societal burdens due to the associated
high direct medical costs of care and indirect costs of lost work productivity.7,8

Although definitive criteria are not available, menstrually-associated migraine attacks tend to
be defined today as those that begin during the time period from the day prior to the start of
menstruation and Day 2 of menstruation.9 Such attacks are common, and more than 50% of
women with migraine report an association between migraine and menstruation, 10 although
in most cases they also have migraine attacks outside the menstrual period. A population-
based study conducted in the Netherlands provided a prevalence rate of 3% for menstrual
migraine.11 Studies indicate that menstrually-associated migraine attacks are the result of
estrogen withdrawal in the late luteal phase of the normal menstrual cycle, but other factors
such as prostaglandin release have also been implicated.9 Menstrually-associated migraine
attacks are treated with standard acute therapies and with specific prophylactic treatments
such as estrogen supplements.9 It used to be widely considered that menstrually-associated
migraine attacks were more severe and less responsive to treatment than non-menstrually-
associated attacks,9,12 but these suppositions have now been disputed. A population-based
epidemiologic study indicated that menstrually-associated migraine attacks were slightly
more painful, but not more disabling, than attacks occurring at other times in the cycle.13
Similar results have been reported for TTH attacks occurring inside and outside the
menstrual period.14 A small study with sumatriptan indicated that the drug was less effective
for menstrually-associated than for non-menstrually-associated migraine attacks (56%
versus 81% of patients reporting relief of their migraine headache 4 hours after treatment,
when treated during and outside the menstrual period, respectively).15 However, more recent
studies have indicated that triptan drug16 and nondrug treatments17 are equally effective in
treating migraine attacks occurring inside and outside the menstrual period. In contrast,
some recent studies indicate that migraine attacks occurring during the menstrual period
tend to be more frequent, more severe, of longer duration and more resistant to treatment
than those occurring at other times of the month.11,18 In clinical practice, menstrually-
associated migraine attacks tend to be treated in the same way as non-menstrually-
associated attacks.

This study was set up to investigate two questions related to women’s health issues:
1. What is the prevalence and severity of headache, depression and bodily pain?
2. Are migraine and other headache attacks occurring during the menstrual period more
or less disabling than those occurring outside it in patients with migraine?

The study extends previous research13 into primary care clinical practice. We recorded
headache characteristics and the disability caused by migraine and other headache attacks
taking place inside and outside the menstrual period. Disability was assessed by recording
the time lost from normal activities and the time spent at less than 50% of normal capabilities
due to headache. Recent research has shown that these assessments form a reliable and
accurate measure of headache-related disability, when incorporated into the Migraine
Disability Assessment (MIDAS) Questionnaire.19

Patients and methods

Patients
All non-hysterectomized women aged 14–50 years at a health centre in Cranleigh, Surrey,
UK, were invited to take part in a two-part study. The study was approved by the local

81
investigational review board, the South West Surrey Local Ethics Committee, and all patients
who took part in the study provided their written informed consent to participate.

Study design
Part 1
All patients were asked to complete and return by mail a questionnaire assessing their
headaches, depression and bodily pain severity (Appendix 1). The point prevalence and
severity of headache and bodily pain and the 5-year prevalence and severity of depression
were assessed and patients with frequent headaches, and disabling and non-disabling
intermittent headaches identified using the questionnaire.

Part 2
Patients who had disabling intermittent headaches and were still having menstrual periods
were asked to complete another questionnaire assessing the disability of all headaches that
occurred inside and outside the menstrual window. Patients were monitored over a 2-month
period, to ensure that a complete menstrual cycle was covered for all patients.

Questionnaires
Headache, depression and bodily pain questionnaire
The questionnaire was developed by the study investigators (Drs Dowson, Kilminster, Clark
and Bundy). Three questions (Questions 1–3) assessed headache, five assessed depression
(Questions 5–9) and four assessed bodily pain severity (Questions 10–13).

The answers to Questions 2 (‘How often do they [headaches] occur?’) and 3 (‘Do the
headaches interfere with your ability to perform normal daily activities?’) were used to
categorize the patients’ headache. Patients who answered ‘a’ to Question 2, with headaches
on most days in the month, were categorized as having frequent headaches. Those who
answered ‘b’, ‘c’ or ‘d’ to Question 2, with up to four or more headaches per month, were
categorized with non-disabling intermittent headache if the headaches did not interfere with
their ability to perform normal activities (Question 3a) and with disabling intermittent
headache if the headaches did interfere with these activities (Question 3b).

The patients’ degree of depression was assessed arbitrarily from Questions 5, 6 and 7,
covering sleep disturbance, mood disturbance and suicidal ideation. Patients who answered
‘no’ to all questions were classified as not being depressed. Those who answered ‘yes’ to
one, two and three of these questions were classified as having mild, moderate and severe
depression, respectively. Depression is therefore here conceived as a continuous variable. In
addition, patients who answered ‘yes’ to Question 9 (defined as those who had received
treatment from a doctor for sleep disturbance, mood disturbance or suicidal ideation) were
classified as having at least moderate depression, depending on their answers to Questions
5, 6 and 7 (i.e. none and one ‘yes’ answer = moderate depression; two and three ‘yes’
answers = severe depression).

Pain severity in any part of the body was assessed from Questions 10, 12 and 13. Patients
who responded ‘yes’ to Question 10 but ‘no’ to Questions 12 and 13 were assessed as
having mild pain. Those who responded ‘yes’ to Questions 10 and 12 but ‘no’ to Question 13
were assessed as having moderate pain. Those who responded ‘yes’ to Questions 10, 12
and 13 were assessed as having severe pain. Question 11 was not counted, but provided
information designed to be useful to the physician.

Menstrual headache questionnaire

Patients with disabling intermittent headache in Part 1 of the study who were still
menstruating were eligible to complete Part 2. Patients reported all their headaches over a
2-month period, using the menstrual headache questionnaire that was mailed to them
(Appendix 2). The study investigators developed the menstrual headache questionnaire,

82
based on the International Headache Society (IHS) diagnostic criteria pertaining at the
time.18 Patients prospectively recorded their menstrual status, time of onset and resolution of
the headaches and other symptoms (self-selected), headache severity, presence of non-
headache symptoms, time lost from normal activities and the further time spent at less than
50% of normal capabilities, use of contraceptives and other drugs and any present pre-
menstrual symptoms.

Headaches were diagnosed by the study physicians as migraine or non-migraine using the
information recorded by the patients in the questionnaire, using their clinical judgment in
conjunction with reference to the relevant IHS diagnostic criteria.18 Headaches were further
subdivided into those that started inside and outside the menstrual period. Data were
analyzed as the number of headache episodes, the time lost from normal activity (total time
and mean time per attack) and the additional time at less than 50% of normal capability (total
time and mean time per attack) for migraine and non-migraine headaches.

Endpoints and analyses

Part 1
The primary endpoint was the prevalence and severity of headaches, depression and bodily
pain. Results were recorded as descriptive statistics only; the number and percentage of
patients with total, mild, moderate or severe symptoms, and subdivided by the age of the
patient. The relationship between age and headache prevalence was analyzed using Chi-
square tests for independent samples. Similar analyses for depression and bodily pain were
not carried out given the small differences between the data samples.

Part 2
The primary endpoint was the disability experienced by the patients, assessed as time lost
from daily activities. The other components of the questionnaire were recorded, but data are
not presented here. Non-parametric rank testing with the Mann-Whitney U Test for related
samples was used to analyze the time lost and time at less than 50% of normal capacity due
to the migraine and non-migraine headaches, comparing attacks from inside and outside the
menstrual period. The level of significance was set at p=0.05.

Results

Patient disposition
Of the 3,470 female patients in the practice who were sent the initial questionnaire, 1,434
(41.3%) completed and returned it (Table 1). Of these, all but five patients were aged
between 17 and 54 years. Approximately equal proportions of patients aged 17–24, 25–34,
35–44 and 45–54 years completed the questionnaire. The proportions of patients in these
age groups were similar to those in the total population in the practice.

Table 1. Patient demography. Number (and percentage) of women in the general practice
who took part in the study.

Number of patients (%)

14–16y 17–24y 25–34y 35–44y 45–54y Total
Patients taking part 5 234 305 498 392 1434
in study (0.3) (16.3) (21.3) (34.7) (27.3) (100)
Patients in the 79 609 744 979 1059 3470
practice (2.3) (17.6) (21.7) (28.2) (30.5) (100)

Prevalence of headache, depression and bodily pain

Headache
Overall, 948 patients (66.1%) completing the initial questionnaire reported current headache
(Table 2). A total of 62 patients (4.3%) reported headache on most days and were

83
categorized as having frequent headache. Two hundred and seventy one patients (18.9%)
had headache episodes up to more than four times a month that interfered with their ability to
perform daily activities, and were categorized with disabling intermittent headache. A total of
615 patients (42.9%) had headache episodes up to more than four times a month that did not
interfere with their ability to perform daily activities, and were categorized with non-disabling
intermittent headache. Similar proportions of patients aged 17–24, 25–34, 35–44 and 45–54
years reported any headache, and disabling and non-disabling intermittent headaches.
However, frequent headaches were reported significantly more often by patients aged 17–24
years than by older patients (Chi-square (3) = 22.9, p < 0.001).

Table 2. Headache. Number (and percentage) of women in the general practice who
reported any headache, frequent headaches*, disabling intermittent headaches** and non-
disabling intermittent headaches***.

Number of patients (%)

17–24y 25–34y 35–44y 45–54y Total
Patients reporting any headache 158 189 338 263 948
(67.5) (62.0) (67.9) (67.1) (66.1)
Patients reporting frequent 24 11 15 12 62
headaches* (10.3) (3.6) (3.0) (3.1) (4.3)
Patients reporting disabling 38 60 85 88 271
intermittent headaches** (16.2) (19.7) (17.1) (22.4) (18.9)
Patients reporting non-disabling 96 118 238 163 615
intermittent headaches*** (41.0) (38.7) (47.8) (41.6) (42.9)
Patients taking part in the study 234 305 498 392 1434
*Putatively diagnosed as CDH.21
**Putatively diagnosed as migraine.21,23
***Putatively diagnosed as TTH.21,23

Depression
Overall, 794 patients (55.4%) completing the initial questionnaire reported the defined
depression symptoms in the previous 5 years. Similar proportions of patients aged 17–24,
25–34, 35–44 and 45–54 years reported depression (Table 3). Approximately 20% of all
patients reported mild or moderate depression, with 15.3% reporting severe symptoms. The
distribution of mild, moderate and severe depression was similar among patients aged 17–
24, 25–34, 35–44 and 45–54 years.

Table 3. Depression. Number (and percentage) of women in the general practice who
reported mild, moderate and severe depression.

Number of patients (%)

17–24y 25–34y 35–44y 45–54y Total
Patients reporting depression 140 178 265 210 794
(mild, moderate or severe) (59.8) (58.4) (53.2) (53.6) (55.4)
Patients reporting mild 51 72 101 67 292
depression (21.8) (23.6) (20.3) (17.1) (20.4)
Patients reporting moderate 52 60 92 78 282
depression (22.2) (19.7) (18.5) (19.9) (19.7)
Patients reporting severe 37 46 72 65 220
depression (15.8) (15.1) (14.5) (16.6) (15.3)
Patients taking part in the study 234 305 498 392 1434

Bodily pain severity

Overall, 40.6% of patients completing the initial questionnaire reported current pain
symptoms in any part of the body, with the incidence generally increasing with increasing
age (Table 4). However, more patients aged 17–24 years reported bodily pain than those

84
aged 25–34 years, with the excess being accounted for by reports of mild pain. Thirteen
percent of all patients reported mild pain, 5.9% moderate pain and 21.7% severe pain. While
the incidence of mild and moderate bodily pain was generally similar among the age groups,
the incidence of severe pain increased with increasing age, from 15.8% for those aged 17–
24 years to 28.6% for those aged 45–54 years.

Table 4. Bodily pain severity. Number (and percentage) of women in the general practice
who reported mild, moderate and severe pain in any part of the body.

Number of patients (%)

17–24y 25–34y 35–44y 45–54y Total
Patients reporting pain (mild, 92 94 207 187 582
moderate or severe) (39.3) (30.8) (41.6) (47.7) (40.6)
Patients reporting mild pain 42 31 62 49 186
(17.9) (10.2) (12.4) (12.5) (13.0)
Patients reporting moderate pain 13 17 29 26 85
(5.6) (5.6) (5.8) (6.6) (5.9)
Patients reporting severe pain 37 46 116 112 311
(15.8) (15.1) (23.3) (28.6) (21.7)
Patients taking part in the study 234 305 498 392 1434

Disability of headaches occurring inside and outside menstruation

Thirty of the 271 patients (11.1%) with disabling intermittent headache in the first part of the
study completed the menstrual headache questionnaire. They detailed a total of 203
headaches that started inside and outside their menstrual periods during the 2-month study
period. Patients reported slightly fewer migraine attacks (89) than non-migraine headaches
(114). Migraine attacks were approximately equally frequent inside (47 attacks) and outside
(42 attacks) the menstrual period. However, non-migraine headache was reported about
three times more often outside the menstrual period (85 attacks) than inside it (29 attacks).

Meaningful mean or median values could not be calculated for the time lost and time at less
than 50% productivity data, due to the data being not normally distributed. Floor effects were
observed as many values were zero. For migraine, the rank order of time lost was
numerically but non-significantly greater for attacks taking place inside the menstrual period
than for those occurring outside the menstrual period (1,344.5 versus 1070.5, p=0.23, Figure
1). The rank order of time at less than 50% productivity was significantly greater for attacks
taking place inside the menstrual period than for those occurring outside the menstrual
period (1,453.6 versus 961.5, p=0.01, Figure 1). For non-migraine headaches, the rank order
of time lost was marginally significantly greater for attacks taking place outside the menstrual
period than for those occurring inside the menstrual period (5,009.0 versus 1,432.0, p=0.06,
Figure 2). The rank order of time at less than 50% productivity was numerically but not
significantly greater for attacks taking place outside the menstrual period than for those
occurring inside the menstrual period (4,902.5 versus 1,538.5, p=0.37, Figure 2). The rank
order of data for migraine headaches during menstruation was similar to that of non-migraine
headache, for both the time lost and time at <50% productivity analyses (Figures 1 and 2).

85
Figure 1. Time lost and time at less than 50% productivity per attack for migraine attacks
occurring inside and outside the menstrual period for patients with migraine (n = 30): analysis
of rank sums.

p=0.01
p=0.23
1600
1453.5
1344.5
1400
1200 1070.5
Rank sums

1000 961.5
Menstrual
800
Nonmenstrual
600
400
200
0
Time lost Time at <50%

Figure 2. Time lost and time at less than 50% productivity per attack for non-migraine
headache attacks occurring inside and outside the menstrual period for patients with
migraine (n = 30): analysis of rank sums.

p=0.37
6000
5009.0 4902.5
5000

4000
Rank sums

Menstrual
3000
p=0.06 Nonmenstrual
2000 1538.5
1432.0

1000

0
Time lost Time at <50%

Discussion

This study provides a snapshot of headache, depression and bodily pain in women attending
a primary care clinical practice. While the data are preliminary, they provide an insight into
clinical issues important to primary care, and suggest future avenues of research. The
patients recruited to the study came from a single UK primary care practice, with a responder
rate of 41.3% completing the initial questionnaire. This is a relatively high response rate;
surveys conducted with the general public more typically elicit a response rate of about 10%.
Respondents were approximately equally distributed among the different age groups.

86
The patients who completed the initial questionnaire reported high levels of headache,
depression and bodily pain. About two-thirds of patients reported any headache, with 4.3%
being categorized with frequent headaches, 18.9% with disabling intermittent headache and
42.9% with non-disabling intermittent headache. While no formal diagnoses were conducted,
there are good reasons for categorizing patients with frequent headaches as having CDH,
those with disabling intermittent headaches as having migraine and those with non-disabling
intermittent headaches as having TTH. CDH is often defined as daily, or near-daily
headache.21 Results from the Spectrum and Landmark Studies indicate that a default
diagnosis of migraine can be given to patients with disabling episodic headaches, and non-
disabling episodic headaches are usually TTH.22,23 Support for this comes from the relative
prevalence of putative CDH, migraine and TTH in this study, which were similar to those
reported for women in previous population-based studies (1-year prevalence: TTH = 86%;1
migraine = 18%;2 CDH approximately 4–5%3), and were generally consistent across all age
groups.

Over 50% of the patients reported symptoms indicative of depression, with 15.3% reporting
severe depression. The prevalence of depression was remarkably similar for all age groups
and severity levels, but was higher than has been previously reported for the general
population (about 12%4). The levels of reported bodily pain were also high, with 40.6% of
patients reporting pain and 21.7% reporting severe pain. Again, all age groups reported a
high prevalence of bodily pain, but reporting of pain overall and severe pain increased with
increasing age of the patient. Overall, these results indicate that headache, depression and
bodily pain are common in the general adult female population.

A low proportion of patients (11.1%) identified with migraine (disabling intermittent headache)
and still having periods responded to the menstrual headache questionnaire. The reasons for
this relatively poor response are not clear, although it is well known that less than half of
migraine sufferers consult with their primary care physicians for care.19 In addition, patients
find it difficult to complete a diary over a period longer than a few weeks.24 The total number
of headache episodes reported was also modest (89 migraine and 114 non-migraine
headache episodes). However, there were some strong trends in the results from the
questionnaire that have implications for clinical practice.

Migraine headaches were relatively more frequent inside than outside the migraine attack, as
shown by the roughly equal numbers of attacks in the approximately 1 week of menstruation
and the 3 weeks outside it. In contrast, non-migraine headaches were approximately equally
distributed inside and outside the menstrual period.

Migraine attacks that occurred inside the menstrual period scored numerically higher than
those occurring outside the menstrual period in terms of rank sums of time lost and time at
less than 50% productivity. These results are in line with a recent study indicating that
menstrually-associated migraine attacks were slightly more painful than non-menstrually-
associated ones.13 However, the differences between menstrually-associated and non-
menstrually associated migraine attacks was not high in the present study, and significance
was only reported for the analysis of time at less than 50% productivity. Results for non-
migraine headache were the opposite of those for migraine, with those occurring outside the
menstrual period scoring numerically, but not significantly, higher than those occurring inside
the menstrual period in terms of the rank sums of time lost and time at less than 50%
productivity. Recent data from the Landmark and Spectrum studies indicate that all
headaches experienced by migraine sufferers with disabling headaches are sensitive to
triptans and may be part of the overall migraine process.23,25 This raises the possibility that
both the migraine and non-migraine headaches analyzed in this part of the study were
essentially migrainous in character, at least in this group of patients with disabling headache.

87
Assessments of disability are important measures, and not only in terms of the response to
headache treatments. They are key measures in aiding communication between patients and
physicians,19 assessing headache patients’ illness severity and treatment needs,17 and
patients’ subjective perceptions of migraine and CDH management.26

Study limitations and suggestions for future research

Part 1
The questionnaires used in the study were not validated tools, but were devised by the
investigators to be simple to use and potentially applicable to primary care clinical practice.
The questionnaires were not tested for reliability or validity, and additional studies would be
required for this. Headache subtypes could be diagnosed from the initial questionnaire, as
described previously in this paper. While the diagnoses were not tested for accuracy, they
follow criteria developed by the IHS,20,21 and are therefore likely to be valid. However, simple
diagnostic screening questionnaires for headache have recently been developed and
validated,27,28 and may therefore be more preferable to use. The initial questionnaire also did
not provide a clear diagnosis for depression, but simply used clinical markers to guide the
physician. Again, it would probably be preferable to use a validated depression questionnaire
in future studies.

The prevalence study design was perhaps not optimal. Additional items that should have
been taken into account included defining the sample in relationship to the population under
study, random sampling, defining an acceptable response rate (>50%), and conducting a
non-response analysis. Despite this, the prevalence of the headache subtypes was close to
expected values, although depression was probably over-estimated. There was the
possibility of the responders being a self-selecting population with significant problems.

Part 2
The menstrual headache questionnaire was based on the IHS diagnostic criteria20 and items
contained in the MIDAS questionnaire.19 Thus, while this questionnaire was also not
validated, it contained items with proven reliability and validity. However, the questionnaire
should be tested for these properties if it is used again.

Some methodological issues were raised by the analyses that have implications for the
design of future studies. The data were confounded by the fact that some of the events were
repeatedly measured in individual patients, and that some patients only had one type of
headache. A much larger sample of patients would be necessary to allow both within and
between patient group comparisons. In the present study, data were analyzed only between
the groups for 30 patients experiencing 203 headache attacks. The results should therefore
be taken with caution. However, non-parametric testing with the Mann-Whitney U Test
showed one significant and one marginally significant result from only four analyses, which
indicates that the results were unlikely to have arisen by chance. Analysis of means or
medians was not a useful measure in the present study because the data were not normally
distributed (in fact they were seriously skewed). Many headache attacks scored zero on the
two analyses, leading to floor effects being observed. Such floor effects have been reported
with use of the MIDAS Questionnaire, on which the analyses were based.29 Future studies
may benefit from using assessments of headache impact which provide a normal distribution
of results, such as the Headache Impact Test30 or the Short Pain Inventory.31

Studies of headache in general practice are urgently required, as clinical experience does
not always follow data from controlled clinical trials.32 We encourage the use of validated
instruments in these studies, which have become available since the inception of the present
study. However, despite the methodological deficiencies in the present study, the clinical
importance of headache, depression and bodily pain were emphasized, and information was
gleaned on the disability associated with migraine and non-migraine headaches inside and
outside the menstrual period. We look forward to further studies on this important topic.

88
Conclusions
This study showed that headache, depression and bodily pain were common in women aged
14–50 years registered at a UK general practice. For migraine sufferers, migraine attacks
that took place during the menstrual period tended to be more disabling than those taking
place outside the menstrual period, but the opposite was true for non-migraine headache.
These results provide a snapshot of headache, depression and bodily pain in a subset of
women attending a UK primary care practice.

Acknowledgements

The authors wish to thank the staff of Cranleigh Health Centre for their assistance during
this study, and the patients who took part. The study was funded by a grant from Merck,
Sharp and Dohme Research Limited.

References
1. Rasmussen BK, Jensen R, Schroll M et al. Epidemiology of headache in a general
population--a prevalence study. J Clin Epidemiol. 1991;44:1147–57.
2. Stewart WF, Lipton RB, Celentano DD et al. Prevalence of migraine headache in the
United States: Relation to age, income, race and other sociodemographic factors.
JAMA 1992;267:64–9.
3. Silberstein SD, Lipton RB. Chronic daily headache. Curr Opin Neurol 2000;13:277–
83.
4. Smith AL, Weissmann MM. Depression. In: Paykel ES ed. Handbook of Affective
Disorders. London: Churchill Livingstone 1992;p111–29.
5. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
6. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine,
and cluster headache with topiramate. Headache 2002;42:796–803.
7. Greenberg PE, Stiglin LE, Finkelstein SN et al. The economic burden of depression in
1990. J Clin Psychiatry 1993;54:405–18.
8. Ferrari MD. The economic burden of migraine to society. Pharmacoeconomics
1998;13:667–76.
9. MacGregor EA. Gynaecological aspects of migraine. Rev Contemp Pharmacother
2000;11:75–90.
10. MacGregor EA, Igarashi H, Wilkinson M. Headaches and hormones: subjective
versus objective assessment. Headache Q 1997;8:126–36.
11. Couturier EG, Bomhof MA, Neven AK et al. Menstrual migraine in a representative
Dutch population sample: prevalence, disability and treatment. Cephalalgia
2003;23:302–8.
12. Solbach P, Sargent J, Coyne L. Menstrual migraine headache: results of a controlled,
experimental, outcome study of non-drug treatments. Headache 1984;24:75–8.
13. Stewart WF, Lipton RB, Chee E et al. Menstrual cycle and headache in a population
sample of migraineurs. Neurology 2000;55:1517–23.
14. Kennan PA, Lindamer LA. Non-migraine headache across the menstrual cycle in
women with and without premenstrual syndrome. Cephalalgia 1992;12:356–9.
15. Gross MLP, Barrie M, Bates D et al. The efficacy of sumatriptan in menstrual
migraine. Eur J Neurol 1995;2:144–5.
16. Silberstein SD, Massiou H, Le Jeunne C et al. Rizatriptan in the treatment of
menstrual migraine. Obstet Gynecol 2000;96:237–42.
17. Gauthier JG, Fournier AL, Roberge C. The differential effects of biofeedback in the
treatment of menstrual and nonmenstrual migraine. Headache 1991;31:82–90.
18. MacGregor EA, Hackshaw A. Prevalence of migraine on each day of the natural
menstrual cycle. Neurology 2004;63:351–3.

89
19. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
20. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8 (Suppl 7):1–92.
21. Headache Classification Subcommittee of the International Headache Society. The
international classification of headache disorders. Cephalalgia 2004;24 (Suppl 1):1–
160.
22. Lipton RB, Cady RK, Stewart WF et al. Diagnostic lessons from the spectrum study.
Neurology 2002;58:S27–31.
23. Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and diagnosis of migraine in
patients consulting their physician with a complaint of headache: data from the
Landmark Study. Headache 2004;44:856–64.
24. Blanchard EB, Hillhouse J, Appelbaum KA et al. What is an adequate length of
baseline in research and clinical practice with chronic headache? Biofeedback Self
Regul 1987;12:323–9.
25. Lipton RB, Stewart WF, Cady RK et al. Sumatriptan for the range of headaches in
migraine sufferers: results of the Spectrum Study. Headache 2000;40:783–91.
26. Peters M, Abu-Saad HH, Vydelingum V et al. Patients' decision-making for migraine
and chronic daily headache management. A qualitative study. Cephalalgia
2003;23:833–41.
27. Dowson AJ, Turner A, Kilminster S et al. Development and validation of the headache
Diagnostic Screening Questionnaire (DSQ): a new questionnaire for the differential
diagnosis of headache for use in primary care. Headache Care 2005;2:111–18.
28. Maizels M, Burchette R. Rapid and sensitive paradigm for screening patients with
headache in primary care settings. Headache 2003;43:441–50.
29. Henry P, Auray JP, Gaudin AF et al. Prevalence and clinical characteristics of
migraine in France. Neurology 2002;59:232–7.
30. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
31. Kilminster SG. Manual for the Short Pain Inventory. West Sussex (UK): Lawrencian
Clinical Publications and Selsey Press Ltd, 1998.
32. Dowson AJ, Kilminster S, Peters M et al. Understanding the evidence: evaluating the
efficacy of migraine medications in clinical practice. Headache Care 2005;2:133–43.

90
Appendix 1. The headache, depression and bodily pain questionnaire

Please circle the answers that you feel apply to you

1. Do you suffer from headaches?
a. No (go to question 4)
b. Yes

2. How often do they occur?

a. Most days in the month
b. More than four times a month
c. One to four times a month
d. Less than once a month

3. Do the headaches interfere with your ability to perform normal daily activities, i.e.
Do you have to stop doing things you would normally do?
a. No
b. Yes

4. Do you have periods?

a. No
b. Yes

5. In the last five years have you ever suffered from sleep disturbance, such as waking
in the early hours of the morning regularly or having difficulty getting off to sleep, that
has lasted for more than two weeks at a time?
a. No
b. Yes

6. In the last five years have you ever had an episode of low mood, poor concentration,
disinterest or tearfulness that has gone on for more than two weeks?
a. No
b. Yes

7. In the last five years have you ever had feelings of worthlessness, guilt or suicidal
ideas that have persisted for more than two weeks?
a. No
b. Yes

8. If you answered Yes to any of questions 5, 6 or 7, did you discuss these feelings with
your doctor?
a. No
b. Yes

9. If you saw your doctor, did you receive any treatment?

a. No
b. Yes

10. Do you suffer from any other type of pain (other than headaches) that occurs
regularly?
a. No
b. Yes

11. Where do you feel the pain and when you get it how long does it last?
Site of pain:-

91
 Duration of pain:-

12. Have you seen a doctor about this pain:

a. No
b. Yes

13. Have you ever received treatment for this pain?

a. No
b. Yes

Thank you for taking the time to answer these questions for us. If you answered Yes
to many of these questions, and especially questions 5, 6 and 7, it may suggest that
you are suffering from one of these conditions. If you have not discussed these
symptoms with your doctor we suggest that you contact him/her at the Health Centre
and he/she may be able to offer you help with your symptoms.

92
Appendix 2. The menstrual headache questionnaire

Date 1 2 3 4 5 6 7 8 9 etc
Menstrual bleeding

Time of onset - first symptom (specify)

- headache

Time of resolution - headache

- last symptom (specify)

Maximum intensity of headache (mild, moderate or

severe)

Other symptoms (specify)

Time lost from normal activity (hours)

Time spent at less than 50% of normal activity

(hours)

Drugs taken

Contraceptive drug (if any)

Pre-menstrual symptoms or intercurrent illness (if

any)

93
3.4

Emotional function with tension-type headache, migraine and

chronic daily headache*

Abstract

Background: The Short Pain Inventory (SPI©) is a prospective, validated questionnaire that
assesses pain severity and mood disturbances for a variety of conditions.
Objectives: In this prospective, within-group, comparative study, the SPI was used to
assess how the mood disturbances and coping abilities change over time in patients with
tension-type headache (TTH), migraine and chronic daily headache (CDH) attending a
primary care headache centre. The healthcare resource utilisation of these patients was also
assessed retrospectively.
Methods: Patients (n = 75) were diagnosed with episodic TTH, migraine or CDH, and
completed a healthcare utilisation questionnaire. Patients completed the SPI 10 times over a
7-day period, starting 1 hour after the onset of their next headache. The SPI data were
analysed statistically as Z scores and coping Z scores for the Total Pain Disturbance, Total
Mood Disturbance, and individual sub-scores for sedation, social interaction, sadness,
anxiety and anger. Data from the healthcare resource utilisation questionnaire were analysed
as descriptive statistics.
Results: All 75 patients completed the healthcare utilisation questionnaire and 42 (56.0%)
completed the SPI for the 7-day period. All headache patients showed considerable mood
disturbances during a headache. In general the disturbances were severe in intensity and of
the order CDH>migraine>TTH. The pattern of mood disturbance was different for each type
of headache, as was the patients’ ability to cope with the mood changes. Patients with CDH
experienced pain and emotional symptoms, particularly sedation, throughout the 7-day
monitoring period. For TTH and migraine, the pain and emotional symptoms resolved within
1–2 days after the headache. The level of healthcare resource utilisation was also in the
order CDH>migraine>TTH, similar to the data reported for the SPI.
Conclusions: Patients with headache had significant emotional symptoms associated with
their headaches. These symptoms resolved within 1–2 days for patients with episodic TTH
and migraine. However, patients with CDH were profoundly affected, and did not improve
physically or emotionally from their headache over a 7-day period. Headache patients
generally experienced higher levels of sedation than did patients with other pain conditions.

* Edited from the full article: Dowson AJ, Bundy M, Kilminster SG. Emotional function with
tension-type headache, migraine and chronic daily headache. Headache Care 2005;
manuscript in preparation.

94
Introduction

Three subtypes of headache, tension-type headache (TTH), migraine and chronic daily
headache (CDH: comprising chronic migraine, chronic TTH, hemicrania continua and new
persistent daily headache in the new International Headache Society (IHS) classification
criteria1), comprise the majority of the benign, primary headaches reported in general clinical
practice. All these headache subtypes are common in the general adult population:
prevalence rates are >50% for TTH; 12% for migraine; and 4% for CDH.2–4 Physicians’
management of these headache subtypes differs markedly in terms of overall strategy,
selection of treatment modalities and duration of therapy. Also, they may be confused with
each other on diagnosis. Migraine may present with symptoms typical of TTH,5 and migraine
may transform to CDH over periods of years or decades.6 Patients with CDH may experience
headaches characteristic of TTH and migraine in the constellation of their symptoms.7
Diagnostic procedures may not, therefore, distinguish clearly between the headache types. A
means of distinguishing the characteristic features, or ‘footprints’ of these headaches would
be of great value in primary care, where time is limited, and the experience of the physician
uncertain.

Several tools have been developed to assess headache features, as an aid to clinical
management. The tools are broadly of three types: scales that specifically measure the
functional ability of the patient, scales that index the migraine pain experience, including
functional ability, and more generic measures.8–13 Of the former type, impact questionnaires
(e.g. the Migraine Disability Assessment [MIDAS] Questionnaire and the Headache Impact
Test [HIT]) assess the effect headaches have on the patient’s ability to work and function
normally over 1- or 3-month periods.8,9 Migraine-specific Quality of Life (QOL) tools assess
the functional status (physical and emotional), well-being and overall health of the patient,
usually over a 1-month period.10–12 Generic QOL tools measure these values also, and also
include a social dimension.13 In addition, recently developed questionnaires have
investigated ways of assessing the efficacy of interventions, using impact-type,14 QOL-type15
or a combination of different types of questions.16 All of these tools assess retrospective
data.

A biopsychosocial model has been proposed for chronic pain, whereby physical symptoms
and emotional distress lead to the suffering that drives the patient to consult a physician.17
The level of pain and emotional disability are directly linked, while the emotional impact can
also be expressed as how well the patient is coping with the illness relative to a certain level
of pain.18 The relationship between pain level, emotional disability and how well the patient
copes with them is relatively little understood and worthy of investigation.

Whilst the symptomatology of most headache subtypes is well understood, the emotional
components of headaches are not. Our interest was to learn about how patients suffering
from TTH, migraine and CDH differ in their pain level and emotional well-being, and the way
they cope with these headache conditions over time. We also wanted to assess how the
differences in emotional disturbance and coping may affect healthcare resource utilisation in
these patient populations. We chose a prospective tool that has been developed specifically
to measure the emotional consequences and associations of physical pain, the Short Pain
Inventory (SPI©).18

In this study, the SPI was used to prospectively assess how the mood disturbances and
coping abilities change over time in patients with TTH, migraine and CDH attending a
primary care headache centre. The healthcare resource utilisation of these patients was also
assessed retrospectively. The study was conducted as part of ongoing use of the SPI in the
UK National Health Service (NHS), and to investigate the spectrum of primary care patients
(including those with TTH).

95
Methods
Patients
Patients who took part in the study were attending a general practice headache clinic at
Cranleigh Health Centre, Surrey, UK (Dr Bundy). The study was approved by the South West
Surrey Local Research Ethics Committee.

Study design
This was a prospective, within-group, comparative study to assess the emotional ‘footprint’ of
headaches, measured by the SPI, over 7 days following the onset of a headache. The
questionnaires were administered by a psychologist (Dr Kilminster). A specialist neurologist
(Dr Dowson) provided a diagnosis for each patient at the study outset, utilising clinical
judgement and the IHS criteria then current.19

Patients completed a retrospective healthcare resource utilisation questionnaire at screening,

covering the previous 6 months. Information was recorded as the number of: consultations
for headache; prescriptions for headache and other conditions; GP visits; GP services used
(out-of-hours telephone calls, emergency telephone calls and home visits); and referrals to a
specialist physician.

Patients had a 6-week window from study onset to experience a headache and complete a
series of SPI questionnaires. Patients self-completed the SPI at home at the following times
over a 7-day period:
• Practise session at the clinic (baseline)
• Within 1 hour after the first headache started or on waking with a headache (Day 1).
For CDH, the definition was the onset of a subjective ‘bad headache’.
• Just after the pain was worst (peak) [Day 1].
• 1 hour after the headache had resolved or on waking with the headache resolved
(Day 1)
• 24 hours after the headache started (Day 2)
• Evening of Day 2 about 8 pm (Day 2)
• Days 3–7 inclusive at about 8 pm.
SPI questionnaires were completed for 7 days, even if the headache lasted for longer than
this time period. Explanation on how to complete the questionnaires was kept to a minimum.

Instruments: the SPI questionnaire

The SPI is a 17-item self-rating questionnaire dedicated to measuring any pain, not only
headache. The questions assess pain intensity, social interaction skills and components of
emotional function. Patients rate each item on a five-point numerical Likert scale: not at all
(score 0), a little (1), moderately (2), very much so (3) and extremely (4). Five subscales
measure sedation, social interaction, anxiety, sadness and anger, and have been fine-tuned
by maximising items that are specific to pain. The SPI takes about one minute to complete,18
and can be accessed via the internet with automatic scoring.20 A scale of Total Mood
Disturbance (TMD, 14 items) can be summated from the SPI, which measures the subtle
mood changes that specifically covary with mild to severe pain (including sadness, anger,
anxiety and sedation scores [social interaction scores are removed because this is a
cognitive-behavioural parameter]). All 17 items can be summated to form the Total Pain
Disturbance (TPD) score. As the physical severity of pain worsens, so generally does each
of the emotional components.21–23 If a patient’s emotional score is exaggerated or flattened
relative to the normal values, then a coping score can also be given as to how well a patient
is coping with their pain, relative to others at the same physical level of pain. In this way an
emotional ‘footprint’ of the headache can be created. Normative coping scores can be
calculated for all items on the SPI Questionnaire. This can be very useful in the individual
management of patients. We have found that around 60% of the variance of physical pain is
explained by the overall mood of the patient.21,22 The SPI has been shown to be far more

96
powerful at discriminating between varying levels of physical pain severity than the McGill
Pain Questionnaire (used as the gold standard measure), with the majority of the pain
variance being captured by the SPI.21

A recently published paper showed that the SPI was a reliable and valid measure of
headache severity, and superior to the HIT. However, the HIT score was better related to
headache diagnosis than was the SPI score.24 As the pain severity increased from none to
extreme, there was a highly significant increase in the SPI subscale scores for sedation,
social interaction, sadness, anxiety and anger, as well as the SPI TMD score. Correlations
between headache severity and the SPI subscale scores ranged from 0.60 to 0.70, and were
0.76 for the TMD score and 0.77 for the TPD score. The correlation between SPI sedation
scores and headache severity was validated in a placebo-controlled, crossover trial of 10mg
temazepam in 12 healthy volunteers, by analysing Critical Flicker Fusion thresholds.24,25

Other studies have demonstrated that the levels of SPI subscales associated with headache
exhibited marked differences when compared with those of other pain states such as dental
pain, osteoarthritis and chronic pain states.22–24 In general, headache patients had the
greatest level of sedation, while dental patients had the greatest levels of anger, sadness
and anxiety. Patients with osteoarthritis had the lowest level of mood disturbance of all the
patient groups. Overall, headache patients had a level of mood disturbance at least as
severe as patients who were attending a secondary care chronic pain clinic.

Endpoints
The primary study endpoints were derived from the SPI questionnaire: the pattern of pain
intensity (assessed as none [Grade 0], mild [Grade 1], moderate [Grade 2], severe [Grade 3]
and extreme [Grade 4]) and emotional items (sedation, social interaction, sadness, anxiety
and anger, total TMD and TPD scores) over the 7 days during and following the headache.
The healthcare utilisation data formed the secondary study endpoint.

Statistical methods
SPI
All spoiled SPI forms were included for the intention-to-treat basis. This was done so as to
reduce the administrator bias and evaluate the SPI as it would be used realistically in clinical
settings. Statistical analysis was by discriminant function analysis, simple t-tests and test-
retest Pearson product moment correlation coefficients. Statistical analysis was carried out
on a Power Macintosh G4/400/200 machine with Statistica V4. All data points were double
validated against source forms. The analysis was carried out by a medical statistician.
Diagnostic labels derived from the IHS classification criteria19 were dummy coded. Summary
and derivative SPI measures were then used as dependent variables in simple t-tests. Data
were 100% validated against source documents.

Simple box plots with mean ± 1 and 1.96 standard errors about the mean were used.
Elsewhere, graphical information was plotted as Z scores. The SPI raw scores were used for
the summary and components. SPI disturbance and coping Z scores were computed. With
the SPI, the extent to which the patient is coping with their pain can be estimated as coping Z
scores, calculated by removing the effects due to physical pain. Put simply, normative coping
values of the mood change were computed relative to patients with either no pain, mild,
moderate, severe and extreme pain. For disturbance score, the normative values were those
of subjects without pain. The SPI Coping score shows how much emotional upset the patient
is in relation to the normal ranges expected in each of the five grades of physical pain. A fully
automated and much simplified electronic version of SPI has been developed and is
accessible via the internet at www.headachetest.co.uk. Previous studies have shown that the
analysis of only four patients was sufficient to distinguish between headache subtypes and
12 were sufficient to assess time of day effects with SPI.20

97
Healthcare resource questionnaire
Data derived from the healthcare resource questionnaire were analysed by descriptive
statistics only.

Results
Patient demography and baseline characteristics
Seventy five patients took part in the study and completed the healthcare utilisation
questionnaire. Their mean age was 44.3 years (SD 9.8), and the majority (97.3%) were
women. At screening, episodic TTH was diagnosed for 15 patients (20%), migraine for 41
patients (54.7%) and CDH for 19 patients (25.3%). Migraine patients were subdivided into
those with low-frequency attacks (≤ 3 attacks per month, n = 30) and high-frequency attacks
(> 3 attacks per month, n = 11). Extreme pain (Grade 4) was reported by 2.6% of patients,
very much pain (Grade 3) by 8.2%, moderate pain (Grade 2) by 10.5%, a little pain (Grade 1)
by 34.2% and no pain (Grade 0) by 44.7% of patients.

Forty two of the 75 patients (56.0%) returned the completed SPI questionnaires within 6
weeks. Six of these patients had TTH (14.3%), 23 had migraine (54.8%) and 13 had CDH
(31.0%). The major dropout rate was in the TTH group. Due to the low frequency of their
headaches, many TTH patients did not experience an initial headache over the 6-week
window.

SPI: Changes in coping scores during, and for 7 days after, the headache
Pain severity
The physical severity of the pain varied considerably over the week following the initial
headache (Figure 1). At the peak of the headache, patients with episodic TTH presented on
average with mild to moderate pain (score 1–2), migraine patients with moderate to severe
pain (score 2–3) and CDH patients with severe pain (score 3). TTH pain largely resolved
within 24 hours and migraine pain within 2 days, but the CDH pain remained at least mild in
intensity for the whole 7-day period.

Figure 1. SPI: Physical pain severity during the headache and over the following week for
patients with episodic TTH, migraine and CDH.

98
Changes in emotional items
The disturbance in the sedation, social interaction, sadness, anxiety and anger items of the
SPI over 7 days during and following the headache are shown in Figure 2 as Z scores.
Coping Z scores for sedation, social interaction, sadness, anxiety and anger items of the SPI
over 7 days during and following the headache are shown in Figure 3.

Figure 2. SPI: Sedation, social interaction, sadness, anxiety and anger Z scores during the
headache and over the following week for patients with episodic TTH, migraine and CDH.

99
100
Figure 3. SPI: Sedation, social interaction, sadness, anxiety and anger coping Z scores
during the headache and over the following week for patients with episodic TTH, migraine
and CDH.

101
102
Sedation disturbance Z scores revealed differences between the patient groups. Although all
headache groups showed severe sedation at the point of worst pain during their headaches,
the footprint of sedation over the remaining week was very different. With migraine, the
sedation had resolved 2 days after the headache. However, patients with CDH and episodic
TTH showed marked levels of sedation throughout the week. With TTH, this resolved by Day
7, but the CDH patients were chronically sedated. Analysing the coping Z scores, TTH
patients appeared to be coping less well than the other groups, with substantial sedation in
excess of the physical pain they experienced.

Despite the fact that all patients appeared sociable at screening, disturbances in social
interaction remained above the normal range throughout the 7-day period for CDH patients.
Two days after their headaches, patients with migraine and TTH had returned to normal
levels of social interaction. Examining the social interaction coping scores showed that, for 5
days of the week, the CDH patients had difficulty coping. Patients with migraine and TTH
returned to normal within 24 hours of the headache.

103
Disturbances in sadness remained elevated for the CDH patients throughout the 7-day
assessment period. However, sadness resolved 2 days after the headache in the migraine
and TTH groups. Coping scores for sadness were not sustained above +0.5Z for any of the
groups. Therefore, much of the sadness induced is likely to be a direct result of the physical
pain endured. There was some suggestion that patients with TTH were coping less well
considering their lower level of pain and lower frequency of headaches compared to the
other groups.

Disturbances in anxiety remained elevated for the CDH patients throughout the 7-day
assessment period. However, anxiety had resolved 2 days after the headache in the
migraine and TTH groups. The SPI anxiety coping Z score showed that patients’ weekly
anxiety remained within normal coping levels. There was a suggestion from the data that
TTH patients might be coping less well with anxiety.

The SPI anger disturbance Z scores showed a similar pattern to the pain severity changes
over the week. Patients with migraine and TTH recovered within 24 hours of the headache,
but CDH patients remained above normal levels for the whole week of assessments. The
SPI anger coping Z scores showed that all the headache patients were coping well with
anger.

Total mood disturbance

The TMD Z scores and coping Z scores, which include the sedation, sadness, anxiety and
anger components, are shown in Figure 4. In general, patients with CDH and migraine had
equivalent and severe mood disturbances during the period of the headache itself. However,
the mood of CDH patients never returned to normal over the 7-day assessment period, while
that of TTH and migraine patients did so within 48 hours of the headache. The coping scores
showed that TTH and CDH patients had mild difficulty coping with their mood disturbances.
The highest levels were seen temporarily with TTH patients. Migraine patients mostly coped
well with their mood disturbances.

Figure 4. Total mood disturbance Z scores and coping Z scores during the headache and
over the following week for patients with episodic TTH, migraine and CDH.

104
The TPD Z score was computed from all 17 SPI subscales. The results closely resembled
those for the TMD discussed above (Figure 4). As before, the CDH patients did not recover
over the 7-day assessment period and their pain was typically +2Z above normal.

Healthcare utilisation
Patients diagnosed with episodic TTH, migraine and CDH showed marked differences in
their utilisation of healthcare resources (Table 1). Patients with TTH used few resources, and
did not consult or receive prescriptions for headache. Most patients with migraine had
consulted for headache, with over three-quarters receiving a prescription for headache from
their GP. Most patients with CDH had consulted for headache, each receiving, on average,
more than one prescription for headache. Patients with CDH consulted their GPs (all causes)
and received more overall prescriptions than patients with migraine or episodic TTH. Out-of-
hours GP services (telephone calls, emergency calls and home visits) were hardly ever used
by these groups of headache patients. Headache referrals were infrequent for episodic TTH
patients, but were used for approximately one-third of migraine and CDH patients. Patients
with low-frequency migraine used fewer resources than those with high-frequency migraine
and CDH, who had a similar pattern of use.

105
Table 1. Healthcare utilisation by patients in the study, reported at baseline for the previous
6-month period.

Number of patients (ratio per patient)

Healthcare Episodic Migraine Low- High- CDH
resource TTH (n = 41) frequency frequency (n = 19)
(n = 15) migraine migraine
(n = 30) (n = 11)
Headache 0 31 (0.76) 19 (0.63) 12 (1.09) 12 (0.63)
consultations
Prescriptions 0 34 (0.83) 10 (0.33) 24 (2.18) 26 (1.37)
for headache
Total 9 (0.6) 143 (3.49) 68 (2.27) 75 (6.82) 131 (6.89)
prescriptions
GP visits 20 (1.33) 108 (2.63) 67 (2.23) 41 (3.73) 76 (4.0)
Out-of-hours 0 1 (0.02) 1 (0.03) 0 0
calls
Emergency 0 0 0 0 0
calls
Home visits 0 1 (0.02) 1 (0.03) 0 2 (0.11)
Referrals 1 (0.07) 13 (0.32) 8 (0.27) 5 (0.45) 6 (0.32)

Discussion

This study demonstrates the link between pain, emotional disability and the associated way
that patients with TTH, migraine and CDH cope with their headaches over time. We have
previously demonstrated that the SPI is a very discriminating measure of headache
severity.24 Results from the current study show that it also discriminates the emotional impact
of different headache subtypes. It is clear that episodic TTH, migraine and CDH leave
different ‘footprints’ in terms of the emotional experience, associated tiredness, disruption to
social interactions and the time course of each of these components.

At peak intensity the pain level as assessed by the SPI was mild to moderate for episodic
TTH, moderate to severe for migraine and severe for CDH. These gradations of intensity are
similar to those generally reported for the separate headache subtypes.1 Patients with the
three headache subtypes all exhibited sedation, disturbances in social interaction, sadness,
anxiety and anger at the peak of the headache. In general, the severity of these emotional
components was more intense for CDH and migraine than for episodic TTH. However, the
pattern of these emotions differed between the three groups in the 7-day period following the
headache.

Analysis of the SPI data demonstrates that CDH is an appropriate term for the condition.
Physical pain was always present over the 7 days, and fluctuated between mild to moderate
pain over the post-headache period. Marked levels of sedation, disturbed social interaction,
sadness, anxiety and anger were also seen over the week following the headache. CDH
patients are clearly chronically fatigued. For half of the week, their level of sedation was due
to impaired coping with tiredness above and beyond the level caused by physical pain. This
may reflect the known association of CDH and Chronic Fatigue Syndrome26 and, certainly,
tiredness is a major feature of the post-dromal period of headache in CDH. CDH patients
also showed chronic disturbances in social interaction and sadness that never recovered
over the 7-day period. For at least part of the week, these disturbances were caused by a
failure to cope above and beyond the level of physical pain endured. On the other hand,
coping scores for anxiety and anger were mostly within normal levels, the high levels of state

106
anxiety and anger disturbance being caused by the physical pain severity. Overall, the level
of emotional disturbance experienced by CDH patients was similar to that seen in patients
attending chronic pain clinics.22 An unfortunate feature of CDH patients is that their mood
disturbance never returns to normal and patients obtain no respite from the pain and its
emotional consequences.

Migraine physical pain improved to mild or no pain within 48 hours of the headache. Although
marked levels of sedation, disturbed social interaction, sadness, anxiety and anger were
seen during the headache, these symptoms typically resolved within 2 days. Coping scores
for these symptoms were at normal levels after the headache. These data indicate that the
patients were coping well with the emotional symptoms, which resulted from the level of
physical pain experienced. Unlike CDH patients, migraine patients are not chronically
sedated.

Physical pain associated with episodic TTH resolved within 24 hours. Marked levels of
sedation were seen during the headache and continued for most of the following week.
Coping scores for sedation were the most disturbed of all the three types of headache,
seemingly being more severe than those related to CDH. This shows that the sedation was
probably not related to physical pain, but was exaggerated by it. It may be a feature of a
delayed recovery from the headache insult or be a precipitating factor for the headache. This,
perhaps surprising feature considering the relatively mild intensity of TTH, is probably
underestimated in the management of these patients. Levels of disturbance in social
interactions, sadness, anxiety and anger were all less intense during the headache than
those associated with migraine and CDH, and resolved within 1–2 days following the
headache. Coping scores for social interaction, anxiety and anger demonstrated that these
symptoms were not above normal ranges, taking into account the level of physical pain.
Coping scores for sadness were sometimes above the normal range for 3 out of the 7 days
post-headache. This suggests mood lability in terms of sadness that is not related to the
physical pain experienced. Coping Z scores for anger tended to be rather low. It is possible
that episodic TTH is typified by a period of anger repression and extreme tiredness, although
the study sample was too small to be definite about this.

In this study, the level of mood disturbance was directly related to the level of healthcare
utilisation in the same group of patients. Direct comparisons cannot be made, as the two
assessments were not contemporaneous. However, the link is logical. Patients with chronic,
severe emotional impact are more likely to require medical services than those in which the
impact is transient. Patients with episodic TTH had a low use of resources, in terms of GP
consultations and prescription medications. Patients with migraine frequently consulted their
GP for care and usually required prescription headache medications. Those with high-
frequency migraine had markedly higher patterns of healthcare utilisation than those with
low-frequency attacks. Patients with CDH consulted their GPs frequently and had high usage
of prescription medications. Patients with high-frequency migraine are at risk of developing
CDH,6 and at least some of the patients in this group may actually have been in this situation.
Mood disturbances may, at least in part, drive patients to consult their GPs for headache
care.

This work has potential clinical implications to the management of headache. Despite its
importance to illness severity, the emotional component of headache is likely to be rarely
elicited by physicians, as it plays no part in the current classification criteria.1 Specialist
physicians may find that the use of SPI forms a useful additional screening and outcome tool
for the evaluation of their patients, particularly those who may later develop depression
and/or anxiety conditions. It may be useful not only to estimate the consequences of pain but
also stress as a causative trigger to headache. However, should primary care physicians
may find it too demanding for everyday use, an automated much simplified electronic version

107
has been developed over the internet at www.headachetest.co.uk.20 This has the benefit of
full electronic delivery of the test and automatic scoring.

The recently developed Migraine Assessment of Current Therapy (Migraine-ACT)

questionnaire contains one question on emotional response (Are you comfortable enough
with your medication to be able to plan your daily activities?).16 This question is suitable for
use in primary care, and ‘no’ answers indicate that the patient has unmet treatment needs. A
single questionnaire item on mood cannot possibly accurately and comprehensively assess
mood, and the questionnaire was not set up to do this. Patients with chronic emotional
problems linked to headache may require referral to a specialist. But one should be mindful
that mood disturbance can be as much a cause as an effect.

This study has certain limitations. Relatively small numbers of patients took part, and a high
proportion of patients did not complete a SPI questionnaire in the 6 weeks allowed. A longer
study duration could have led to a reduction in the drop out rate.

In conclusion, patients with episodic TTH, migraine and CDH all had considerable mood
disturbances during their headaches. The more pain they experienced, the greater was the
level of mood disturbance. The pattern of mood disturbance was different for each type of
headache, as is the patients’ ability to cope with the mood changes. Patients with episodic
TTH and migraine had significant emotional symptoms associated with their headaches,
which resolved within 1–2 days. However, patients with CDH were most profoundly affected,
and did not recover physically or emotionally from their headache over a 7-day period. The
association of mood disturbances with common headache subtypes may explain in part the
reported co-morbidity of headache with psychiatric disorders.27 Patients with chronic
headaches have clear emotional needs, both for their headache and other disorders, which
may not be currently appreciated or addressed by healthcare services. Future large studies
with SPI in headache are warranted to further define and quantify the emotional component
of headache impact.

References

1. Headache Classification Subcommittee of the International Headache Society. The

international classification of headache disorders. Cephalalgia 2004;24 (Suppl 1):1–
160.
2. Rasmussen BK, Jensen R, Schroll M et al. Epidemiology of headache in a general
population—a prevalence study. J Clin Epidemiol 1991;44:1147–57.
3. Stewart WF, Lipton RB, Celentano DD et al. Prevalence of migraine headache in the
United States. Relation to age, income, race, and other sociodemographic factors.
JAMA 1992;267:64–9.
4. Silberstein SD, Lipton RB. Chronic daily headache Curr Opin Neurol 2000;13:277–83.
5. Lipton RB, Stewart WF, Cady RK et al. Sumatriptan for the range of headaches in
migraine sufferers: results of the spectrum study. Headache 2000;40:783–91.
6. Mathew NT, Stubits E, Nigam MR. Transformation of episodic migraine into daily
headache: analysis of factors. Headache 1982;22:66–8.
7. Limmroth V, Kazarawa S, Fritsche G et al. Features of medication overuse headache
following overuse of different acute headache drugs. Neurology 2002;59:1011–4.
8. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of
the Migraine Disability Assessment (MIDAS) score. Neurology 1999; 53:988–94.
9. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
10. Jhingran P, Osterhaus JT, Miller DW et al. Development and validation of the
Migraine-Specific Quality of Life Questionnaire. Headache 1998;38:295–302.

108
11. Wagner TH, Patrick DL, Galer BS et al. A new instrument to assess the long-term
quality of life effects from migraine: development and psychometric testing of the
MSQOL. Headache 1996;36:484–92.
12. El Hasnaoui A, Vray M, Richard A et al. Assessing the severity of migraine:
Development of the MIGSEV scale. Headache 2003;43:628–35.
13. Stewart AL, Greenfield S, Hays RD et al. Functional status and well-being of patients
with chronic conditions: Results from the Medical Outcomes Study. JAMA
1989;262:903–13.
14. Chatterton ML, Lofland JH, Shechter A et al. Reliability and validity of the Migraine
Therapy Assessment Questionnaire. Headache 2002;42:1006–15.
15. Hartmaier SL, Santanello NC, Epstein RS et al. Development of a brief 24-hour
migraine-specific quality of life questionnaire. Headache 1995;35:320–9.
16. Dowson AJ, Tepper SJ, Baos V et al. Identifying patients who require a change in
their acute migraine treatment: The Migraine Assessment of Current Therapy
(Migraine-ACT) Questionnaire. Curr Med Res Opin 2004;20:1125–35.
17. Loeser JD. What is chronic pain? Theor Med 1991;12:213–25.
18. Manual. The Short Pain Inventory. Selsey, West Sussex: Lawrencian Clinical Ltd.
England.
19. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8 (Suppl 7):1–96.
20. The SPI Online Test for headache pain. www.headachetest.co.uk
21. Kilminster SG, Mould GP. Comparison of the internal reliability and validity of the
McGill Pain Questionnaire and Short Pain Inventory. Int J Pharm Med 2002;16:87–
95.
22. Kilminster SG, Power M, Fozard JR. Survey of pain in two medical and dental clinics,
with non patient controls using the Short Pain Inventory. Int J Pharm Med
2000;14:137–47.
23. Kilminster SG, Mould GP. Comparison of diclofenac spray and gel on knee joints of
patients with osteoarthritic pain. Clin Drug Invest 1999;18:345–54.
24. Kilminster SG, Dowson A, Bundy M. The Headache Impact Test® and the Short Pain
Inventory©: Outcome measures compared. Int J Pharm Med 2003;17:23–32.
25. Puranik S, Fozard J, Paremain G et al. Randomised single blind study to evaluate the
effects of Action Potential Simulator Therapy compared with placebo in patients with
chronic back pain. Pain Clinic 2002;14:69–73.
26. Peres MF, Zukerman E, Young WB et al. Fatigue in chronic migraine patients.
Cephalalgia 2002;22:720–4.
27. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and
co-morbidities. Neurology 2001;56 (Suppl 1):4–12.

109
4

The clinical utility of assessing disability in

primary care clinical practice

110
4.1

Assessing the impact of migraine and other headaches*

Abstract

Migraine is a remarkably disabling condition, although unpredictable and heterogeneous in

frequency, duration and severity. It can be difficult to manage in primary care, where it is
under-recognised, under-diagnosed and under-treated. Proposals have been made that
migraine care could be improved by incorporating assessments of migraine impact into
management strategies. Research has shown that measuring headache-related disability,
together with assessments of pain intensity, headache frequency, tiredness, mood
alterations and cognition can be used to assess the impact of migraine on sufferers’ lives and
society. From this research two simple and brief impact tools were developed; the Migraine
Disability Assessment (MIDAS) Questionnaire and the Headache Impact Test (HIT). Both
tools are scientifically valid measures of headache severity and have the potential to improve
communication between patients and their physicians, assess headache severity and act as
outcome measures to monitor treatment efficacy. Each of these tools offers its own
advantages. For example, HIT was designed for greater accessibility (on the internet at
www.headachetest.com and www.amIhealthy.com and as a paper-based form known as
HIT-6) and has a wider coverage of the spectrum of headache than MIDAS. Impact tools are
also being increasingly recommended as part of generalised headache management
guidelines to produce an individualised treatment plan for each patient in concert with other
clinical assessments. It is not possible as yet to unequivocally recommend the optimal impact
tool for use in primary care, but it should be usable by GPs, pharmacists, nurses and
patients, and for research purposes.

* Edited from the full published article: Dowson AJ. Assessing the impact of migraine. Curr
Med Res Opin 2001;17:298–309.

111
Introduction

Migraine is a common, debilitating neurological disorder that affects about 12% of the
general population. It is more prevalent in women than in men and in Caucasians than in
Black and Asian races.1 Migraine attacks consist of moderate to severe headaches, which
are typically throbbing, one-sided and aggravated by physical activity. The headache is
usually accompanied by photophobia and/or phonophobia and nausea, and less frequently
by vomiting and aura symptoms. Attacks occur on average about once or twice a month, last
for about 24 hours, and are separated by symptom-free intervals.2 Migraine attacks result in
significant reductions in sufferers’ health-related quality of life (HRQoL) compared with
normal healthy subjects.3 However, migraine is an unpredictable, heterogeneous disorder
and attacks vary widely in frequency, duration, severity and reported symptoms.4

Due to this variability, migraine can be difficult to manage in primary care. In clinical practice,
migraine is under-estimated, under-diagnosed and under-treated, leading to many patients
dropping out from care (Figure 1). A recent study in the UK showed that 86% of migraine
sufferers had consulted their physician at some time for treatment. However, 37% of
sufferers had dropped out, leaving 49% as currently consulting.5 Studies in the USA and
France indicated that about half of all migraine sufferers who consulted their general
practitioners did not receive a correct diagnosis.6,7 Patients may not receive appropriate
treatment even when they do consult and are diagnosed correctly. International and UK
studies have shown that only a minority of migraine sufferers take prescription medications8,9
and patient satisfaction with their usual analgesic acute medications is low.9,10

Figure 1. Barriers to migraine care. For effective migraine care, sufferers need to consult
their physician and be diagnosed, treated and followed-up effectively. Migraine sufferers are
lost to care through under-consultation, under-diagnosis and under-treatment of their
attacks.5–10

Yes Yes Yes Yes

Migraine Ongoing
patients in Appropriately assessment Good
Consulting Diagnosed
need of treated of control outcome
care

No No No No

Motivate Improve Improve Encourage

patient to diagnosis treatment follow-up
seek care

A range of initiatives has been set up to improve migraine care, by encouraging migraine
patients to consult their physicians and for physicians to improve their diagnostic and
treatment strategies. Several of these involve assessing the impact of migraine on sufferers’
lives, and using this information to facilitate management strategies. This article reviews the
rationale for, and development of, tools that assess migraine impact and their potential roles
in general practice.

112
Research into the impact of migraine

In some instances, the impact of migraine can be considered as the objective effects of the
illness on sufferers’ lifestyles, including their work and leisure activities, rather than subjective
effects expressed as symptoms and HRQoL. Impact has great similarities to the World
Health Organization’s definition of disability: “a restriction or lack (resulting from an
impairment) of ability to perform an activity in the manner or within the range considered
normal for a human being”.11 Studies from the USA, Canada and Japan have shown that
migraine causes significant disability in its sufferers, with two thirds or more reporting at least
mild disability and one third or more reporting moderate to severe disability.4,12,13 In a UK
study, two thirds of migraine sufferers reported that migraine disrupted their lives, with three
quarters having to lie down during attacks (Table 1).14 A second study indicated that between
one third and two thirds of migraine sufferers in the UK (an estimated 1.9–3.8 million people)
felt that they were not in control of their migraine and the way it affected their day-to-day
lives.9

Table 1. Migraine-related disability reported from a study conducted in the United Kingdom.14

Disability Proportion of sufferers

Physical functioning
Always have to lie down 76
Not in control of life 34
Disruption of life 67
Employment
Usually miss work 50
Difficulty performing work 72
Cancel appointments/meetings 67
Rely on other people 45
Perceived effect on promotion 15
Unpaid work
Postpone household chores 90
Family and leisure activities
Relations with family and friends affected 54

The consequences of migraine impact are seen in patients’ lifestyles, including employment,
unpaid work and family and leisure activities. The loss of these activities has been quantified
in a series of studies. In the USA, each working migraine sufferer missed an average of 4.4
days of work per year and the equivalent of 12 further days due to reduced productivity
during attacks.15 In the UK, half of sufferers reported missing work and over two thirds
reported difficulty performing work during attacks (Table 1).14 Migraine may even lead to
unemployment. In a Health Maintenance Organisation in the USA, the unemployment rate
was 2- to 4-fold higher in severely affected migraine sufferers than in the general
population.16 School and college work is also affected in young migraine sufferers. In a
Scottish study, children with migraine were absent from school for significantly longer periods
than those without migraine (7.8 days versus 3.7 days per year, P<0.0001).17 This personal
burden of migraine is reflected in an economic burden on society. Indirect costs (due to work
absence and reduced work productivity) are very large for migraine, being estimated at £1
Billion or more in the UK.18 These costs are very much higher than the direct costs due to
medical care, last estimated at £23 million in 1990.19 Although direct medical costs for
migraine have increased since 1990, partly due to the introduction of the triptan drugs,
indirect costs still provide the main economic burden of the illness.

113
Migraine also affects unpaid work and family and leisure activities. In a UK study, 90% of
migraine sufferers reported that they postponed their household work during an attack (Table
1).14 Several studies have shown that migraine attacks commonly result in the cancellation of
social events, and affect relationships with partners, children, friends and other people.14,20,21

Migraine is therefore a remarkably disabling condition, with most sufferers reporting

significant impact associated with their attacks in all areas of their lifestyles. Recent research
has focused on the use of impact assessments to measure the severity of migraine in
comparison with other assessments.

Assessing the impact of migraine

Migraine attacks vary in severity from moderate pain, with no activity limitations, to severe
pain with prolonged incapacitation.4 To measure the impact of migraine, parameters must be
defined that capture the personal burden on the sufferer and the economic burden on
society. Following this, a scientifically valid tool needs to be developed to capture this
information in a way that is simple to use and clinically relevant.

Rationale for using impact tools

Pain intensity is the most important aspect of migraine to the individual sufferer. Sufferers
report headache pain more frequently than all other migraine symptoms and most patients
consulting their physicians do so for pain relief.12 However, economic studies have shown
that headache-related disability is the most important determinant of migraine’s societal
impact in economic terms.22 Assessments of pain and disability were therefore included in
initial studies assessing the impact of migraine.

The Chronic Pain Index (CPI) was developed to measure the severity of a number of pain
conditions, including headache, in primary care patients.16 Results showed that pain intensity
and disability measures formed a reliable hierarchical scale whereby pain intensity scaled the
lower range and disability the upper range of severity. Four severity grades were identified
that covered the whole pain spectrum. When the CPI was tested on 740 primary care
headache patients over a 2-year period, the severity grade was directly related to the impact
of the headache on the individual and the direct and indirect economic costs (Figure 2).16,23
There was approximately 3–4 fold greater individual impact, direct costs and indirect costs
(assessed as the unemployment rate) for patients in the highest severity grade compared
with those in the lowest grade. These results were confirmed in a study conducted in
migraine patients who assessed the pain and disability associated with their most recent
attack.4

114
Figure 2. Relationship of the Chronic Pain Index severity grade to the individual and societal
costs of headache when tested on a group of 740 primary care headache patients over 2
years. a. Pain impact on the individual assessed as a composite of activity limitations,
depression and poor-to-fair self-rated HRQoL. b. Impact on society assessed as direct
medical costs of care per year. c. Impact on society assessed as indirect costs over 2 years
(assessed as the unemployment rate).16,23 Grade I = low pain intensity and low disability;
Grade II = high pain intensity and low disability; Grade III = high disability which was
moderately limiting; Grade IV = high disability which was severely limiting.

a.

Pain Impact (activity limitations, depression and poor-to-fair self-rated HRQoL)

60
Extent of disability

40

20

Grade IV Grade II
Grade III Grade I
0
1 month 1 year 2 years

b.

Total cost of headache care per year per patient

1000
Mean cost of headache care ($US)

800

600

400

200

0
I II III IV
Migraine severity grade at baseline

115
 c.

Unemployment rate
30
Severity grade at baseline
Grade IV Grade II
Grade III Grade I

20
Unemployed (%)

10

0
Baseline Year 1 Year 2

Using the CPI as a model, the Headache Impact Questionnaire (HImQ) was developed to
measure headache impact. This was a relatively complex, 16-item questionnaire assessing
headache frequency, pain intensity and disability, the latter measured as total lost time in
employment, household work and non-work activities. Studies showed that the HImQ was a
scientifically reliable and valid measure of migraine severity. 24,25 However, it was complex to
score, requiring mathematical equations, and was not intuitive to use as it was based on a
composite of pain, frequency and disability measures. It was therefore not suitable for use in
primary care, but has proved to be an excellent research tool.

This research did indicate that it was possible to assess the impact of migraine using
scientifically valid questionnaires. The next challenge was to develop simple but accurate
and comprehensive questionnaires that could be used in the primary care setting.

Development of impact tools for use in primary care

Two tools have been developed to assess the impact of migraine, both partly based on the
HImQ. However, they use quite different strategies to achieve their goals. The Migraine
Disability Assessment (MIDAS) Questionnaire assesses impact as a static measure of
disability, while the Headache Impact Test (HIT) uses a global assessment of impact,
including headache-related disability. It should be noted that both questionnaires assess
headache in general, rather than migraine specifically, and so have potential application to
the whole field of headache management.

The Migraine Disability Assessment (MIDAS) Questionnaire

MIDAS is a paper-based questionnaire, designed to be accessible at physicians’ surgeries
and pharmacies. Headache sufferers answer five questions in three activity domains
covering the previous 3-month period (Figure 3).26 They score the number of lost days due to
headache in employment, household work and family and social activities. Sufferers also
report the number of additional days with significant limitations to activity (defined as at least
50% reduced productivity) in the employment and household work domains. The total MIDAS
score is obtained by summing the answers to the five questions as lost days due to
headache. This can sometimes be higher than the actual number of lost headache days due
to any one day being counted in more than one domain. The score is categorised into four

116
severity grades: Grade I = 0–5 (defined as minimal or infrequent disability); Grade II = 6–10
(mild or infrequent disability); Grade III = 11–20 (moderate disability); Grade IV = 21 and over
(severe disability); Two other questions (A and B) are not scored, but were designed to
provide the physician with clinically relevant information on headache frequency and pain
intensity. Information on MIDAS can be accessed at www.migraine-disability.net.

MIDAS was tested extensively and shown to have face validity (i.e. meets physicians’
conceptions of important clinical criteria) and to be reliable, accurate, easy to use and score
and intuitively meaningful to physicians.27–29 These features support its suitability for use in
clinical practice. MIDAS has been proposed as an aid to communication between patients
and healthcare professionals, as an aid to referral for primary care physicians, as a means of
specifying treatment by using it to stratify patients according to their treatment needs, as an
outcome measure and as a public health initiative to coordinate public policies for migraine
management.8

The MIDAS Questionnaire has several strengths. It has proved to be an effective aid to
communication between physicians and their migraine patients with regard to disability, a
subject often overlooked in conventional consultations.30 MIDAS is now widely used for this
purpose by headache specialists and neurologists. Clinical studies have shown that MIDAS
is sensitive to change, and can be used as an outcome measure in the follow-up of patients
with migraine or CDH during treatment.31,32 However, these data require corroboration in
large scale clinical trials.

117
Figure 3. The Migraine Disability Assessment (MIDAS) Questionnaire.26 The MIDAS Program
was developed by Innovative Medical Research Inc, with sponsorship and assistance from
AstraZeneca.

However, MIDAS also has several weaknesses. It does not assess the full spectrum of
headache, covering only about 35% of the range between moderate and severe intensity.33
This may be due to its reliance on a single measure (disability). Additionally, the MIDAS
grade score may not indicate the true medical need of patients. In the author’s clinical
practice, many needy patients scored as Grade I, theoretically with little disability. A US study
has shown that about 10% of MIDAS Grade I headache sufferers had more than six severely
painful headaches per year.34 On the other hand, patients with frequent headaches, e.g.
chronic daily headache, tend to all score as Grade IV. Some physicians have suggested that

118
MIDAS is weighted towards the measurement of headache frequency over disability.
Relatively few physicians currently use MIDAS as a tool to specify treatment according to the
patient’s disability grade, despite some evidence for this from a large clinical trial.35

The Headache Impact Test (HIT)

HIT was first developed as a web-based test, designed to be accessible to all physicians and
headache sufferers through the Internet (at www.headachetest.com and
www.amIhealthy.com). Questions are not printed on forms in advance. Rather, it is a
dynamic questionnaire (Figure 4), with items derived from four validated headache
questionnaires sampling all areas of headache impact (the Headache Disability Inventory
[HDI], the HImQ, the MIDAS Questionnaire and the Migraine-Specific Quality of Life
Questionnaire [MSQ]).36 Patients are questioned until clinical standards of score precision
are met. Internet-HIT matches the questions asked to each patient’s severity level. In
practice, five questions are sufficient to grade the majority of headache sufferers.36 Clinical
standards of accuracy were met by 98% of migraine sufferers, and 97%, 87% and 61% of
people with severe, moderate and mild headaches, respectively, completing five or fewer
questions. In an extensive testing study in over 19,000 headache sufferers in the USA,
Internet-HIT was shown to be reliable and valid, and covered the whole spectrum of
headache.33,37,38 Coverage was similar to that reported for HImQ and much greater than that
of MIDAS.33 Results showed that migraine had greater impact on sufferers’ lives than most
other headaches.37 Internet-HIT differentiated sufferers on the basis of diagnosis and
characteristics such as headache severity and frequency.37 It took only 1–2 minutes to
complete, and its use motivated headache sufferers to seek medical care and facilitated
headache-related communication between patients and their physicians.39

Figure 4. The web-based Headache Impact Test (Internet-HIT).36,37 HIT was developed by
QualityMetric Inc and GlaxoSmithKline.

HIT-6 is a paper-based, short-form questionnaire based on the Internet-HIT question pool,

designed for people without access to the Internet (Figure 5). Six questions cover pain
severity, loss of work and recreational activities, tiredness, mood alterations and cognition.
Each question is scored on a five-point scale, with the scores being added to produce the
final score.40 HIT-6 scores are categorised into four grades, representing minimal, mild,
moderate and severe impact due to headache. Studies have shown that HIT-6 is reliable and
valid,40 and can be used to help diagnose headaches and select management strategies
according to headache severity.41,42 Also, HIT-6 scores were directly related to self-reported
hours of workplace productivity loss due to headache.43 Internet-HIT and HIT-6 scores
compared well to each other when the two forms of the questionnaire were tested on a group
of headache sufferers.44

119
Figure 5. The paper-based short form Headache Impact Test (HIT-6).40 HIT was developed
by QualityMetric Inc and GlaxoSmithKline.

120
Both the MIDAS and HIT tools provide scientifically reliable and valid measures of headache
impact, and show great promise for wide utility in clinical practice. However, both forms of
HIT appear to have certain advantages over MIDAS. HIT has the brevity of a short-form
questionnaire but has the accuracy required for measuring individual patients at all levels of
impact. This may be due to its wider range of questions compared to MIDAS, which scores
time lost due to headache only. HIT is also more widely accessible than MIDAS, designed to
be used by everyone through the Internet and as a paper-based form. However, HIT was
developed more recently than MIDAS and has so far not been used in clinical practice to the
same extent. So far, only MIDAS has exhibited sensitivity to change and can therefore be
used to monitor the outcome of therapy.31,32 HIT has not been tested for this, but may be
expected to also have this feature. Future work will define further the place of each of these
impact tools in primary care, and such studies are currently underway with HIT.

Strategies for managing migraine using impact measures

Although impact measures for migraine are a recent development, they have been adopted
widely by many opinion leaders in the field of headache. Both MIDAS and HIT have been
proposed to have a wide potential for use in clinical practice, to improve communication
between physicians and their patients, to assess illness severity and as public health
measures.8,39,41,42 Impact measures have been incorporated in several initiatives to develop
guidelines for the management of migraine in primary care, including those issued by the
Migraine in Primary Care Advisors (MIPCA) in the UK45 and the Headache Consortium46 and
the Primary Care Network47 in the USA.

These three sets of guidelines have several features in common. They use assessments of
migraine impact or disability in the initial evaluation of patients and recommend that
treatment should be tailored to the patient’s individual needs. Additionally, patients who have
disabling migraine are recommended to have access to migraine-specific therapies from the
outset. Assessing migraine impact has obvious applications in these areas of management
and future migraine guidelines are likely to incorporate impact tools significantly.

The Migraine in Primary Care Advisors’ (MIPCA) Guidelines

In 1997, MIPCA was the first organisation to advocate assessing the impact of migraine as
part of its management strategy, before clinically valid impact tools had been developed.
Fully revised and evidence-based guidelines were published in 2002,45 and updated in
2004.48 MIPCA advocates an individualised approach to care, treatment being prescribed
according to each patient’s needs. Factors considered include the nature of the patient’s
attacks, the impact of headache on the individual’s life and the demands of the patient’s
lifestyle.

The MIPCA guidelines are illustrated in Figure 6. At the initial consultation, the physician is
recommended to conduct a diagnostic assessment and to take a careful history covering the
nature of the headaches, previous treatments taken and the impact on the patient’s life.
Patients who experience up to four attacks per month are given acute therapy with a simple
analgesic, with or without an anti-emetic (for mild-to-moderate migraine) or an oral triptan (for
moderate-to-severe migraine). Rescue medication is also given for when the initial treatment
fails. Nasal spray or subcutaneous triptan formulations may be considered if the patient has
difficulties with oral therapies or requires a fast therapeutic effect due to the demands of their
lifestyle.

121
Figure 6. The MIPCA management guidelines for migraine in the UK.45

ƒDetailed history, patient education

and commitment Initial consultation
ƒDiagnostic screening and
differential diagnosis
ƒAssess illness severity
ƒAttack frequency and duration
ƒPain severity
ƒImpact (MIDAS or HIT)
ƒNon-headache symptoms
ƒPatient history and preferences

Intermittent Intermittent
mild-to-moderate migraine moderate-to-severe migraine
(+/- aura) (+/- aura)

Behavioural/ Initial treatment

complementary therapies

Rescue
Aspirin/NSAID (large dose) Oral triptan
Aspirin/paracetamol
Rescue
plus anti-emetic
Second dose/Nasal
spray/subcutaneous
triptan

If unsuccessful Follow-up
Second dose/ treatment
Oral triptan
Alternative oral triptan
Rescue
Nasal spray/subcutaneous
triptan
Symptomatic treatment

If unsuccessful
Frequent headache Migraine Consider prophylaxis +
(i.e. ≥4 attacks per month) acute treatment for
breakthrough migraine
attacks
If If unsuccessful
Chronic daily management
headache (CDH)? unsuccessful
Consider referral
Copyright MIPCA 2004, all rights reserved

If the initial therapy is unsuccessful, an alternative therapy should be provided. For patients
who fail on multiple acute therapies, and for migraine patients with four or more headaches
per month, prophylactic treatment is recommended, together with additional acute treatment
for breakthrough attacks. Migraine patients who fail on this treatment, and those diagnosed
with chronic daily headache (CDH), may require referral to a specialist physician.

Commentary

Migraine is a remarkably disabling condition, but is not always managed effectively in primary
care. Barriers to migraine care exist for both general practitioners (GPs) and patients, leading

122
to under-consultation, under-diagnosis and under-treatment. Patients often do not rate non-
headache symptoms as medical and tend not to communicate the disability they suffer. In
two studies with a total of 105 physicians in North America and Europe, only about one third
of patients were reported to volunteer information about their headache-related disability.30 It
can be difficult for the physician to go through protocols with patients. Headache diaries,
though useful, need reinforcement with other measures and GPs need to know that migraine
causes disability. In the same North American / European studies discussed above,30
physicians reported that they recorded symptoms related to diagnosis (e.g. pain intensity and
associated symptoms) rather than information on headache-related disability. Similar results
have been reported from the UK. In a recent study of 33 UK GPs, only 54% took the impact
of migraine into consideration on diagnosis as opposed to 81% who assessed family history,
93% who assessed symptomatology and 90% who assessed trigger factors (AJ Dowson,
unpublished results). Only 24% of the physicians felt that migraine always had an impact on
the patient’s life. Such assessments of impact were subjective, as none of the GPs was
aware of the MIDAS or HIT impact tools. Overall, measuring the impact of migraine is likely
to aid the initial assessment of migraine, and improve treatment delivery.

Research has shown that the impact of migraine and other headaches can be measured
using assessments of some or all of pain severity, headache frequency, limitations to work
and leisure activities, tiredness, mood alterations and cognition. Tools that address
headache impact are now available and are being brought into clinical use. The MIDAS
Questionnaire assesses disability only, as lost time in employment, unpaid work and family
and leisure activities, whereas HIT assesses a global range of impact, including pain
intensity, disability and other items, which can be tailored to the individual in Internet HIT.

Neither MIDAS nor HIT is used widely in UK general practice. MIDAS is used in specialist
practices, where it has proved to be effective in improving communication between
physicians and their patients, and as a measure of treatment outcome. However, it has not
proved as useful when tested in UK general practice, partly due to the overlapping nature of
the questions. Patients too, are unsure of it. Little or no feedback was received from Migraine
Action Association (MAA: the UK headache patient support group) members when MIDAS
was published in the MAA Newsletter and tested in pharmacies in the UK. HIT has not been
used significantly to date in UK clinical practice.

An impact tool can be recommended for the primary care of migraine and its desirable
attributes identified. It should be a self-administered form that is able to be summarised
briefly. Open questions should be used, rather than a ‘ticking boxes’ approach. The tool
should be used in the physician’s initial assessment to aid in the initiation of appropriate
management strategies. It should also be able to measure the efficacy of medications and
needs to be discriminating, e.g. impact testing may be used as a tool to measure if a
treatment is working for a patient, rather than merely if it is not. If the treatment is working,
the patient should remain on that same treatment. In practice, the impact tool should be
integrated with other assessments of migraine symptoms. Finally, impact tools need to be
refined and shown to work in general practice, where they can be used to review patient
costs and potential cost savings of treatments. From the patients’ viewpoint, it is important to
ask questions about migraine impact, but this should be taken as part of the whole clinical
judgement of patients’ needs.

Migraine impact is probably the most important headache assessment for the patient. Impact
tools should therefore be widely accessible to patients at home, in pharmacies and GPs’
waiting rooms and with nurse consultants, e.g. in well woman clinics. They should also have
positive effects on the GP’s practice, highlighting that migraine is a disabling illness and
raising its profile in primary care. However, education of both patients and physicians on
migraine impact and the use of impact tools will be necessary for their successful

123
introduction. The initial launch of impact tools is probably best achieved through headache
clinics and GPs with an existing interest in treating headache.

Both MIDAS and HIT meet many of the above desirable attributes for an impact tool suitable
for use in primary care. They are both brief, simple to use and score, and have demonstrated
scientific validity and clinical utility. HIT has greater accessibility (by design for use on the
internet and as a paper-based form) and a greater coverage of the spectrum of headache
patients, irrespective of their illness severity. MIDAS is sensitive to change and is
increasingly being used as an outcome tool in clinical studies.31,32 However, with MIDAS,
there is the possibility of patients with infrequent but disabling headache being categorised
as having minimal disability, and so being overlooked.

Recently published guidelines for the management of migraine have included assessments
of migraine impact in the evaluation of migraine severity and treatment choice, in conjunction
with traditional history taking and symptom evaluation. These guidelines advocate an
individualised approach to care, where patients are managed according to their illness
severity and lifestyle needs. Patients are educated and encouraged to play a key role in the
care decisions made. However, previously published guidelines for migraine management in
many countries, including the UK,49 USA50 and Germany51 make no mention of assessing
migraine impact in the evaluation of patients. Nevertheless, impact assessments have a
central role in the management of migraine and should be used in the development of new
guidelines for primary care in the UK.

In conclusion, the assessment of migraine impact is important in UK primary care. The use of
impact tools can improve communication between physicians and their patients, aid in the
assessment of migraine severity and the production of an individualised treatment plan in
concert with other clinical assessments, and in the monitoring of the response to therapy. In
the future, impact tools will be used in generalised headache management guidelines for the
UK. As yet, it is not possible to unequivocally recommend the optimal impact tool, but it
should be usable by GPs, pharmacists, nurses and patients, and should also be useful for
purposes of research.

Acknowledgements

This paper was presented at a meeting of the Migraine in Primary Care Advisors (MIPCA) on
14 June 2001 at the Berystede Hotel, Ascot, Berkshire, UK, sponsored by GlaxoSmithKline.
Meeting attendees were Dr Tony Bland, Chester, UK; Dr Sue Bradford, Sheffield, UK; Dr
Frances Carter, London, UK; Dr Heather Fearon, Runcorn, UK; Dr Ahmet Fuat, Darlington,
UK; Dr Trevor Rees, Solihull, UK; Ms Heather Spitall, Birmingham, UK; Ms Ann Turner,
Migraine Action Association, Northampton, UK; Dr David Watson, Aberdeen, UK. The author
thanks all the meeting attendees for their valuable contributions to the issues discussed in
this article

References
1. Breslau N, Rasmussen BK. The impact of migraine: epidemiology, risk factors, and
co-morbidities. Neurology 2001;56(Suppl 1):S4–12.
2. Micieli G. Suffering in silence. In: Edmeads J, ed. Migraine: A Brighter Future.
Worthing: Cambridge Medical Publications, 1993;1–12.
3. Osterhaus JT, Townsend RJ, Gandek B et al. Measuring the functional status and
well-being of patients with migraine headache. Headache 1994;34:337–43.
4. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity. Disability, pain intensity,
and attack frequency and duration. Neurology 1994;44(Suppl 4):S24–39.
5. Lipton RB, Stewart WF, Liberman J et al. Patterns of healthcare utilization for
migraine in England. Cephalalgia 1999;19:305. Abstract.
6. Stang PE, Von Korff M. The diagnosis of headache in primary care: factors in the

124
 agreement of clinical and standardized diagnoses. Headache 1994;34:138–42.
7. Richard A, Massiou H, Herrmann MA. Frequency and profile of migraine patients
seen by general practitioners. Semin Hop Paris 1999;75:440–6.
8. Edmeads J, Láinez JM, Brandes JL et al. Potential of the Migraine Disability
Assessment (MIDAS) Questionnaire as a public health initiative and in clinical
practice. Neurology 2001;56(Suppl 1):S29–34.
9. Dowson A, Jagger S. The UK migraine patient survey: quality of life and treatment.
Curr Med Res Opin 1999;15:241–53.
10. Lipton RB, Stewart WF. Acute migraine therapy: do doctors understand what
migraine patients want from therapy? Headache 1999;39(Suppl 2):20–26.
11. National Academy of Sciences/Institute of Medicine (NAS/IOM). Disability in America:
toward a national agenda for prevention. Washington DC: NAS Press, 1991.
12. Edmeads J, Findlay H, Tugwell P et al. Impact of migraine and tension-type
headache on life-style, consulting behaviour, and medication use: a Canadian
population survey. Can J Neurol Sci 1993;20:131–7.
13. Sakai F, Igarashi H. Epidemiology of migraine in Japan. Cephalalgia 1997;17:15–22.
14. Clarke CE, MacMillan L, Sondhi S et al. Economic and social impact of migraine. Q J
Med 1996;89:77–84.
15. Von Korff M, Stewart WF, Simon DJ et al. Migraine and reduced work performance: a
population-based diary study. Neurology 1998;50:1741–5.
16. Von Korff M, Ormel J, Keefe FJ et al. Grading the severity of chronic pain. Pain
1992;50:133–49.
17. Abu-Arefeh I, Russell G. Prevalence of headache and migraine in schoolchildren.
BMJ 1994;309:765–9.
18. Ferrari MD. The economic burden of migraine to society. PharmacoEconomics
1998;13:667–76.
19. Office of Health Economics. Compendium of Health Statistics. London: OHE, 1989.
20. Smith RC. Impact of migraine on the family. Headache 1996;36:278. Abstract.
21. Kryst S, Scherl ER. Social and personal impact of headache in Kentucky. In: Olesen
J, ed. Headache Classification and Epidemiology. New York: Raven Press Ltd, 1994;
p345–50.
22. de Lissovoy G, Lazarus SS. The economic cost of migraine. Present state of
knowledge. Neurology 1994;44(Suppl 4):56–62.
23. Von Korff MR, Stang PE. Prediction of the psychosocial and behavioral outcomes of
headache. In: Olesen J, ed. Headache Classification and Epidemiology. New York:
Raven Press Ltd, 1994; p367–71.
24. Stewart WF, Lipton RB, Simon D et al. Reliability of an illness severity measure for
headache in a population sample of migraine sufferers. Cephalalgia 1998;18:44–51.
25. Stewart WF, Lipton RB, Simon D et al. Validity of an illness severity measure in a
population sample of migraine headache sufferers. Pain 1999;79:291–301.
26. Stewart WF, Lipton RB, Kolodner K et al. Reliability of the migraine disability
assessment score in a population-based sample of headache sufferers. Cephalalgia
1999;19:107–14.
27. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of
the Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.
28. Stewart WF, Lipton RB, Kolodner KB et al. Validity of the Migraine Disability
Assessment (MIDAS) score in comparison to a diary-based measure in a population
sample of migraine sufferers. Pain 2000;88:41–52.
29. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56(Suppl 1):S20–8.
30. Holmes WF, MacGregor EA, Sawyer JPC et al. Information about migraine disability
influences physicians’ perceptions of illness severity and treatment needs. Headache
2001;41:343–50.
31. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,

125
 including the use of triptans, in Japan. Headache Care 2004;1:115–21.
32. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine,
and cluster headache with topiramate. Headache 2002;42:796–803.
33. Ware JE, Bjorner JB, Kosinski M. Practical implications of item response theory and
computerized adaptive testing. Medical Care 2000;38:S73–82.
34. Lipton RB, Sheftell F, Stewart WF et al. The MIDAS perception study: headache
characteristics and treatment at low MIDAS grades. Cephalalgia 2001;21:326–7.
Abstract.
35. Lipton RB, Stewart WF, Stone AM et al. Stratified care vs step care strategies for
migraine: results of the Disability in Strategies of Care (DISC) Study. JAMA
2000;284:2599–605.
36. Bjorner JB, Kosinski M, Ware JE Jr. Calibration of an item pool for assessing the
burden of headaches: an application of item response theory to the headache impact
test (HIT). Qual Life Res 2003;12:913–33.
37. Bayliss MS, Dewey JE, Dunlap I et al. A study of the feasibility of Internet
administration of a computerized health survey: the headache impact test (HIT). Qual
Life Res 2003;12:953–61.
38. Bjorner JB, Kosinski M, Ware JE Jr. Using item response theory to calibrate the
Headache Impact Test (HIT) to the metric of traditional headache scales. Qual Life
Res 2003;12:981–1002.
39. Pryse-Phillips W. Evaluating migraine disability: the headache impact test instrument
in context. Can J Neurol Sci 2002;29 (Suppl 2):S11–5.
40. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
41. Nachit-Ouinekh F, Dartigues J-F, Henry P et al. Use of the headache impact test
(HIT-6) in general practice: relationship with quality of life and severity. Eur J Neurol
2005;12:189–93.
42. Vuillaume De Diego E, Lanteri-Minet M. Recognition and management of migraine in
primary care: influence of functional impact measured by the headache impact test
(HIT). Cephalalgia 2005;25:184–90.
43. Bayliss MS, Kosinski M, Diamond M et al. HIT-6 scores discriminate among
headache sufferers differing in headache-associated workplace productivity loss.
Cephalalgia 2001;21:333–4. Abstract.
44. Dahlof C, Tepper S, Pryse-Phillips W et al. Comparability of scores on the Headache
Impact Test™ (HIT) and HIT-6™. Cephalalgia 2001;21:334–5. Abstract.
45. Dowson AJ, Lipscombe S, Sender J et al. New guidelines for the management of
migraine in primary care. Curr Med Res Opin 2002;18:414–39.
46. Silberstein SD, for the US Headache Consortium. Practice parameter: evidence-
based guidelines for migraine headache (an evidence-based review). Report of the
Quality Standards Subcommittee of the American Academy of Neurology. Neurology
2000;55:754–63.
47. Bedell AW, Cady RK, Diamond ML et al. Patient-centered strategies for effective
management of migraine. Springfield: Primary care Network, 2000.
48. Migraine in Primary Care Advisors. Guidelines for the management of migraine in
primary care, 2nd Edition. Guidelines 2004;22:127–31.
49. Steiner TJ, MacGregor EA, Davies PTG. Guidelines for all doctors in the diagnosis
and management of migraine and tension-type headache. British Association for the
Study of Headache 2000; www.bash.org.uk
50. Lewis TA, Solomon GD. Advances in migraine management. Cleveland Clin J Med
1995;62:148–55.
51. Diener HC, Brune K, Gerber WD et al. Therapy of acute migraine attacks and
migraine prophylaxis—guidelines of the German Migraine and Headache Society [in
German]. Nervenheilkunde 1997;16:500–10.

126
4.2

The outcome of headache management following nurse

intervention: assessment in clinical practice using the Migraine
Disability Assessment (MIDAS) Questionnaire*

Abstract

Objectives: This audit assessed the outcome of a nurse intervention strategy in patients
presenting with headache to a primary care surgery.
Methods: All patients aged 18–65 years attending the surgery who reported headache in the
previous 3 months were assessed by a nurse and completed a Migraine Disability
Assessment (MIDAS) Questionnaire and a questionnaire investigating headache features
and physician consultations. All patients were given oral and written advice on headache
management by the nurse. Patients of MIDAS Grade I/II (minimal-mild disability) with severe
or frequent headaches and all Grade III–IV (moderate-severe disabling headache) patients
were also offered an appointment with a headache specialist physician. Patients were
reviewed after 3 and 6 months with the same questionnaires.
Results: A total of 195 patients took part in the study. At baseline, 136 (69.7%) were MIDAS
Grade I/II and 59 (30.3%) were Grade III/IV. Compared with baseline, patients reported
significant mean reductions at 6 months in: total MIDAS scores (5.9 versus 9.6; p = 0.014);
headache frequency (MIDAS A score: 7.0 versus 12.6; p = 0.009); headache severity
(MIDAS B score: 5.0 versus 6.0; p = 0.003); and physician consultations for headache (3-
month period: 0.05 versus 0.30; p = 0.05).
Conclusions: Advice on headache management given by a nurse can potentially lead to
significantly improved patient outcomes. MIDAS appeared to be a sensitive outcome
measure for reduction in disability in headache sufferers. These results warrant further
investigation in randomised, controlled clinical studies.

* Edited from the full published article: Dowson AJ, Salt R, Kilminster S. The outcome of
headache management following nurse intervention: assessment in clinical practice using
the Migraine Disability Assessment (MIDAS) Questionnaire. Headache Care 2004;1:177–81.

127
Introduction

Headache management is currently suboptimal in primary care. Migraine patients remain

under-diagnosed and under-treated, with many patients not consulting a physician for care,1
and this state has improved little over the past decade or more.2 Patients with chronic daily
headache (CDH) are frequently not managed in primary care, but are referred to specialist
headache clinics.3 Sufferers of tension-type headache rarely consult a physician for care.4
Recently published guidelines for migraine management in the UK advocate a primary care
team approach to headache management, utilising practice nurses and other healthcare
professionals, as well as the physician.5 A UK pilot study has suggested that headache
education provided by nurses can be effective in reducing the severity of headaches.6
Following advice given by a primary care nurse to a group of headache sufferers, the
severity of headache and associated disability declined over a 6-month period compared to a
control group who did not receive this advice. In addition, a retrospective study in the USA
indicated that the provision of headache advice by Advanced Practice Nurses resulted in
improved quality of life in adult migraine patients.7

The practice nurse may therefore have a key role to play in the management of headache
patients. This audit was set up to investigate the outcome of a nurse intervention strategy in
patients presenting with headache to a primary care clinic. The main outcome measure used
was the Migraine Disability Assessment (MIDAS) questionnaire. MIDAS assesses the
disability associated with headache, by measuring the days lost over a 3-month period in
employment or education, unpaid work and family and leisure activities.8 MIDAS is a reliable
and valid measure of disability,9 and has been shown to be sensitive to changes in outcome
in clinical studies.10,11 MIDAS is scored as the number of days of missed activity and scores
are categorised as Grade I (score 0–5; minimal or infrequent disability), Grade II (score 6–10;
mild or infrequent disability), Grade III (score 11–20; moderate disability), or Grade IV (score
over 20; severe disability).9

Patients and methods

All patients aged 18–65 years attending the primary care Merrow Park Surgery in Surrey, UK
for any reason were interviewed. Those who reported headache in the previous 3 months
were eligible for inclusion in the audit. Patients were assessed by a primary care nurse with
experience in headache (RS) and completed a MIDAS Questionnaire and a questionnaire
investigating headache features and physician consultations (Table 1).

Table 1. Nurse-administered questionnaire investigating patients’ headache features and

physician consultations.

1. In the last 3 months how many times have you consulted your GP?
2. How many of these consultations were specifically about headaches?
3. Do you think your headaches affect your mood?
4. Do you think your headaches make you tired or lethargic?
5. Do you feel an improvement in your headaches compared with 3 months ago?
6. Do you feel that overall you need to see your GP less now?

The design of the audit is illustrated in Figure 1.

128
Figure 1. Design of the audit: management decisions during the study.
IHS = International Headache Society.

Patients
screened n = 676

Nurse advice
Advice sheet n = 195
Websites

MIDAS

MIDAS Grade I and II MIDAS Grade I and II (frequency

(frequency <2/month >2/month or high intensity >5)
and low intensity <5) MIDAS Grade III and IV

n = 136 n = 59

Offer to see
physician
No further advice (IHS diagnosis,
examination,
treatment)

Refer if needed

All patients were given oral and written advice on headache management by the nurse,
covering trigger factors, use of acute medications, treatment targets, prevention of frequent
attacks and consulting with a physician if chronic daily headache was suspected. Patients
were referred to the UK patient support group, The Migraine Action Association, for further
information. Patients with MIDAS Grades I–II (minimal or mild disability), with infrequent and
low intensity attacks, received no further specific headache management advice. However,
patients with MIDAS Grades III–IV (moderate-severe disabling headache), plus those with
MIDAS Grades I–II and either frequent or severe intensity attacks, were offered an
appointment with a headache specialist physician (AJD), for diagnosis and appropriate
interventions. Patients were further reviewed by the nurse after 90 and 180 days, when they
again completed the MIDAS questionnaire and the questionnaire investigating headache
features and physician consultations (Table 1).

Endpoints from the audit included:

• The total MIDAS score (days lost from daily activities over the previous 3-month
period).
• MIDAS Question A score (headache frequency over the previous 3-month period).
• MIDAS Question B score (average headache severity over the previous 3-month
period, using a 10-point scale where 0 = no pain and 10 = pain as bad as it can be).
• Patients’ headache features and physician consultation (from the questionnaire in
Table 1).

129
Summary statistics were calculated as changes from baseline at 90 and 180 days for each
item and were analysed by t-tests for dependent samples. The definition of significance was
set at p ≤ 0.05. Additional non-parametric analyses, using Wilcoxon rank tests, were also
conducted on these data. This was required, as MIDAS data are not necessarily normally
distributed due to floor effects (patients can score 0 on the questionnaire).12

Results
Baseline characteristics
A total of 676 patients were interviewed by the nurse, of whom 477 did not report headaches,
195 (28.8%) had headaches and took part in the audit, and four declined to participate. At
baseline, 136 patients (69.7%) were MIDAS Grade I/II with low headache severity and
frequency and 59 (30.3%) were Grade I/II with high headache severity and frequency or
Grade III/IV. These 59 patients were offered an appointment with the headache specialist,
but only 17 actually had the consultation. One hundred and ninety four patients had a
recorded MIDAS score at baseline, and this number decreased to 87 and 70 patients at 90
and 180 days, respectively.

Changes in outcome scores at 90 and 180 days

Figures 2–5 illustrate the changes from baseline in the audit endpoints after 90 and 180
days. The MIDAS items all cover the 3-month period immediately preceding completion of
the questionnaire. Compared with baseline, patients reported significant mean reductions in
total MIDAS score at 180 days (5.9 versus 9.6 days; p = 0.014, Figure 2). Compared with
baseline, patients reported significant mean reductions in headache frequency (MIDAS
Question A score) at 180 days (7.0 versus 12.6 headaches; p = 0.009, Figure 3). Compared
with baseline, patients reported significant mean reductions in headache severity (MIDAS
Question B score) at 180 days (5.0 versus 6.0; p = 0.003, Figure 4). Similar results were
reported when the data were analysed non-parametrically.

Compared with baseline, patients reported no significant changes at 90 and 180 days in the
total number of times they consulted a GP, but did report significant mean reductions in the
number of consultations for headache in the previous 3-month period at 180 days (0.05
versus 0.30 consultations; p = 0.05, Figure 5). Similar results were reported when the data
were analysed non-parametrically.

130
Figure 2. Total MIDAS scores (total activity days lost due to headache in the previous 3
months) at baseline (n = 194), and 90 (n = 87) and 180 (n = 70) days. Results are presented
as means, ± SE (boxes) and ± 1.96 SE (bars).

131
Figure 3. MIDAS Question A scores (number of days with headache in the previous 3-month
period) at baseline (n = 192) and at 90 (n = 86) and 180 (n = 68) days. Results are presented
as means, ± SE (boxes) and ± 1.96 SE (bars).

132
Figure 4. MIDAS Question B scores (average headache severity in the previous 3 months,
where 0 = no pain at all, and 10 = pain as bad as it can be) at baseline (n = 194) and at 90 (n
= 87) and 180 (n = 68) days. Results are presented as means, ± SE (boxes) and ± 1.96 SE
(bars).

133
Figure 5. Average number of consultations for headache per patient in the previous 3-month
period: data for baseline (n = 194), and 90 (n = 85) and 180 (n = 70) days. Results are
presented as means, ± SE (boxes) and ± 1.96 SE (bars).

Throughout the study, approximately 90% of patients self-reported that headaches affected
mood and resulted in feelings of tiredness or lethargy. However, patients reported some
improvement in headaches and the need to see their GP at 90 and 180 days compared with
baseline (Table 2). However, no significant changes were reported in any of these items
during the study (Table 2).

134
Table 2. Patient perceptions of headaches at baseline, and 90 and 180 days: percentage of
patients reporting ‘Yes’ to the questions.

Percentage of patients
Question Day 0 Day 90 Day 180
Do you think your headaches affect your mood? 91.2 87.8 90.0
(NS) (NS)
Do you think your headaches make you tired or 90.2 87.1 93.1
lethargic? (NS) (NS)

Do you feel an improvement in your headaches 38.2 38.3 45.6

compared with 3 months ago? (NS) (NS)

Do you feel that overall you need to see your GP NA 48.5 62.5
less now? (NS)

NA = Not applicable
NS = not significant

Discussion

Results from this audit illustrate the positive role that the practice nurse can have in the
management of headache in primary care. Recent UK guidelines provide clear roles for
nurses in the screening and follow-up of migraine patients.5,13 Recommendations state that
nurses should undertake information dissemination and recording of data using headache
history and impact questionnaires and headache diaries. Previous, small studies have
indicated that the provision of information by nurses can have a positive effect on headache,
even when no other interventions are provided.6,7

Patients who took part in the audit were markedly affected by headaches, with approximately
90% reporting that headaches affected mood and caused tiredness. The baseline mean
MIDAS score was 9.6, indicating mild to moderate headache-related disability. On average,
patients experienced about 4 headaches per month. These patients had significantly
improved headache outcomes at 180 days compared with baseline, in terms of headache-
related disability, headache frequency and severity. Additionally, they perceived an
improvement in headache and reported a reduced need for consulting a physician for
headache. As very few patients consulted a physician during the audit, this improvement can
be attributed with a reasonable degree of confidence to the advice given by the nurse on
headache management.

MIDAS has been shown to be an accurate measure of headache severity9 and to be

sensitive to interventions, with drug treatments resulting in significant reductions in total
MIDAS score.10,11 In this audit, the mean MIDAS score was significantly reduced after 180
days, indicating headache improvement, which in turn correlated with the patients’ self
reporting of headache improvement and a reduced need to consult with a GP. MIDAS
therefore seems to be a sensitive outcome measure for headache sufferers.

The audit does have certain shortcomings. As no control group was included, the reductions
in MIDAS and other items could be due to a placebo effect. However, there was good
evidence of significant improvements in MIDAS items and significant reductions in headache
consultations, as parametric and non-parametric analyses exhibited the same patterns of
results. Also, the numbers of patients were reduced due to drop outs as the visits proceeded,
which may have caused some bias. The expected bias would result in an improvement. This
is because, psychologically, patients are more likely to continue with an investigation if they
are generally feeling well or better. Overall, this was an interesting exploratory study that

135
provided positive results. It paves the road for future randomised, controlled studies
investigating the role of nurses in headache management.

In conclusion, this audit indicated that advice on headache management given by a nurse
could potentially lead to significantly improved patient outcomes. MIDAS appeared to be a
sensitive outcome measure for reduction in disability in headache sufferers.

Acknowledgement

This audit was sponsored in part by an unrestricted educational grant from AstraZeneca.

References
1. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
2. Lipton RB, Scher AI, Steiner TJ et al. Patterns of health care utilization for migraine in
England and the United States. Neurology 2003;60:441–8.
3. Dowson AJ. Analysis of the patients attending a specialist UK headache clinic over a
3-year period. Headache 2003;43:14–8.
4. Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and diagnosis of migraine in
patients consulting their physician with a complaint of headache: data from the
Landmark Study. Headache 2004;44:856–64.
5. Dowson AJ, Lipscombe S, Sender J et al. New guidelines for the management of
migraine in primary care. Curr Med Res Opin 2002;18:414–39.
6. Main A, Abu-Saad H, Salt R et al. Management by nurses of primary headache: a
pilot study. Curr Med Res Opin 2002;18:471–8.
7. Allen LL, Haririfar M, Cohen J et al. Quality of life and locus of control of migraineurs.
Clin Excell Nurse Pract 2000;41–9.
8. Stewart WF, Lipton RB, Kolodner K et al. Reliability of the migraine disability
assessment score in a population-based sample of headache sufferers. Cephalalgia
1999;19:107–14.
9. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56 (Suppl 1):S20–8.
10. Otsuka N, Sakai F, Iigaya M et al. MIDAS — assessments of migraine management
including the use of triptans in Japan. Headache Care 2004;1:115–21.
11. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine,
and cluster headache with topiramate. Headache 2002;42:796–803.
12. Dowson AJ, Bundy M, Kilminster SG. Emotional function with tension-type headache,
migraine and chronic daily headache. Headache Care 2005; manuscript in
preparation.
13. MacBean H, Leech J, Dungay J et al. Management of migraine by nurses. Practice
Nursing 2004;15:346–50.

136
4.3

Outcome measures compared: Headache Impact Test (HIT) and

Short Pain Inventory© (SPI)*

Summary

The psychometric properties of the Headache Impact Test (HIT-DYNA®) were compared with
the 17-item Short Pain Inventory (SPI©) in 75 patients presenting to primary care with
tension-type headache, migraine or chronic daily headache. Also compared were the abilities
of the SPI and HIT in terms of discriminating headache severity and diagnosis.

The HIT score correlated 0.28 with the severity of the headaches whereas the SPI correlated
0.76 on its Total Mood Disturbance score. The correlations of the HIT items with pain
severity never rose above 0.26 whereas the correlations of the SPI items were all typically
about 0.7. The SPI significantly discriminated with headache severity levels (none, mild,
moderate, severe, extreme), with t values of 2–20. The HIT scores were far less powerful at
discriminating severity, with t test values from 2–5. HIT scores at screening significantly
discriminated with diagnosis, with the resolution between tension-type headache and chronic
daily headache reaching 1/100 million. The SPI summary scores and subscales did not
discriminate with diagnosis. These differences between HIT and SPI were confirmed by
factor analysis.

The results showed the HIT to be poorly related to the severity of the pain but very closely
related to the diagnostic label. In contrast to this, the SPI was very closely related to severity
but not related to diagnosis.

* Summarised and edited from the published article: Kilminster SG, Dowson A, Bundy M.
The Headache Impact Test® and the Short Pain Inventory©: Outcome measures compared.
Int J Pharm Med 2003;17:23–32.

137
Introduction

Pain outcome measures for clinical trials are broadly of two types; scales that measure the
functional impact of the specific disease entity, and scales that index the general pain
experience. The Oswestry Back Pain Questionnaire,1 and the Headache Impact Test (HIT®)2
typify questionnaires of the former type whereas the McGill Pain Questionnaire (MPQ)3 and
the Short Pain Inventory (SPI©)4 represent the latter. Whatever methods used, and no matter
how clever and complex their psychometric development, tests must be reliable and valid
under specified conditions. The basic requirements include published internal reliability in the
form of Cronbach alpha, split-half reliability, and test-retest reliability under specified time
intervals. Additionally, the test should state clearly if it measures the here and now (state),
the general (trait), or the retrospective (over the preceding weeks or months). The test should
be able to discriminate and resolve fine detail on a scale of what it purports to measure
(validity). The test data should show empirically whether it is a valid diagnostic or severity
measure.

The HIT is a retrospective (4-week window) tool to measure the impact headaches have on a
person's ability to function in their everyday activities.5,6 The Dynamic Headache Impact Test
(HIT-DYNA®) was designed to be administered via the Internet and was developed from
questions included in four widely-used measures of headache impact (the Migraine-Specific
Quality of Life Questionnaire [MSQ]; the Headache Disability Index [HDI]; the Headache
Impact Questionnaire [HImQ] and the Migraine Disability Assessment Questionnaire
[MIDAS]) using item response theory psychometric methods.7 The MSQ, HDI, HImQ and
MIDAS are all well referenced and psychometrically robust tests.8–11 From a dynamic
computerised HIT scoring system available on the web, in a few minutes the programme
yields a description of the impact headaches are having on a patient’s life and their ability to
function.12 Many of the items on the HIT are clearly retrospective, relating to the previous 4
weeks. In general, only five questions are required to obtain a HIT score.

The SPI is a prospective 17-item self-rating questionnaire dedicated to measuring any pain,
not only headache. Patients rate each item on a five point Likert scale: not at all, a little,
moderately, very much so and extremely. Five subscales measuring sedation, social
interaction, anxiety, sadness and anger have been fine-tuned by emphasising items that
were specific to pain. The SPI takes about one minute to complete and assesses the
headache as it is at the time of completion. A scale of Total Mood Disturbance (TMD) can be
summated from the SPI which measures the subtle mood changes that specifically covary
with mild to severe pain. As the physical severity of pain worsens so generally does each of
the emotional components.4,13,14 If a patient’s emotional score is exaggerated or flattened
relative to the normative values, then a coping score can also be given as to how well a
patient is coping with their pain, relative to others at the same physical level of pain. This
can be very useful in the individual management of patients. Testing showed the SPI to be
psychometrically robust.4,13,14

The psychometric properties of the computerised HIT (HIT-DYNA) were compared with the
17-item SPI in patients presenting to primary care with headache. Also compared were the
severity of the headaches and the diagnostic labels given by a specialist neurologist, with the
ability of the SPI and HIT in terms of discriminating between diagnoses and headache
severity.

138
Methods
Patients
Patients who took part in the study were attending a general practice headache clinic at
Cranleigh Health Centre, Surrey, UK (Dr Bundy). The study was approved by the South West
Surrey Local Research Ethics Committee.

Design
This was a within-group comparative study carried out at one general practice (Cranleigh
Health Centre headache clinic) comparing the discriminant validity of the computerized HIT
(HIT-DYNA) and the SPI. The questionnaires were administered by a psychologist (SGK). A
specialist neurologist (AD) provided a diagnosis for each patient based on the International
Headache Society (IHS) criteria pertaining at the time,15 and on his own clinical experience.
The neurologist or his nurse supervised the Internet testing for the HIT evaluation. The SPI
was self-completed at home at the following times:

1. Practise session at the practice. Day 0

2. Within 1 hour after starting headache or on waking. Day 1
3. Just after pain is worst (peak) Day 1
4. 1 hour after headache has gone or on waking Day 1
5. 24 hours after headache started. Day 2
6. Evening of Day 2 about 8 pm Day 2
7. Days 3–7 inclusive about 8 pm Days 3-7

Explanation on how to complete the questionnaires was kept to a minimum and all spoiled
forms were entered into the analysis on an intention-to-treat basis. This was done so as to
reduce the administrator bias and evaluate the SPI as it would be used realistically in clinical
screening settings.

Statistical methods
Analysis was by discriminant function analysis, simple t-tests, and test-retest Pearson
product moment correlation coefficients. Statistical analysis was carried out on a Power
Macintosh G4/400/200 machine with Statistica V4. All data points were double-validated
against source forms. The analysis was carried out by a medical statistician (SGK).

For the discriminant validity analysis, the SPI recorded headache severity (‘pain right now’)
as no pain (0), mild pain (1), moderate pain (2), severe pain (3), and extreme pain (4). Thus
pain groups were divided by pain severity as screening samples. The SPI raw scores were
used for the summary and components. The diagnostic labels derived from the IHS
classification were dummy-coded. Summary and derivative SPI measures were then used as
dependent variables in simple t-tests. Discriminant function analysis is the same as simple t-
tests in this case where there is only one independent and one dependent variable. Factor
analyses with orthogonal and Varimax rotation of the principal components was carried out to
examine the structure and redundancy of the two tests.

Results
Demography
A total of 75 patients were recruited to the study, diagnosed with episodic TTH (15 patients
[20%]), migraine (41 patients [54.7%]) and CDH (19 patients [25.3%]). The mean age of the
patients was 44.3 (SD 9.8). The majority (73, 97.3%) were women. At screening, extreme
pain (4) was reported by two patients (2.6%), severe pain (3) by six (8.2%), moderate pain
(2) by eight (10.5%), mild pain (1) by 26 (34.2%) and no pain by 33 (44.7%). Forty two
patients returned the completed questionnaires within 6 weeks Six of these patients had TTH
(14.3%), 23 had migraine (54.8%) and 13 had CDH (31.0%). Drop outs (n = 33) were mostly
due to patients not experiencing a headache during the 6-week window.

139
Differentiation of pain severity by HIT and SPI
Figure 1 shows the student t tests computed for headache severity comparisons of all
headaches for the summary SPI TMD and HIT scores. As headache severity increased from
none (0) to extreme (4) there was a highly significant increase in SPI TMD ranging from 4–
20. There was a relative diminishment of t for 0 v 4 compared to 0 v 1, 0 v 2 and 0 v 3. This
is probably explained by the low number of patients with extreme pain (n = 2). Nevertheless,
the results showed the SPI TMD score had excellent discrimination at resolving the fine
detail differences in pain severity. The HIT scores were far less powerful at discriminating
headache severity, with t test values from 2–5 only.

Figure 1. Discriminant t-values of the SPI Total Mood Disturbance (TMD) and HIT at
resolving the differences in pain severity (0 = no pain; 1 = mild pain; 2 = moderate pain; 3 =
severe pain; 4 = extreme pain).

Student t
22
HIT
TMD
18

SPI Total Mood Disturbance

14

10

HIT
2

-2
0 0v1 0v2 0v3 0v4 1v2 2v3 4v5
HEADACHE SEVERITY COMPARISONS

The best correlation of the HIT with headache severity was 0.28. The relationship was much
stronger with all of the SPI subscales and the correlation of the TMD score was 0.76.The
nature of these relationships is also displayed in Figures 2 and 3. Similar results to the SPI
TMD score were observed for SPI social interaction, sedation, sadness, anger and anxiety
sub-scores.

140
Figure 2. Dynamic HIT scores relationship to headache pain severity (0 = no pain; 1 = mild
pain; 2 = moderate pain; 3 = severe pain; 4 = extreme pain)..

HIT Box & Whisker plot

66
+1.96*SE
64 +SE
Mean
62 -SE
-1.96*SE
60

58

56

54

52

50
0 1 2 3 4
HEADACHE PAIN SEVERITY

Figure 3. SPI Total Mood Disturbance (TMD) scores relationship to headache pain severity
(0 = no pain; 1 = mild pain; 2 = moderate pain; 3 = severe pain; 4 = extreme pain)..

SPI: TOTAL MOOD DISTURBANCE (TMD)

55
+1.96*SE
50 +SE
Mean
45
-SE
40 -1.96*SE

35

30

25

20

15

10
0 1 2 3 4
HEADACHE SEVERITY

141
Differentiation of diagnosis with HIT and SPI
There were originally four main groups diagnosed by the neurologist: chronic daily headache
(CDH), high-frequency migraine (MIG-H), low-frequency migraine (MIG-L) and tension-type
headache (TTH). However this was simplified into CDH, TTH and total migraine (MIG)
groups due to sample size limitations.

The study used t tests to see if the HIT or SPI scores could reveal a statistically significant
difference between the neurologist-validated diagnoses of CDH, TTH and migraine on the
initial screening visit. There was significant relationship (p < 0.05) between the HIT score and
diagnosis (Figure 4), but this was not observed with the SPI.

Figure 4: HIT relationship with diagnosis (CDH = chronic daily headache; MIG = migraine;
TTH = episodic tension-type headache).

HIT Box & Whisker plot

68
+1.96*SE
+SE
62
Mean
-SE
56 -1.96*SE

50

44

38

32
CDH MIG TTH
DIAGNOSIS

Factor structure of the outcome measures and diagnosis

The independence of the HIT and SPI were checked by Quartimax and Varimax factor
analyses. Various analyses gave strikingly the same factor structure. The factor structure
revealed that the major variance was accounted for by the SPI Total Pain Disturbance. All
the SPI mood subscales also loaded highly on the first factor as did the physical severity of
the headache, with an explained variance of 58%. The HIT score and the diagnosis (even if it
included high or low migraine frequency) did not relate to the Total Pain Disturbance.

The second smaller factor which accounted for only 22% of the variance was a diagnostic
variable and loaded highly with the HIT.

Discussion

The results of this study indicate that the HIT is very good at measuring the diagnosis and is
superior in this to the SPI. The SPI is very good at indexing the severity of the pain
experience, while the HIT is relatively poor at this. These data show the HIT and SPI to be
complementary in measuring headache.

142
Clinicians should be vigilant in conceptualising the difference between a diagnostic and a
severity scale. The SPI measures the mood disturbance and social interaction associated
with the physical severity of pain. This 17-item tool is a state-dependent test which measures
the here and now. It does not rely upon the retrospective recollections of the past as do tests
like the MIDAS and HIT. The SPI measures the pain now; it does not enquire as to the cause
of the pain but into its effects. Beyond this, the SPI does not have intentions to give causal
statements about the medical or biological derivatives of pain.

The HIT is a tool to measure the impact headaches have on a person’s ability to function at
work, at home, at school and in social situations. It only purports to measure headache as an
outcome measure and clearly states it is not a diagnostic. As a retrospective view of the
patient it probably does measure headache outcome, but it is a blunt measure of severity. It
is perhaps best used before long term anti-migraine treatments are initiated and again at the
end of an extended treatment session. However, it cannot be used for point assessments of
efficacy.

The HIT did not correlate well with headache severity nor was it as powerful as SPI at
resolving differences in the fine gradations of headache severity. The SPI was more closely
correlated with headache severity and each of the subscales showed excellent discrimination
between the small changes in severity (Figure 1). The HIT does not capture emotional or
sedation status of the patient, both of which are important aspects to the patient. The SPI is
clearly not a diagnostic tool, but is a measure of headache severity. However, the HIT scores
were closely associated with the headache diagnosis at screening. Various factor analyses
were carried out to examine the factor structure and redundancy of the SPI and HIT. Two
factors were identified
1) Headache severity, best measured by the SPI
2) Diagnosis, best measured by a neurologist or the HIT.

The study has limitations, mostly related to the relatively small numbers of patients analysed.
For example, in the analysis of discriminant t values (Figure 1) there was a relative
diminishment of t for none versus extreme headache as compared to none versus mild, none
versus moderate and none versus severe, where the relationship was approximately linear.
This was probably explained by the low number of patients with extreme pain (n = 2). It is
therefore probably best to consider the statistical analyses as exploratory, requiring
confirmation in a large study with equivalent numbers of patients in each severity group.

The HIT does discriminate between headache subtypes very well and, should a neurologist
be not available at screening, this is a rather good diagnostic test. The SPI is not a diagnostic
test but is a very discriminating outcome measure of headache severity and the
consequential emotional impact. Accordingly, the data support using the HIT and SPI in
different ways. As outcome measures, the HIT and SPI are very different and in some sense
complementary. Clinical trialists should take care choosing and using their outcome
measures appropriately.

References

1. Puranik S, Fozard J, Paremain G et al. A randomised single blind study to evaluate

the effects of Action Potential Simulator Therapy compared with placebo in patients
with chronic back pain. Pain Clinic 2002;14:69–73 .
2. www.amIhealthy.com and www.qualitymetric.com
3. Melzack R. The McGill Pain Questionnaire: Major properties and scoring methods.
Pain 1975;1:277–91.

143
4. Kilminster SG, Power M, Fozard JR. Survey of pain in two medical and dental clinics,
with non patient controls using the Short Pain Inventory. Int J Pharm Med
2000;14:137–47.
5. www.headachetest.com
6. Bayliss MS, Dewey JE, Dunlap I et al. A study of the feasibility of Internet
administration of a computerized health survey: the headache impact test (HIT). Qual
Life Res 2003;12:953–61.
7. Ware JE, Bjorner JB, Kosinski M. Practical implications of item response theory (IRT)
and computer adaptive testing – a brief summary of ongoing studies of widely used
headache impact scales. Med Care 2000;38 (Suppl II):II-73–II-82.
8. Jhingram P, Osterhaus JT, Miller DW et al. Development and validation of a
migraine-specific quality of life questionnaire. Headache 1998;38:295–302
9. Jacobson GP, Ramadan NM, Norris L et al. Headache Disability Inventory (HDI):
Short-term test-retest reliability and spouse perceptions. Headache 1995;35:534–9.
10. Stewart WF, Lipton RB, Simon D et al. Validity of an illness severity measure of
headache in a population sample of migraine sufferers. Pain 1999;79:291–301.
11. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of
the Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.
12. www.amihealthy.com/Static/HITInfo.asp
13. Kilminster SG, Mould P. Comparison of the internal reliability and validity of the McGill
Pain Questionnaire and Short Pain Inventory. Int J Pharm Med 2002;16:87–95.
14. Kilminster SG, Mould GP. Comparison of diclofenac spray and gel on knee joints of
patients with osteoarthritic pain. Clin Drug Invest 1999;18:345–54.
15. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias, and
facial pain. Cephalalgia 1988;8 (Suppl 7):1–96.

144
4.4

Development and validation of the Headache Diagnostic Screening

Questionnaire (DSQ): a new questionnaire for the differential
diagnosis of headache for use in primary care*
Abstract

This prospective, open investigation was set up to validate the 8-item Headache Diagnostic
Screening Questionnaire (DSQ), containing questions screening for possible sinister
symptoms, and evaluating headache impact, frequency and duration, the use of symptomatic
medications and the presence of migraine aura symptoms. Patients completed the DSQ on
their own and with a pharmacist’s help. The same pharmacist, a GP and a headache
specialist conducted a differential headache diagnosis according to their usual practices.
Diagnoses were deduced from the DSQ using an algorithm by a healthcare professional,
who was blind to the other diagnoses. The primary endpoint was the sensitivity of the DSQ,
measured as the proportion of correct diagnoses assessed from the patient, pharmacist and
GP responses, compared with the gold standard of specialist diagnosis, for each headache
subtype. Eighty seven patients (80.5% women, mean age 52.9 years) completed the study,
all completing the questionnaire without help within a few minutes. Most patients were
diagnosed by the specialist with migraine (with or without aura) or chronic daily headache
(with or without medication overuse headache [MOH]). The overall sensitivity of the patient-
completed DSQ was good (59.3%) and similar to the pharmacist-completed DSQ (66.6%)
and GP diagnosis (59.4%). The accuracy of the DSQ-derived diagnoses was greatest for
total migraine, migraine with aura, MOH and migraine without aura. The accuracy of the
pharmacists’ headache diagnosis was improved markedly when they used the DSQ,
particularly with regard to diagnosing migraine with and without aura and MOH. In
conclusion, the Headache DSQ is a brief, rapid to use and accurate diagnostic screening
questionnaire for headache, and is especially sensitive for migraine and MOH. It can be
recommended for screening new headache patients at and below the primary care level.

* Edited from the published article: Dowson AJ, Turner A, Kilminster S et al. Development
and validation of the headache Diagnostic Screening Questionnaire (DSQ): a new
questionnaire for the differential diagnosis of headache for use in primary care. Headache
Care 2005;2:111–18.

145
Introduction

Despite the many recent advances in drug therapy, management of migraine remains
suboptimal in primary care. Patients frequently do not consult with a physician for treatment,
and physicians in turn often fail to make a correct diagnosis and prescribe appropriate
treatments.1,2 The result is that sufferers end up receiving little medical care, relying on over-
the-counter medications, which are often ineffective.3 This situation has not improved
markedly over the past decade or so.2,4

Improving the differential diagnosis of headache is one potential way of improving headache
management. The International Headache Society (IHS) publishes criteria for headache
diagnosis that are comprehensive, but too lengthy for everyday clinical use.5 Recently, brief
screening questionnaires for migraine have been developed, which are accurate and rapid
and easy to use, but are specific for migraine only.6,7 What is needed in primary care is a
brief, easy to use questionnaire which screens for the constellation of headache subtypes.
The diagnosis can then be confirmed if necessary by additional questioning.

The UK Migraine in Primary Care Advisors (MIPCA) primary care group have developed
evidence-based guidelines for migraine and chronic headache management in primary
care.8,9 As part of these initiatives, a 4-item diagnostic questionnaire was developed to
screen patients for the common headache subtypes of migraine with and without aura,
episodic tension-type headache (TTH), chronic daily headache (CDH) and medication
overuse headache (MOH), while features of potential secondary (sinister) headaches and
cluster headache are elicited by additional questioning. Figure 1 shows the questionnaire
and the diagnostic algorithm derived from it. The questions are all based on the IHS
classification criteria5 and other evidence sources:
1. Episodic TTH typically has low impact on patients’ daily activities, while migraine and
CDH have high impact. Studies have indicated that any high-impact, episodic
headache can be given a default diagnosis of migraine.10
2. Chronic headaches are defined as those occurring on > 15 days per month, while
episodic headaches occur on ≤ 15 days per month.5 Chronic headaches of duration > 4
hours are generally classified as CDH, while shorter-duration headaches may be
classified in several subtypes (including cluster headache), depending on their duration
and presenting symptoms.5 CDH is defined arbitrarily as any headache of duration > 4
hours that occurs on > 15 days per month. In the new IHS classification, these
headaches equate to chronic migraine, chronic TTH, hemicrania continua and new
persistent daily headache.5 In practice, patients are often seen with symptoms of mixed
chronic migraine and chronic TTH.
3. CDH can be categorised as MOH if it is associated with the overuse of symptomatic
medications (e.g. analgesics, ergots or triptans) on ≥ 10 days per month.5
4. Attacks of migraine with aura are associated with reversible sensory symptoms (mostly
visual) which occur immediately before the headache starts.5

146
Figure 1. The MIPCA diagnostic screening questionnaire for headache and its associated
diagnostic algorithm.8

Exclude sinister Patient presenting

Headache with headache

Q1. What is the impact of the headache

low on the sufferer’s daily life and activities?
Episodic TTH
High
Q2. How many days of headache
Migraine/CDH does the patient have every month?
> 15 ≤ 15

Consider short-lasting Migraine

CDH chronic headaches

Q3. For patients with CDH, on how many Q4. For patients with migraine, does the
days per week does the patient take attack start with reversible homonymous
symptomatic medications? visual symptoms with or without
unilateral sensory symptoms?
<2 ≥2
Yes No
No medication Consider medication
overuse With aura Without aura
overuse

In an initiative to improve headache diagnosis, MIPCA and the Migraine Action Association
(MAA: the UK patient support group) have jointly developed a checklist designed for use by
people with headaches that aims to help the GP diagnose and manage the presenting
headache subtype. This checklist has been modified slightly for use by pharmacists to help
them with the diagnosis and processing of patients with headache. However, the first eight
questions for the patient and pharmacist questionnaires are identical and act as a diagnostic
screen. The questions are closely based on the MIPCA 4-item questionnaire. This diagnostic
screen needs to be validated before it can be recommended for use in everyday clinical
practice. The objective of the current investigation was to validate the 8-item Headache
Diagnostic Screening Questionnaire (DSQ), comparing data from completions of the
questionnaire by patients and pharmacists with a conventional GP diagnosis and the gold
standard of diagnosis by a headache specialist.

Patients and methods

Patient population
The aim was to enrol 100 patients with migraine, episodic TTH, CDH (with or without MOH)
and cluster headache. Patients living local to the Surrey area were recruited from the MAA
database and from those attending Dr Dowson’s own specialist clinic patients. Patients were
men or women who had at least a 1-year history of headache. Patients were excluded if they
had a history of significant psychiatric or any other major illness, or a history of abuse of
alcohol or other recreational drugs. All patients provided their written consent before
participating in the investigation, by signing an informed consent form.

Design of the investigation

This was an open, prospective, single-visit investigation, conducted in a single centre in the
UK. The investigational design is shown in Figure 2.

147
Figure 2. Design of the investigation.

Patient completes Patient

headache DSQ
validation

1:1 assignment
Primary care physician Pharmacist conducts
conducts differential diagnosis and patient
diagnosis completes headache
DSQ with the pharmacist
Pharmacist/GP
Pharmacist conducts validation
diagnosis and patient Primary care physician
completes headache conducts differential
DSQ with the pharmacist diagnosis

Headache specialist Confirmatory

conducts differential diagnosis
diagnosis

Patients completed the DSQ twice, once on their own and once in conjunction with a
pharmacist. Patients also underwent a differential diagnosis procedure with the pharmacist
and a GP. Alternate patients saw the pharmacist first then the GP, or the GP first and then
the pharmacist, in a non-randomised sequential fashion. Finally, a confirmatory diagnostic
procedure with a headache specialist was conducted. Table I shows the 8-item Headache
DSQ, and the algorithm used to define the diagnosis that is deduced from it.

148
Table 1. The MIPCA/MAA 8-item headache diagnostic screening questionnaire (DSQ).

1. Are all the headaches usually similar in their symptoms, frequency and duration?
Yes / No
2. Have you had headaches for longer than 6 months?
Yes / No
3. Are you aged between 5 and 50 years?
Yes / No

4. Does the headache interfere to a noticeable extent with your normal daily life (work,
education and social activities)?
Yes / No
5. On average, how many days with headache do you have per month?
Less than 1 / 1 / 1–4 / 5–15 / 15–30 / Every day
6. On average, how long do your headaches last?
Less than 15 minutes / 15 minutes to 1 hour / 1–2 hours / 2–4 hours /
over 4 hours / My headaches are always there
7. On average, on how many days per week do you take analgesic medications?
Less than 1 / 1 / Up to 2 / 2 or more / Every day
8. Do changes in your senses (sight, taste, smell or touch) occur before the headache
starts?
Yes / No

NB. If the patient answers No to any of Questions 1, 2, or 3, they may possibly have sinister headache.
They should be advised to seek immediate medical advice from their GP.
If the patient answers Yes to questions 1, 2 and 3, they should complete the remainder of the
questionnaire.

Diagnostic algorithm

• A ‘no’ answer to Questions 1, 2 or 3 indicates the possibility of secondary (or sinister)

headaches. These patients should be investigated further and do not complete the
remaining questions.

For patients who answer ‘yes’ to Questions 1–3:

• Question 4:
o ‘No’ = episodic tension-type headache
o ‘Yes’ = migraine or chronic headache
• Question 5:
o <1; 1; 1–4 and 5–15 days = migraine
o 15–30 days and every day = chronic headache
• Question 6: For patients with chronic headaches only:
o <15 minutes = investigate further
o 15 minutes to 1 hour = possible cluster headache
o 1–2 and 2–4 hours = investigate further
o Over 4 hours and my headaches are always there = chronic daily headache
(CDH)
• Question 7: For patients with CDH only:
o <1; 1; up to 2 = CDH without medication overuse
o 2 or more and every day = CDH with medication overuse (MOH)
• Question 8: For patients with migraine only:
o Yes = migraine with aura
o No = migraine without aura.

149
At baseline, patients completed an informed consent form and a demography questionnaire
(age, gender and race). They then completed the DSQ, without help or any divulging of the
purpose of the questionnaire (patient-completed DSQ). Each patient was then assigned in a
1:1 ratio to see a pharmacist or GP. Consecutive patients were assigned to see the
pharmacist or GP first, before seeing the other healthcare professional. The pharmacist
conducted a diagnostic assessment of the patient, without using the DSQ (pharmacist
diagnosis). They then completed the DSQ with the patient, with additional questioning and
advice where necessary (pharmacist-completed DSQ). The GP conducted a standard
differential headache diagnostic procedure with the patient (GP diagnosis), and were
provided with summaries of the IHS criteria for the diagnosis of migraine, episodic TTH,
CDH, MOH and cluster headache. Finally, the patient consulted with a headache specialist
(specialist diagnosis), who conducted a confirmatory diagnostic procedure (the gold standard
diagnosis). The patients remained blind throughout the investigation to the purpose of these
assessments, and the evaluators were all blinded to the results of each others’ assessments.

Endpoints
All DSQs and pharmacist, GP and specialist diagnoses were collected. The returned DSQs
were assessed by a trained healthcare professional, who was blind to the pharmacist and
physician diagnoses and provided diagnoses according to the DSQ algorithm (Table 1). The
primary endpoint was the sensitivity of the DSQ, measured as the proportion of correct
diagnoses assessed from the patient, pharmacist and GP responses, compared with the gold
standard of specialist diagnosis, for each headache subtype. Sensitivity is defined as the
proportion of those patients with the illness correctly identified as positive by the test.
Specificity was not assessed in the investigation.

Statistical analyses
Introduction
This investigation was designed to assess the accuracy (validity) of the 8-item DSQ when
used by the patient alone, or by the patient in consultation with a pharmacist. Diagnostic data
derived from the completions of the checklist were cross-tabulated with those from diagnoses
deduced by the pharmacist, GP and headache specialist. All data were 100% triple validated
against source documents. The analyses were conducted using the Statistica V4 (Tulsa
USA) statistical package.

Analyses
Baseline demographic assessments were analysed as descriptive statistics only. The
deduction of DSQ-based diagnoses was conducted by a trained professional using the
diagnostic algorithm (Table 2). These diagnoses were formally validated against the gold
standard diagnosis. The accuracy of diagnoses derived from the DSQ was analysed by
cross-tabulating diagnostic labels on the following data sets:
• Patient-completed DSQ versus pharmacist diagnosis
• Patient-completed DSQ versus pharmacist-completed DSQ
• Patient-completed DSQ versus GP diagnosis
• Patient-completed DSQ versus specialist diagnosis
• Pharmacist diagnosis versus GP diagnosis
• Pharmacist diagnosis versus specialist diagnosis.
• Pharmacist-completed DSQ versus pharmacist diagnosis
• Pharmacist-completed DSQ versus GP diagnosis
• Pharmacist-completed DSQ versus specialist diagnosis
• GP diagnosis versus specialist diagnosis.

The different diagnostic sensitivities were compared using the Wilcoxon matched-pairs test,
as ‘Z’ scores and ‘T’ statistics, with the level of significance set at p < 0.05.

150
The study statistician (SK) calculated that a sample size of 100 patients (assuming 25 per
diagnostic group [migraine, episodic TTH, CDH and cluster headache]) was appropriate to
assess sensitivity, as it is similar to the numbers used in other studies of this design.

Results
Patient characteristics
Eighty seven patients took part in the investigation, 44 who saw the pharmacist first and 43
who saw the GP first. All patients completed the five diagnostic procedures. Most patients
were women (n = 70, 80.5%), all were Caucasian in racial origin and the mean age was 52.9
years (range 23–77 years). The vast majority of the patients were diagnosed by the specialist
with migraine (with and without aura; n = 48, 55.2%) or CDH (with and without MOH; n = 32,
36.8%). Only four patients were diagnosed with cluster headache and one with TTH. All
patients completed the DSQ without help within a few minutes.

Sensitivity analysis
Sample sizes were adequate for the patients with migraine (n = 48) and total CDH (n = 32),
but were very small for those with episodic TTH (n = 1) and cluster headache (n = 4). The
sensitivity analyses are shown in Table 2. Overall, the pharmacist diagnoses had the poorest
sensitivities across all diagnostic groups (29.3% sensitivity). However, the pharmacist-
completed DSQ sensitivity was markedly higher, at 66.6%, and similar to that of the patient-
completed DSQ (59.3%) and GP diagnoses (59.4%). The diagnostic sensitivity for ‘total
migraine’ was high in all groups (83.3%–93.8%) and similar, although lower, for ‘total CDH’
(43.8–50.0%). The sensitivity for diagnosing migraine with aura, migraine without aura and
MOH was substantially higher in the DSQ groups (patient-completed and pharmacist-
completed) than in the pharmacist and GP diagnosis groups. There were too few patients in
the episodic TTH and cluster headache categories for analysis.

Table 2. Sensitivity of each diagnostic assignment analysed from the patient-completed

DSQ, pharmacist-completed DSQ, pharmacist diagnosis and GP diagnosis for episodic TTH,
CDH, MOH, migraine and cluster headache compared with the gold standard specialist
assignment.

Sensitivity
Number of patients (%)
Diagnosis Number Patient- Pharmacist- Pharmacist GP
of completed completed diagnosis diagnosis
patients DSQ DSQ
TTH 1 1 (100.0) 1 (100.0) 0 (0) 1 (100.0)
CDH 17 2 (11.8) 2 (11.8) 4 (23.5) 5 (29.4)
Migraine 11 9 (81.8) 10 (90.9) 1 (9.1) 7 (63.6)
with aura
Migraine 32 19 (59.4) 21 (65.6) 0 (0) 16 (50.0)
without aura
MOH 15 9 (60.0) 11 (73.3) 0 (0) 2 (13.3)
Total CDH 32 15 (46.9) 16 (50.0) 14 (43.8) 16 (50.0)
(CDH +
MOH)
Total 48 43 (89.6) 44 (91.7) 40 (83.3) 45 (93.8)
migraine*
Cluster 4 1 (25.0) 2 (50.0) 3 (75.0) 3 (75.0)
headache
Mean 59.3 66.6 29.3 59.4

151
* The ‘migraine with aura’ and migraine without aura’ groups do not add up to the ‘total
migraine’ group because a group labelled ‘migraine’ was only identified by the specialist, GP
and pharmacist diagnoses, and was included in the overall ‘total migraine’ group.

Analysis of the Wilcoxon matched-pairs tests of the different diagnostic groups is shown in
Table 3. The data was organised so as to match pairs of diagnostic groups, validated with
the specialist diagnosis. The diagnosis from the patient-completed DSQ was significantly
less accurate than that from the pharmacist-completed DSQ (p = 0.03). However, there were
no significant differences between the patient-completed DSQ and the pharmacist and GP
diagnoses. Both were equally good by reference to the gold standard (specialist). The only
variation to this was that CDH was diagnosed less accurately on the pharmacist-completed
DSQ.

Table 3. Analysis of the Wilcoxon matched-pair tests for accuracy from the patient-completed
DSQ, pharmacist-completed DSQ, and pharmacist, GP and specialist diagnoses.

Wilcoxon comparison Wilcoxon Wilcoxon p-value

T statistic Z score
Patient-completed DSQ versus 7.00 1.54 0.12
pharmacist diagnosis
Patient-completed DSQ versus 0.00 2.20 0.03
pharmacist-completed DSQ
Patient-completed DSQ versus GP 12.50 0.25 0.80
diagnosis
Pharmacist-completed DSQ versus 5.00 1.82 0.07
pharmacist diagnosis
Pharmacist-completed DSQ versus GP 6.50 0.84 0.40
diagnosis
GP diagnosis versus pharmacist 0.00 2.37 0.02
diagnosis
Specialist diagnosis versus patient- 0.00 2.37 0.02
completed DSQ
Specialist diagnosis versus pharmacist- 0.00 2.37 0.02
completed DSQ
Specialist diagnosis versus pharmacist 0.00 2.52 0.01
diagnosis
Specialist diagnosis versus GP diagnosis 0.00 2.20 0.03

There was a trend for the pharmacist-completed DSQ to be more accurate than the
pharmacist diagnosis, although the difference did not reach statistical significance
(p = 0.07). The pharmacist-completed DSQ was of similar accuracy to that of the GP
diagnosis. The GP diagnosis was significantly more accurate than the pharmacist diagnosis
(p = 0.02). The patient-and pharmacist-completed DSQ (p = 0.02 for both analyses), the
pharmacist diagnosis (p = 0.01) and the GP diagnosis (p = 0.03) were all significantly less
accurate than the specialist diagnosis (the gold standard used in this study).

Discussion

The 8-item DSQ was designed to be a diagnostic aid for headache to be used at the patient’s
first consultation with a healthcare professional, who may be a GP, nurse, pharmacist or
other practitioner (e.g. opticians and dentists often see people with headache). Its principal
use is proposed to be an aid to deciding the appropriate first management decisions. The
questionnaire was designed to be brief and simple to use, so that patients could complete it
on their own. Although the design of the DSQ was based on sound theoretical criteria,5 it had

152
to be demonstrated to be accurate in terms of diagnostic sensitivity before being
recommended for general use.

The patients who took part in the investigation were mostly women, and were somewhat
older than those recruited in most migraine studies.11 This may be due to the high proportion
of CDH sufferers recruited, whose age distribution tends not to be skewed as is the typical
young to middle-aged migraine sufferer.12,13 Both migraine and CDH are more prevalent in
women than in men.12,13 All the patients completed the initial patient-completed DSQ with
little difficulty within a few minutes, and all questions were answered. This demonstrates the
general ease of use of the questionnaire.

Overall, both the patient- and pharmacist-completed DSQs exhibited good sensitivity, of
around 60% for both measures, values very similar to that of the GP diagnosis. The best
sensitivity was demonstrated for total migraine (about 90%), migraine with aura (about 80–
90%), MOH (about 60–75%) and migraine without aura (about 60–65%). The sensitivity for
total CDH (CDH with and without MOH) was somewhat lower at about 50%, driven by a poor
sensitivity for CDH without MOH (about 12%). Results for episodic TTH and cluster
headache could not be interpreted due to the small number of patients with these conditions.
These results indicate that the DSQ is an effective tool for the differential diagnosis of
headache. Interestingly, the accuracy of the pharmacists’ headache diagnosis was improved
markedly when they used the DSQ, particularly with regard to diagnosing migraine with and
without aura and MOH. It would be interesting to see if GPs’ diagnostic accuracy could
improve to the specialist level if they had access to the DSQ.

As would be expected, all groups were significantly less accurate in diagnosis than the
headache specialist diagnoses. However, when pharmacists had access to the DSQ, their
diagnostic accuracy was significantly greater than that of patients using the DSQ and
approached that of the GP diagnosis. These results indicate that, with this aid, pharmacists
can diagnose headache accurately, which is of clinical importance as they may see more
headache sufferers than do GPs.1 Use of the DSQ therefore has the potential to improve
headache management in the pharmacy setting, from which many effective medications are
available without a prescription. In turn, the burden of managing headache on busy GPs may
be relieved.

Some weaknesses were observed in the results. The study was not properly randomised,
with consecutive patients being openly assigned to seeing the pharmacist or GP first. The
study was not large enough to determine whether the order of seeing these healthcare
professionals had any effect on the results. However, the pharmacists and GPs were blind to
each other’s assessments. The entry criteria for numbers of patients with episodic TTH and
cluster headache were not met by the patients recruited from the databases of the MAA and
Dr Dowson. For this reason the specificity of the DSQ was not assessed. The lack of patients
with TTH and cluster headache is perhaps not surprising bearing in mind what we know
about these headaches. Cluster headache is an uncommon condition, affecting < 0.5% of the
population,14 and is not seen often in general practice. Episodic TTH is very common, but
sufferers rarely consult for the condition, preferring to self-treat with OTC medications.3 Such
sufferers are, however, very likely to consult pharmacists. It would be useful to repeat this
validation exercise in a specialist headache clinic (where cluster headache patients are more
common) and in a pharmacy (where episodic TTH sufferers may be encountered frequently).
The accuracy of the DSQ for CDH without MOH was poor, and patients tended to
underestimate the duration and/or the frequency of their headaches. Additionally, migraine
patients sometimes confused aura symptoms with prodromes when answering question 8.
Based on feedback from patients who took part in the study, we suggest that the following
small changes to the questions may help to improve the accuracy of the DSQ:
• Question 1: Has the pattern of your headaches been generally stable (i.e. no change
or only small changes in frequency and severity) over the past few months?

153
 • Question 6: On average how long do your headaches last, if left untreated?
• Question 8: Do changes in your senses (sight, taste, smell or touch) occur in the
period immediately before the headache starts?
Table 4 shows the revised DSQ, including these changes. Future testing of the DSQ for
sensitivity and specificity should use this version of the questionnaire.

Table 4. The revised 8-item headache diagnostic screening questionnaire (DSQ) produced
following completion of the investigation.

1. Has the pattern of your headaches been generally stable (i.e. no change or only small
changes in frequency and severity) over the past few months?
Yes / No
2. Have you had headaches for longer than 6 months?
Yes / No
3. Are you aged between 5 and 50 years?
Yes / No
4. Does the headache interfere to a noticeable extent with your normal daily life (work,
education and social activities)?
Yes / No
5. On average, how many days with headache do you have per month?
Less than 1 / 1 / 1–4 / 5–15 / 15–30 / Every day
6. On average, how long do your headaches last, if left untreated?
Less than 15 minutes / 15 minutes to 1 hour / 1–2 hours / 2–4 hours /
over 4 hours / My headaches are always there
7. On average, on how many days per week do you take analgesic medications?
Less than 1 / 1 / Up to 2 / 2 or more / Every day
8. Do changes in your senses (sight, taste, smell or touch) occur in the period
immediately before the headache starts?
Yes / No

NB. If the patient answers No to any of Questions 1, 2, or 3, they may possibly have sinister headache.
They should be advised to seek immediate medical advice from their GP.
If the patient answers Yes to questions 1, 2 and 3, they should complete the remainder of the
questionnaire.

The DSQ has considerable potential clinical utility in the management of headache in primary
care. Patients should be able to obtain a copy at their first point of care, whether a GP
surgery, pharmacist, optician, dentist or other provider. They can complete the DSQ, either at
home by themselves or with the help of the practitioner. Both patients and pharmacists find
the questionnaire easy to complete. The practitioner should review the DSQ and provide
appropriate advice. Episodic TTH can usually be managed successfully by the sufferer or by
a pharmacist.15 Patients with migraine and CDH probably need to be advised to see a GP for
care, although the pharmacist may be able to manage those with uncomplicated, mild-to-
moderate migraine symptoms.8,9 Using the questionnaire, there is the potential for earlier and
better headache diagnoses. Following pharmacist intervention, patients should be
empowered to obtain help from the GP sooner rather than later. Improved migraine
management has the potential to reduce the high economic cost of the illness to society.16

It is interesting to compare the DSQ with other screening questionnaires that have been
recently developed for headache. Migraine was diagnosed sensitively and specifically using
three questions on headache-related disability, nausea and sensitivity to light,6 or three
questions on disability, headache duration and the lack of new-onset headaches in the
previous 6 months.7 Both these questionnaires restricted themselves to diagnosing migraine

154
only. Maizels and Burchette have developed the Brief Headache Screen to diagnose
different headache subtypes.17 Migraine, daily headache syndromes and medication overuse
could be distinguished using three questions: the frequency of disabling headache, the
presence of other, mild headaches and the use of symptomatic medications. There is
considerable overlap between the DSQ and these other questionnaires in terms of assessing
disability, headache frequency and duration, and the use of headache medications, which
reinforces the validity of these assessments.

In conclusion, the headache DSQ is a brief, rapid to use and accurate diagnostic screening
questionnaire for headache, and is especially sensitive for migraine and MOH. It can be
recommended for screening new headache patients at and below the primary care level.

Acknowledgements

We are especially grateful to the patients who gave their time and commitment to take part in
this investigation. We are also pleased to acknowledge the help of the GPs (Drs Susan
Boucher, Michael Bundy, Annette de la Court, Frances Edgar, Helen Grisewood, Trevor
Rees and David Watson) and pharmacists (Ms Gillian Donkin, Alison Ranftler and Victoria
Withington) who conducted the diagnostic assessments. This investigation was sponsored in
part by an unrestricted educational grant from AstraZeneca.

References

1. Lipton RB, Scher A, Steiner TJ et al. Patterns of healthcare utilization for migraine in
England and in the United States. Neurology 2003;60:441–8.
2. Lipton RB, Diamond S, Reed M et al. Migraine diagnosis and treatment: results from
the American Migraine Study II. Headache 2001;41:638–45.
3. Edmeads J, Findlay H, Tugwell P et al. Impact of migraine and tension-type headache
on life-style, consulting behaviour, and medication use: a Canadian population survey.
Can J Neurol Sci 1993;20:131–7.
4. Lipton RB, Stewart WF, Simon D. Medical consultation for migraine: results from the
American Migraine Study. Headache 1998;38:87–96.
5. Headache Classification Subcommittee of the International Headache Society. The
international classification of headache disorders, 2nd Edition. Cephalalgia 2004;24
(Suppl 1):1–160.
6. Lipton RB, Dodick D, Sadovsky R et al. A self-administered screener for migraine in
primary care: The ID Migraine validation study. Neurology 2003;61:375–82.
7. Cady RK, Borchert LD, Spalding W et al. Simple and efficient recognition of migraine
with 3-question headache screen. Headache 2004;44:323–7.
8. Dowson AJ, Lipscombe S, Sender J et al. New guidelines for the management of
migraine in primary care. Curr Med Res Opin 2002;18:414–39.
9. Dowson AJ, Bradford S, Lipscombe S et al. Managing chronic headaches in the clinic.
Int J Clin Pract 2004; 58:1142–51.
10. Tepper SJ, Dahlöf CGH, Dowson A et al. Prevalence and diagnosis of migraine in
patients consulting their physician with a complaint of headache: data from the
Landmark Study. Headache 2004;44:856–64.
11. The Oral Sumatriptan Dose-Defining Study Group. Sumatriptan – an oral dose-
defining study. Eur Neurol 1991;31:300–5.
12. Silberstein SD, Lipton RB. Chronic daily headache. Curr Opin Neurol 2000;13:277–83.
13. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and
co-morbidities. Neurology 2001;56 (Suppl 1):4–12.
14. Sjaastad O, Bakketeig LS. Cluster headache prevalence: Vågå study of headache
epidemiology. Cephalalgia 2003;23:528–33.

155
15. Dowson AJ. Migraine and Other Headache: Your Questions Answered. Edinburgh:
Churchill Livingstone, 2003.
16. Steiner TJ, Scher AI, Stewart WF et al. The prevalence and disability burden of adult
migraine in England and their relationship to age, gender and ethnicity. Cephalalgia
2003;23:519–27.
17. Maizels M, Burchette R. Rapid and sensitive paradigm for screening patients with
headache in primary care settings. Headache 2003;43:441–50.

156
4.5

Identifying patients who require a change in their current acute

migraine treatment: the Migraine Assessment of Current Therapy
(Migraine-ACT) questionnaire*
Abstract
Background: Currently available measures of the efficacy of acute migraine medications are not
frequently used in primary care. They may be too burdensome and complicated for routine use.
Objectives: To design and test a new, easy to use, 4-item assessment tool, the Migraine Assessment
of Current Therapy (Migraine-ACT) questionnaire for use by clinicians, to quickly evaluate how a
recently prescribed acute medication is working, and to identify patients who require a change of their
current acute treatment.
Methods: A 27-item Migraine-ACT questionnaire was developed by an international advisory board of
headache specialists. Questions were formulated in four domains: headache impact, global
assessment of relief, consistency of response and emotional response. All these are clinically
important measures of migraine severity and treatment outcome. All questions were dichotomous and
answered by yes or no. Patients (n = 185) attending secondary care headache clinics who were
diagnosed with migraine according to International Headache Society criteria entered a multinational,
prospective, observational study to investigate the test-retest reliability and construct validity of the 27-
item Migraine-ACT. Patients completed the Migraine-ACT on two occasions, separated by a 1-week
interval, and test-retest reliability was assessed by Pearson product moment and Spearman rank
measures. Construct validity was assessed by correlating patients’ answers to the 27-item Migraine-
ACT with those to other questionnaires (individual domains and total scores) conceptually related to it;
the Short-Form 36 quality of life questionnaire (SF-36), the Migraine Disability Assessment (MIDAS)
questionnaire and the Migraine Therapy Assessment Questionnaire (MTAQ). Discriminatory t-tests
were used to identify the four Migraine-ACT questions (one in each domain) which discriminated best
between the domains of the SF-36, MIDAS, and MTAQ. These four items constituted the final 4-item
Migraine-ACT.
Results: The test-retest reliability of the 27 Migraine-ACT questions ranged from good to excellent,
and correlation coefficients were highly significant for all items. The consistency of reporting the yes
and no answers was also excellent. Correlations of Migraine-ACT items with SF-36 and MIDAS items
and SF-36, MIDAS and MTAQ total scores indicated that the following were the most discriminating
items, in the respective four domains, and constitute the final Migraine-ACT questionnaire:
• Consistency of response - Does your migraine medication work consistently, in the
majority of your attacks?
• Global assessment of relief - Does the headache pain disappear within 2 hours?
• Impact - Are you able to function normally within 2 hours?
• Emotional response - Are you comfortable enough with your medication to be able to
plan your daily activities
The 4-item Migraine-ACT was shown to be highly reliable (Spearman / Pearson measure r = 0.82).
The individual questions, and the total 4-item Migraine-ACT score, showed good correlation with items
of the SF-36, MIDAS and MTAQ questionnaires, particularly with the total MTAQ and SF-36 scores.
Conclusions: The 4-item Migraine-ACT questionnaire is an assessment tool for use by primary care
physicians to identify patients who require a change in their current acute migraine treatment. It is brief
and simple to complete and score, and has demonstrated reliability, accuracy and simplicity. Migraine-
ACT can therefore be recommended for everyday clinical use by clinicians.

* Edited from the published article: Dowson AJ, Tepper SJ, Baos V, Baudet F, D’Amico D,
Kilminster S. Identifying patients who require a change in their current acute migraine
treatment: the Migraine Assessment of Current Therapy (Migraine-ACT) questionnaire. Curr
Med Res Opin 2004;20:1125–35.

157
Introduction
The Migraine Assessment of Current Therapy (Migraine-ACT) is designed to be a brief
assessment tool used in initial assessments and follow-up to help the clinician determine
quickly and easily whether a change in the acute therapy of their migraine patients is
required. The 27-item version was developed by an advisory board of headache specialists
from the UK, Germany, Italy, Spain and the USA. This paper describes an international study
designed to assess the test-retest reliability and construct validity of the items constituting the
27-item Migraine-ACT, and use these data to select the four items that will constitute the final
questionnaire.

Several recent initiatives have resulted in the development of novel questionnaires that can
be used to assess the efficacy of migraine therapies:
• Impact tools, e.g. the Migraine Disability Assessment (MIDAS) Questionnaire1 and
the Headache Impact Test (HIT).2
• Quality of Life (QOL) questionnaires, which can be generic or migraine-specific.3
• Disease management tools, e.g. the Migraine Therapy Assessment Questionnaire
(MTAQ).4

All these questionnaires have been shown to be reliable and valid tools in populations of
headache sufferers, and all have potential in assessing the outcome from interventions.4–6
However, experience from clinical practice has demonstrated some problems with the
questionnaires that have that limited their uptake there:
• Patients often find them too complicated to use without supervision.
• Physicians are frequently unaware of the questionnaires, and find them too long and
unwieldy for everyday use.
• Both MIDAS ands HIT-6 are brief questionnaires,1,2 but both patients and physicians
require education for their use, which is generally not available.
• When asked, many physicians state that they prefer to ask a single question on
impact, rather than a series of questions.

There is a clear need for a brief, simple, patient-friendly, re-evaluation tool to be used in
triage and as a good means of evaluating the efficacy of medications used for the
management of migraine in primary care. Such simple screening tools have proved valuable
in other therapy areas. The CAGE alcohol dependence questionnaire contains only four
yes/no type questions, with two ‘yes’ responses constituting a positive screening test result.7
A three-question yes/no type questionnaire was developed for asthma to provide the patient
and physician with a rapid assessment of outcome, and as a trigger for intervention.8 We set
out to develop a similar questionnaire for migraine, the Migraine Assessment of Current
Therapy (Migraine-ACT). As originally developed by an international advisory board (AJD,
SJT and VB), Migraine-ACT consisted of 27 questions (Table 1) in four domains selected
from a much larger question pool on the basis of consensus and the group’s extensive
clinical experience to be of significant importance on the basis of clinical evidence: impact;9,10
global assessment of relief;11,12 consistency of response;12,13 and emotional response.14
Questions are answered by the patient as yes or no. The aim is to produce a 4-item test, with
one question in each of the four domains above.

158
Table 1. The initial 27-item Migraine Assessment of Current Therapy (Migraine-ACT)
questionnaire.

Each question starts with the text ‘When you take your treatment’
Impact
Question Code Yes No
When you take your treatment:
Is the impact of the headache significantly relieved within 2 hours? IMP 1
Is the disability caused by the headache significantly relieved within 2 hours? IMP 2
Are you able to resume your normal activity (i.e. work or family-leisure-social IMP 3
activity, etc.) within 2 hours?
Are you able to function normally within 2 hours? IMP 4
Do you usually not need to lie down after 2 hours? IMP 5
Can you get back to what you were doing within 2 hours? IMP 6
Are you able to experience normal enjoyment of your daily life within 2 hours? IMP 7
Are you able to cope normally with everyday life within 2 hours? IMP 8
Are you able to concentrate normally within 2 hours? IMP 9
Do you feel that your sense of emotional well-being returns to normal within 2 IMP 10
hours?
Have any feelings of tiredness, irritability, sadness, anger and anxiety IMP 11
disappeared within 2 hours?

Global assessment of relief

Question Code Yes No
When you take your treatment:
Does the headache pain disappear within 2 hours? GAR 1
Is the headache pain relieved after 2 hours? GAR 2
Does the headache pain disappear within 2 hours and not return within 24 GAR 3
hours?
Does the headache pain disappear within 2 hours and not return within 48 GAR 4
hours?
Does the headache pain start to disappear within 30 minutes? GAR 5
Do your non-headache symptoms (e.g. nausea and sensitivity to light, sound or GAR 6
smells) disappear within 2 hours?
Are your non-headache symptoms (e.g. nausea and sensitivity to light, sound GAR 7
or smells) relieved within 2 hours?
Are you clear of all your symptoms within 2 hours? GAR 8
Does your migraine headache pain stop within 2 hours? GAR 9

Consistency of response
Question Code Yes No
When you take your treatment: CONS 1
Does your migraine medication work attack after attack?
Does your migraine medication work reliably in the majority of your attacks? CONS 2
Does your migraine medication work consistently, in the majority of your CONS 3
attacks?

Emotional response
Question Code Yes No
When you take your treatment: EMOT 1
Does your migraine medication work consistently enough that you can plan
your future activities?
Are you confident that the migraine medication you are taking will treat the next EMOT 2
attack effectively?
Are you comfortable enough with your medication to be able to plan your daily EMOT 3
activities?
Do you feel in control of your migraine enough so that you feel there will be no EMOT 4
disruption to your daily activities?

159
Patients and Methods
Patients
Patients at five secondary care headache centres, in the UK, Spain, Germany, Italy and the
USA, took part in the study. All patients were attending the centres for migraine treatment.
Each centre enrolled 40 patients with migraine diagnosed according to the International
Headache Society (IHS) criteria.15,16 Eligible patients were men or women aged between 18–
65 years, with at least a 1-year history of migraine, experiencing on average 1–4 attacks per
month, and a minimum of 24 hours between each attack. Patients were able to distinguish
migraine attacks from other types of headache. Patients were excluded from the study if they
had a history of significant psychiatric illness, or any other illness that could affect the
interpretation of the results, e.g. major depression or anxiety disorder, or had a history of
abuse of alcohol, or other recreational drugs.

Study design
This was an open, prospective, multinational, observational, two-visit study, conducted in the
UK, Spain, the USA, Germany and Italy, using questionnaires which were translated into the
local languages. Patients were assessed at baseline and completed a series of outcome
questionnaires; the 27-item Migraine-ACT, the Short Form-36 (SF-36) quality of life (QOL)
questionnaire,3 the MIDAS Questionnaire1 and the MTAQ Questionnaire.4 The validity of the
Migraine-ACT Questionnaire was assessed from the responses to these questionnaires.
Additionally, we evaluated test-retest reliability, thus the same patients completed the 27-
item Migraine-ACT Questionnaire again 1 week later.

Patients provided their written consent before participating in the study, by signing a consent
form. The study was conducted in accordance with the Declaration of Helsinki, as revised by
the most recent meetings of the World Medical Assembly. The study was conducted in the
spirit of, and according to, the principles of Good Clinical Practice (GCP), as embodied in the
ICH Harmonised Tripartite Guideline for GCP.

Study procedures
Migraine patients (n = 200; 40 per centre) attended the clinic, or were contacted by
telephone, for demography baseline assessments (age, gender, race, weight and height) and
completion of the outcome questionnaires: 27-item Migraine-ACT; SF-36 QOL questionnaire;
MIDAS Questionnaire; and the MTAQ Questionnaire (Visit 1). All patients (n = 40 per centre)
who completed the 27-item Migraine-ACT attended the clinic or were contacted by telephone
1 week later to complete the Migraine-ACT Questionnaire for a second time (Visit 2). During
this time, no changes were allowed in the patients’ migraine treatments. Any patients who
withdrew from the study were recorded, and the reasons for withdrawal noted. Patients who
withdrew were included in the analyses of results.

Endpoints
The primary study endpoints were test-retest reliability (a comparison of the Migraine-ACT
data from the 27 individual questions from Visits 1 [baseline] and 2 [1 week]), and construct
validity (a comparison of the Migraine-ACT data from the 27 individual questions with
relevant data from the SF-36 and MIDAS (domains and total scores) and MTAQ
questionnaires (total score only) at Visit 1 [baseline]). The construct validity data were used
to identify the four items (one in each domain) to be contained in the final Migraine-ACT
questionnaire.

Statistical analyses
Introduction
This study was designed to assess the test-retest reliability and construct validity of the 27
items constituting the original Migraine-ACT questionnaire. These data were then used to
construct the final 4-item version of the questionnaire. Test-rest reliability was assessed by

160
comparing patients’ responses to the individual Migraine-ACT questions when completed at
baseline (Visit 1) and 1 week later (Visit 2). Validity was assessed by comparing patients’
responses to the individual Migraine-ACT questions at baseline (Visit 1) with those to
components of questionnaires related to the conception of Migraine-ACT, i.e. the SF-36,
MIDAS and MTAQ questionnaires. All analyses were conducted with Statistica (Tulsa, USA)
software with statistical hypotheses tested for two-tailed tests of type 1 error at 0.05. Data
were 100% validated against source documents.

Baseline analyses
All baseline demographic assessments were analysed as descriptive statistics only (means
and standard deviations).

Test-retest reliability
Test-retest reliability of the individual Migraine-ACT questions was assessed using Pearson’s
product moment and Spearman rank measures. A minimum test-retest product moment of r
= 0.30 (detected statistically at alpha2 p = 0.05) would be n = 40 per centre. A test–retest or r
statistic of 0.75 or over represents excellent agreement, 0.40 to 0.75 represents intermediate
to good agreement, and below 0.40 represents poor agreement.17 The whole process was
repeated and confirmed on the summed 4-item score.

Construct validity
To assess the construct validity of the individual Migraine-ACT questions, the associations
were determined between patients’ responses to the individual Migraine-ACT questions and
instruments related conceptually to the Migraine-ACT questionnaire. Correlations were made
between Migraine-ACT and the MIDAS, SF-36 and MTAQ questionnaires (criterion
variables). Discrimination was analysed by simple t-tests of means from dichotomised
Migraine-ACT item scores. Factor analysis was used to identify and score the key features of
the criterion variables and the Migraine-ACT questions were correlated and discriminated
with these factors. Separate analyses checked for centre effects (i.e. within each centre
[country]), and for the total scores derived from all countries combined (i.e. overall analysis).

Item selection
The questions to be included in the final 4-item Migraine-ACT questionnaire were identified
from the correlation coefficients and discrimination of all 27 questions with summary and
derived subscale scores and factor analysed criterion scores from MIDAS, MTAQ and SF36.
In addition, the questions all had to exhibit at least good test-retest reliability. Regression was
used to compare the summation of the individual final four items versus weighted items.
Summation of the best four items was made after various analyses to find the best structure
and redundancy of terms from the MIDAS, MTAQ and SF36.

Orthogonal and Oblique rotations of the principal components derived the maximum variance
factors associated with a reduced data set. Regression-derived new composite variables
were constructed from the factor analyses.

Results

Demography and baseline characteristics

One hundred and eighty five patients took part in the study. All patients were assessed at
Visit 1, with 143 (77%) completing the questionnaire on both occasions. The majority of the
patients (68%) were women. All but one of the patients was of white racial origin. The mean
age was 44 years (range 14 to 87 years), with the vast majority (172, 93%) aged 18–65
years.

161
Test-retest reliability
Spearman rank order measures for all 27 Migraine-ACT items are shown in Table 2. The
analyses were nonparametric and pairwise deleted. There were few dropouts; between 176
and 185 of the 185 patients taking part in the study provided data at Visits 1 and 2.
Spearman rank order measures ranged from 0.54 to 0.79 for the individual 27 Migraine-ACT
items, indicating good-to-excellent reliability. Summated Spearman measures for the impact
(IMP1–11), global assessment of relief (GAR1–9), consistency of response (CONS 1–3) and
emotional response (EMOT1–4) items exhibited excellent reliability, and the total summation
of all 27 items was extremely reliable, at r = 0.99 (Table 2).

Table 2. Test-retest reliability of the 27-item Migraine-ACT Questionnaire. Spearman rank

order measures are shown below for all 27 Migraine-ACT items. Analyses were
nonparametric and pairwise deleted.

Migraine-ACT questions n r t (n - 2) p-value

IMP1 184 .68 12.67 <.0001
IMP2 182 .61 10.23 <.0001
IMP3 182 .77 16.25 <.0001
IMP4 185 .72 13.96 <.0001
IMP5 183 .54 8.61 <.0001
IMP6 181 .79 16.96 <.0001
IMP7 184 .73 14.31 <.0001
IMP8 182 .70 13.10 <.0001
IMP9 182 .74 14.76 <.0001
IMP10 183 .62 10.66 <.0001
IMP11 183 .62 10.64 <.0001
GAR1 181 .58 9.44 <.0001
GAR2 183 .63 11.04 <.0001
GAR3 179 .61 10.18 <.0001
GAR4 179 .58 9.44 <.0001
GAR5 183 .60 10.15 <.0001
GAR6 182 .60 10.15 <.0001
GAR7 182 .62 10.56 <.0001
GAR8 179 .58 9.53 <.0001
GAR9 177 .72 13.68 <.0001
CONS1 182 .68 12.43 <.0001
CONS2 184 .70 13.15 <.0001
CONS3 182 .61 10.30 <.0001
EMOT1 180 .74 14.76 <.0001
EMOT2 179 .77 16.04 <.0001
EMOT3 178 .78 16.40 <.0001
EMOT4 176 .62 10.56 <.0001
Total IMPACT (IMP 1–11) 168 1.00 137.78 <.0001
Total GAR (GAR 1–9) 164 .78 16.08 <.0001
Total CONS (CONS 1–3) 180 .73 14.33 <.0001
Total EMOT (EMOT 1–4) 172 .80 17.41 <.0001
TOTAL Migraine-ACT 143 .99 75.62 <.0001
(27 items)

Analyses of consistency of response were conducted to investigate the reliability data further.
The proportion of patients who consistently answered ‘yes’ or ‘no’ at both Visits 1 and 2 is
shown in Table 3. The consistency of response was excellent for all 27 Migraine-ACT items.
Similar consistency was shown in the distribution of ‘yes’ and ‘no’ answers over the two
visits.

162
Table 3. The consistency of the ‘yes’/‘no’ responses between Visit 1 and Visit 2 for the 27
Migraine-ACT items: proportion of patients who responded ‘yes’ or ‘no’ at Visit 1 who had the
same responses at Visit 2.

Consistency of response (% patients)

Migraine-ACT item ‘No’ answers ‘Yes’ answers
IMP 1 76.74 90.28
IMP 2 75.00 84.00
IMP 3 87.14 87.07
IMP 4 83.15 85.86
IMP 5 66.67 83.04
IMP 6 77.78 93.50
IMP 7 84.85 84.27
IMP 8 80.49 85.58
IMP 9 88.07 81.82
IMP 10 80.00 78.21
IMP 11 88.43 69.23
GAR 1 79.75 73.83
GAR 2 83.33 88.89
GAR 3 90.43 62.32
GAR 4 86.07 67.21
GAR 5 83.58 75.47
GAR 6 77.91 79.80
GAR 7 71.43 88.81
GAR 8 86.40 66.10
GAR 9 85.06 82.11
CONS 1 82.54 83.06
CONS 2 78.79 93.78
CONS 3 70.00 86.86
EMOT 1 84.75 85.83
EMOT 2 83.08 81.36
EMOT 3 84.00 89.55
EMOT 4 79.79 75.56

Construct validity and item selection

Discriminatory t-test scores between the individual 27 Migraine-ACT questions and 15 items
contained within the SF-36, MIDAS and MTAQ questionnaires are shown at the end of the
article in Table 4. Eleven items related to the SF-36 questionnaire (physical function, role-
physical, bodily pain, general health, vitality, social functioning, role-emotional, mental heath,
total physical, total mental and total SF-36 scores). Three items related to the MIDAS
questionnaire (MIDAS A score [headache frequency], MIDAS B score [headache severity]
and total MIDAS score [time lost for everyday activities]) and one item related to the MTAQ
score (total MTAQ score). Finally, the total score from all 15 variables was correlated with the
four best Migraine-ACT items. A t-discrimination score of ≥ 2 indicates a significant
correlation.

Correlations of Migraine-ACT items with SF-36 and MIDAS items and SF-36, MIDAS and
MTAQ total scores indicated that the most discriminating items in the four domains of the
Migraine-ACT questionnaire were IMP 4, GAR 1, CONS 3 and EMOT 3. The discriminating
factors for these four items were all highly significant. The best correlation was observed
between the Migraine-ACT items and the MTAQ total score. These data indicate that the four
items constituting the final Migraine-ACT questionnaire should be:
• Impact: Are you able to function normally within 2 hours? (IMP 4)

163
 • Global assessment of relief: Does the headache pain disappear within 2 hours?
(GAR 1)
• Consistency of response: Does your migraine medication work consistently, in
the majority of your attacks? (CONS 3)
• Emotional response: Are you comfortable enough with your medication to be
able to plan your daily activities? (EMOT 3)

The results showed that the four items in the Migraine-ACT discriminated most with the
MTAQ score, and to a slightly lesser degree with the total SF-36 score. Factor analysis
showed that four items described most of the variance in the data: the total SF-36 score
described 42.01%; the SF-36 role emotional score described 9.49%; the total MTAQ score
described 8.93%; and the MIDAS B score (pain severity) described 6.96%. Mean t-
discrimination scores from the 15-item analysis above were compared with those from the
factor analysis (Table 5), and were very similar, indicating a high level of correlation.

Table 5. Discrimination analysis (discriminatory t-tests) of the four items constituting the final
Migraine-ACT Questionnaire: comparison of t-test values calculated from all 15 items
contained within the SF-36, MIDAS and MTAQ questionnaires, and the four items derived
from factor analysis.

Mean t-discrimination
Migraine-ACT item Total score derived from all Total score derived from
15 SF-36, MIDAS and factor analysis
MTAQ items
IMP 4 3.07 3.97
GAR 1 2.70 3.14
CONS 3 2.08 2.71
EMOT 3 3.16 4.18

Analysis of the 4-item Migraine-ACT questionnaire

The 4-item Migraine-ACT questionnaire was further analysed in several ways. The individual
questions, and the total Migraine-ACT score, were correlated with subscales and total scores
of the SF-36, MIDAS and MTAQ questionnaires using Pearson product moment measures
(Table 6). The Migraine-ACT items and total score correlated well with the other items,
particularly so with the total MTAQ and SF-36 domain and total scores. Regression of MTAQ
and total SF-36 on Migraine-ACT were both highly significant (MTAQ: beta = -0.58, SE =
0.06, p < 0.0001; SF-36: beta = 0.36, SE = 0.07, p < 0.0001). The 4-item Migraine-ACT
questionnaire was examined for test-retest reliability, and was shown to be very reliable by
the Pearson and Spearman measures (r = 0.82 for both calculations).

164
Table 6. Analysis of the 4-item Migraine-ACT questionnaire: calculation of Pearson product
moment measures with the SF-36, MIDAS and MTAQ questionnaires.

Pearson product moment measures

IMP4 GAR1 CONS3 EMOT3 Total
Migraine-
ACT score
SF-36 Bodily Pain 0.345 0.24 0.25 0.33 0.38
SF-36 General 0.18 0.15 0.25 0.24
Health
SF-36 Vitality 0.22 0.25 0.22 0.28
SF-36 Social 0.31 0.30 0.21 0.34 0.38
Functioning
SF-36 Mental 0.20 0.21 0.21 0.25
Health
SF-36 Total 0.31 0.25 0.20 0.29 0.34
Physical
SF-36 Total Mental 0.27 0.27 0.18 0.29 0.33
SF-36 Total 0.31 0.28 0.20 0.31 0.36
MIDAS B score -0.24 -0.20 -0.18 -0.18 -0.26
Total MIDAS score -0.28 -0.29 -0.31 -0.32
Total MTAQ score -0.50 -0.39 -0.38 -0.59 -0.60

In scoring the Migraine-ACT questionnaire, it is desirable to produce simple cut-off points

that facilitate interpretation and use. The simplest scoring method is to score all ‘yes’
answers as ‘1’ and all ‘no’ answers as ‘0’, and sum the answers. This produces a 5-item
score, ranging from 0–4. A score of ‘4’ might indicate that the patient has no medication
needs, while a score of ‘0’ may indicate very high medication needs. Alternatively, the items
can be individually weighted, resulting in a more complex scoring scheme. However,
regression analysis indicated that weighting the scoring only increased its discriminatory
power by 3% as compared to the MTAQ questionnaire. The simpler method was therefore
practically as good as a more complicated weighting method. Table 7 shows how the simple
version Migraine-ACT scores compare with those of MIDAS, MTAQ and SF-36 total scores.
A decreasing Migraine-ACT score (indicating increased medication needs) correlated with
increasing MIDAS and MTAQ scores (indicating increased management needs) and a
decreasing SF-36 score (indicating reduced quality of life). The mean Migraine-ACT score for
patients (n = 181) participating in this study was 2.55 (SE = 0.11).

Table 7. Interpretation of the total Migraine-ACT score: relation to MIDAS, MTAQ and SF-36
scores

Migraine-ACT score MIDAS score MTAQ score SF-36 score

4 16 2.25 54
3 22 2.95 48
1 34 4.37 46
0 40 5 44

MIDAS scores range from 0 to over 100. There are four grades: I = no or minimal disability
(score 0–5); II = mild disability (score 6–10); III = moderate disability (score 11–20); and IV =
severe disability (score > 20). Most migraine sufferers score as Grade III or IV.1
MTAQ scores range from 0–8, with a higher score indicating an increasing number of
management issues.4
SF-36 scores range from 0–100, with increasing score correlating with better health. A score
of 50 is the average for the US population.23

165
Discussion

Any new tool assessing clinically important endpoints needs to exhibit good test-retest
reliability and validity before it can be recommended for use. Good test-retest reliability
means that the tool assesses the data consistently over time. Validity denotes the accuracy
of the tool in assessing what it is meant to measure. Validity is commonly assessed by
criterion (gold standard) or construct validity (comparing the accuracy of the new tool
compared to other tools that are related to the same disease area). Additional attributes are
necessary if a tool is designed for use in primary care. The tool needs to be brief, and simple
to complete and score. It should also exhibit face validity, being intuitively meaningful to the
target audience. All these factors were included in the design and testing of the Migraine-
ACT questionnaire.

Migraine-ACT was designed from the outset to have high face validity. The question domains
were chosen to be clinically relevant. Migraine impact, the global response to medication,
consistency of response and the emotional component of the headache are all important in
assessing the severity of illness and the response to treatment.9–14 The questionnaire was
also designed to be brief, with only four questions, one in each domain. Simplicity was
assured by ensuring that all questions were answered by ‘yes’ or ‘no’.

The individual 27 Migraine-ACT questions all exhibited good-to-excellent test-retest

reliability. However, additional analyses of reliability were required due to the variables being
dichotomous (yes/no) rather than continuous, with resultant low degrees of freedom in the
analyses. We therefore analysed the consistency of response in the patients’ answers at the
two study visits. Analyses of consistency (the proportion of patients who answered ‘yes’ or
‘no’ at both study visits and the proportion of ‘yes’ and ‘no’ answers for each question at the
study visits) provided a high consistency in the responses (typically over 80% agreement).
Taken together with the good test-retest reliability data, it can be concluded that the
Migraine-ACT questions are highly reliable. While the 1-week gap between the first and
second completions of the questionnaire may be considered to be short, this was done to
minimise potential changes in the patients’ migraine attacks, which could have altered the
results.

For the assessment of construct validity, the 27 items constituting the original Migraine-ACT
were correlated with a series of questionnaires that have been shown to be sensitive to
outcome in headache. SF-36 is a generic QOL questionnaire that can be used for any illness
state. It has been shown to accurately assess the severity of migraine, and effective
treatment results in an increased score, indicating improved QOL.18 MIDAS is a headache-
specific questionnaire that assesses disability as time lost from daily activities, as well as the
headache frequency and severity.1 It is a reliable1,19 and valid20 measure, and is a sensitive
outcome measure in migraine and other headaches.21,22 MTAQ is a migraine-specific
questionnaire that is a reliable and valid screening tool,4 although its utility as an outcome
measure has not yet been demonstrated. Discriminatory t-tests showed that many of the
Migraine-ACT items correlated well with the total SF-36 score and some of its subscales, the
MIDAS total score and MIDAS B scores, and the total MTAQ score. Of all these items,
Migraine-ACT items correlated best with the total SF-36 and MTAQ scores. The four items
constituting the final Migraine-ACT questionnaire were selected on the basis of those items
(one in each domain) that correlated best with the total score from all 15 items from the SF-
36, MIDAS and MTAQ variables. Factor analysis confirmed these data.

The four-item Migraine-ACT questionnaire is shown in Figure 8 at the end of this article. The
order of the questions is designed so that migraine and headache items are introduced first,
while the more generic questions are answered later. Each ‘yes’ answer is scored as ‘1’ and
each ‘no’ answer as ‘0’. The total Migraine-ACT score is achieved by summing the answers

166
to the four items. We suggest that a score of 4 corresponds to the physician not needing to
re-evaluate the patient while 0 corresponds to a very high need for re-evaluation. However,
further studies are required to confirm this relationship and define the cut-off point that
indicates the need for evaluation and/or a change in treatment. Regression analysis showed
that there was no significant advantage in weighting the answers, which would have
increased the complexity of the questionnaire markedly. The result is a questionnaire that is
intuitive and simple to complete and score, yet correlates highly with both generic (SF-36)
and disease-specific (MTAQ) measures of health status. Further testing showed that the 4-
item Migraine-ACT is highly reliable and valid.

We envisage the main use of the Migraine-ACT as a rapid and accurate screener to identify
patients whose current acute treatment regimen needs to be changed. It is intended for use
by the primary care physician or clinician in everyday clinical practice. Currently-used tools
do not fulfil this role because they are too complex for routine use.1–4 There were clear but
opposite relationships between the 4-item Migraine-ACT with the SF-36 general health
measure, and the headache-specific MIDAS and the highly migraine-specific MTAQ. This
would be expected, as illness and health are opposite constructs. The 4-item Migraine-ACT
score ranges from 0–4 and indexes a window into the total SF-36 score ranging from 44–54.
The higher the score, the better off is the patient. Similarly the specific MIDAS and MTAQ
measure illness with higher scores being worse for the patient. The 4-item 0–4 score indexes
a window into MIDAS from 16–40 and into MTAQ ranging from 2–5. The MIDAS scores,
ranging from mid-Grade III to high Grade IV are what would be expected in a population of
secondary care patients having significant headache-related disability.1 The 4-item Migraine-
ACT is appreciably shorter, simpler and easier to complete than any of the three other
questionnaires above.

Further studies are required to define the clinical utility of Migraine-ACT, to define the score
that indicates a change in treatment is needed, and to investigate its sensitivity to change for
use as an outcome tool in comparison to changes in SF-36, MIDAS and MTAQ. This study
was conducted in secondary care referral centres. Studies in primary care practice are
therefore particularly needed to demonstrate its utility in the everyday management of
migraine. However, we believe that the main use of Migraine-ACT will be in clinical practice,
and naturalistic studies conducted in this domain will determine the true utility of the
questionnaire.

Conclusions

The 4-item Migraine-ACT questionnaire is a rapid, reliable, brief, simple to complete and
score assessment tool, which can be recommended for everyday clinical use by the clinician
to identify patients who require a change in their current acute migraine treatment.

References
1. Stewart WF, Lipton RB, Kolodner K et al. Reliability of the migraine disability
assessment score in a population-based sample of headache sufferers. Cephalalgia
1999;19:107–14.
2. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
3. Osterhaus JT, Townsend RJ, Gandek B et al. Measuring the functional status and
well-being of patients with migraine headache. Headache 1994;34:337–43.
4. Chatterton ML, Lofland JH, Shechter A et al. Reliability and validity of the Migraine
Therapy Assessment Questionnaire. Headache 2002;42:1006–15.
5. Dowson AJ. Assessing the impact of migraine. Curr Med Res Opin 2001;17:298–309.
6. Solomon GD, Skobieranda FG, Genzen JR. Quality of life assessment among
migraine patients treated with sumatriptan. Headache 1995;35:449–54.

167
7. Mayfield D, McLeod G, Hall P. The CAGE questionnaire: validation of a new
alcoholism screening instrument. Am J Psychiatry 1974;131:1121–3.
8. Keeley D. Measuring clinical outcome in asthma. J R Coll Physicians 2000;34:9–11.
9. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
10. Lipton RB, Dodick D, Sadovsky R et al. A self-administered screener for migraine in
primary care. Neurology 2003;61:375–82.
11. International Headache Society Clinical Trials Subcommittee. Guidelines for
controlled trials of drugs in migraine: second edition. Cephalalgia 2000;20:765–86.
12. Ferrari MD, Roon KI, Lipton RB et al. Oral triptans (serotonin 5-HT1B/1D agonists) in
acute migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358:1668–75.
13. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity. Disability, pain intensity,
and attack frequency and duration. Neurology 1994;44 (Suppl 4):24–39.
14. Kilminster SG, Dowson A, Bundy M. Headache Impact Test (HIT) and Short Pain
Inventory©. Outcome measures compared: Int J Pharm Med 2003;17:23–32.
15. Headache Classification Committee of the International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial neuralgias and
facial pain. Cephalalgia 1988;8 (Suppl 7):1–92.
16. Headache Classification Subcommittee of the International Headache Society. The
international classification of headache disorders, 2nd edition. Cephalalgia 2004;24
(Suppl 1):1–160.
17. Nunnally JC, Bernstein IH. Psychometric theory. New York: McGraw-Hill; 1994.
18. Jhingran P, Cady RK, Rubino J et al. Improvements in health-related quality of life
with sumatriptan treatment for migraine. J Fam Pract 1996;42:36–42.
19. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of
the Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.
20. Stewart WF, Lipton RB, Kolodner KB et al. Validity of the Migraine Disability
Assessment (MIDAS) score in comparison to a diary-based measure in a population
sample of migraine sufferers. Pain 2000;88:41–52.
21. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
22. D’Amico D, Usai S, Grazzi L et al. The impact of primary headaches on patients’
lives: Italian experience with the MIDAS and SF-36 questionnaires. Headache Care
2004;1:123–8.
23. Anon. Norm-based scoring of SF-36 scales. In: SF-36 Summary Measures Manual.
Lincoln, RI, USA: QualityMetric Inc. 2003; p 22–35.

168
 Table 4. Construct validity of the 27-item Migraine-ACT questionnaire: discrimination analysis (discriminatory t-tests) between the 27 Migraine-
ACT items and the 15 items contained within the SF-36, MIDAS and MTAQ questionnaires.
Mean t-discrimination
Migraine- SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 SF-36 MIDAS MIDAS Total Total Total
ACT Physical Role- Bodily General Vitality Social Role Mental Total Total Total A Score B Score MIDAS MTAQ score
question Function Physical Pain Health Functioning Emotional Health Physical Mental Score Score from all
15
variables
IMP 1 -0.00 0.51 3.15 1.72 2.16 3.38 -0.89 1.63 2.28 2.19 2.17 -1.11 -2.12 -2.63 -5.06 1.98
IMP 2 1.08 0.61 3.07 2.52 1.77 2.67 0.61 1.76 2.69 2.50 2.63 -1.57 -2.30 -3.05 -6.17 2.24
IMP 3 1.20 0.85 3.72 2.96 2.45 3.91 0.98 3.19 3.38 3.89 3.77 -0.99 -4.42 -3.61 -6.75 2.96
IMP 4 -1.06 -1.53 -5.04 -2.39 -3.05 -4.48 -1.05 -2.70 -4.16 -3.61 -4.15 1.17 3.37 3.62 7.31 3.07
IMP 5 2.15 -1.18 0.52 0.22 -0.27 -0.17 -0.57 0.18 0.56 -0.09 0.35 -1.22 -0.28 1.53 0.66 0.69
IMP 6 -0.07 0.50 3.21 2.07 1.28 3.69 0.59 2.43 2.07 2.75 2.60 -0.81 -3.39 -2.39 -6.09 2.17
IMP 7 0.18 -0.68 -3.57 -2.07 -3.10 -3.50 -1.71 -2.81 -2.68 -3.59 -3.13 0.92 2.79 3.10 5.81 2.56
IMP 8 0.55 0.86 3.84 0.93 1.81 4.42 0.86 2.88 2.11 2.70 2.57 -1.09 -2.91 -3.22 -5.76 2.39
IMP 9 0.096 -0.82 -3.87 -1.72 -2.58 -3.30 -0.43 -2.38 -2.54 -2.84 -2.75 1.29 3.06 4.02 4.92 2.35
IMP 10 0.30 -1.58 -3.36 -1.21 -2.71 -3.71 -2.41 -4.25 -2.19 -3.59 -3.23 1.69 2.34 3.52 5.94 2.67
IMP 11 -0.27 -1.71 -2.37 -2.37 -3.85 -3.87 -1.72 -4.49 -2.98 -4.04 -3.45 1.04 3.55 2.90 4.06 2.76
GAR 1 0.57 1.11 3.37 1.93 3.44 4.32 0.72 2.82 3.28 3.67 3.74 -1.77 -2.70 -3.76 -5.40 2.70
GAR 2 -1.57 0.30 1.20 -0.10 -0.23 1.49 -2.18 0.45 -0.15 -0.21 0.03 0.16 0.00 -0.96 -4.30 0.84
GAR 3 1.86 0.11 -2.31 -1.32 -2.05 -1.54 0.67 -1.55 -1.21 -2.14 -1.92 0.87 2.26 2.28 3.21 1.62
GAR 4 -0.80 -0.22 2.47 2.61 2.13 2.21 0.50 2.28 1.75 2.85 2.51 -1.94 0.39 -2.15 -4.62 1.98
GAR 5 0.75 0.23 2.02 2.24 2.37 2.00 0.44 1.03 3.19 2.78 3.10 -0.48 -1.60 -2.67 -4.45 1.86
GAR 6 0.22 0.14 2.13 1.44 2.11 2.51 0.42 2.36 2.15 2.33 2.43 -2.25 -1.45 -2.61 -4.07 1.83
GAR 7 -0.48 0.85 0.92 -0.15 0.39 1,51 -0.60 2.04 0.44 0.71 0.93 -1.94 -1.21 -0.50 -3.21 1.04

169
GAR 8 0.27 -1.29 -3.40 -1.99 -2.37 -2.79 -1.19 -2.74 -2.36 -3.09 -2.96 1.07 3.90 3.38 4.52 2.33
GAR 9 0.61 0.04 2.46 1.21 2.13 2.92 -0.66 2.11 1.98 2.17 2.30 -0.49 -2.67 -2.99 -4.35 1.89
CONS 1 -0.15 0.63 3.35 1.33 2.19 2.44 0.90 2.08 1.97 2.26 2.12 -1.73 -2.66 -2.05 -4.90 2.06
CONS 2 1.13 0.20 2.84 2.22 1.26 2.88 0.15 1.75 2.19 1.75 1.98 0.27 -2.96 -1.94 -5.91 1.93
CONS 3 0.29 -1.27 3.53 2.05 1.84 2.88 -0.62 1.91 2.64 2.40 2.62 -1.05 -2.39 -1.27 -5.19 2.08
EMOT 1 -0.09 -0.82 -3.73 -3.26 -2.02 -3.61 -1.01 -2.31 -3.20 -3.43 -3.28 0.10 2.65 3.64 6.44 2.54
EMOT 2 -0.77 0.37 4.24 1.88 0.75 3.34 1.00 1.30 2.29 1.92 2.37 0.03 -2.24 -2.72 -6.25 2.09
EMOT 3 0.04 -0.55 -4.64 -3.41 -2.98 -4.90 -1.10 -2.88 -3.86 -3.98 -4.10 0.71 2.40 4.11 9.10 3.16
EMOT 4 -1.37 -0.78 -6.05 -3.31 -2.86 -4.75 -1.62 -1.90 -4.23 -3.35 -3.55 1.06 1.84 4.80 6.74 3.12
Figure 1. The 4-item Migraine-ACT questionnaire recommended for use in primary care.

Please answer all four questions below, as ‘yes’ or ‘no’, by placing a tick in the relevant box.

Question Yes No
When you take your treatment:
Does your migraine medication work consistently, in the
majority of your attacks?

When you take your treatment:

Does the headache pain disappear within 2 hours?

When you take your treatment:

Are you able to function normally within 2 hours?

When you take your treatment:

Are you comfortable enough with your medication to be
able to plan your daily activities?

Migraine-ACT Score

Scoring the Migraine-ACT questionnaire

One or more ‘no’ answers may indicate the need to change treatment. An increasing number of
‘no’ answers indicates increasing treatment needs.

© Practical Solutions in Medicine Ltd. 2004. Permission is freely given for the use of the
Migraine-ACT questionnaire by individuals for the purposes of research and to improve
treatment of migraine patients. As a courtesy, researchers and commercial parties are
asked to inform the copyright holder in advance before using the questionnaire.

170
4.6

The Migraine Assessment of Current Therapy (Migraine-ACT)

questionnaire: investigation of reliability, validity and clinical utility in
a multinational study*

Abstract

The 4-item Migraine-ACT questionnaire is an assessment tool for use by primary care
physicians to identify patients who require a change in their current acute migraine treatment. It
has been shown to be easy to use, and to be reliable and accurate in its assessments. The
Migraine-ACT study database was analysed further, providing additional information on the
reliability, validity and potential clinical utility of the questionnaire.
Reliability was assessed by recording the distribution of Migraine-ACT scores recorded at
baseline and 1 week later (test-retest reliability). Analyses of consistency of Migraine-ACT
scores were conducted on the total sample of patients and for the five separate centres
(Germany, Italy, Spain, UK and USA), using Pearson and Spearman correlations. Validity was
assessed by comparing the t-discrimination values for clinically-relevant questions within
domains of the original 27-item questionnaire. Reliability and validity were also assessed by
constructing an ‘alternative’ (Form B) Migraine-ACT questionnaire, derived from an analysis of
the second best items in each domain in the original study data. Clinical utility was assessed
using Pearson pairwise correlations to compare Migraine-ACT scores with clinically-defined
criteria as analysed by the SF-36 Quality of Life questionnaire, the Migraine Disability
Assessment (MIDAS) questionnaire and the Migraine Therapy Assessment (MTAQ)
questionnaire.
The distribution of Migraine-ACT scores between the two completions of the questionnaire was
consistent for the total sample (test-retest reliability r = 0.81) and between the individual
countries (r = 0.61–0.92). In this study, the validity (assessed as t-discrimination) of the
Migraine-ACT ‘impact’ and ‘global assessment of relief’ questions were markedly higher than
those of other endpoints used in migraine clinical studies. The Form B Migraine-ACT
questionnaire was almost as reliable and accurate as the original Form A questionnaire. The
distribution of Migraine-ACT scores was: 0 = 12.6%, 1 = 13.7%, 2 = 14.7%, 3 = 20.5% and 4 =
38.4%. The change in Migraine-ACT score correlated with, and had a linear relationship with
changes in SF-36, MIDAS and MTAQ scores, and indicated that a Migraine-ACT score of ≤ 2
corresponded with a need to consider changing the patient’s acute medication. About 40% of
the migraine patients in the study scored ≤ 2 and may have had significant unmet treatment
needs.
These data confirm the good reliability and validity of the Migraine-ACT questionnaire and
provide further evidence for its utility in clinical practice.

*Summarised from the article: Kilminster S, Dowson AJ, Tepper S, Baos V, Baudet F, D’Amico
D. The Migraine Assessment of Current Therapy (Migraine-ACT) questionnaire: investigation of
reliability, validity and clinical utility in a multinational study. Headache 2005; Submitted for
publication.

171
Introduction

A 4-item questionnaire was developed, the Migraine Assessment of Current Therapy (Migraine-
ACT) questionnaire (see Chapter 3.5 and Table 1) as a means to identify patients attending
primary care clinics who require a change in their current acute migraine treatment. Four
clinically important domains of the migraine experience were assessed: consistency of
response,1,2 global assessment of relief,2,3 headache impact4,5 and emotional response.6 A
multicentre, international study was conducted with Migraine-ACT (the original Migraine-ACT
questionnaire was developed in English then translated into Spanish, German and Italian for use
in the study. The translations were verified by blind back-translation into English). The study
demonstrated that Migraine-ACT was easy to use, and was reliable and accurate in its
assessments and suitable for use by the primary care physician. Scoring the questionnaire is by
simple summing of the ‘yes’ scores (range 0–4).7 This article describes further analyses of the
Migraine-ACT study database, providing additional information on the reliability, validity and
clinical utility of the questionnaire.

Table 1. The 4-item Migraine-ACT questionnaire recommended for use in primary care (Form
A).7

Please answer all four questions below, as ‘yes’ or ‘no’, by placing a tick in the relevant box.
Question Yes No
When you take your treatment:
Does your migraine medication work consistently, in
the majority of your attacks?

When you take your treatment:

Does the headache pain disappear within 2 hours?

When you take your treatment:

Are you able to function normally within 2 hours?

When you take your treatment:

Are you comfortable enough with your medication to
be able to plan your daily activities?

Migraine-ACT Score

Scoring the Migraine-ACT questionnaire

One or more ‘no’ answers may indicate the need to change treatment. An increasing number of
‘no’ answers indicates increasing treatment needs.

© Practical Solutions in Medicine Ltd. 2004. Permission is freely given for the use of the Migraine-
ACT questionnaire by individuals for the purposes of research and to improve treatment of
migraine patients. As a courtesy, researchers and commercial parties are asked to inform the
copyright holder in advance before using the questionnaire (email address: migraineact@psim-
limited.com).

172
Methods
Patients and design
Details of the patient inclusion and exclusion criteria and the overall study design are described
in Chapter 3.5.

Reliability
Test-retest reliability was previously assessed by Pearson and Spearman measures.7 Further
analyses of reliability were conducted in terms of the distribution of Migraine-ACT scores
recorded at baseline and 1 week later for the total sample of patients and the individual
countries. Results were expressed as the proportion of patients at each score. Test-retest
reliability was assessed as the Pearson correlation and the Spearman rank correlation for
comparisons of data from the total patient sample and the individual countries.

Validity
Construct validity of Migraine-ACT was previously assessed7 using discriminatory t-tests to
relate dichotomous ‘yes’ and ‘no’ responses scored 1 and 0 in patients’ answers to the 27-item
Migraine-ACT with those to the Short-Form 36 quality of life questionnaire (SF-36),8 the Migraine
Disability Assessment (MIDAS) questionnaire9 and the Migraine Therapy Assessment
Questionnaire (MTAQ) [see ‘Methods’ chapter].10 Further analyses of validity were conducted in
terms of correlating the t-discrimination values for clinically-relevant questions within domains of
the original 27-item questionnaire. t-discrimination values were compared for the following items:
• Impact: Are you able to function normally within 2 hours (IMP 4)?; Do you usually not
need to lie down after 2 hours (IMP 5)?; Are you able to concentrate normally within 2
hours (IMP 9)?; Have any feelings of tiredness, irritability, sadness, anger and anxiety
disappeared within 2 hours (IMP 11)?.
• Global assessment of relief: Does the headache pain disappear within 2 hours (GAR
1)?; Is the headache pain relieved after 2 hours (GAR 2)?; Does the headache pain
disappear within 2 hours and not return within 24 hours (GAR 3)?; Does the headache
pain disappear within 2 hours and not return within 48 hours (GAR 4)?; Does the
headache pain start to disappear within 30 minutes (GAR 5)?; Do your non-headache
symptoms (e.g. nausea and sensitivity to light, sound or smells) disappear within 2
hours (GAR 6)?; Are you clear of all your symptoms within 2 hours (GAR 8)?

An ‘alternative’ Migraine-ACT questionnaire was also constructed, derived from an analysis of

the second best items (Form B) for each domain in the original study data. This questionnaire
was examined for reliability (Spearman rank correlations) and validity (t-discrimination scores).
The overall reliability of the original questionnaire (Form A) and the Form B questionnaire were
compared using Pearson pairwise correlations.

Clinical utility
The clinical utility of the Migraine-ACT questionnaire was examined in several ways to relate
Migraine-ACT scores to medical need. Firstly, the distribution of Migraine-ACT scores recorded
above in the reliability analysis was examined. Secondly, the Migraine-ACT score was correlated
to the total scores obtained for the SF-36, MIDAS and MTAQ questionnaires, using Pearson
pairwise correlations. Following this, Migraine-ACT scores were compared to clinically-defined
criteria as analysed by the SF-36, MIDAS and MTAQ questionnaires.

Results
Reliability
The distribution of Migraine-ACT scores at baseline and 1 week later for the different countries
and the total patient population are shown in Table 2. Analyses of these data and the Pearson

173
and Spearman test-retest correlations (Table 3) showed that the data between the two
completions of Migraine-ACT were consistent for the total population and for the individual
countries. All correlations were significant to p < 0.01. Pearson correlations ranged from 0.61 to
0.92, and similar values were reported for the Spearman rank correlations. The correlation for
the total patient population was r = 0.81 for both analyses.

Table 2. Consistency of the 4-item Migraine-ACT questionnaire: proportion of patients assigned

to each Migraine-ACT score during the two completions of the questionnaire (total sample and
data for each country). Scores range from 0 (all answers are ‘No’) to 4 (all answers are ‘Yes’).

Total sample
1st completion (n = 190) 2nd completion (n = 185)
Migraine-ACT n Patients (%) n Patients (%)
score
4 73 38.4 68 36.8
3 39 20.5 30 16.2
2 28 14.7 36 19.5
1 26 13.7 26 14.1
0 24 12.6 25 13.5

UK
1st completion (n = 40) 2nd completion (n = 39)
Migraine-ACT n Patients (%) n Patients (%)
score
4 16 40.0 16 41.0
3 9 22.5 6 15.4
2 5 12.5 7 17.9
1 4 10.0 4 10.3
0 6 15.0 6 15.4

Spain
1st completion (n = 40) 2nd completion (n = 40)
Migraine-ACT n Patients (%) n Patients (%)
score
4 15 37.5 14 35.0
3 6 15.0 7 17.5
2 3 7.5 7 17.5
1 5 12.5 3 7.5
0 11 27.5 9 22.5

174
USA
1st completion (n = 40) 2nd completion (n = 40)
Migraine-ACT n Patients (%) n Patients (%)
score
4 20 50.0 18 45.0
3 8 20.0 5 12.5
2 7 17.5 10 25.0
1 5 12.5 5 12.5
0 0 0 2 5.0

Germany
1st completion (n = 30) 2nd completion (n = 26)
Migraine-ACT n Patients (%) n Patients (%)
score
4 7 23.3 9 34.6
3 7 23.3 3 11.5
2 7 23.3 4 15.4
1 8 26.7 6 23.1
0 1 3.3 4 15.4

Italy
1st completion (n = 40) 2nd completion (n = 40)
Migraine-ACT n Patients (%) n Patients (%)
score
4 15 37.5 11 27.5
3 9 22.5 9 22.5
2 6 15.0 8 20.0
1 4 10.0 8 20.0
0 6 15.0 4 10.0

Table 3. Test-retest reliability of the Migraine-ACT questionnaire: Pearson and Spearman rank
correlations for the completions of the Migraine-ACT questionnaire at baseline and 1 week later:
analysis by country and for the whole patient sample.

Country Pearson Spearman rank

correlation correlation
UK 0.91 0.90
Spain 0.92 0.92
USA 0.61 0.65
Germany 0.74 0.74
Italy 0.75 0.77
Total 0.81 0.81
patient sample

All comparisons were significant to p < 0.01

175
Validity
Figure 1 shows the analysis of t-discrimination scores for assessments of impact and the global
assessment of relief. In examining the relative discriminatory value of measures of migraine
impact, the ability to function normally within 2 hours (IMP 4) was markedly more discriminatory
than not needing to lie down after 2 hours (IMP 5), being able to concentrate normally within 2
hours (IMP 9) and the disappearance of emotional symptoms within 2 hours (IMP 11). In
examining the relative discriminatory value of measures of global assessment of relief, the
assessment of headache pain disappearance at 2 hours (GAR 1) was markedly more
discriminatory than that of headache relief at 2 hours (GAR 2), sustained pain-free rates over 24
hours (GAR 3) or 48 hours (GAR 4), early onset of pain disappearance (GAR 5), disappearance
of non-headache symptoms (GAR 6) and disappearance of all symptoms (GAR 8).

Figure 1. Comparison of t-discrimination values for Migraine-ACT items equivalent to commonly-

used endpoints in migraine studies. a. Impact: ability to function normally within 2 hours (IMP 4),
no need to lie down after 2 hours (IMP 5), able to concentrate normally within 2 hours (IMP 9)
and symptoms of tiredness, irritability, sadness, anger and anxiety disappear within 2 hours (IMP
11). b. Global assessment of relief. pain-free at 2 hours (GAR 1), 24 (GAR 3) and 48 (GAR 4)
hours, headache relief at 2 hours (GAR 2), start of headache relief within 30 minutes (GAR 5),
free of non-headache symptoms at 2 hours (GAR 6) and clear of all symptoms within 2 hours
(GAR 8).

a.

3.5
3.07
3 2.76

2.5 2.35
t-discrimination

1.5

1 0.69

0.5

0
IMP 4 IMP 5 IMP 9 IMP11

176
 b.

3 2.70

2.5 2.33
1.98
t-discrimination

2 1.86 1.83
1.62
1.5

1 0.84

0.5

0
GAR 1 GAR 2 GAR 3 GAR 4 GAR 5 GAR 6 GAR 8

Table 4 shows the ‘alternative’ (Form B) 4-item questionnaire that was derived from the original
27-item Migraine-ACT by selecting the second most discriminatory question in each domain.
Calculated values for reliability (Spearman rank correlation) and validity (t-discrimination scores)
for each question were similar to those of the original 4-item questionnaire. Pearson pairwise
correlations showed that the overall test-retest reliability of the Form B questionnaire was high, (r
= 0.81) and was virtually the same as that of the original Form A questionnaire (r = 0.82). The
correlation was significant to p < 0.01.

177
Table 4. An ‘alternative’ (Form B) 4-item Migraine-ACT questionnaire and its reliability and
validity, compared with those of the original questionnaire. Values in parentheses show the
equivalent data for the original Migraine-ACT questionnaire (Form A7).

Question Reliability Validity

(Spearman (t-discrimination)
rank
measure)
When you take your treatment:
Does your migraine medication work attack 0.68 2.06
after attack? (0.61) (2.08)

When you take your treatment:

Are you clear of all your symptoms within 2 0.58 2.33
hours? (0.58) (2.70)

When you take your treatment:

Are you able to resume your normal activity 0.77 2.96
(i.e. work or family-leisure-social activity, etc.) (0.72) (3.07)
within 2 hours?

When you take your treatment:

Do you feel in control of your migraine enough 0.63 3.12
so that you feel there will be no disruption to (0.78) (3.16)
your daily activities?

Clinical utility
The distribution of Migraine-ACT scores for the total population was: 0 = 12.6%, 1 = 13.7%, 2 =
14.7%, 3 = 20.5% and 4 = 38.4%. The distribution of scores between the two completions of the
questionnaire was consistent between countries (Table 2).

Pearson pairwise correlations for the comparison of Migraine-ACT scores with the total SF-36,
MIDAS and MTAQ scores are shown in Table 5. All correlations were significant at p < 0.05. The
data showed that Migraine-ACT closely correlated with SF-36, MIDAS and MTAQ total scores
both times it was completed (at baseline and after 1 week). The tendency was for a greater
correlation (stronger relationship) with MTAQ > MIDAS > SF-36. The pattern of correlations
showed great construct consistency.

Table 5. Pearson pairwise correlations for the comparisons of Migraine-ACT, SF-36, MIDAS and
MTAQ total scores: comparisons of data at baseline and 1 week.

Total SF-36 score Total MIDAS score Total MTAQ score

Migraine-ACT 0.36 -0.32 -0.60
score (baseline)
Migraine-ACT 0.34 -0.39 -0.58
score
(1 week)

All correlations were significant at p < 0.05.

178
Figure 2 shows how the Migraine-ACT scores corresponded to SF-36, MIDAS and MTAQ
scores. The relationship was essentially linear in all cases. As the Migraine-ACT score
decreased from 4 to 0, the SF-36 score deceased from 53.6 to 44.0, the MIDAS score increased
from 16.0 to 39.3 and the MTAQ score increased from 2.3 to 5.0.

Figure 2. Relationship of Migraine-ACT scores to clinically-defined criteria as analysed by scores

on the SF-36, MIDAS and MTAQ questionnaires.

60

50
Score (arbitrary units)

40
SF-36
30 MIDAS
MTAQ
20

10

0
4 3 2 1 0

Increasing medical need

Discussion

The aims of this post hoc analysis of the original Migraine-ACT study were to provide further
data on the robustness of the results, and to give the physician insights into the clinical utility of
the questionnaire and how it might best be used in the clinic. To do this, new analyses of
reliability and validity were conducted and examined for clinically relevant data.

The original assessments of test-retest reliability used appropriate statistical methodology

(Pearson product moment and Spearman rank measures). This methodology has been used in
examining the reliability of, among others, the MIDAS Questionnaire, both in its original English
version,5 and in scientific translation into other languages.11 The 27 items of the original
Migraine-ACT questionnaire all exhibited good-to-excellent reliability over 1 week, and the final
4-item Migraine-ACT was a highly reliable measure (r = 0.82).7 Data presented here
demonstrate that there was high consistency in the Migraine-ACT scores reported from the
individual countries and the total population. Test-retest reliability ranged from 0.61–0.92
according to country. Additionally, an ‘alternative’ 4-item Form B questionnaire derived from the
same 27 questions was also highly reliable (r = 0.81). These data reinforce the good reliability of
the 4-item Migraine-ACT questionnaire.

179
The assessment of construct validity (accuracy) commonly involves the comparison of data from
a new tool with those from standard existing tools (‘gold standard’ measures).12 In the original
study, Migraine-ACT was compared with the SF-36 QOL questionnaire, the MIDAS disability
questionnaire and the MTAQ disease management questionnaire, and was shown to correlate
well with these questionnaires, especially with the total MTAQ and SF-36 scores.7 Data
presented here showed that a second Form B 4-item questionnaire could be derived from the
original 27 questions in which the questions were almost as accurate as the 4-item Migraine-
ACT questionnaire. The two sets of questions exhibited considerable redundancy, indicating that
they accurately measured the same specific domains. Additionally, three of the original
questions assessing global assessment of relief (GAR 1, 4 and 8) and three of the impact
questions (IMP 4, 9 and 11) showed similar levels of discrimination.

This conclusion is reinforced when examining the t-discrimination scores associated with
different types of questions used in the ‘impact’ and ‘global assessment of relief’ domains.
Migraine is associated with different areas of impact, including:
• The ability to function normally, which can be assessed with the MIDAS questionnaire.13
• The need to lie down during attacks, which was assessed as important in the studies that
developed the Headache Impact Test.14
• Cognitive changes, which are frequently reported during attacks, and have been
assessed in one study.15
• Emotional changes, which have been identified as important components of migraine,
tension-type headache and chronic daily headache.6,16
The results reported here indicate that assessing the ability to function normally after treatment
is markedly more discriminating in assessing impact that the other three measures. Numerous
studies have indicated that the disruption of patients’ everyday activities is the key measure of
illness severity in migraine patients.4,5,13,17,18

Over the years, much research has attempted to define the optimal measure of relief of migraine
symptoms, usually using a 2-hour endpoint:
• The disappearance of the headache, which is used as the gold standard measure by the
International Headache Society.3
• The relief of headache, defined as an improvement from severe or moderate to mild or
none, and used as the primary endpoint in most migraine studies to this day.19
• Disappearance of the headache which is sustained for 24 or 48 hours after treatment.2,20
• Onset of headache disappearance, usually defined as within 30 minutes after
treatment.21
• Disappearance of non-headache symptoms of the migraine attack.19
• Disappearance of all symptoms of the migraine attack.20
The results reported here indicate that assessing the disappearance of the headache at 2 hours
after treatment is markedly more discriminating in assessing relief than any of the other
measures above. The results therefore agree with recent studies that have indicated the pain-
free measure is the most important clinical endpoint in migraine studies.2,3,20 Migraine-ACT
therefore contains two highly valid measures for assessing the efficacy of treatment. Domains of
consistency of response and emotional response have not been assessed to anything like the
same extent as impact and global response, but are both clinically relevant in assessing the
constellation of migraine symptoms.2,6

The clinical relevance of these data is considerable. The original Migraine-ACT questionnaire
was developed in English then translated into Spanish, German and Italian for use in the study.

180
The translations were verified by blind back-translation into English. The high values obtained
for consistency and reliability in the distribution of the Migraine-ACT grade scores attest to the
robustness of the questionnaire for use in different countries. Migraine-ACT has since been
translated into Dutch (Appendix 1) and its use is advocated in the Netherlands.22

The relationship of Migraine-ACT scores to scores on the SF-36, MIDAS and MTAQ
questionnaires provides further evidence of clinical utility. Pearson pairwise correlations
confirmed that the Migraine-ACT score was closely related to SF-36, MIDAS and MTAQ total
scores. The linear relationship between the range of Migraine-ACT scores and scores for the
three other questionnaires is noteworthy and demonstrates an even gradation of Migraine-ACT
scoring. The mean SF-36 score in the US population is 508 and patients with Migraine-ACT
scores of 2, 1 and 0 would therefore seem to have lower quality of life than the general
population (SF-36 scores of 48.9, 46.5 and 44.0, respectively). The MIDAS scores ranged from
16.0 (Migraine-ACT score 4) to 39.3 (Migraine-ACT score 0). MIDAS scores of 11–20 (Grade III)
equate with moderate disability, while scores of > 20 (Grade IV) equate to severe disability.9 The
mean MIDAS score for migraine patients is typically in the Grade III range.9 MTAQ scores
ranged from 2.3 (Migraine-ACT score 4) to 5.0 (Migraine-ACT score 0). MTAQ scores range
from 0–8 depending on the number of management issues identified for the patient. A score of ≥
2 is used as the threshold for when follow up is warranted.10

From these data, it can be deduced that a Migraine-ACT score of 2, 1 and 0 identifies patients
for whom a change in acute treatment may be warranted. These patients have reduced quality
of life, severe headache-related disability and a considerable number of outstanding
management issues. In this study, 41.0% of patients met this criterion, and the results were
relatively similar between the countries (UK = 37.5%, Spain = 47.5%, USA = 30.0%, Germany =
53.3%, Italy = 40.0%, Table 2). It might be expected that the proportion of patients presenting to
primary care physicians with Migraine-ACT scores of ≤ 2 would be even higher, due to the
likelihood of them having received suboptimal care.23–27

The general limitations of the Migraine-ACT study are described in Chapter 3.5. From the
analyses described here, further work is desirable to investigate the clinical significance of the
Migraine-ACT scoring scheme. In addition, further studies are warranted to investigate Migraine-
ACT as an outcome tool to monitor the effect of interventions.

In conclusion, these data confirm the good reliability and validity of the Migraine-ACT
questionnaire and provide further evidence for its utility in clinical practice. It is suggested that a
score of ≤ 2 may indicate that a change in the patient’s acute medication is warranted and a
score of ≤ 1 may indicate that the change is mandated.

References
1. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity. Disability, pain intensity,
and attack frequency and duration. Neurology 1994;44 (Suppl 4):24–39.
2. Ferrari MD, Roon KI, Lipton RB et al. Oral triptans (serotonin 5-HT1B/1D agonists) in acute
migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358:1668–75.
3. International Headache Society Clinical Trials Subcommittee. Guidelines for controlled
trials of drugs in migraine: second edition. Cephalalgia 2000;20:765–86.
4. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
5. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of the
Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.

181
6. Kilminster SG, Dowson A, Bundy M. Headache Impact Test (HIT) and Short Pain
Inventory©. Outcome measures compared: Int J Pharm Med 2003;17:23–32.
7. Dowson AJ, Tepper SJ, Baos V et al. Identifying patients who require a change in their
acute migraine treatment: The Migraine Assessment of Current Therapy (Migraine-ACT)
Questionnaire. Curr Med Res Opin 2004;20:1125–35.
8. Ware JE, Snow KK, Kosinski M et al. SF-36 Health Survey manual and interpretation
guide. Boston: New England Medical Center, 1993.
9. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56 (Suppl 1):S20–28.
10. Chatterton ML, Lofland JH, Shechter A et al. Reliability and validity of the Migraine
Therapy Assessment Questionnaire. Headache 2002;42:1006–15.
11. D’Amico D, Mosconi P, Genco S. The Migraine Disability Assessment (MIDAS)
questionnaire: translation and reliability of Italian version. Cephalalgia 2001;21:947–52.
12. Stewart WF, Lipton RB, Simon D et al. Validity of an illness severity measure in a
population sample of migraine headache sufferers. Pain 1999;79:291–301.
13. Lipton RB, Stewart WF, Sawyer J et al. Clinical utility of an instrument assessing
migraine disability: the Migraine Disability Assessment (MIDAS) questionnaire.
Headache 2001;41:854–61.
14. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
15. Evers S, Ruschenschmidt J, Frese A et al. Impact of antimigraine compounds on
cognitive processing: a placebo-controlled crossover study. Headache 2003;43:1102–8.
16. Dowson AJ, Kilminster SG, Bundy M et al. Emotional function with tension-type
headache, migraine and chronic daily headache. Headache Care 2005 (manuscript in
preparation).
17. Lipton RB, Dodick D, Sadovsky R et al. A self-administered screener for migraine in
primary care: The ID Migraine validation study. Neurology 2003;61:375–82.
18. Cady RK, Borchert LD, Spalding W et al. Simple and efficient recognition of migraine with
3-question headache screen. Headache 2004;44:323–7.
19. Pilgrim AJ. Methodology of clinical trials of sumatriptan in migraine and cluster headache.
Eur Neurol 1991;31:295–9.
20. Adelman JU, Lipton RB, Ferrari MD et al. Comparison of rizatriptan and other triptans on
stringent measures of efficacy. Neurology 2001;57:1377–83.
21. Charlesworth BR, Dowson AJ, Purdy A et al. Speed of onset and efficacy of zolmitriptan
nasal spray in the treatment of migraine: a randomised, double-blind, placebo-controlled,
dose-ranging study versus zolmitriptan tablet. CNS Drugs 2003;17:653–67.
22. Anon. Checklist migraine. www.hoofdpijn.nl.
23. Edmeads J, Findlay H, Tugwell P et al. Impact of migraine and tension-type headache on
life-style, consulting behaviour, and medication use: a Canadian population survey. Can
J Neurol Sci 1993;20:131–7.
24. Stang PE, Von Korff M. The diagnosis of headache in primary care: factors in the
agreement of clinical and standardized diagnoses. Headache 1994;34:138–42.
25. Celentano DD, Stewart WF, Lipton RB et al. Medication use and disability among
migraineurs: a national probability sample survey. Headache 1992;32:223–8.
26. MacGregor EA, Brandes J, Eickermann A. Migraine prevalence and treatment patterns:
the global Migraine and Zolmitriptan Evaluation survey. Headache 2003;43:19–26.
27. Dueland AN, Leira R, De Vries J et al. A multinational study of migraine treatment
patterns among young, active women: the Migraine Effect on Life and Treatment (MELT)
study. Headache Care 2004;1:19–21.

182
5

General discussion

The main aim of these studies was to investigate the clinical significance of headache-related
disability and impact; the clinical importance of assessing disability, the means of recognising
the patients with severe disability and the development of new ways to assess headache-related
disability and their utility in clinical practice. The main methodology used in the thesis was critical
reviews of the literature and open, naturalistic studies designed to mimic the situation in
everyday primary care clinical practice. In this chapter the main study findings are discussed and
related to the published clinical literature. The research and clinical implications deduced from
the data are outlined. Finally, the methodology of the studies is criticised and future studies and
directions for further research outlined.

5.1 Main results

Clinical importance of assessing headache-related disability
For evidence-based medicine, the randomised, double-blind, controlled clinical study remains
the gold standard for evaluating the clinical profile of drugs.1,2 However, as discussed in Chapter
3.1, several criteria need to be met before the study can be categorised as high quality Grade A
clinical evidence. Studies need to be randomised, double-blind and to contain enough patients in
each treatment group to allow for sufficient power for the statistical analyses. Patient withdrawals
should be low (≤ 10%) and equivalent in the separate treatment groups. Outcome measures
should be quantifiable and clinically meaningful, and statistical tests should be appropriate for
the populations and endpoints studied.3 These criteria should be mandatory for all controlled
studies (Figure 1). Additionally for headache, a placebo group should always be included to
control for the variable placebo response that characterises the condition.4

Figure 1. Criteria for the design of Grade A clinical studies.

Randomised,
double-blind design

Placebo / comparator
controlled

Appropriate patient
group selected

Diagnosis by
defined criteria

Study powered
appropriately

Validated, clinically
relevant endpoints

Tolerability /
safety assessed

183
Not surprisingly, these stringent conditions are not always met, and post hoc and meta-analyses
are sometimes used if clinical evidence is inconclusive and when comparisons between drugs
are desired but a proper controlled study has not yet been conducted. For headache, such
analyses have generated controversy, rather than contributing substantially to understanding the
efficacy of acute migraine therapies (with inconsistencies in the studies included and analyses
conducted).5–8 Their chief utility should be to suggest avenues for future clinical studies and they
should not be used on their own as a means of selecting acute treatments for migraine.3

However, even results from double-blind, controlled clinical trials may not always be optimal in
terms of reflecting the situation in clinical practice. For example, in migraine controlled clinical
studies show only small differences between the oral triptans, whereas patients describe clear, if
individual, preferences between them.3 The conundrum is that individual patients respond
differently to medications that have similar efficacy profiles, and reinforce the need for an
approach to care that is tailored to the needs of each patient. The basis for these differences in
effect may be due to biological (e.g. neuroreceptor profile) and/or clinical (e.g. co-morbidities and
the presence of single or multiple headache conditions) factors. Patients differ in how they cope
with pain, their sensitivity to pain and in other ways that lead to headache being a
heterogeneous rather than an isolated condition. Only very large studies using vast numbers of
patients can detect statistically significant differences between treatments in such a
heterogeneous environment. However, the differences are often small and not necessarily
clinically relevant. This fundamental methodological problem is encountered in other medical
fields, e.g. palliative care and psychiatric disorders.

The development of new, more sensitive endpoints to assess migraine relief may therefore be
needed. Proposed new endpoints tend to be global measures that combine assessments of
efficacy and tolerability, including assessments of patient preference and satisfaction, the impact
on patients’ daily lives and the patient’s quality of life. Of these three types of endpoints, patient
preference / satisfaction and headache impact can provide simple and intuitive assessments,9,10
while quality of life is a more complex measure wherein the units of assessment do not relate
directly to clinically meaningful measures.11 Such endpoints may be more valid and sensitive
than the traditional headache-based ones, although this has to be demonstrated in clinical
studies.

Assessing patient preference

Chapter 3.2 reviews and discusses the data on using patient preference / satisfaction as an
outcome measure in migraine studies.9 Many studies have used patient preference as
secondary analyses of efficacy. Overall, the results show that patients can distinguish
differences in efficacy between the triptans and non-triptan medications for migraine, and also
between the different triptans.9 These data are remarkable, as it has proved difficult to show
significant differences between acute migraine medications using the usual primary endpoint of
relief of headache.12–20 More limited data for patient satisfaction show similar trends.9 A limitation
of these data is that, as secondary analyses, they were not subject to rigorous statistical
analyses.

Patients’ assessment of their preference for their migraine treatments may therefore be a valid
and sensitive measure of clinical efficacy that is relevant to real-life clinical practice, taking into
account both efficacy and tolerability. The results also indicate that patient preference may be a
more sensitive measure of efficacy and clinical utility than conventional clinical trial endpoints,
although this remains to be proven in rigorously designed clinical studies. Another limitation of
the data is that patients’ preferences are individual and not predictable beforehand, possibly

184
being based on individual biological and/or clinical factors.9 Despite this, physicians should
consider using assessments of patient preference when reviewing the efficacy of acute migraine
medications, based on the author’s clinical experience. It is important that new instruments to
assess preference/ satisfaction are appropriately developed and tested. It would be good to
have a standard instrument that is used in all studies to make the findings more easily
comparable.

Assessing headache-related disability

Headache can be a significantly disabling condition, as defined by the World Health
Organization as a condition resulting in limitations to daily activities.21 Migraine is particularly well
categorised as being disabling. Numerous studies (reviewed in Chapter 4.1) demonstrate that
migraine results in sufferers being unable to conduct their everyday daily activities (employment,
household tasks and family and leisure activities) during their attacks.22 The studies of Lipton,
Stewart and colleagues elegantly demonstrate that disability rather than pain severity is the key
measure of migraine severity.23 Tension-type headache (TTH) is generally thought to be a
relatively non-disabling condition in terms of pain and headache-related disability.24 The study
assessing emotional impact of headache with the SPI questionnaire (Chapter 3.4) indicated that
the pain associated with TTH was indeed transient and mild-to-moderate intensity. However, the
results showed a marked chronic sedation associated with TTH25 that has been noted once
before.26 These data may indicate that the disability associated with TTH is currently under-
estimated. The same study showed that chronic daily headache (CDH) was associated with
chronic pain and emotional impact that was severe in intensity.25 These data confirm the well-
recognised severe disability associated with CDH,27 but extend the impact into the emotional
domain. From all these data, it is clear that physicians should ask headache patients about their
level of disability before deciding on a management plan, although it seems that few currently do
so.28 In fact, information on disability does change the physician’s perception of illness severity
and subsequent handling of management.28

There are currently limited data on the use of disability or impact tools as endpoints in clinical
studies. However, two studies have shown that the MIDAS questionnaire is sensitive to change,
and can be used to monitor patients with migraine or CDH in the clinic.29,30 The HIT
questionnaire has been much less studied in this regard and, as described in Chapter 4.3, may
be more useful as a diagnostic tool than as a measure of headache severity.31 Both MIDAS and
HIT capture retrospective data, over 3 months and 1 month, retrospectively.22 The SPI
questionnaire collects prospective data,32 and may therefore be a better clinical study endpoint
than either MIDAS or HIT, although this can only be demonstrated with further studies.

Recognising headache-related disability

As already described, assessing headache-related disability is important in the clinical
evaluation of headache. It is therefore important that the physician is able to recognise patients
who are disabled and to manage them accordingly. The level of disability also needs to be
monitored over time, and this aspect of recognising disability was assessed in two studies in this
thesis:
• Investigating the level of disability associated with headaches occurring inside and
outside the menstrual period
• Investigating the emotional impact of different headache subtypes over time.

Disability associated with headaches occurring inside and outside the menstrual period
The study described in Chapter 3.3 provided a snapshot of headache in women attending a
primary care clinical practice, and also provided information on their level of depression and
bodily pain.33 Patients who completed an initial questionnaire reported high levels of headache,

185
depression and bodily pain. About two-thirds of patients reported headache, over 50%
symptoms indicative of depression, and about 40% bodily pain. The symptoms were reported
across all age groups. Overall, the results indicate that headache, depression and bodily pain
are common in the general adult female population and that many patients clearly have multiple
symptoms. Physicians would benefit from asking patients about their levels of pain and mood
disturbance. Headache is frequently co-morbid with psychiatric disorders (particularly major
depression and anxiety),34 and patients with multiple co-morbidities are likely to require more
care than those with a single condition.

Migraine sufferers in this study were invited to complete a disability questionnaire encompassing
some of the items now contained in the MIDAS questionnaire over a 2-month period, covering at
least a full menstrual period. However, this was not a MIDAS study, being designed before the
full development of the MIDAS questionnaire. The data showed that migraine headaches were
relatively more frequent inside than outside the menstrual period, while non-migraine headaches
were approximately equally distributed inside and outside the menstrual period. This data for the
distribution of migraine attacks is in line with recently published data.35 The study also showed
that migraine attacks that occurred inside the menstrual period were somewhat more disabling
than those occurring outside the menstrual period in terms of rank sums of the MIDAS items of
time lost and time at less than 50% productivity. These results are in line with a recent study
indicating that menstrually-associated migraine attacks were only slightly more painful than non-
menstrually-associated ones,36 but not with another study indicating that attacks during
menstruation were markedly more disabling than those occurring outside the menstrual period.35
Results for non-migraine headache were the opposite of those for migraine, with attacks
occurring outside the menstrual period scoring numerically, but not significantly, higher than
those occurring inside the menstrual period in terms of the rank sums of time lost and time at
less than 50% productivity. These data indicate that the pattern of headache-related disability
can differ at different time periods in the patient’s life and also with respect to the different
headache subtypes experienced by the patient. They emphasise the importance of a correct
diagnosis (or diagnoses) and the monitoring of patients over time with a headache diary or other
form of questionnaire.

The emotional impact of different headache subtypes over time

The study described in Chapter 3.4 categorised the emotional impact of episodic TTH, migraine
and CDH with the SPI questionnaire.25 This allowed the analysis of the emotional ‘footprint’ of
each headache subtype in terms of the emotional experience, associated sedation, disruption to
social interactions and the time course of each of these components. Patients recorded their
headaches with the SPI at peak intensity and at intervals for 7 days afterwards.

At peak intensity the pain level was mild-to-moderate for episodic TTH, and patients exhibited
sedation, disturbances in social interaction, sadness, anxiety and anger. These symptoms all
resolved within 24 hours, except for sedation, for which marked levels persisted for most of the
following week. Coping scores for sedation were the most disturbed of all the three types of
headache, even including CDH. This is surprising, considering the relatively mild intensity of
TTH. Sedation may be a precipitating factor for TTH or a feature of the recovery process that is
unrelated to the headache severity. Also, coping scores for sadness were sometimes above the
normal range for 3 out of the 7 days post-headache. The results indicate that TTH patients have
mood lability and may mean that the illness severity of patients is currently underestimated by
physicians.

At peak intensity the pain level was moderate-to-severe for migraine and patients exhibited
sedation, disturbances in social interaction, sadness, anxiety and anger. In general, the severity

186
of these emotional components was more intense for migraine than for TTH. All these symptoms
typically resolved within 2 days of the headache and coping scores were at normal levels after
the headache. These data indicated that the patients coped well with the emotional symptoms,
which resulted from the level of physical pain experienced.

At peak intensity the pain level was severe for CDH and patients exhibited sedation,
disturbances in social interaction, sadness, anxiety and anger. In general, the severity of these
emotional components was more intense for CDH than for TTH. However, these symptoms were
maintained over the 7-day period following the headache. Pain remained at a mild-to-moderate
level. Marked levels of sedation, disturbed social interaction, sadness, anxiety and anger were
also seen during this period. The coping scores for sedation, social interaction and sadness
were above the level caused by the pain. The result is that CDH patients are chronically
fatigued. CDH is associated with Chronic Fatigue Syndrome37 and sedation is a major feature of
the post-dromal period of CDH. Overall, the level of emotional disturbance experienced by CDH
patients was similar to that seen in patients attending chronic pain clinics.38 An unfortunate
feature of CDH patients is that their mood disturbance never returns to normal and patients
obtain no respite from the pain and its emotional consequences. This pattern of emotional
impact associated with CDH is similar to that reported in a qualitative study.39

Overall, patients with TTH, migraine and CDH all had considerable mood disturbances during
their headaches. The more severe the pain they experienced, the greater was the level of mood
disturbance. The level of mood disturbance was also directly related to the healthcare utilisation
of the patients. The pattern of mood disturbance was different for each type of headache, as
was the patients’ ability to cope with the mood changes. The association of mood disturbances
with common headache subtypes may explain in part the reported co-morbidity of headache
with psychiatric disorders.34 Physicians should elicit the emotional needs of their headache
patients, which may not be currently appreciated or addressed by healthcare services, as they
play no part in the current classification criteria for headache.24

Assessing headache-related disability

Chapters 3.3 and 3.4 show that disability is a key feature of headache, and can differ in intensity
and quality in different headache subtypes and in individual patients over time.25,33 Assessing
headache-related disability should therefore have utility in monitoring patients, as an outcome
measure and possibly for other assessments. A review of the methodological robustness and
clinical utility of disability and impact questionnaires was conducted for the thesis (Chapter 4.1)22
as well as a series of studies investigating the utility of disability questions in diagnosis
(Chapters 4.3 and 4.4)31,40 and as outcome measures for patient assessment (Chapters 4.2, 4.5
and 4.6).41–43

Reviewing disability measures for headache

Research (reviewed in Chapter 4.1) has shown that headache-related disability and impact can
be measured using assessments of some or all of pain severity, headache frequency, limitations
to work and leisure activities, tiredness, mood alterations and cognition. Disability-based
measures include the MIDAS, HIT and SPI questionnaires, all of which are now available for
clinical use.22

The MIDAS questionnaire retrospectively measures headache-related disability as time lost (as
the number of days) in employment, unpaid work and family and leisure activities. Numerous
studies have shown that it is reliable,10 valid,10 sensitive to change29,30 and exhibits considerable
clinical utility in the clinic.44 HIT (either in the Internet45 or paper-based HIT-646 forms)
retrospectively assesses a global range of impact, including pain intensity, disability and other

187
items. HIT is also reliable and valid,45,46 and has some evidence of clinical utility for assessing
diagnosis and headache severity in the clinic.47,48 Chapter 4.3 of this thesis indicates that HIT is
a good diagnostic tool but a relatively poor tool for assessing headache severity.31 The SPI
questionnaire prospectively assesses pain intensity and emotional impact of the headache over
time.32 It is also reliable and valid,49 and can distinguish the emotional components of different
pain states50 and, as shown in Chapter 3.4, those of different headache subtypes.25 Synthesising
the available evidence, it would seem that HIT can be used for headache diagnosis, and MIDAS
and SPI for assessing headache severity and as outcome measures in clinical studies. MIDAS
has been shown to be a useful outcome tool in different headache disorders,29,30 and we await
similar evidence from studies with SPI.

The assessment of headache-related disability is a key item of recently published migraine

guidelines in the USA1 and migraine and CDH guidelines in the UK.51,52 It is used to improve
communication between physicians and their patients and, with other measures, to assess
headache severity to tailor treatment to each patient’s individual needs. However, although used
a lot in specialist clinics, disability tools have not yet penetrated significantly into primary care.
They seem to be relatively unknown and the clinical experience of the author shows them to be
too complex for patients to complete unaided. Clearly, educational initiatives are needed to
promote the use of disability tools in primary care. However, physicians should enquire about
disability and a simple question such as: How do your headaches interfere with your normal
daily activities? may be sufficient to start a discussion and obtain relevant clinical information
from the patient.

Utility of disability questions in headache diagnosis

Already discussed are the results from Chapter 4.3 investigating the clinical utility of the HIT and
SPI questionnaires. The results showed that the Internet version of HIT is very good at
differentiating the diagnosis of TTH, migraine and CDH.31 HIT was not designed to be a
diagnostic tool, but to measure the impact headaches have on a person’s ability to function at
work, at home, at school and in social situations. Indeed, the developers state that HIT is not a
diagnostic tool.45 We can only speculate on the factors within the HIT questionnaire that are
associated with diagnosis. HIT assesses a constellation of factors: headache-related pain
severity and disability, the need to lie down, tiredness and mood alterations. As SPI assesses
emotional function and is not a diagnostic tool, it seems unlikely that tiredness and mood
alterations correlate with diagnosis. However, pain severity24 and disability27,47 do differ between
the three headache subtypes, and these seem the most likely drivers of diagnosis, at least in the
HIT questionnaire. Whatever the case, HIT does discriminate between headache subtypes very
well and, should a neurologist be not available, is a good diagnostic test.

Chapter 4.4 describes the development and testing of the 8-item DSQ, designed to be an
inclusive diagnostic aid for headache to be used at the patient’s first consultation with a
healthcare professional.40 The items assessed were headache frequency, duration and related
disability, together with frequency of use of analgesic medications, prevalence of aura symptoms
and a series of questions eliciting potential sinister headaches. The study showed that the DSQ
was simple to complete, either by the patient on their own or in concert with a healthcare
professional. Overall, the DSQ exhibited good diagnostic sensitivity (of about 60%) when
completed by the patient or the pharmacist, especially for total migraine, migraine with and
without aura and MOH. These results indicate that the DSQ is an effective inclusive tool for the
differential diagnosis of headache. Interestingly, the accuracy of the pharmacists’ headache
diagnosis was improved markedly when they used the DSQ, providing evidence of clinical utility.

188
Again, it was not possible to identify the DSQ items that correlated best with diagnosis.
Compared with the HIT questionnaire, an item on disability was present, while one on pain
severity was not. It is tempting to speculate that the assessment of headache-related disability,
in concert with other headache features, is the main predictor of diagnosis. Such a pattern is
seen in other recently-developed diagnostic questionnaires for headache. With the inclusive
Brief Headache Screen, migraine, daily headache syndromes and medication overuse were
distinguished using three questions: the frequency of disabling headache, the presence of other,
mild headaches and the use of symptomatic medications.53 Migraine exclusively could be
diagnosed sensitively and specifically using three questions on headache-related disability,
nausea and sensitivity to light,54 or three questions on disability, headache duration and the lack
of new-onset headaches in the previous 6 months.55 There is considerable overlap between the
DSQ and these other inclusive and exclusive questionnaires in terms of assessing disability,
together with other headache features and symptoms, which reinforces the content and external
validity of the DSQ questionnaire.23

The DSQ is therefore a useful headache screening questionnaire for use when the patient first
seeks care for the condition from a healthcare professional, who may be a GP, nurse,
pharmacist or other practitioner (e.g. opticians and dentists often see people with headache). Its
principal use is proposed to be an aid to deciding the appropriate first management decisions. It
may be particularly suitable for use by pharmacists, which is of clinical importance as they may
see more headache sufferers than do GPs.56

Utility of disability questions as outcome measures for patient assessment

Two studies are included in this thesis investigating the use of disability questions to assess
patient outcome.
1. Chapter 4.2: the clinical utility of the MIDAS questionnaire when used by nurses
at patient screening41
2. Chapters 4.5 and 4.6: the methodological robustness and clinical utility of the
Migraine-ACT questionnaire, which contains two items on disability.42,43

In Chapter 4.2, the MIDAS questionnaire was used to monitor new patients with mild-to-
moderate headache at baseline and for 6 months after being provided with headache advice by
a specialist nurse.41 The patients had significantly improved headache outcomes 6 months later
in terms of headache-related disability, headache frequency and severity, all assessed with
MIDAS. Additionally, they perceived an improvement in headache and reported a reduced need
for consulting a physician for headache. As very few patients consulted a physician during the
study, this improvement could be attributed to the advice given by the nurse on headache
management, despite the absence of a control group. The study also adds to the published data
indicating that MIDAS may be a sensitive outcome measure for headache sufferers.29,30

It is worth noting that the simple provision of advice can lead to a sustained significant
improvement in headache, although previous small studies have suggested the same effect.57,58
Recently published headache guidelines have advocated the provision of advice and information
to patients at headache screening,51,52 and this study illustrates how effective this strategy can
be. The practice nurse has a positive role in the management of headache in primary care,
particularly with respect to screening and follow-up, and guidelines dedicated for nurse use have
been published.59 These recommendations state that nurses should undertake information
dissemination and recording of data using headache history and impact questionnaires and
headache diaries.

189
Chapters 4.5 and 4.6 outline the development of a new outcome questionnaire, Migraine-ACT,
to assess the need to change a migraine patient’s acute medication.42,43 This somewhat limited
functionality is, however, clinically important due to the fact that acute medications are usually
under-used, despite the availability of effective treatments today.60–63 Four clinically-relevant
domains were included in the questionnaire: impact,23,54 global assessment of relief,5,64
consistency of response5,65 and emotional response.25,31 Twenty seven questions were
developed and reduced to four (one in each domain) by testing for reliability and validity. The
tool was designed from the outset to be brief, and simple to complete and score by yes/no
answers. The use of clinically-relevant domains provides prior evidence of face validity.

The individual 27 Migraine-ACT questions all exhibited good-to-excellent test-retest reliability in

the total population and within the five countries taking part in the study. Assessment of
construct validity demonstrated that many of the 27 Migraine-ACT items correlated with the SF-
36 QOL questionnaire,11 the MIDAS questionnaire10 and the MTAQ disease management
questionnaire66 and allowed selection of the best four items in the final questionnaire. The four-
item Migraine-ACT questionnaire was highly reliable and valid. In assessing the impact items,
the ability to function normally after treatment (a measure of headache-related disability) was
markedly more discriminating than other measures. Numerous studies have previously indicated
that the disruption of patients’ everyday activities is the key measure of illness severity in
migraine patients.23,54,55,67,68 In addition, the resolution of headache at 2 hours provided a more
discriminating assessment of global response than other items tested, agreeing with data from
other studies.5,64,69

The total Migraine-ACT score is achieved by summing the answers to the four items without
weighting (‘yes’ = 1 and ‘no’ = 0). A score of 4 corresponds to the physician not needing to re-
evaluate the patient while 0 corresponds to a very high need for re-evaluation. Scores of 2, 1
and 0 correlated with lower than average quality of life,11 MIDAS scores equivalent to severe
disability10 and MTAQ scores above the threshold for when follow up is deemed warranted.66
From these data, it was deduced that a Migraine-ACT score of 2, 1 and 0 identifies patients for
whom a change in acute treatment may be warranted. These patients have reduced quality of
life, severe headache-related disability and a considerable number of outstanding management
issues. It was suggested that a score of ≤ 2 may indicate that a change in the patient’s acute
medication is warranted and a score of ≤ 1 may indicate that the change is mandated.43

The 4-item Migraine-ACT questionnaire is therefore a rapid, reliable, brief, simple to complete
and score assessment tool, which can be recommended for everyday clinical use by the clinician
to identify patients who require a change in their current acute migraine treatment. Table 1
summarises the key features of the questionnaire.

190
Table 1. Features of the Migraine-ACT questionnaire.

Property Feature of Migraine-ACT questionnaire

Objective Assess the need to change a migraine
patient’s acute medication
Questions Four yes/no questions on headache-
related disability, global response,
consistency of response and emotional
impact
Scoring 5-point scale (0–4), with a score ≤ 2
indicating that a change in acute
medication is warranted
Reliability, validity and potential clinical Good
utility
Who conducts the questionnaire Physician (primary care or specialist)

5.2 Implications: methodological, research and clinical

The importance of assessing headache-related disability was first identified by Lipton, Stewart
and colleagues several years ago. This research led to the understanding that disability is the
key driver of headache severity,23 and to the development of the MIDAS questionnaire.10 The
literature searches and studies outlined in this thesis have extended the knowledge of disability
assessments and expanded their potential use in research and in the clinic in several areas:
Methodological
1. The type of naturalistic, longitudinal clinical studies used in this thesis are suitable for use
in primary care and produce clinically-meaningful results, as described above. The
potential advantages of this type of study over the conventional controlled or open-label
clinical study are:
a. Real-life patients are more representative of the general patient population than
the selected groups of patients seen in conventional clinical studies, where
children, women pregnant or wishing to become pregnant, patients with some co-
morbidities and older people are usually excluded.
b. Fewer patients may be required in order to achieve the clinical effect.
c. Commercially-available drugs, doses and formulations are used according to
everyday clinical practices, allowing real-life comparisons.
d. The long-term effectiveness of treatment is investigated, rather than the efficacy
at a point in time.

Research
1. Conventional clinical study endpoints for headache may be relatively insensitive.3 MIDAS
is now frequently used as an endpoint and we now have the Migraine-ACT
questionnaire, based on disability and symptom relief, which shows promise as a new
study endpoint.
2. There is an emotional component associated with headache-related disability,
encompassing sedation, poor social interaction, sadness, anger and anxiety. The pattern
of the emotional disturbances varies in different headache subtypes.
3. All disability and impact measures are not the same. HIT is more a diagnostic tool than a
severity tool. MIDAS and SPI are tools that assess headache severity and not diagnosis.
Clinical
1. The pattern of disability related to headaches can differ at different time periods in the
patient’s life and also with respect to the different headache subtypes experienced by the

191
 patient. This emphasises the importance of the physician’s correct diagnosis (or
diagnoses) and their prospective monitoring of patients over time.
2. Physicians should assess headache-related disability when they evaluate their headache
patients. One of the available questionnaires (MIDAS, HIT or SPI) can be used, or just a
simple question about the disturbance to daily activities may help to start discussions and
reveal clinical information.
3. Assessments of disability play key roles in the management of headache:
a. Improving communication between the healthcare professional and the patient.
b. As part of the diagnostic procedure. Perhaps few physicians will use HIT alone for
diagnosis, but HIT, or a general question on disability can provide valuable
diagnostic information and is useful for case finding.
c. As part of the assessment of headache severity, helping to tailor treatment to
each patient’s individual needs.
e. As an integral part of primary care guidelines for headache, so guiding treatment
choice.
f. Assessing disability is a suitable job for the practice nurse, and forms a key part
of their role in headache management.
4. The DSQ is a sensitive inclusive tool for the screening of headache sufferers when they
first present for care. It contains one question on disability, together with other questions
on different headache features. It has considerable potential clinical utility in the
management of headache in primary care. Patients should be able to obtain a copy at
their first point of care, whether a GP surgery, pharmacist, optician, dentist or other
provider. They can complete the DSQ, either at home by themselves or with the help of
the practitioner. The practitioner then reviews the DSQ and provides appropriate advice.
Using the questionnaire, there is the potential for earlier and better headache diagnoses.
The DSQ may be particularly useful to pharmacists. Following pharmacist intervention,
patients can either be treated immediately or, if appropriate, be empowered to obtain
help from the GP sooner rather than later.
5. The MIDAS questionnaire is also suitable for first-line use in the clinic. Patients identified
with a marked level of disability can have improvements in their headaches simply by
receiving appropriate information from a nurse or other healthcare professional.
6. The Migraine-ACT questionnaire has been developed as a rapid, reliable and accurate
screener to identify patients whose current acute treatment regimen needs to be
changed. It contains four questions on headache-related disability, emotional impact, and
response to treatment. It is intended for use by the primary care physician in everyday
clinical practice. A score of 2, 1 or 0 is suggested as being indicative of the need to
change medications.

5.3 Methodological issues and suggestions for future research

As well as advantages, the naturalistic studies conducted for this thesis have several potential
disadvantages that need to be taken into account when evaluating the results:
1. The studies were relatively small in terms of numbers of patients and sometimes in
duration.
2. The open design means that results could not be unequivocally assigned to the
intervention used. Other factors, e.g. a placebo response, could bias the results.
3. The studies may not have been powered appropriately.
4. Asymmetries of design may have occurred.
5. Some of the questionnaires used were not validated.
Each study is now considered in turn for its methodological robustness and the potential for
future research.

192
Chapter 3.3: Disability associated with headaches occurring inside and outside the
menstrual period
This study investigated the prevalence of headache, depression and bodily pain in women
attending a UK general practice. The disability associated with migraine and other headache
attacks occurring during and outside the menstrual period was assessed in those women with
migraine. Results showed that the patients reported a high prevalence of headache, depression
and bodily pain. For migraine sufferers, migraine attacks that took place during the menstrual
period tended to be more disabling than those taking place outside the menstrual period, but the
opposite was true for non-migraine headache.

The questionnaires (headache, depression and bodily pain) used in the first part of this study
were not validated tools, but were devised by the investigators to be simple to use and
potentially applicable to primary care clinical practice. They were not tested for reliability or
validity, and additional studies would be required for this. Despite this, headache subtype and
the severity of depression and bodily pain were identified for each patient. The prevalence of the
headache subtypes was close to expected values, although depression was probably over-
estimated. Future work on this topic should probably use validated diagnostic questionnaires for
headache (e.g. the DSQ)40 and depression (e.g. the Beck Depression Inventory).70

The design of this prevalence study was also not optimal. Additional items that should have
been taken into account included defining the sample in relationship to the population under
study, random sampling, defining an acceptable response rate (> 50%), and conducting a non-
response analysis. There was the possibility of the responders being a self-selecting population
with significant problems, although the similarities in age range between the responders and the
total population argues against this.

Part 2
The menstrual headache questionnaire used in Part 2 of the study was also not validated.
However, it was based on the IHS diagnostic criteria pertaining at the time71 and contained items
now used in the MIDAS questionnaire.10 It therefore contained items with proven reliability and
validity. However, the questionnaire should be tested for reliability if it is used again. This in fact
seems unlikely to happen, as the MIDAS questionnaire could be used in its place.

Some methodological issues were raised by the study analyses that have implications for the
design of future studies. The data were confounded by the fact that some of the events were
repeatedly measured in individual patients, and that some patients only had one type of
headache. A much larger sample of patients would be necessary to allow both within and
between patient group comparisons. The results should be taken with caution as data were
analysed between the groups for only 30 patients experiencing 203 headache attacks. However,
non-parametric testing with the Mann-Whitney U Test showed one significant and one marginally
significant result from only four analyses, which indicates that the results were unlikely to have
arisen by chance. Analysis of means or medians was not a useful measure in the present study
because the data were not normally distributed. Many headache attacks scored zero on the two
analyses, leading to floor effects being observed. Such floor effects have been reported with use
of the MIDAS Questionnaire, on which the analyses were based.72 Future studies may benefit
from using assessments of headache impact which provide a normal distribution of results, such
as the Headache Impact Test45,46 or the Short Pain Inventory.32

This study is best considered a pilot investigation of questionnaires to assess headache,

depression, bodily pain and the disability associated with headache. While the methodology

193
used has been superseded by more recent developments, it served to investigate new ways to
measure headache-related disability at a time when validated questionnaires were not available
for headache. Such studies are necessary in the development of new tools and their publication
serves scientific accountability.

Chapter 3.4: the emotional impact of different headache subtypes over time
In this prospective, within-group, comparative study, the SPI was used to assess how the mood
disturbances and coping abilities change over time in patients with episodic TTH, migraine and
CDH attending a primary care headache centre. The healthcare resource utilisation of these
patients was also assessed retrospectively. The results showed that patients with headache had
significant emotional symptoms associated with their headaches. These symptoms resolved
within 1–2 days for patients with episodic TTH and migraine. However, patients with CDH were
profoundly affected, and did not improve physically or emotionally from their headache over a 7-
day period. Headache patients generally experienced higher levels of sedation than did patients
with other pain conditions.

The limitations to this study are discussed below in the text on Chapter 4.3

Chapter 4.3: comparison of HIT and SPI

This sub-analysis of study 3.4 above compared the psychometric properties of the Headache
Impact Test (HIT-DYNA®) with the 17-item Short Pain Inventory (SPI©) and the abilities of the
SPI and HIT in terms of discriminating headache severity and diagnosis. The results showed the
HIT to be poorly related to the severity of the pain but very closely related to the diagnostic label.
In contrast to this, the SPI was very closely related to severity but not related to diagnosis.

In the two chapters based on this study (3.4 and 4.3), relatively small numbers of patients took
part, and a high proportion of patients did not experience a headache (especially those with
TTH) and hence did not complete a SPI questionnaire in the 6 weeks allowed. A longer study
duration could have led to a reduction in the drop out rate. The healthcare utilisation analysis
was performed with retrospective collected data, for practical reasons. Of course, a prospective
design would have been more powerful. Patients from a primary care headache centre were
recruited so that, as far as possible, criteria for a naturalistic study were met.

SPI is now a well-established questionnaire that can be used to assess any pain state.32,38,50 It is
available to interested healthcare professionals at www.headachetest.co.uk and can be
completed on-line. Future studies should include SPI as an endpoint in controlled studies of
headache treatments, in comparison to more conventional endpoints (including other disability
questionnaires).

Chapter 4.2: the clinical utility of the MIDAS questionnaire when used by nurses at patient
screening
This audit assessed the outcome of a nurse intervention strategy in patients presenting with
headache to a primary care surgery. The results showed that advice on headache management
given by a nurse led to significant improvements in the patients’ outcomes. MIDAS appeared to
be a sensitive outcome measure for reduction in disability in headache sufferers.

The main shortcoming of this study was that it was uncontrolled. As no control group was
included, the reductions in MIDAS and other items could have been due to a placebo effect.
Nevertheless, there was good evidence of significant improvements in MIDAS items and
significant reductions in headache consultations, as parametric and non-parametric analyses
exhibited the same patterns of results.

194
A further shortcoming was that the numbers of patients reduced as the visits proceeded due to
drop outs. This could have caused some bias, leading to an improvement in the patients’
condition. This is because, psychologically, patients are more likely to continue with an
investigation if they are generally feeling well or better than if they are deteriorating or have no
change in their condition. However, the results were promising enough to pave the way for
future randomised, controlled studies investigating the role of nurses in headache management.

Chapter 4.4: utility of disability questions in headache diagnosis

This prospective, open investigation was set up to validate the 8-item Headache Diagnostic
Screening Questionnaire (DSQ), containing questions screening for possible sinister symptoms,
and evaluating headache impact, frequency and duration, the use of symptomatic medications
and the presence of migraine aura symptoms. The results showed that the Headache DSQ is a
brief, rapid to use and accurate diagnostic screening questionnaire for headache, and is
especially sensitive for migraine and MOH. It can be recommended for screening new headache
patients at and below the primary care level.

Some weaknesses were observed in this study. The study was not properly randomised, with
alternate patients being openly assigned to seeing the pharmacist or GP first. The study was not
large enough to determine whether the order of seeing these healthcare professionals had any
effect on the results. However, the pharmacists and GPs were blind to each other’s
assessments. The entry criteria for numbers of patients with episodic TTH and cluster headache
were not met by the patients recruited into the study. For this reason the specificity of the DSQ
could not be assessed.

The accuracy of the DSQ for CDH without MOH was poor, and patients tended to underestimate
the duration and/or the frequency of their headaches. Additionally, migraine patients sometimes
confused aura symptoms with prodromes when answering question 8 (Do changes in your
senses (sight, taste, smell or touch) occur before the headache starts?). Small changes were
made to the questionnaire to rectify these points, as part of its ongoing development. A follow-up
randomised study with the modified DSQ is warranted to assess its sensitivity and specificity.
Further studies with the DSQ could also be conducted in specialist headache clinics (where
cluster headache patients are more common and could be assessed) and in pharmacies (where
episodic TTH sufferers may be encountered frequently and could be assessed). In addition, it
would be interesting to see if GPs’ diagnostic accuracy could improve to the specialist level if
they had access to the DSQ.

However, the value of this study is that the Headache DSQ as developed is a brief, rapid to use
and accurate diagnostic screening questionnaire for headache, and is especially sensitive for
migraine and MOH. The DSQ, and other questionnaires, demonstrate the utility of incorporating
assessments of disability into diagnostic questionnaires. It is recommended for screening new
headache patients at and below the primary care level, although future studies will more fully
define its place in the clinic.

Chapters 4.5 and 4.6: development and testing of the Migraine-ACT questionnaire
This study designed and tested a new 4-item assessment tool, the Migraine Assessment of
Current Therapy (Migraine-ACT) questionnaire for use by clinicians, to quickly evaluate how a
recently prescribed acute medication is working, and to identify patients who require a change of
their current acute treatment. The results showed that Migraine-ACT, containing four questions
on headache impact (measured as headache-related disability), global response and
consistency of response to therapy and emotional impact, was reliable, valid, easy to use and

195
score, and exhibited considerable potential clinical utility. Migraine-ACT can be recommended
for everyday clinical use by clinicians.

The Migraine-ACT study used rigorous methodology to confirm the reliability and validity of the
questionnaire. The result is a robust questionnaire that is suitable for use in research and in the
clinic. However, additional studies are required to further examine the clinical utility of Migraine-
ACT, to define the score that indicates a change in treatment is needed, and to investigate its
sensitivity to change for use as an outcome tool in comparison to other endpoints. The study
was conducted in secondary care referral centres. Studies in primary care practice are
particularly needed to demonstrate the utility of Migraine-ACT in the everyday management of
migraine. The main use of Migraine-ACT is likely to be in primary care, and naturalistic studies
conducted in this domain will determine the true utility of the questionnaire. Pharmacists and
nurses may also find the questionnaire useful and studies with these professionals would also
be valuable.

5.4 Final remarks

The research and studies contained in this thesis have confirmed the clinical importance of
assessing disability when managing headache patients. Disability assessments play key roles in
the screening, diagnosis and treatment of these patients. While there is no definitive method to
measure disability, several tools are available and have utility in diagnosis (HIT), assessing
headache severity and monitoring treatments (MIDAS) and in assessing emotional disability
(SPI). Disability assessments also play key roles in the development of other questionnaires.
The DSQ sensitively screens for diagnosis using questions on disability and other headache
features, features consistent with those of other diagnostic questionnaires. Migraine-ACT
assesses the need to change acute medications using questions on headache-related disability,
emotional impact, and overall response and consistency of response. While these
questionnaires require further development, physicians are now closer to the goals of having
definitive diagnostic and outcome measures in their armamentarium.

References

1. Silberstein SD, for the US Headache Consortium. Practice parameter: evidence-based

guidelines for migraine headache (an evidence-based review). Report of the Quality
Standards Subcommittee of the American Academy of Neurology. Neurology
2000;55:754–62.
2. Pilgrim AJ. Methodology of clinical trials of sumatriptan in migraine and cluster headache.
Eur Neurol 1991;31:295–9.
3. Dowson AJ, Kilminster S, Peters M et al. Understanding the evidence: evaluating the
efficacy of migraine medications in clinical practice. Headache Care 2005;2:133–43.
4. Loder E, Goldstein R, Biondi D. Placebo effects in oral triptan trials: the scientific and
ethical rationale for continued use of placebo controls. Cephalalgia 2005;25:124–31.
5. Ferrari MD, Roon KI, Lipton RB et al. Oral triptans (serotonin 5-HT1B/1D agonists) in acute
migraine treatment: a meta-analysis of 53 trials. Lancet 2001;358:1668–75.
6. Belsey JD. Reconciling effectiveness and tolerability in oral triptan therapy: a quantitative
approach to decision making in migraine management. J Clin Res 2001;4:105–27.
7. Williams P, Reeder CE. A comparison of the cost-effectiveness of almotriptan and
sumatriptan in the treatment of acute migraine using a composite efficacy/tolerability end
point. J Manag Care Pharm. 2004;10:259–65.
8. Fox AW, Keywood C, Sheftell FD et al. Consequences of transforming measures of
efficacy for acute therapies: 5-HT1B/1D as a worked example. Headache 2004;44:48–52.

196
9. Dowson AJ, Tepper SJ, Dahlöf C. Patients’ preference for triptans and other medications
as a tool for assessing the efficacy of acute treatments for migraine. J Headache Pain
2005;6:112–20.
10. Stewart WF, Lipton RB, Dowson AJ et al. Development and testing of the Migraine
Disability Assessment (MIDAS) Questionnaire to assess headache-related disability.
Neurology 2001;56(Suppl1):S20–8.
11. Ware JE, Snow KK, Kosinski M et al. SF-36 Health Survey manual and interpretation
guide. Boston: New England Medical Center, 1993.
12. Myllylä VV, Havanka H, Herrala L et al. Tolfenamic acid rapid release versus sumatriptan
in the acute treatment of migraine: comparable effect in a double-blind, randomized,
controlled, parallel-group study. Headache 1998;38:201–7.
13. Dowson A, Ball K, Haworth D. Comparison of a fixed combination of domperidone and
paracetamol (Domperamol) with sumatriptan 50 mg in moderate to severe migraine: a
randomised UK primary care study. Curr Med Res Opin 2000;16:190–7.
14. Oral sumatriptan and Aspirin-plus-Metoclopramide Comparative Study Group. A study to
compare oral sumatriptan with oral aspirin plus oral metoclopramide in the acute
treatment of migraine. Eur Neurol 1992;32:177–84.
15. Tfelt-Hansen P, Henry P, Mulder LJ et al. The effectiveness of combined oral lysine
acetylsalicylate and metoclopramide compared with oral sumatriptan for migraine. Lancet
1995;346:923–6.
16. Freitag FG, Cady R, DiSerio F et al. Comparative study of a combination of
isometheptene mucate, dichloralphenazone with acetaminophen and sumatriptan
succinate in the treatment of migraine. Headache 2001;41:391–8.
17. Goldstein J, Silberstein SD, Elkind AH et al. A placebo-controlled comparison of the
combination of acetaminophen, aspirin, and caffeine with sumatriptan succinate in the
early treatment of migraine: Results from the Asset trial. Neurology 2003;60 (Suppl 1):
A171. (Abstract)
18. Loder E, Brandes JL, Silberstein S et al. Preference comparison of rizatriptan ODT 10
mg and sumatriptan 50 mg tablet in migraine. Headache 2001;41:745–53.
19. Pascual J, Munoz R, Leira R. An open preference study with sumatriptan 50 mg and
zolmitriptan 2.5 mg in 100 migraine patients. Cephalalgia 2001;21:680–4.
20. Pascual J, Bussone G, Hernandez JF et al. Comparison of preference for rizatriptan 10
mg wafer versus sumatriptan 50 mg tablet in migraine. Eur Neurol 2001;45:275–83.
21. National Academy of Sciences/Institute of Medicine (NAS/IOM). Disability in America:
toward a national agenda for prevention. Washington DC: NAS Press, 1991.
22. Dowson AJ. Assessing the impact of migraine. Curr Med Res Opin 2001;17:298–309.
23. Lipton RB, Goadsby PJ, Sawyer JPC et al. Migraine: diagnosis and assessment of
disability. Rev Contemp Pharmacother 2000;11:63–73.
24. Headache Classification Committee of the International Headache Society. The
international classification of headache disorders; 2nd Edition. Cephalalgia 2004;24(Suppl
1):1–160.
25. Dowson AJ, Bundy M, Kilminster SG. Assessing patients: coping with headache,
emotional function and healthcare utilisation associated with episodic tension-type
headache, migraine and chronic daily headache. Headache Care 2005; manuscript in
preparation.
26. Rasmussen BK. Migraine and tension-type headache in a general population:
psychosocial factors. Int J Epidemiol. 1992;21:1138–43.
27. Bigal ME, Rapoport AM, Lipton RB et al. Assessment of migraine disability using the
migraine disability assessment (MIDAS) questionnaire: a comparison of chronic migraine
with episodic migraine. Headache 2003;43:336–42.
28. Holmes WF, MacGregor EA, Sawyer JPC et al. Information about migraine disability

197
 influences physicians’ perceptions of illness severity and treatment needs. Headache
2001;41:343–50.
29. Otsuka N, Sakai F, Iigaya M et al. MIDAS assessments of migraine management,
including the use of triptans, in Japan. Headache Care 2004;1:115–21.
30. Mathew NT, Kailasam J, Meadors L. Prophylaxis of migraine, transformed migraine, and
cluster headache with topiramate. Headache 2002;42:796–803.
31. Kilminster SG, Dowson A, Bundy M. The Headache Impact Test® and the Short Pain
Inventory©: Outcome measures compared. Int J Pharm Med 2003;17:23–32.
32. Manual. The Short Pain Inventory. Selsey, West Sussex: Lawrencian Clinical Ltd.
England.
33. Dowson AJ, Kilminster SG, Salt R et al. Disability associated with headaches occurring
inside and outside the menstrual period in those with migraine: A general practice study.
Headache 2005;45:274–82.
34. Breslau N, Rasmussen BK. The impact of migraine: Epidemiology, risk factors, and co-
morbidities. Neurology 2001;56 (Suppl 1):4–12.
35. MacGregor EA, Hackshaw A. Prevalence of migraine on each day of the natural
menstrual cycle. Neurology 2004;63:351–3.
36. Stewart WF, Lipton RB, Chee E et al. Menstrual cycle and headache in a population
sample of migraineurs. Neurology 2000;55:1517–23.
37. Peres MF, Zukerman E, Young WB et al. Fatigue in chronic migraine patients.
Cephalalgia 2002;22:720–4.
38. Kilminster SG, Power M, Fozard JR. Survey of pain in two medical and dental clinics,
with non patient controls using the Short Pain Inventory. Int J Pharm Med 2000;14:137–
47.
39. Peters M, Abu-Saad HH, Vydelingam V et al. Migraine and chronic daily headache
management: a qualitative study of patient’s perceptions. Scand J Caring Sci
2004;18:294–303.
40. Dowson AJ, Turner A, Kilminster S et al. Development and validation of the headache
Diagnostic Screening Questionnaire (DSQ): a new questionnaire for the differential
diagnosis of headache for use in primary care. Headache Care 2005;2:111–18.
41. Dowson AJ, Salt R, Kilminster S. The outcome of headache management following nurse
intervention: assessment in clinical practice using the Migraine Disability Assessment
(MIDAS) Questionnaire. Headache Care 2004;1:177–81.
42. Dowson AJ, Tepper SJ, Baos V et al. Identifying patients who require a change in their
current acute migraine treatment: the Migraine Assessment of Current Therapy
(Migraine-ACT) questionnaire. Curr Med Res Opin 2004;20:1125–35.
43. Kilminster S, Dowson AJ, Tepper SJ et al. The Migraine Assessment of Current Therapy
(Migraine-ACT) questionnaire: investigation of reliability, validity and clinical utility in a
multinational study. Headache 2005; submitted for publication.
44. Edmeads J, Láinez JM, Brandes JL et al. Potential of the Migraine Disability Assessment
(MIDAS) Questionnaire as a public health initiative and in clinical practice. Neurology
2001;56 (Suppl 1):S29–34.
45. Bayliss MS, Dewey JE, Dunlap I et al. A study of the feasibility of Internet administration
of a computerized health survey: the headache impact test (HIT). Qual Life Res
2003;12:953–61.
46. Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring
headache impact: the HIT-6. Qual Life Res 2003;12:963–74.
47. Nachit-Ouinekh F, Dartigues J-F, Henry P et al. Use of the headache impact test (HIT-6)
in general practice: relationship with quality of life and severity. Eur J Neurol
2005;12:189–93.
48. Vuillaume De Diego E, Lanteri-Minet M. Recognition and management of migraine in

198
 primary care: influence of functional impact measured by the headache impact test (HIT).
Cephalalgia 2005;25:184–90.
49. Kilminster SG, Mould P. Comparison of the internal reliability and validity of the McGill
Pain Questionnaire and Short Pain Inventory. Int J Pharm Med 2002;16:87–95.
50. Kilminster SG, Power M, Fozard JR. Survey of pain in two medical and dental clinics,
with non patient controls using the Short Pain Inventory. Int J Pharm Med 2000;14:137–
47.
51. Dowson AJ, Lipscombe S, Sender J et al. New guidelines for the management of
migraine in primary care. Curr Med Res Opin 2002;18:414–39.
52. Dowson AJ, Bradford S, Lipscombe S et al. Managing chronic headaches in the clinic. Int
J Clin Pract 2004;58:1142–51.
53. Maizels M, Burchette R. Rapid and sensitive paradigm for screening patients with
headache in primary care settings. Headache 2003;43:441–50.
54. Lipton RB, Dodick D, Sadovsky R et al. A self-administered screener for migraine in
primary care: The ID Migraine validation study. Neurology 2003;61:375–82.
55. Cady RK, Borchert LD, Spalding W et al. Simple and efficient recognition of migraine with
3-question headache screen. Headache 2004;44:323–7.
56. Lipton RB, Scher A, Steiner TJ et al. Patterns of healthcare utilization for migraine in
England and in the United States. Neurology 2003;60:441–8.
57. Main A, Abu-Saad H, Salt R et al. Management by nurses of primary headache: a pilot
study. Curr Med Res Opin 2002;18:471–8.
58. Allen LL, Haririfar M, Cohen J et al. Quality of life and locus of control of migraineurs. Clin
Excell Nurse Pract 2000;41–9.
59. MacBean H, Leech J, Dungay J et al. Management of migraine by nurses. Practice
Nursing 2004;15:346–50.
60. Edmeads J, Findlay H, Tugwell P et al. Impact of migraine and tension-type headache on
life-style, consulting behaviour, and medication use: a Canadian population survey. Can
J Neurol Sci 1993;20:131–7.
61. Lipton RB, Scher AI, Kolodner K et al. Migraine in the United States: epidemiology and
patterns of health care use. Neurology 2002;58:885–94.
62. Celentano DD, Stewart WF, Lipton RB et al. Medication use and disability among
migraineurs: a national probability sample survey. Headache 1992;32:223–8.
63. Forward SP, McGrath PJ, MacKinnon D et al. Medication patterns of recurrent headache
sufferers: a community study. Cephalalgia 1998;18:146–51.
64. International Headache Society Clinical Trials Subcommittee. Guidelines for controlled
trials of drugs in migraine: second edition. Cephalalgia 2000;20:765–86.
65. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity. Disability, pain intensity,
and attack frequency and duration. Neurology 1994;44 (Suppl 4):24–39.
66. Chatterton ML, Lofland JH, Shechter A et al. Reliability and validity of the Migraine
Therapy Assessment Questionnaire. Headache 2002;42:1006–15.
67. Stewart WF, Lipton RB, Whyte J et al. An international study to assess reliability of the
Migraine Disability Assessment (MIDAS) score. Neurology 1999;53:988–94.
68. Lipton RB, Stewart WF, Sawyer J et al. Clinical utility of an instrument assessing
migraine disability: the Migraine Disability Assessment (MIDAS) questionnaire.
Headache 2001;41:854–61.
69. Adelman JU, Lipton RB, Ferrari MD et al. Comparison of rizatriptan and other triptans on
stringent measures of efficacy. Neurology 2001;57:1377–83.
70. Beck AT, Beamesderfer A. Assessment of depression: the depression inventory. Mod
Probl Pharmacopsychiatry 1974;7:151–69.

199
71. Headache Classification Committee of the International Headache Society. Classification
and diagnostic criteria for headache disorders, cranial neuralgias and facial pain.
Cephalalgia 1988;8 (Suppl 7):1–92.
72. Henry P, Auray JP, Gaudin AF et al. Prevalence and clinical characteristics of migraine in
France. Neurology 2002;59:232–7.

200
6

Summary

Introduction
The main aim of this thesis is to investigate the clinical significance of headache-related
disability; the clinical importance of assessing disability, the means of recognising the patients
with severe disability and the development of new ways to assess headache-related disability
and their utility in clinical practice. The main methodology used in the thesis was critical reviews
of the literature and open, naturalistic studies designed to mimic the situation in everyday
primary care clinical practice.

Clinical significance of headache-related disability

In Chapter 3.1, the factors involved in the understanding of clinical study evidence are reviewed,
and new study endpoints discussed that may be relevant to headache clinical practice. Large,
randomised, double-blind, controlled clinical trials form the gold standard of clinical evidence for
evaluating new medications. Meta-analyses and post hoc analyses of existing trials are
increasingly used to compare therapies, but should not be used on their own to evaluate
therapies. Controlled clinical trials evaluating triptans in the acute treatment of migraine
demonstrate clear superiority of the drugs over placebo, but only small differences are reported
in studies comparing individual triptans and triptans with non-triptan medications. In contrast,
patients can clearly distinguish between triptans and non-triptans and between different triptans
in clinical practice. This incongruity may be due to the relative insensitivity of the standard
clinical trial endpoint, relief of headache. New and more sensitive clinical trial endpoints are
required for use in clinical studies that reflect everyday general practice (naturalistic studies).
Among the potential endpoints are assessments of patient preference, impact on the patient’s
daily life and quality of life. However, novel endpoints should be developed, in collaboration with
patients, which can summarise the whole migraine experience.

Chapter 3.2 reviews the evidence that patient preference may be a clinical study endpoint that
is useful in headache clinical practice. Patient preference assesses a global measure of efficacy
and tolerability, and may be a valid and sensitive means of distinguishing between the triptans.
In a series of studies, patients consistently expressed a clear preference for triptans over their
usual non-triptan acute medications, e.g. analgesics and ergotamine. Direct comparator studies
of patient preference with oral triptans showed that patients could distinguish between different
triptans, and between different formulations of the same triptan. Patients could even distinguish
between the three oral doses of sumatriptan. The most frequently provided reasons for
preference were speed of response and overall effectiveness. Patient preference may therefore
be a sensitive and valid clinical trial endpoint and physicians should consider using it when
reviewing the efficacy of acute migraine medications.

Chapter 3.3 describes a study that investigated the prevalence of headache, depression and
bodily pain in women attending a UK general practice. The disability associated with migraine
and other headache attacks occurring during and outside the menstrual period was also
assessed in those women with migraine. Women aged 14–50 years completed two
questionnaires. The first questionnaire assessed the point prevalence of headache and bodily

201
pain, and the 5-year prevalence of depression in the total population. The second questionnaire
assessed the disability of all headaches over a 2-month period (to capture a complete menstrual
cycle) for patients reporting migraine who were still menstruating. Disability was assessed as the
time lost and time spent at less than 50% productivity in normal activities due to headache, and
analysed as rank sums using the Mann-Whitney U Test. The first part of the study showed that
the prevalence of headache (66.1%), depression (55.4%) and bodily pain (40.6%) were high in
this population of women. In the second part of the study, for migraine, the rank order of time at
less than 50% productivity was significantly greater for attacks taking place inside the menstrual
period than for those occurring outside the menstrual period (p=0.01). For non-migraine
headaches, the time lost appeared to be numerically greater for attacks taking place outside the
menstrual period than for those occurring inside the menstrual period, and the comparison
approached significance (p=0.06). In conclusion, the patients reported a high prevalence of
headache, depression and bodily pain. For migraine sufferers, migraine attacks that took place
during the menstrual period tended to be more disabling than those taking place outside the
menstrual period, but the opposite was true for non-migraine headache.

Chapter 3.4 describes a prospective, within-group, comparative study, in which the Short Pain
Inventory (SPI©) was used to assess how the mood disturbances and coping abilities changed
over time in patients with episodic tension-type headache (TTH), migraine and chronic daily
headache (CDH) attending a primary care headache centre. The healthcare resource utilisation
of these patients was also assessed retrospectively.
Patients completed the SPI 10 times over a 7-day period, starting 1 hour after the onset of their
next headache. The SPI data were analysed statistically as Z scores and coping Z scores for the
Total Pain Disturbance, Total Mood Disturbance, and individual sub-scores for sedation, social
interaction, sadness, anxiety and anger. Data from the healthcare resource utilisation
questionnaire were analysed as descriptive statistics. Seventy five patients completed the
healthcare utilisation questionnaire and 42 (56.0%) completed the SPI for the 7-day period. All
headache patients showed considerable mood disturbances during a headache. In general the
disturbances were severe in intensity and of the order CDH>migraine>TTH. The pattern of mood
disturbance was different for each type of headache, as was the patients’ ability to cope with the
mood changes. Patients with CDH experienced pain and emotional symptoms, particularly
sedation, throughout the 7-day monitoring period. For TTH and migraine, the pain and emotional
symptoms resolved within 1–2 days after the headache. The level of healthcare resource
utilisation was also in the order CDH>migraine>TTH, similar to the data reported for the SPI. In
conclusion, patients with headache had significant emotional symptoms associated with their
headaches. These symptoms resolved within 1–2 days for patients with episodic TTH and
migraine. However, patients with CDH were profoundly affected, and did not improve physically
or emotionally from their headache over a 7-day period. Headache patients generally
experienced higher levels of sedation than did patients with other pain conditions.

Assessing headache-related disability

Chapter 4.1 reviews the disability caused by migraine and other headache disorders. Migraine
is a remarkably disabling condition, although unpredictable and heterogeneous in frequency,
duration and severity. It can be difficult to manage in primary care, where it is under-recognised,
under-diagnosed and under-treated. Proposals have been made that migraine care could be
improved by incorporating assessments of disability into management strategies. Research has
shown that measuring headache-related disability, together with assessments of pain intensity,
headache frequency, tiredness, mood alterations and cognition can be used to assess the
impact of migraine on sufferers’ lives and society. From this research two simple and brief
impact tools were developed; the Migraine Disability Assessment (MIDAS) Questionnaire and
the Headache Impact Test (HIT). Both tools are scientifically valid measures of headache

202
severity and have the potential to improve communication between patients and their
physicians, assess headache severity and act as outcome measures to monitor treatment
efficacy. Disability-based tools are also being increasingly recommended as part of generalised
headache management guidelines to produce an individualised treatment plan for each patient
in concert with other clinical assessments. It is not possible as yet to unequivocally recommend
the optimal impact tool for use in primary care, but it should be usable by GPs, pharmacists,
nurses and patients, and for research purposes.

Chapter 4.2 describes an audit that assessed the outcome of a nurse intervention strategy in
patients presenting with headache to a primary care surgery. Patients aged 18–65 years
attending the surgery who reported headache in the previous 3 months were assessed by a
nurse and completed a MIDAS Questionnaire and a questionnaire investigating headache
features and physician consultations. All patients were given oral and written advice on
headache management by the nurse. Patients were reviewed after 3 and 6 months with the
same questionnaires. A total of 195 patients took part in the study. At baseline, 136 (69.7%)
were MIDAS Grade I/II and 59 (30.3%) were Grade III/IV. Compared with baseline, patients
reported significant mean reductions at 6 months in: total MIDAS scores (5.9 versus 9.6; p =
0.014); headache frequency (MIDAS A score: 7.0 versus 12.6; p = 0.009); headache severity
(MIDAS B score: 5.0 versus 6.0; p = 0.003); and physician consultations for headache (3-month
period: 0.05 versus 0.30; p = 0.05). In conclusion, advice on headache management given by a
nurse can potentially lead to significantly improved patient outcomes. MIDAS appeared to be a
sensitive outcome measure for reduction in disability in headache sufferers. These results
warrant further investigation in randomised, controlled clinical studies.

Chapter 4.3 describes a sub-analysis of Study 3.4 investigating the psychometric properties of
Internet HIT (HIT-DYNA®) compared with the SPI©. Also compared were the abilities of the SPI
and HIT in terms of discriminating headache severity and diagnosis. Result showed that the HIT
score correlated 0.28 with the severity of the headaches whereas the SPI correlated 0.76 on its
Total Mood Disturbance score. The correlations of the HIT items with pain severity never rose
above 0.26 whereas the correlations of the SPI items were all typically about 0.7. The SPI
significantly discriminated with headache severity levels (none, mild, moderate, severe,
extreme), with t values of 2–20. The HIT scores were far less powerful at discriminating severity,
with t test values from 2–5. HIT scores at screening significantly discriminated with diagnosis,
with the resolution between TTH and CDH reaching 1/100 million. The SPI summary scores and
subscales did not discriminate with diagnosis. These differences between HIT and SPI were
confirmed by factor analysis. In conclusion, the results showed the HIT to be poorly related to
the severity of the pain but very closely related to the diagnostic label. In contrast to this, the SPI
was very closely related to severity but not related to diagnosis.

Chapter 4.4 describes a prospective, open investigation designed to validate a new inclusive
diagnostic questionnaire for headache, the 8-item Headache Diagnostic Screening
Questionnaire (DSQ), containing questions screening for possible sinister symptoms, and
evaluating headache-related disability, frequency and duration, the use of symptomatic
medications and the presence of migraine aura symptoms. Patients completed the DSQ on their
own and with a pharmacist’s help. The same pharmacist, a GP and a headache specialist
conducted a differential headache diagnosis according to their usual practices. Diagnoses were
deduced from the DSQ using an algorithm by a healthcare professional, who was blind to the
other diagnoses. The primary endpoint was the sensitivity of the DSQ, measured as the
proportion of correct diagnoses assessed from the patient, pharmacist and GP responses,
compared with the gold standard of specialist diagnosis, for each headache subtype. Eighty
seven patients (80.5% women, mean age 52.9 years) completed the study, all completing the

203
questionnaire without help within a few minutes. Most patients were diagnosed by the specialist
with migraine (with or without aura) or CDH (with or without medication overuse headache
[MOH]). The overall sensitivity of the patient-completed DSQ was good (59.3%) and similar to
the pharmacist-completed DSQ (66.6%) and GP diagnosis (59.4%). The accuracy of the DSQ-
derived diagnoses was greatest for total migraine, migraine with aura, MOH and migraine
without aura. The accuracy of the pharmacists’ headache diagnosis was improved markedly
when they used the DSQ, particularly with regard to diagnosing migraine with and without aura
and MOH. In conclusion, the Headache DSQ is a brief, rapid to use and accurate diagnostic
screening questionnaire for headache, and is especially sensitive for migraine and MOH. It can
be recommended for screening new headache patients at and below the primary care level.

Chapters 4.5 and 4.6 describe the design and testing of a new 4-item assessment tool, the
Migraine Assessment of Current Therapy (Migraine-ACT) questionnaire for use by clinicians, to
quickly evaluate how a recently prescribed acute medication is working, and to identify patients
who require a change of their current acute treatment. A 27-item Migraine-ACT questionnaire
was developed by an international advisory board of headache specialists. Questions were
formulated in four domains: headache impact, global assessment of relief, consistency of
response and emotional response, all with yes (score = 1) or no (score = 0) answers. Migraine
patients (n = 185) attending secondary care headache clinics entered a multinational,
prospective, observational study to investigate the test-retest reliability and construct validity of
the 27-item Migraine-ACT. Patients completed the Migraine-ACT on two occasions, separated
by a 1-week interval, and test-retest reliability was assessed by Pearson product moment and
Spearman rank measures. Construct validity was assessed by correlating patients’ answers to
the 27-item Migraine-ACT with those to other questionnaires (individual domains and total
scores) conceptually related to it; the Short-Form 36 quality of life questionnaire (SF-36), the
Migraine Disability Assessment (MIDAS) questionnaire and the Migraine Therapy Assessment
Questionnaire (MTAQ). Discriminatory t-tests were used to identify the four Migraine-ACT
questions (one in each domain) which discriminated best between the domains of the SF-36,
MIDAS, and MTAQ. These four items constituted the final 4-item Migraine-ACT. The test-retest
reliability of the 27 Migraine-ACT questions ranged from good to excellent, and correlation
coefficients were highly significant for all items. The consistency of reporting the yes and no
answers was also excellent. Correlations of Migraine-ACT items with SF-36 and MIDAS items
and SF-36, MIDAS and MTAQ total scores indicated that the following were the most
discriminating items, in the respective four domains, and constitute the final Migraine-ACT
questionnaire:
• Consistency of response - Does your migraine medication work consistently, in the
majority of your attacks?
• Global assessment of relief - Does the headache pain disappear within 2 hours?
• Impact - Are you able to function normally within 2 hours?
• Emotional response - Are you comfortable enough with your medication to be able
to plan your daily activities
The 4-item Migraine-ACT was shown to be highly reliable (Spearman / Pearson measure r =
0.82). The individual questions, and the total 4-item Migraine-ACT score, showed good
correlation with items of the SF-36, MIDAS and MTAQ questionnaires, particularly with the total
MTAQ and SF-36 scores. Additional testing confirmed the good reliability, validity and potential
clinical utility of the 4-item questionnaire. In conclusion, the 4-item Migraine-ACT questionnaire
is brief and simple to complete and score, and has demonstrated reliability, accuracy and
simplicity. A score of ≤ 2 indicates that a change in medication may be warranted. Migraine-ACT
can therefore be recommended for everyday clinical use by clinicians.

204
Chapter 5 describes the main discussion points arising from this research. The searches and
studies contained in this thesis have confirmed the clinical importance of assessing disability
when managing headache patients. Disability-based assessments play key roles in the
screening, diagnosis and treatment of these patients. While there is no definitive method to
measure disability, several tools are available and have utility in diagnosis (HIT), assessing
headache severity and monitoring treatments (MIDAS) and in assessing emotional disability
(SPI). Disability assessments also play key roles in the development of other questionnaires.
The DSQ sensitively screens for diagnosis using questions on disability and other headache
features, features consistent with those of other diagnostic questionnaires. Migraine-ACT
assesses the need to change acute medications using questions on headache-related disability,
emotional impact, and overall response and consistency of response.

The main limitations of the naturalistic studies were that they were open-label, small in terms of
numbers of patients and sometimes duration, resulting in lack of power and possible design
asymmetries. Larger, controlled studies are now required to take this research forward.
However, the type of naturalistic, prospective, open, longitudinal study described here is
eminently suited for research in primary care.

In conclusion, the main clinical applications of this work are:

• The validation of naturalistic studies for use in primary care.
• The use of disability-based measures for assessing headache patients in screening,
diagnosis, to assess severity, and as part of guidelines for treatment.
• The development of a disability-based tool (DSQ) for diagnosing headaches.
• The development of a disability-based tool (Migraine-ACT) for assessing the need to
change a migraine patient’s acute medication.

205