ORIGINAL ARTICLE

Adherence to hand hygiene guidelines – significance of measuring

fidelity
Anne Korhonen, Helena Ojanper€
a, Teija Puhto, Raija J€
arvinen, Pirjo Kejonen and Arja Holopainen

Aims and objectives. The aim was to evaluate the usability of fidelity measures in
compliance evaluation of hand hygiene. What does this paper contribute
Background. Adherence to hand hygiene guidelines is important in terms of to wider global clinical
patient safety. Compliance measures seldom describe how exactly the guidelines community?
are followed. • Combining fidelity indicators
Design and methods. A cross-sectional observation study in a university hospital with evaluation of the compli-
setting was conducted. Direct observation by trained staff was performed using a ance rate of hand hygiene may
reveal the gaps between optimal
standardised observation form supplemented by fidelity criteria. A total of 830
and actual hand hygiene prac-
occasions were observed in 13 units. Descriptive statistics (frequency, mean, per- tices.
centages and range) were used as well as compliance rate by using a standard • The appropriate length of hand
web-based tool. In addition, the binomial standard normal deviate test was con- rubbing is a key indicator of fide-
ducted for comparing different methods used in evaluation of hand hygiene and lity.
in comparison between professional groups. • Although evaluating fidelity may
be seen as resource consuming, it
Results. Measuring fidelity to guidelines was revealed to be useful in uncovering
constitutes a beneficial organisa-
gaps in hand hygiene practices. The main gap related to too short duration of tional contribution on account of
hand rubbing. Thus, although compliance with hand hygiene guidelines measured the high financial and human
using a standard web-based tool was satisfactory, the degree of how exactly the costs caused by healthcare-re-
guidelines were followed seemed to be critical. lated infections.
Conclusions. Combining the measurement of fidelity to guidelines with the com-
pliance rate is beneficial in revealing inconsistency between optimal and actual
hand hygiene behaviour.
Relevance to clinical practice. Evaluating fidelity measures is useful in terms of
revealing the gaps between optimal and actual performance in hand hygiene.
Fidelity measures are suitable in different healthcare contexts and easy to measure
according to the relevant indicators of fidelity, such as the length of hand rub-
bing. Knowing the gap facilitates improvements in clinical practice.

Key words: compliance evaluation, fidelity measures, hand hygiene, healthcare

staff, observation study

Accepted for publication: 31 May 2015

Authors: Anne Korhonen, PhD, Researcher, Nursing Research Foun- University Hospital, Oulu; Arja Holopainen, PhD, Research Direc-
dation, Helsinki; Helena Ojanper€a, MSc, RN, Head Nurse Infection tor, Nursing Research Foundation, Helsinki, Finland
Control Nurse, Oulu University Hospital, Oulu; Teija Puhto, MD, Correspondence: Anne Korhonen, Researcher, Nursing Research
Specialist in Infectious Diseases, Oulu University Hospital, Oulu; Foundation, Asemamiehenkatu 2, 00520 Helsinki, Finland.
Raija J€arvinen, RN, Infection Control Nurse, Oulu University Hospi- Telephone: +358 44 529 0047.
tal, Oulu; Pirjo Kejonen, MSc, PhD, Chief Nursing Officer, Oulu E-mail: [email protected]

© 2015 John Wiley & Sons Ltd

Journal of Clinical Nursing, 24, 3197–3205, doi: 10.1111/jocn.12969 3197
A Korhonen et al.

national level demand and a national infection control reg-

Introduction ister, called the Finnish Hospital Infection Program (SIRO)
Good hand hygiene prevents healthcare-associated infec- ensures the comparability of hospital infection outcomes
tions effectively (Allegranzi & Pittet 2009, WHO 2009) (Kanerva et al. 2009). Organisation level structures, such as
and has become an imperative in health care in terms of high-tech facilities and infection control departments in
patient safety. Hand hygiene is a simple individual action. hospitals, ensure processes in terms of disseminating knowl-
However, improvement of performance requires complex edge about hand hygiene practices to the staff. In addition,
system changes across the organisations (VanDeusen Lukas the education of healthcare staff is comprehensive by nature
et al. 2010). Hand hygiene performance is usually mea- and there are no uneducated personnel working in patient
sured in terms of occurrence of hand disinfection and is care in hospitals. In spite of these features, Finnish health-
defined in terms of a compliance rate (Allegranzi & Pittet care shares worldwide concerns relating to hand hygiene
2009, Randle et al. 2012). Hand washing, hand rubbing practices among healthcare professionals. This study aims
and using gloves for aseptic tasks are usual methods for to evaluate the usability of fidelity measures in compliance
hand hygiene. The focus of this study is on hand rubbing, evaluation in hand hygiene to find the gaps between actual
which is important as a determinative one of the preven- and optimal practices.
tion of infections in health care. Despite emphasising hand
rubbing, hand washing is recommended in certain situa-
Background
tions. Several common factors hinder the realisation of
good hand hygiene, such as lack of appropriate hand
Hand hygiene as a source of economic burden in
hygiene practices and use of jewellery and long or artificial
hospitals
nails in health care (Allegranzi & Pittet 2009, WHO
2009). Hospital acquired infections (HAI) are costly in terms of
Guidelines, such as those developed by WHO (2009), organisational and patient outcomes. Infections increase
synthesise evidence with the aim of assisting practitioners’ morbidity, mortality, care-related costs and human suffer-
decision-making and improving the practices and quality of ing (Boyce & Pittet 2002, WHO 2009). The financial costs
care. Measuring hand hygiene compliance rates does not of HAIs cause an excess burden on healthcare systems all
necessarily indicate how closely the guidelines are followed, over the world. In Finland, for example, the cost is esti-
although fidelity to guidelines is an intrinsic factor in suc- mated to be some EUR 200–500 million per year (The
cess of hand sanitation. All too often, studies measure Association of Finnish Local and Regional Authorities
merely whether or hand washing or sanitisation behaviours 2010). According to Stone et al. (2009), converting the
were observed, not whether or not the behaviours followed costs of HAIs in the USA in 1992 using the consumer price
guidelines. There are many published reports documenting index revealed that these costs are $6
65 billion in 2007
the flaws in hand hygiene practices, and numerous publica- currency. Special interest should be paid to the actual pre-
tions describe solutions to the problem. Nevertheless, an vention of these costs, as the proportion of HAIs has been
unsatisfactory level of compliance still exists. There is little estimated to be preventable in 15–30% (Grol & Grimshaw
research available measuring fidelity to hand hygiene guide- 2003) or 10–70% (Harbarth et al. 2003) of cases. In Fin-
lines. Yet, failure to properly wash or sanitise hands results land, Kanerva et al. (2009) estimated that the annual HAI-
in the transmission of infectious agents. related deaths based on national prevalence data from 2005
Different terms are used in studying hand hygiene. The totalled 271 deaths per million population, of which 136
term ‘compliance’ refers to the number of hand hygiene patients would not have died if they had not developed
actions per the number of opportunities for hand hygiene. HAI.
The term ‘fidelity’ describes the extent to which the key ele- Good hand hygiene is an effective way to prevent costs
ments of intervention, i.e. hand hygiene guidelines, are fol- caused by HAI. Despite the harm caused by flaws in hand
lowed and thus deepen attention on the qualitative hygiene, adherence to guidelines by healthcare staff remains
elements of performance. a challenge (Erasmus et al. 2010, Valim et al. 2014). Con-
The study context is a university hospital setting in Fin- cerns related to low adherence are heightened by ever
land where many structural factors facilitate patient safety increasing evidence of the consequences of poor hand
in terms of evidence-based hand hygiene practices. The hygiene, not mentioning the financial burden on healthcare
Health Care Act (1326/2010) emphasises safety in care as a systems.

© 2015 John Wiley & Sons Ltd

3198 Journal of Clinical Nursing, 24, 3197–3205
Original article Evaluating fidelity to hand hygiene guidelines

Fidelity to evidence-based guidelines is needed to ensure

Elements of good hand hygiene practice
that staff use effective practices in hand hygiene. Fidelity
International guidelines relating to effective hand hygiene ratings have been used to analyse the extent to which the
procedures are consistent and are presented by the World key elements of the intervention were in place and also to
Health Organisation (WHO 2009). The principles stated in track changes over time (VanDeusen Lukas et al. 2010,
the guidelines have been available for decades stating that Fuller et al. 2012). Fidelity assessment has both administra-
hand disinfection is needed before and after patient contact, tive and research purposes (Mowbray et al. 2003) as in this
before aseptic tasks, after body fluid exposure and after study, which assesses fidelity to hand hygiene guidelines for
touching patient surroundings, as well as between different patient safety.
care activities on the same patient (Boyce & Pittet 2002, Different tools are available for evaluating compliance to
Harrington et al. 2007, WHO 2009). hand hygiene guidelines. Both the WHO and Joanna Briggs
It is also recommended to use alcohol-based products for Institute (JBI) have developed a tool for this purpose. The
hand disinfection purposes. In clinical situations, the effi- JBI tool, the so-called PACES programme (Practical Appli-
cacy of these products relates to contact time, the volume cation of Clinical Evidence System), was developed for
of products used and whether the hands are wet when alco- web-based use in clinical settings for the evaluation of
hol is used (Boyce & Pittet 2002, Goroncy-Bermes et al. implementing evidence into practice (Harrington et al.
2010). Alcohol-based disinfection products are recom- 2007). Both tools developed by the WHO and the JBI are
mended due to their better microbiological efficacy, the in common use (Harvey et al. 2012) in addition to locally
shorter time needed to achieve the desired effects and easy developed tools (VanDeusen Lukas et al. 2010, Oh et al.
access to the point of care (Allegranzi & Pittet 2009). Cur- 2012). In this study, we paid attention to fidelity to hand
rently, the recommended duration of hand rubbing is hygiene guidelines to obtain a more precise measurement of
30 seconds, which is consistent with the European stan- the evidence based on the duration of hand rubbing and in
dard. The standard (EN 1500) is for evaluating the efficacy using jewellery or long or artificial nails. We compared the
of hand hygiene disinfection products (Goroncy-Bermes two approaches to evaluate hand hygiene performance as
et al. 2010). compliance and fidelity.

Evaluating hand hygiene performance Methods

Hand hygiene has been evaluated in many ways, such as A cross-sectional approach was used in evaluating the
frequency or hand hygiene technique, staff attitudes and usability of fidelity evaluation as compared with compliance
knowledge. Indirect measures, such as the consumption of evaluation. A structured observation was used aiming to
alcohol-based hand hygiene products and microbiologic gain insight into hand hygiene practices in daily work.
measures are also used (Scheithauer et al. 2009). In addi- Structured observation refers to data collection with obser-
tion, compliance rate and fidelity have been used as the per- vation form based on clearly defined data collection meth-
spectives of evaluation. ods. The role of the observer differs from detached to full
Hand hygiene performance is usually evaluated in terms participation in the activities that they observe (Caldwell &
of compliance (Allegranzi & Pittet 2009, Randle et al. Atwal 2005). In this study, nonparticipant, direct and struc-
2012, Valim et al. 2014). The concept of compliance refers tured observation was used in natural settings.
to the number of hand hygiene actions per the number of Direct observation is common in evaluating hand hygiene
opportunities for hand hygiene (McAteer et al. 2008, Oh in health care (Steed et al. 2011, Oh et al. 2012). The
et al. 2012). The opposite concept of noncompliance was method has been ranked as the most effective in measuring
defined by Picheansathian et al. (2008) as failure to practice hand hygiene performance (McAteer et al. 2008). The cred-
hand hygiene as described in the guidelines. However, it is ibility of observation is based on the observer’s ability to
seldom explicitly described how the concept of compliance see, record, interpret and evaluate information. It depends
is operationalised in studies. as much on the behaviour of the observer as of that of the
Another view on evaluating hand hygiene performance people being observed (Scheithauer et al. 2009). Using
relates to fidelity to guidelines. Fidelity gives a more precise observation as a data collection method is criticised in
explanation of evidence-based practice (Mowbray et al. terms of the time and staff needed (McAteer et al. 2008).
2003, O’Connor et al. 2007), thus going further than In addition, the methodological quality of observation is
merely revealing the occurrence of the phenomenon. discussed in terms of the lack of information concerning

© 2015 John Wiley & Sons Ltd

Journal of Clinical Nursing, 24, 3197–3205 3199
A Korhonen et al.

inter-rater reliability and the lack of clear standard operat- whether the practices are consistent with the guidelines,
ing procedures needed to assess reliability and to ensure the three of the original criteria (numbers 1, 4 and 5) were
widespread replication or potential risk of the Hawthorne open to sub-criteria (Table 1); The purpose was to clarify
effect (Randle et al. 2012). Many of these issues are pre- too broad criteria. The sub-criteria for fidelity were based
ventable in observing hand hygiene performance. on evidence (WHO 2009, Goroncy-Bermes et al. 2010).
Due to the difficulties of measuring whether hand rubbing
is routine by nature (original PACES criterion), fidelity to
The observation form
evidence was determined as the duration of hand rubbing
An expert panel consisting of the staff of the hospital infec- before patient contact and after touching a patient’s sur-
tion control unit compiled the observation form based on roundings. In the 4th criterion, ‘different care activities’
the evaluation criteria developed by the JBI. The PACES (original PACES criterion) were defined as ‘before clean
programme consists of seven criteria, five of which (related procedures’ and ‘after body fluid exposure’. Fidelity to the
to compliance to hand rubbing) were selected for this study. 5th criterion relating to staffs’ hand hygiene education
The original PACES criterion relating to hand rubbing (original PACES criterion) was defined as the minimum
being routine by nature was found challenging, because the knowledge to be gained in hand hygiene education, such as
word ‘routine’ was not clearly defined. Thus, it may mean not wearing rings, watches or artificial or long (i.e. exceed-
that something was done to decontaminate the hands, but ing fingertips) fingernails (Trick et al. 2003, WHO 2009).
may not have followed the guidelines. This was due to healthcare education include aseptic beha-
In this study, the definition of ‘routine’ in the first PACES viour training and also to continuous attention to hand
criterion was operationalised as a sum variable combining hygiene by the infection control unit of the hospitals.
the results of the criteria numbers 2 (Hands are decontami- The observation form was to be used manually. It con-
nated before patient contact), 3 (Hands are decontaminated tained the identification number of the unit and observed
after patient contact) and 4 (Hands are decontaminated staff in terms of subgroups of physicians and the other staff
between different care activities) (Table 1). To investigate including all nurses, physiotherapists and laboratory and

Table 1 The selected original criteria (Practical Application of Clinical Evidence System) and the sub-criteria for fidelity to evidence-based
practices

Joanna Briggs Institute criteria for

compliance rate Sub-criteria for fidelity References

1. An alcohol-based hand rub is routinely* Length of hand rubbing: Boyce and Pittet (2002)
used for hand hygiene unless hands are (1) Before direct patient contact Goroncy-Bermes et al. (2010)
visibly soiled (2) After touching the patient’s EN 1500 standard
surroundings
2. Hands are decontaminated immediately Hands are decontaminated immediately Boyce and Pittet (2002)
before each and every episode of direct before each and every episode of direct WHO (2009)
patient contact or care and contact with patient contact
inanimate objects including equipment
3. Hands are decontaminated immediately Hands are decontaminated immediately Boyce and Pittet (2002)
after contact with an individual patient after contact with an individual patient WHO (2009)
and/or all inanimate objects including
equipment.
4. Hands are decontaminated with an Hands are decontaminated: Boyce and Pittet (2002)
alcohol-based hand rub (unless hands are (1) Before clean/aseptic procedure WHO (2009)
visibly soiled) between different care (2) After touching body fluid exposure
activities for the same patient
5. Staff has received education about hand During care activities, are there any Boyce and Pittet (2002)
hygiene (1) rings on fingers? Trick et al. (2003)
(2) artificial or long natural nails (i.e. WHO (2009)
exceeding fingertips)?
(3) watches?

*Measured as sum variable of criteria 2, 3 and 4.

© 2015 John Wiley & Sons Ltd

3200 Journal of Clinical Nursing, 24, 3197–3205
Original article Evaluating fidelity to hand hygiene guidelines

radiology personnel as background information (referred as rion), the sum variable of the ‘Routine use of hand rub’
‘nurses’). All items on the observation form, except those was formed by combining the criteria numbers 2, 3 and 4
relating to time measurement, were dichotomous (Yes/No). (Table 1), counting the frequency of Yes choices and divid-
The content validity of the form was evaluated by an expert ing the sum by total options. The other criteria were anal-
panel. Inter-rater reliability was evaluated by simultaneous ysed independently as frequencies and percentages and
observations conducted by two nurses experienced in infec- illustrated as graphic bars by the programme.
tion control, being 0
85. The main differences between the The second analysis concerned fidelity where descriptive
observations were in measuring the length of time of hand statistics (fr, mean, % and range) were used as a basis for
rubbing. Excluding this, inter-rater reliability was 0
88. evaluating the practice. In addition, the binomial standard
In the observation process, one case was defined as an normal deviate (SND) test was conducted for comparing
occasion where hand hygiene practices must occur, such as different methods (compliance rate vs. fidelity to guidelines)
dispensing medication, touching the patient or his/her envi- used in the evaluation of hand hygiene. The test is conve-
ronment. As we were interested in hand hygiene perfor- nient for dichotomous variables, where the distribution is
mance, the same people could be observed repeatedly. the probability of distribution of the number of ‘successes’
in independent yes/no trials (Polit & Beck 2012; 407). In
this study, for example, the value ‘yes’ when measuring the
Data collection
compliance rate (i.e. something was done) and in the case
For data collection, 16 observers from 13 different units of of fidelity (i.e. hand rubbing lasted 30 seconds or more
the hospital were selected by senior personnel, consisting of meaning that it happened according to the guideline)
head nurses and these in expert positions in the hospital. referred to success measured by two different methods. In
The potential observers for recruitment had participated in the first PACES criterion, the length of hand rubbing lasting
a training day relating to the introduction of the JBI pro- 30 seconds or more was interpreted as hand rubbing being
gramme for implementing evidence into practice. The routine by nature. In the fourth (Hands are decontaminated
observers were offered 1
5 hours training, comprising between different care activities) and fifth (Staff has
induction into optimal hand hygiene practices, an introduc- received education about hand hygiene) criteria, a sum vari-
tion to the observation form, discussion on observation as a able was formed to the sub-categories; in the fourth crite-
data collection method and the ethics of observing. No for- rion, both options (before aseptic procedure and after
mal observation training was offered due to the long expe- touching body fluids, Table 1) were calculated together.
rience of the observers and the lack of interpretations For the fifth criterion, the variable was scaled as follows
needed in observations (Parahoo 2006). Data were gathered based on the fact that from the point of view of infection
in December 2011. All observers, except those in radiology control each part of the variable has equivalent value. In
and laboratory units, observed hand hygiene performance this study, each of them can get the value of 1, which
in a neighbouring unit to avoid observers’ habituation to means that whether the sum of the parts varies between 1–
certain routines in their own unit, which might have com- 3, it indicates that there are one or all parts of the variable
promised their sensitivity while observing. The observers in use (rings, watches or unsuitable nails). The value 0 indi-
selected 5 four-hour episodes during the two-week study cated that the criterion was accepted, that is there were no
period based on their own duties. In practice, the data were rings, watches or unsuitable nails in use. Hand hygiene
collected in different units (surgical, medical, paediatric, practices were also examined by the type of provider,
radiological and laboratory unit) excluding all intensive physicians vs. nurses and others. The results of the compli-
care units. The occasions observed were common daily car- ance measure and measuring fidelity were entered and pre-
ing situation, such as giving medication, caring for patients sented by the PACES programme graphics (Fig. 1).
and medical rounds. The exceptional situations, as resusci-
tations or other demanding procedures were excluded.
Ethical considerations

The research was conducted in accordance with national

Data analysis
legislation and the Declaration of Helsinki (2013). Adminis-
The data were analysed in two phases. The first related to trative approval for the study was given by the hospital
the automatic analysis by the PACES programme as per- instead of approval from an official ethics committee based
centage rates for compliance. In defining whether the hand on Finnish law (Medical Research Act 2010/794). All
disinfection was routinely performed (the first PACES crite- hospital units were informed about the study in the hospital

© 2015 John Wiley & Sons Ltd

Journal of Clinical Nursing, 24, 3197–3205 3201
A Korhonen et al.

Table 2 A sum variable of ‘The routine use of hand rub’ illustrated

by the sum of the Practical Application of Clinical Evidence System
criteria number 2, 3 and 4 and divided by subgroups

Observed situations Total of

observed
Yes No
situations
Actors f % f % N

Nurses 1590 81 361 19 1951

Physicians 201 57 150 43 351
All together 1791 78 511 22 2302

difference between the professional groups was not statisti-

cal significant (p = 0
5, 95% CI 0
03–0
05). For example,
Figure 1 The results by the original Practical Application of Clini-
cal Evidence System criteria in term of compliance (black bars) and
the mean duration of hand rubbing was on average
results in terms of fidelity to guidelines (grey bars). 15
3 seconds (range 1–60 seconds) for nurses, who disin-
fected their hands before touching the patient (n = 538,
bulletin and through information given by the chief nursing 76%) and on average 14
5 seconds (range 1–57 seconds)
officers and head nurses. In addition, the observers for physicians who disinfected their hands before touching
introduced themselves and the aim of the observation, its the patient (n = 66, 55%). A difference was also found in
anonymity and confidentiality to the staff and to the the criterion relating to staff education, showing a 3% dif-
patients who were present every time they entered the units. ference (p < 0
0001, 95% CI 0
02–0
04) indicating that
Oral permission was given by those who were observed. jewellery, watches or long or artificial nails were observed.

Results Discussion
A total of 830 occasions were observed consisting of 2302 The study aimed to evaluate the usability of fidelity mea-
different actions, where hand hygiene should be performed. sures in compliance evaluation of hand hygiene. Evaluating
Among these situations, the actors were nurses (n = 709) fidelity to guidelines revealed that it is possible to uncover
and physicians (n = 121). In general, hand hygiene compli- gaps in hand hygiene practices. The results demonstrated
ance varied between 77–100% depending on the criteria. that compliance with hand hygiene guidelines was routine
The best compliance rate was found in the criterion of staff and satisfactory by nature, indicating that an alcohol-based
education (Fig. 1, black bars). solution was in general applied to hands. In the case of
The differences between the two evaluation methods used fidelity evaluation, the length of hand rubbing was unsatis-
(compliance rate vs. fidelity to guidelines) were statistically factory in terms of the short duration of hand rubbing
significant (p < 0
0001, 95% CI 0
66–0
7) in the first criterion before touching the patient and after touching the patient’s
used an alcohol hand rub routinely. Similarly, a statistically surroundings. In addition, in the last criterion in PACES,
significant difference (p < 0
0001, 95% CI 0
02–0
04) was concerning hand hygiene education being interpreted as
found in the last criterion – absence of jewellery and artificial knowledge to be gained, a small, but statistical significant
or long nails (Fig. 1) but not among the other criteria used. difference was found, which indicated that some rings,
Evaluating hand disinfection by compliance rate revealed, watches or long or artificial nails were observed. Thus, the
that it was routine by nature more often among the group information gained in hand hygiene education was not fully
‘nurses’, which consisted of nurses, physiotherapists and used in some instances.
laboratory and radiology personnel, than among the physi- In the literature, the overall compliance rate in hand
cians. Considering all observations together revealed that in hygiene performance has usually been found to be below
22% of occasions, hand hygiene would not be defined as 50% (Valim et al. 2014), indicating that the results of this
routine by nature (Table 2). study seem to be good. The compliance rate referring to
The results relating to fidelity to guidelines indicated dif- actions per opportunities (Steed et al. 2011) is a useful indi-
ferences in performance (Fig. 1, grey bars). The overall cator for measuring to what extent hand rubbing is used in
duration of hand rubbing (the first criterion) was 30 sec- general. The good compliance rate found in this study may
onds or more in 10% of the cases (n = 118). However, the reflect active efforts by the hospital infection control unit in

© 2015 John Wiley & Sons Ltd

3202 Journal of Clinical Nursing, 24, 3197–3205
Original article Evaluating fidelity to hand hygiene guidelines

emphasising the meaning of hand disinfection. It is also continue to develop institutional policies based on expert
suggested that the compliance rate may have been affected opinions. An international expert opinion is available in
by observation. In addition, hand rubbing has been empha- guidelines developed by the WHO (2009). In addition, the
sised in Finnish health care universally for years. Evaluating restriction of finger rings, long and artificial nails and
fidelity would be useful if the focus of interest is on watches is recommended by another guidelines (Pratt et al.
whether this solution is used appropriately. 2001). The WHO (2009) emphasises the importance of
Interpreting PACES criteria as sub-criteria in terms of hand hygiene in patient safety and that active monitoring
fidelity to guidelines revealed ineffective hand-rubbing of hand hygiene practices is therefore important.
behaviour. Hand-rubbing behaviour was routine by nat- Direct observation proved to be a suitable and efficient
ure in terms of compliance rate while the duration of method for evaluating hand hygiene practices (Boyce & Pit-
hand rubbing was too short in terms of fidelity to guide- tet 2002, McAteer et al. 2008) and it is a method used in
lines. According to the European standard, the effective PACES. According to Morgan et al. (2012), automated
length of hand rubbing has been defined as 30 seconds hand hygiene count devices may be more effective in mea-
(Goroncy-Bermes et al. 2010). More studies are needed suring unit-wide compliance than direct observation. A
for explaining the shortest duration of effective hand rub- variety of methods is needed and the selection will depend
bing. In addition, interpreting staff education in terms of on the purpose of the evaluation. Observation makes it pos-
the outcomes of education and measured by not using sible to observe appropriate fidelity concerning the length
rings, watches and having long nails, caused a small but of hand rubbing, as in our study, for example. Even the
statistically significant decrease in fidelity evaluation com- Hawthorne effect is considered to be one of the main prob-
pared with the compliance rate. The result is clinically lems in observational study (Parahoo 2006), we conducted
significant as a means of patient safety, as using rings, open observations as an ethical principle (Kohli et al.
watches and having long or artificial nails increase 2009). In addition, we suggested an open observation may
patient risk for HAIs (Allegranzi & Pittet 2009, WHO act as a simultaneous intervention facilitating hand hygiene
2009) and undermines the other staff members’ efforts to performance. Although the observations of practices have
follow the guidelines. been found to be resource consuming (Morgan et al. 2012),
A recent Cochrane review (Arrowsmith & Taylor 2012) the costs caused by care-related infections are more expen-
revealed that there is a lack of evidence to determine sive for the healthcare organisations and the patients than
whether using rings or nail polish affects the rate of wound the costs caused by observation.
infections. The review included only one RCT (published
1994) relating to using rings and nail polish of OR staff on
Limitations
postoperative wound infection. Fagernes and Lingaas
(2011) studied the hands of healthcare workers (n = 465) The main limitation of the study related to the lack of a
for the total number and presence of certain type of bacte- formal training of the observers. Although the observers
ria. Using one plain ring increased the carriage rate of had a long work experience, they had little experience of
Enterobacteriacea (odds ratio 2
71, 95% CI 1
42–5
20, observation as a data collection method, which may cause
p = 0
003). The increased bacteria count of Staphylococcus failures by the observers (Caldwell & Atwal 2005, Parahoo
aureus was found in fingernails longer than 2 mm (odds 2006). Even the observers’ training has been recommended
ratio 2
17, 95% CI 1
29–3
66, p = 0
004). No effect of nail to ensure proper data collection (Parahoo 2006), the formal
polish was found. In another study (Jeans et al. 2010) it training was not offered due the observers’ professional
was found that wearing a wrist watch increased bacterial qualification and experience in health care. In addition, the
contamination of the wrists but if the wrists are not manip- sub-criteria for fidelity to guidelines may be interpreted dif-
ulated; no increased hand contamination was found. This ferently in other contexts, depending on the aim of evalua-
was not in line with the study of Fagernes and Lingaas tion and context-specific factors. However, we based the
(2011), where it was found that using wrist watch was fidelity criteria on current evidence and infection control
associated with an increased bacterial count on hands com- expertise. In addition, the professional education in Finland
pared with hands without a watch (95% CI 1
73–6
07, consists of the principles of infection prevention and no
p = 0
001). Thus, due the small amount of studies which uneducated personnel work in direct patient care. Conse-
evaluate the impact of rings, nail polish and wrist watches, quently, special caution is required when extrapolating the
the authors (Fagernes & Lingaas 2011, Arrowsmith & Tay- results of fidelity assessment to other contexts even if the
lor 2012) recommended that healthcare organisations must results are consistent with previous knowledge.

© 2015 John Wiley & Sons Ltd

Journal of Clinical Nursing, 24, 3197–3205 3203
A Korhonen et al.

The validity and reliability of an observation study

depends on the degree of interpreting the focus of interest
Relevance to clinical practice
into clear items to be observed (Parahoo 2006). We used the The measurement of fidelity to guidelines is a suitable
criteria developed by the JBI. These criteria are based on evi- method for revealing an inconsistency between optimal and
dence and are consistent with those developed by the WHO actual hand hygiene practices. Fidelity measures seem to be
(2009). However, the term ‘routine’ used in the first PACES relevant to clinical issues in different healthcare contexts.
criterion might be confusing and difficult to measure as it is. Knowing the gap facilitates continuous improvement in
Thus, the criterion was measured forming a sum variable clinical practice.
developed according to the expert panel. The sub-criteria for
evaluating fidelity were based on current evidence and an
expert panel of the hospital infection control unit. The obser-
Acknowledgements
vation form was evaluated by experts in the hospital infec- The authors give their gratitude to The Federation for
tion control unit to ensure its content validity (Parahoo Nurse Education, Finland for their financial support.
2006). The reliability of an observational study relates to the
consistency by which the observers may categorise their
observations. The items to be observed were unambiguous
Contributions
by nature and thus did not require any interpretation, which Conception and study design: AK, HO, TP, RJ, PK, AH;
has been found to increase the consistency of observations data analysis: AK; manuscript preparation: AK and dis-
(Caldwell & Atwal 2005, Parahoo 2006). In spite of this, we cussing, revising, approving the article: HO, TP, RJ, PK,
found that the observation form needs to be revised by add- AH.
ing time measures to all of the appropriate PACES criteria to
enhance comprehensive measurement of fidelity.
Funding
Financial support was given to the study by The Federation
Conclusions
for Nurse Education, Finland.
Combining the measurement of fidelity to guidelines with
the compliance rate is beneficial in revealing an inconsis-
tency between common hand hygiene behaviour and the
Conflict of interest
guidelines. The authors declare that they have no conflict of interests.

References
Allegranzi B & Pittet D (2009) Role of collect data in nursing research. Nurse Fuller C, Michie S, Savage J, McAteer J,
hand-hygiene in health care-associated Researcher 13, 42–54. Besser S, Charlett A, Hayward A,
infection prevention. Journal of Hospi- Declaration of Helsinki (2013) Ethical prin- Cookson BD, Cooper BS, Duckworth
tal Infection 73, 305–315. ciples for medical research involving G, Jeanes A, Roberts J, Teare L &
Arrowsmith VA & Taylor R (2012) Re- human subjects. Available at: http:// Stone S (2012) The feedback interven-
moval of nail polish and finger rings www.ub.edu/recerca/Bioetica/doc/De- tion trial (FIT) – Improving hand
to prevent surgical infection. The claracio_Helsinki_2013.pdf (accessed 2 hygiene compliance in UK healthcare
Cochrane Database of System- January 2014). workers: a stepped wedge cluster ran-
atic Reviews Issue 5, Art. No: CD Erasmus V, Daha TJ, Brug H, Richardus domized controlled trial. Public
003325. JH, Behrendt MD, Vos MC & van Library of Science One 7, e416–e417.
Boyce JM & Pittet D (2002) Guideline for Beek F (2010) Systematic review of Goroncy-Bermes P, Koburger T & Meyer
hand hygiene in health-care settings. studies on compliance with hand B (2010) Impact of the amount of
Recommendations of the healthcazre hygiene guidelines in hospital care. In- hand rub applied in hygienic hand dis-
infection control practices advisory fection Control and Hospital Epidemi- infection on the reduction of microbi-
committee and the HICPAC/SHEA/ ology 31, 283–294. ological counts on hands. Journal of
APIC/IDSA hand hygiene task force. Fagernes M & Lingaas E (2011) Factors Hospital Infection 74, 212–218.
Morbidity and Mortality Weekly interfering with the microflora on Grol R & Grimshaw J (2003) From best
Report 51, RR-16. hands: a regression analysis of samples evidence to best practice: effective
Caldwell K & Atwal A (2005) Non-partic- from 465 healthcare workers. Journal implementation of change in patients’
ipant observation: using video tapes to of Advanced Nursing 67, 297–307. care. Lancet 362, 1225–1230.

© 2015 John Wiley & Sons Ltd

3204 Journal of Clinical Nursing, 24, 3197–3205
Original article Evaluating fidelity to hand hygiene guidelines

Harbarth S, Sax H & Gastmeier P (2003) O’Connor C, Small SA & Cooney SM Steed C, Kelly JW, Blackhurst D, Boeker S,
The preventable proportion of nosoco- (2007) Program fidelity and adaptation: Diller T, Alper P & Larson E (2011)
mial infections: an overview of pub- meeting local needs without compro- Hospital hand hygiene opportunities:
lished reports. Journal of Hospital mising program effectiveness. What where and when (HOW2)? The
Infection 54, 258–266. works, Wisconsin – Research to Prac- HOW2 Benchmark study. American
Harrington L, Lesh K, Doell L & Ward SK tice Series. 4. Available at: http://what- Journal of Infection Control 39, 19–26.
(2007) Reliability and validity of hand works.uwex.edu/attachment/whatworks_ Stone PW, Kunches L & Hirschhorn L
hygiene measures. Journal for Health- 04.pdf (accessed 17 October 2012). (2009) Cost of hospital-associated
care Quality 29, 20–29. Oh E, Hamzah HBM, Yan CC & Ang E infections in Massachusetts. American
Harvey G, Kitson A & Munn Z (2012) Pro- (2012) Enhancing hand hygiene in a Journal of Infection Control 37, 210–
moting continence in nursing homes in polyclinic in Singapore. International 214.
four European countries: the use of Journal of Evidence-Based Healthcare The Association of Finnish Local and
PACES as a mechanism for improving 10, 204–210. Regional Authorities (2010) Sairaa-
the uptake of evidence-based recommen- Parahoo K (2006) Nursing Research. Prin- lainfektioista viidennes torjuttavissa
dations. International Journal of Evi- ciples Process and Issues, 2nd edn. (In Finnish). Available at: http://www.
dence-Based Healthcare 10, 388–396. Palgrave Macmillan, Basingstoke, kunnat.net/fi/Kuntaliitto/media/tiedot-
Jeans AR, Moore J, Nicol C, Bates C & New York, NY, pp. 347–374. teet/2010/10/Sivut/Sairaalainfektiois-
Read RC (2010) Wristwatch use and Picheansathian W, Pearson A & Suchaxaya ta-viidennes-torjuttavissa.aspx (access-
hospital-acquired infection. Journal of P (2008) The effectiveness of a promo- ed 15 October 2013).
Hospital Infection 74, 16–21. tion programme on hand hygiene The Health Care Act (1326/2010). . Avail-
Kanerva M, Ollgren J, Virtanen MJ & compliance and nosocomial infections able at: http://www.finlex.fi/fi/laki/
Lyytik€ainen O (2009) Estimating the in a neonatal intensive care unit. Inter- ajantasa/2010/20101326 (accessed 15
annual burden of health care-associ- national Journal of Nursing Practice October 2013).
ated infections in Finnish adult acute 14, 315–321. Trick WE, Vernon MO, Hayes RA,
care hospitals. American Journal of Polit DF & Beck CT (2012) Inferential Nathan C, Rice TW, Peterson BJ, Se-
Infection Control 37, 227–230. statistics. In Nursing Research. Gener- greti J, Welbel SF, Solomon SL & We-
Kohli E, Ptak J, Smith R, Taylor E, Talbot ating and Assessing Evidence for instein RA (2003) Impact of ring
EA & Kirkland KB (2009) Variability Nursing Practice, 9th edn (Polit DF & wearing on hand contamination and
in the Hawthorne effect with regard Beck CT). Wolters Kluwer, Lippincott, comparison of hand hygiene agents in
to hand-hygiene performance in high- Williams & Wilkins, Philadelphia, PA, a hospital. Clinical Infectious Diseases
and low-performing inpatient care pp. 404–432. 36, 1383–1390.
units. Infection Control and Hospital Pratt RJ, Pellowe C, Loveday HP, Robin- Valim MD, Marziale MHP, Richart-Marti-
Epidemiology 30, 222–225. son N, Smith GW, the epic guideline nez M & Sanjuan-Quiles A (2014) In-
McAteer J, Stone S, Fuller C, Charlett A, development team, Barrett S, Davey P, strument for evaluating compliance
Cookson B, Slade R, Michie S & Harper P, Loveday C, McDoudall C, with infection control practices and fac-
NOSEC/FIT group (2008) Develop- Mulhall A, Privett S, Smales C, Taylor tors that affect it: an integrative review.
ment of an observational measure of L, Weller B & Wilcox M (2001) The Journal of Clinical Nursing 23, 1502–
health care worker hand-hygiene epic project: developing national 1519.
behavior: the hand-hygiene observa- evidence-based guidelines for prevent- VanDeusen Lukas C, Engle RL, Holmes
tion tool (HHOT). Journal of Hospital ing healthcare-associated infections. SK, Parker VA, Petzel RA, Seibert
Infection 68, 222–229. Phase 1: guidelines for preventing MN, Schwartz M & Sullivan JL
Medical Research Act 2010/794. Available hospital-acquired infections. Journal (2010) Strengthening organizations to
at: http://www.finlex.fi/fi/laki/kaannok- of Hospital Infection 47(Suppl.), S1– implement evidence-based clinical
set/1999/en19990488 (accessed 15 S81. practices. Health Care Management
December 2013). Randle J, Firth J & Vaughan N (2012) An Review 35, 235–245.
Morgan DJ, Pineles L, Shardell M, Young observational study of hand hygiene WHO (World Health Organization) (2009)
A, Ellingson K, Jernigan JA, Day HR, compliance in paediatric wards. Jour- WHO Guidelines on Hand Hygiene in
Thom KA, Harris AD & Perencevitch nal of Clinical Nursing 22, 2586– Health Care. First Global Patient Safety
EN (2012) Automated hand hygiene 2592. Challenge. Clean Care is Safer Care.
count devices may better measure Scheithauer S, Haefner H, Schwantz T, WHO Library Cataloguing-in-Publica-
compliance than human observation. Schulze-Steinen H, Schiefer J, Koch A, tion Data, Geneva, Switzerland. Avail-
American Journal of Infection Control Engels A & Lemmen SW (2009) Com- able at: http://whqlibdoc.who.int/
40, 955–999. pliance with hand hygiene on surgical, publications/2009/9789241597906_
Mowbray CT, Holter MC, Teague GB & medical, and neurologic intensive care eng.pdf?bcsi_scan_7cec9f03dcb71299
Bybee D (2003) Fidelity criteria: devel- units: direct observations versus calcu- =MqwhW5/SakPNGhSQfEvcsgsLjYo6
opment, measurement and validation. lated disinfectant usage. American AAAA70XROg==&bcsi_scan_filename=
The American Journal of Evaluation Journal of Infection Control 37, 835– 9789241597906_eng.pdf (accessed 15
24, 315–340. 841. March 2012).

© 2015 John Wiley & Sons Ltd

Journal of Clinical Nursing, 24, 3197–3205 3205