Health Psychology © 2010 American Psychological Association

2010, Vol. 29, No. 1, 56 – 64 0278-6133/10/$12.00 DOI: 10.1037/a0017787

Using the Common Sense Model of Illness Perceptions to Examine

Osteoarthritis Change: A 6-Year Longitudinal Study

Ad A. Kaptein, Jessica Bijsterbosch, Sarah E. Hampson

and Margreet Scharloo Oregon Research Institute, Eugene, Oregon
Leiden University Medical Center, Leiden, the Netherlands

Herman M. Kroon and Margreet Kloppenburg

Leiden University Medical Center, Leiden, the Netherlands

Objective: To examine the association between changes in common sense models and changes in
functional status over a 6-year follow-up in patients with osteoarthritis. Design: At baseline and
follow-up, osteoarthritis outpatients (N ⫽ 241) recruited from a university medical center completed the
Illness Perception Questionnaire—Revised (IPQ-R), the Australian/Canadian Osteoarthritis Hand Index,
and the Western Ontario and McMasters Universities Osteoarthritis Index. Also, their physician-assessed
pain intensity, and biomedical, and clinical measures of medical severity of osteoarthritis were recorded.
Main outcome measures: Functional disability, pain intensity. Results: Over 6 years, functional
disability and pain intensity increased. The IPQ-R dimensions of timeline, personal control, and illness
coherence became more negative, and emotional representations became less negative (i.e., more
accepting). Patients identified as sharing a similar profile of negative changes on the IPQ-R had
significantly worse functioning on 2 of 3 outcomes, independent of objectively measured osteoarthritis
severity. Conclusions: Changes in illness perceptions were associated with changes in outcomes.
Interventions to prevent increasingly negative patterns of illness perceptions over time, with an
emphasis on strengthening control cognitions, may benefit functional status outcomes in patients
with osteoarthritis.

Keywords: common sense model, illness perceptions, longitudinal design, osteoarthritis, self-
management

The outcome of medical care for patients with chronic physical change over time. We are aware of only three previous longitudi-
illness is determined to a considerable extent by nonmedical fac- nal studies in which changes in illness perceptions were examined
tors (e.g., Leventhal, Weinman, Leventhal, & Phillips, 2008). together with change in health status. Foster et al. (2008) found
According to the common sense model (CSM), illness perceptions that the changes seen in several dimensions of the Illness Percep-
(both cognitive and emotional) and coping responses are determi- tion Questionnaire—Revised (IPQ-R; Moss-Morris, Weinman,
nants of medical outcomes (Leventhal, Brissette, & Leventhal, Petrie, Horne, Cameron, & Buick, 2002; see also the Illness
2003). There is considerable evidence in support of various aspects Perception Questionnaire Web site at http://www.uib.no/ipq/
of the CSM, although studies of processes by which illness per- index.html) were different in patients with low back pain who had
ceptions change and the health consequences of these changes a good clinical outcome compared with those who had a poor
remain relatively rare (Hagger & Orbell, 2003). The present study outcome at 6-month follow-up. Furze, Lewin, Murberg, Bull, and
examined the association between changes in illness perceptions Thompson (2005) found that change in beliefs about angina was
and changes in functional status over a 6-year follow-up period for
the most significant predictor for physical status at 1-year follow-
patients with osteoarthritis (OA).
up. In a large sample of recently diagnosed patients with Type 2
Longitudinal studies of illness perceptions for a chronic illness
diabetes, self-management and a patient education program led to
create the opportunity to examine whether illness perceptions
changes in illness perceptions, with consequent changes in quality
of life and metabolic control at the 3-month follow-up (Skinner et
Ad A. Kaptein and Margreet Scharloo, Unit of Psychology, Leiden al., 2006).
University Medical Center, Leiden, the Netherlands; Jessica Bijsterbosch Our study also enabled the exploration of a new theoretical issue
and Margreet Kloppenburg, Department of Rheumatology, Leiden Univer- regarding illness perceptions, namely the examination of clusters
sity Medical Center, Leiden, the Netherlands; Sarah E. Hampson, Oregon of persons characterized by similar change profiles across dimen-
Research Institute, Eugene, Oregon; Herman M. Kroon, Department of
sions of illness perception and the relation of these clusters to
Radiology, Leiden University Medical Center, Leiden, the Netherlands.
Correspondence concerning this article should be addressed to Ad A.
changes on various outcomes. The developers of the CSM have
Kaptein, Unit of Psychology, Leiden University Medical Center (LUMC), emphasized the potential value of examining interrelations be-
P.O. Box 9600 [post zone J9, r.84], 2300 RC Leiden, the Netherlands. tween combinations of illness perceptions as predictors of out-
E-mail: [email protected] comes in patients with chronic physical illness (Leventhal et al.,

56
 EXAMINING OSTEOARTHRITIS CHANGE 57

2003). Clatworthy, Hankins, Buick, Weinman, and Horne (2007) Method

took up this challenge and maintained that “people do not hold
illness representations in isolation, they are part of a schema . . . Participants and Recruitment
when it comes to the analysis, it may be more appropriate to use
a method that takes into account all aspects of a patient’s illness The GARP study population comprises Caucasian sib-pairs of
schema. . . . cluster analysis enables the identification of groups of Dutch ancestry with familial OA at multiple sites. Details on the
people who share similar illness perceptions, and the utility of the recruitment, selection, and inclusion have been published else-
CSM in predicting coping and outcome from these beliefs can still where (Riyazi et al., 2005). Patients were included in the study
be tested” (p. 125). An objective of our study, therefore, was to through rheumatology and orthopedic outpatient clinics or through
determine whether there would be differences on outcomes be- practices of general practitioners (family physicians). Patients with
tween groups of patients identified as sharing similar patterns of secondary OA, familial syndromes with a clear Mendelian inher-
change in illness perceptions. itance pattern, or a shortened life expectancy were excluded. The
OA is one of the most common chronic conditions in elderly GARP study was approved by the Medical Ethics Committee of
persons in developed societies, with a significant impact on their the Leiden University Medical Center, Leiden, the Netherlands.
quality of life (e.g., Theis, Helmick, & Hootman, 2007). Current OA diagnosis. All patients had familial OA. The OA had to
treatment for OA includes pharmacological therapy to alleviate the have a polyarticular or generalized nature, defined as OA at
impact of inflammation and pain, physiotherapy to facilitate ac- multiple sites. Patients were eligible for inclusion if they had
tivities of daily living, and psychosocial interventions to reduce the symptomatic OA at multiple joint sites in the hand or with OA in
negative psychosocial effects and to encourage social participation two or more of the following joint sites: hand, spine, knee, or hip.
in society (Dieppe & Lohmander, 2005; Kratz, Davis, & Zautra, Patients with just one symptomatic joint site with OA were re-
2007; Newman, Steed, & Mulligan, 2004). We are aware of 13 quired to have structural abnormalities (radiographic OA or bony
previous empirical studies in which illness perceptions of OA swelling) in at least one other joint site. This phenotype is in
patients were addressed (Appelt, Burant, Siminoff, Kwoh, & Ibra- accordance with the definition by Kellgren and Lawrence of gen-
him, 2007; Ballantyne, Gignac, & Hawker, 2007; Botha-Scheepers eralized OA (Kellgren & Moore, 1952; Lawrence, 1963). The
generalized nature of the disease was not the same in all patients;
et al., 2006; Elder, 1973; Ferreira & Sherman, 2007; Gignac, Cott,
for example, a combination of hand and spine or of knee and hand.
& Badley, 2002; Hampson & Glasgow, 1996; Hampson, Glasgow,
The frequency of all combinations was described in Riyazi et al.
& Zeiss, 1994; Hill, Dziedzic, Thomas, Baker, & Croft, 2007;
(2005). More patients had involvement of hands (about 70%) than
Hudak et al., 2002; Orbell, Johnston, Rowley, Espley, & Davey,
knees (approximately 30%) and hips (approximately 25%), but all
1998; Peat, Greig, Wood, Wilkie, Thomas, & Croft, 2005; Toye,
patients had generalized OA.
Barlow, Wright, & Lamb, 2006). These studies corroborate the
Symptomatic OA in the knee and hip was defined with the
CSM by demonstrating that OA patients’ illness perceptions are
American College of Rheumatology (ACR) criteria for knee and
associated with limitations in daily activities, well-being, health
hip OA (Altman et al., 1991). Knee OA was defined as pain or
status, and quality of life. A pattern emerged across these various
stiffness on most days of the previous month and osteophytes at
studies to indicate that more negative perceptions of OA were
joint margins of the tibiofemoral joints. Hip OA was defined as
associated with more functional disability. However, these studies pain or stiffness in the groin and hip region on most days of the
shared the limitation of being cross-sectional, precluding infer- previous month in addition to femoral or acetabular osteophytes of
ences about causes and effects. joint space narrowing on radiograph. Symptomatic hand OA was
In the recent Genetics, ARthritis and Progression (GARP) study defined according to the ACR criteria (Altman et al., 1990) as pain
(Riyazi et al., 2005) illness perceptions were assessed at entry and or stiffness on most days of the previous month in addition to three
6 years later. The aim of the GARP cohort study is to identify of the following criteria: bony swelling of 2 or more of the 10
determinants of OA susceptibility and progression (Riyazi et al., selected joints (bilateral distal interphalangeal joints 2 ⫹ 3, bilat-
2005). Given the longitudinal design of the GARP study and the eral proximal interphalangeal joints 2 ⫹ 3, and carpometacarpal 1
detailed and objective assessments of biomedical and clinical joint), bony swelling of 2 or more distal joints, fewer than three
characteristics, this study allowed examination of the association swollen metacarpal joints, and deformity of at least one of the 10
between changes in illness perceptions and changes in functional selected joints. Symptomatic OA of the spine was defined as pain
status over an extended follow-up period, controlling for various or stiffness in the spine on most days of the previous month in
indicators of health status. Although OA is a chronic condition, addition to a Kellgren–Lawrence score of 2 in at least one disk or
treatment and self-management activities can prevent further de- one apophyseal joint.
cline in, or even improve, functional status. Over a 6-year follow- Of the 384 patients evaluated at baseline (August 2000 –March
up, there is ample opportunity for illness perceptions to change in 2003), 317 (82.6%) gave informed consent to participate. Of the
response to changes in health status and for health status to change eligible patients, 241 completed the IPQ-R at baseline and
in response to coping activities prompted by illness perceptions. In follow-up (April 2007–May 2008). The mean follow-up time was
furtherance of Leventhal et al.’s (2003) and Clatworthy et al.’s 6.0 years (SD ⫽ 0.4; Riyazi, Rosendaal, Slagboom, Kroon, Breed-
(2007) work, we hypothesized that a group of patients sharing veld, & Kloppenburg, 2008).
similar positive changes in illness perceptions would have reduc- Measures. Sociodemographic characteristics (e.g., age, gen-
tions in functional impairments, whereas the patients with negative der, marital status, body mass index [BMI], education) were col-
changes in illness perceptions would have a greater degree of lected at baseline. Three biomedical measures were used to assess
functional impairment. severity of OA: The Australian/Canadian Osteoarthritis Hand In-
58 KAPTEIN ET AL.

dex (AUSCAN) assesses hand pain, stiffness, and function by method as researched and advised for research in illness percep-
self-report (Bellamy et al., 2002); the Kellgren–Lawrence scale is tions by Clatworthy et al. (2007). All change scores were stan-
a measure of radiologically assessed degree of OA (Kellgren, dardized to z scores before clustering. Ward’s clustering method
1963); and the Western Ontario and McMaster Universities Os- was conducted to determine the centroids and number of groups,
teoarthritis Index (WOMAC) assesses lower extremity pain, stiff- followed by K-means analysis. Squared Euclidian distance was
ness, and function in OA of the knee or hip by self-report (Bel- selected as the similarity measure, and the cluster centroids and
lamy, Buchanan, Goldsmith, Campbell, & Stitt, 1988). Pain numbers of clusters determined by Ward’s method were used for
intensity was assessed during a physical examination in response to the K-means analysis. The dendrogram and agglomeration sched-
lateral pressure or passive movement of the joint, (0 ⫽ no pain, 1 ⫽ ule of the initial Ward’s clustering method suggested that it would
complaining of pain, 2 ⫽ complaining of pain and wincing, 3 ⫽ be appropriate to set the K-means clustering solution to produce
complaining of pain and withdrawal of the joint) in the hands, two clusters.
knees, hips, and spine, and on a dichotomous scale (0 ⫽ no pain, Independent t tests were used to investigate differences in
1 ⫽ pain) in the acromioclavicular joints, sternoclavicular joints, IPQ-R change scores between both cluster groups.
elbows, ankles, and metatarsal phalangeal joints. This pain inten- We performed three repeated measures analyses of covari-
sity score (range ⫽ 0 –145) is a modification of the articular index ance (ANCOVAs) to test the effects of cluster group on changes
for the assessment of OA described by Doyle, Dieppe, Scott, and in pain intensity, AUSCAN score, and WOMAC score. The
Huskisson (1981). factors in these analyses were cluster group (Cluster 1: patients
We assessed CSMs of OA using the IPQ-R (Moss-Morris et al., identified as having more negative illness perceptions over
2002). In the instructions, patients were asked to answer the time; and Cluster 2: patients identified as having more positive
questions with regard to their OA, as suggested by the designers of illness perceptions over time), time (baseline and 6-year follow-
the IPQ-R (for details, see the IPQ-R at http://www.uib.no/ipq/ up), and potentially confounding variables entered as covari-
index.html). The IPQ-R measures illness perceptions, emotional ates: age, gender, BMI, Kellgren–Lawrence score at baseline,
representations, and perceived causes, and it assesses patients’ and, additionally, pain intensity (at baseline and at 6 years) for
beliefs about (1) the identity of the disease (labels and symptoms the dependent variables AUSCAN and WOMAC. The reported
describing the illness [14 items]; in the instruction, “illness” was values for the strength of the associations between independent
substituted with “osteoarthritis”); (2) whether the timeline is acute and dependent variables in the MANOVAs and ANCOVAs are
or chronic (6 items); (3) the consequences of the disease (the partial etas squared (␩2).
severity of the illness and the impact of the disease on life in
general, self-image, finance, and family members [6 items]); (4)
the degree of personal control over OA (6 items); (5) the extent to Results
which treatment controls or cures the disease (5 items); (6) illness
coherence (the degree to which patients believe they understand Sample
their illness, 5 items), (7) the cyclical nature of the disease (the
likely variability of the disease and/or symptoms, 4 items), and (8) At the time of the present study, 241 patients completed the
the emotional representation of the disease (negative emotions IPQ-R, AUSCAN, and WOMAC at baseline and follow-up. Pa-
experienced due to OA, 6 items). The Causes subscale assesses the tient baseline characteristics are shown in Table 1. The majority of
degree to which the patient attributes the cause of the disease to participants were older women, with a BMI at the lower end of
psychological factors, risk, immune function, and accident or overweight, representing a range of educational achievement.
chance. As in the Identity scale, in the fragment “Causes of my
illness,” “osteoarthritis” replaced “illness.” All items were rated on
5-point Likert-type scales ranging from strongly disagree (1) to
strongly agree (5). Items were coded so that high scores represent Table 1
strong beliefs on these particular dimensions. Higher scores indi- Patients’ Baseline Demographic Characteristics
cate a stronger belief that the experienced symptoms are part of the Demographic Mean, frequency, or %
patient’s illness, in the chronicity of OA, in serious negative
consequences of OA, in the patient’s own ability to control symp- N 241
toms, in the effectiveness of treatment for controlling OA, in the Age, M, (and SD) 59.0 (7.5)
Gender (% female) 82.2
coherence of OA, in the cyclical nature of OA, and a stronger Marital status
negative emotional response to OA. Married/living together 186
Statistical analysis. Two repeated measures of multivariate Single 55
analyses of variance (MANOVAs) were conducted to compare BMI, M, (and SD) 26.8 (4.7)
IPQ-R scores and disease progression at baseline with scores at Education
Elementary school 27
follow-up. Cluster analysis was used to classify patients into Junior high school 76
subgroups according to their change in illness perceptions from High school 85
baseline to 6-year follow-up. Simple change scores (follow-up College/university 53
score minus baseline score) of the illness perceptions dimensions M (and SD) for Kellgren–Lawrence score 43.9 (20.0)
Range 0–180
“identity,” “timeline chronic,” “timeline cyclical,” “conse-
quences,” “personal control,” “treatment control,” and “emotional Note. BMI ⫽ body mass index. Kellgren–Lawrence is a measure of
representations” were used to perform the two-stage clustering radiographically defined degree of osteoarthritis severity.
 EXAMINING OSTEOARTHRITIS CHANGE 59

Mean scores on the IPQ-R dimensions, AUSCAN, WOMAC, change in illness perceptions. Increases in identity; chronic time-
and physician-reported pain intensity at baseline and at follow-up line; consequences; and decreases in personal control, treatment
are presented in Table 2. control, and emotional representations (cluster Group 1) describe
Change on IPQ-R dimensions and disease progression. We an illness model that becomes more negative over time (Clatwor-
conducted a repeated measures MANOVA to investigate differ- thy et al., 2007; Hagger & Orbell, 2003; Leventhal et al., 2003).
ences over time in scores on the IPQ-R dimensions. All dimen- Decreases in identity, chronic timeline, consequences, emotional
sions and the perceived causes were entered as dependent vari- representations, and increases in personal control and treatment
ables. There was a statistically significant difference over time on control (Cluster 2), represent an illness model that can be defined
the combined dependent variables, F(12, 224) ⫽ 3.66, p ⬍ .01, as positive. Both clusters had negative change scores on emotional
Wilks’s ␭ ⫽ 0.84, multivariate ␩2 ⫽ .16. When the results for the representations, indicating a tendency for both to get less negative
dependent variables were considered separately, five IPQ-R di- over time. However, the positive cluster became significantly less
mensions differed significantly between baseline and follow-up. negative than the negative cluster, which is consistent with the
For the entire sample, beliefs changed to a significantly more theoretical model (Clatworthy et al., 2007; Hagger & Orbell, 2003;
chronic timeline, F(1, 235) ⫽ 8.28, p ⫽ .004, ␩2 ⫽ .03; less Leventhal et al., 2003).
personal control over the illness, F(1.235) ⫽ 8.69, p ⫽ .004, ␩2 ⫽ Differences between cluster groups on functional status.
.04; increased sense of coherence, F(1, 235) ⫽ 10.72, p ⫽ .001, Pain intensity. A 2 (time) ⫻ 2 (cluster group) mixed-model
␩2 ⫽ .04; a reduction in the belief in OA as cyclical, F(1, 235) ⫽ ANCOVA revealed that the main effects for cluster group, F(1,
4.91, p ⫽ .028, ␩2 ⫽ .02; and a less strong negative emotional 203) ⫽ 1.39, p ⬎ .05, ␩2 ⫽ .01; and time, F(1, 203) ⫽ 2.80, p ⬎
response to OA (i.e., more positive), F(1, 235) ⫽ 11.58, p ⫽ .001, .05, ␩2 ⫽ .01; were not significant (see Figure 1). Thus, there were
␩2 ⫽ .05. No significant differences between baseline and no overall differences in the pain intensity scores of the negative
follow-up were found on the other IPQ-R dimensions or on the cluster group (M ⫽ 8.54), compared with the positive cluster group
IPQ-R questions that explore perceived causes of OA. (M ⫽ 10.01). Pain intensity scores at follow-up (M ⫽ 10.76) were
A repeated measures MANOVA was also conducted to inves- not significantly higher than at baseline (M ⫽ 7.80). Of the
tigate differences over time in disease progression. AUSCAN, potentially confounding variables (age, gender, BMI, Kellgren–
WOMAC, and pain intensity scores were entered as dependent Lawrence score), only the Time ⫻ Gender interaction was signif-
variables. There was a statically significant difference over time on icant, F(1, 203) ⫽ 3.90, p ⬍ .05, ␩2 ⫽ 0.02; suggesting a sharper
the combined dependent variables, F(3, 206) ⫽ 11.41, p ⬍ .001, rise in pain intensity for females across both groups.
Wilks’s ␭ ⫽ 0.86, multivariate ␩2 ⫽ .14. When the results for the AUSCAN. A significant Time ⫻ Cluster Group effect was
dependent variables were considered separately, scores on the obtained, F(1, 201) ⫽ 9.96, p ⬍ .01, ␩2 ⫽ .05. Examination of
AUSCAN, F(1, 208) ⫽ 10.31, p ⫽ .002, ␩2 ⫽ .05; and pain the cell means indicated that, although there was an increase in
intensity, F(1, 208) ⫽ 31.85, p ⬍ .0001, ␩2 ⫽ .13; indicated an AUSCAN scores for the negative cluster group from baseline
increased (negative) impact on daily functioning and pain. No (M ⫽ 17.65) to follow-up (M ⫽ 22.86), the positive cluster group
significant differences were observed for the sample as a whole on did not change in AUSCAN scores from baseline (M ⫽ 21.26) to
WOMAC scores. follow-up (M ⫽ 21.60; see Figure 2). At baseline, the negative
Table 3 shows the mean IPQ-R change scores for the two cluster group had significantly better AUSCAN scores than did the
subgroups of patients classified according to their profile of positive cluster group, t(238) ⫽ ⫺1.99, p ⬍ .05. Other significant

Table 2
Descriptive Statistics for Baseline and 6-Year Follow-Up Illness Perceptions and Disease Progression

Baseline Follow-up
Illness perception
dimension Range M SD M SD Fa p

Identity 0–14 5.3 2.5 5.2 2.2 0.60 .438

Timeline acute/chronic 6–30 25.4 3.7 26.2 3.4 8.28 .004
Consequences 6–30 16.8 4.6 16.5 4.6 0.87 .351
Personal control 6–30 18.8 3.5 18.0 3.8 8.69 .004
Treatment control 5–25 13.9 2.8 13.6 3.0 2.50 .115
Illness coherence 5–25 17.9 4.1 18.6 4.0 10.72 .001
Timeline cyclical 4–20 14.3 3.1 13.8 3.2 4.91 .028
Emotional representations 6–30 14.3 5.2 13.3 5.4 11.58 .001
Psychological attribution 6–30 12.7 4.3 12.4 4.4 0.69 .407
Risk attribution 7–35 17.7 3.3 18.0 3.6 1.40 .237
Immune function attribution 3–15 6.7 2.0 6.4 2.2 2.69 .102
Accident/chance attribution 2–10 4.9 1.6 4.9 1.6 0.05 .823
AUSCAN total score 0–60 19.5 14.2 22.2 14.1 10.31 .002
WOMAC total score 0–100 27.2 22.9 28.9 23.1 0.28 .598
Pain intensity 0–145 7.9 8.3 10.8 9.5 31.85 .000

Note. AUSCAN ⫽ Australian/Canadian Osteoarthritis Hand Index; WOMAC ⫽ Western Ontario and McMaster Universities Osteoarthritis Index.
a
A repeated measures multivariate analysis of variance was conducted to investigate differences over time.
60 KAPTEIN ET AL.

Table 3
Mean Differences in IPQ-R Change Scoresa Between Both Cluster Groups

Cluster 1: Illness model Cluster 2: Illness model

more negative over time more positive over time
(n ⫽ 114) (n ⫽ 126)
Illness perception
dimension M SD M SD F p

Identity 0.45 2.35 ⫺0.71 2.39 3.793 .000

Timeline acute/chronic 3.01 3.42 ⫺1.24 3.24 9.882 .000
Consequences 1.81 4.28 ⫺2.31 4.06 7.648 .000
Personal control ⫺2.76 3.30 0.99 3.46 ⫺8.582 .000
Treatment control ⫺2.19 2.70 1.38 2.59 ⫺10.436 .000
Illness coherence 0.48 3.17 0.95 3.60 ⫺1.077 .283
Timeline cyclical ⫺0.52 3.69 ⫺0.42 3.12 ⫺0.214 .831
Emotional representations ⫺0.06 4.14 ⫺1.91 4.96 3.113 .002

Note. IPQ-R ⫽ Illness Perception Questionnaire—Revised.

a
Simple change scores ⫽ follow-up score ⫺ baseline score.

effects emerged for Kellgren–Lawrence scores, F(1, 201) ⫽ 8.74, Other significant effects emerged for BMI, F(1, 200) ⫽ 32.89,
p ⬍ .01, ␩2 ⫽ .04; for baseline pain scores, F(1, 201) ⫽ 19.17, p ⬍ p ⬍ .001, ␩2 ⫽ .14; for baseline pain scores, F(1, 200) ⫽ 8.22, p ⬍
.001, ␩2 ⫽ .09; and for follow-up pain scores, F(1, 201) ⫽ 41.16, .01, ␩2 ⫽ 0.04; and for follow-up pain scores, F(1, 200) ⫽ 37.44,
p ⬍ .001, ␩2 ⫽ .17; showing more negative AUSCAN scores p ⬍ .001, ␩2 ⫽ .16; showing more negative WOMAC scores
across both time points for patients with higher Kellgren– across both time points for patients with higher BMI scores and
Lawrence scores and higher pain intensity scores. higher pain intensity scores.
WOMAC. A significant Time ⫻ Cluster group effect was Although the two patient clusters were not significantly associ-
obtained, F(1, 200) ⫽ 9.43, p ⬍ .01, ␩2 ⫽ .05. Examination of the ated with changes over time in physician-reported pain intensity,
cell means indicated that, although there was an increase in they were associated with modest but meaningful changes at
WOMAC scores for the negative cluster group from baseline (M ⫽ follow-up in AUSCAN and WOMAC scores. As hypothesized, the
25.51) to follow-up (M ⫽ 31.42), the positive cluster group did cluster with a more positive illness model was associated with
slightly improve in WOMAC scores from baseline (M ⫽ 28.97) to better outcomes, and the cluster with a more negative illness model
follow-up (M ⫽ 26.85; see Figure 3). At baseline, the negative was associated with poorer outcomes on the two functional im-
cluster group had slightly (nonsignificant) better WOMAC scores pairment scales, AUSCAN and WOMAC. These results corrobo-
than did the positive cluster group. rate the validity of the two-cluster solution for the IPQ-R dimen-

Figure 1. Change in pain intensity from baseline to 6-year follow-up for the two cluster groups.
 EXAMINING OSTEOARTHRITIS CHANGE 61

Figure 2. Change in Australian/Canadian Osteoarthritis Hand Index (AUSCAN) score from baseline to 6-year
follow-up for the two cluster groups.

sions presented here and suggest that these clusters may be long-term follow-up is indicative of a reciprocal process between
associated with clinically meaningful changes in functional im- illness representations and illness outcomes as proposed by the
pairment. CSM (Leventhal et al., 2003). The present findings for OA are
comparable with those of previous studies of low back pain (Foster
Discussion et al., 2008), angina (Furze et al., 2005), venous thrombosis
(Kaptein, van Korlaar, Cameron, Vossen, van der Meer, &
The results of this prospective study with a 6-year follow-up Rosendaal, 2007), and diabetes (Skinner et al., 2006). Together,
add to the limited number of empirical studies in which longi- these results have important clinical implications. They suggest
tudinal changes in IPQ-R dimensions were examined. They that identifying illness dimensions on which patients hold beliefs
advanced our knowledge of changes in CSMs of OA over time, indicative of poor outcomes and intervening to change these be-
suggesting which IPQ-R dimensions remain stable and which liefs may have beneficial effects on the course of a chronic disease
ones change. For OA, it appears that attributions of causality (Clatworthy et al., 2007; Hagger & Orbell, 2003; Newman et al.,
remain relatively unaffected by the passage of time. However, 2004). As noted by Clatworthy et al. (2007):
over time, OA is increasingly perceived as a relatively chronic
condition, as less cyclical, and as less amenable to personal [A]s the focus of illness perception research moves toward interven-
control, independent of objectively assessed illness severity. tion development, there is a further practical advantage to grouping
Moreover, the identification of two patient clusters, each with people in this way. Groups of people with schemata associated with
poor coping or outcome would be ideal targets for interventions. The
similar change profiles across the dimensions of illness percep-
cluster analysis would not only identify these groups but would also
tions as recommended by Clatworthy et al. (2007), yielded
provide information on the types of beliefs held by the groups that
additional meaningful associations between change in illness may need to be addressed in an intervention. (p. 126)
perceptions and change in functional status. Consistent with the
conclusions from Hagger and Orbell’s (2003) meta-analysis of Strengths of the present study include the comparatively
illness perceptions, a deterioration in functional abilities over large sample size compared with previous research on OA
time was associated with a pattern of change on illness percep- illness perceptions, the unusually long follow-up period, and
tions associated with poor outcomes: more passive and chronic the relatively low level of subject attrition. The present sample
views, perceiving less control, and experiencing a higher emo- was comparable with the samples of OA patients in the studies
tional load regarding the illness. mentioned in the Introduction with regard to sociodemographic
Demonstrating that change to a more negative illness represen- and other medical characteristics. The measure of the illness
tation is associated with deterioration of functional status across perceptions used here reflected the same theoretical base (the
62 KAPTEIN ET AL.

Figure 3. Change in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score from
baseline to 6-year follow-up for the two cluster groups.

CSM of Leventhal et al., 2003) as many of these studies. Such trol their OA and the effectiveness of their medical treatment
comparability increases the external validity and, hence, the and reduce perceived symptoms and the perceived physical,
generalizability of our findings. social, and emotional consequences of the disease— could re-
Limitations include the absence of a measure of functional sult in less self-reported functional disability. Future research
status that was not based on self-report. However, the AUSCAN on patients with OA should focus on identifying more precisely
and WOMAC are widely used to assess the impact of OA in daily which patterns of illness perceptions are associated with more
life and are considered the gold standard in research on OA specific outcome measures and on developing interventions
patients. Moreover, unlike in many previous studies, pain intensity designed to change these patterns of beliefs.
was measured objectively and controlled for in all analyses. As-
sessment of change on both illness perceptions and functional References
status at one or more times during the follow-up period could have
yielded even more interesting results, enabling the examination of Altman, R., Alarcon, G., Appelrouth, D., Bloch, D., Borenstein, D., Brandt,
correlated change across time and investigated cross-lagged cor- K., Brown, C., et al. (1990). The American College of Rheumatology
criteria for the classification and reporting of osteoarthritis of the hand.
relations. Multiple assessments are recommended for future stud-
Arthritis and Rheumatism, 33, 1601–1610.
ies. Altman, R., Alarcon, G., Appelrouth, D., Bloch, D., Borenstein, D., Brandt,
The potential of interventions to change illness perceptions K., Brown, C., et al. (1991). The American College of Rheumatology
and examine effects thereof on disease outcomes is only just criteria for the classification and reporting of osteoarthritis of the hip.
beginning to be recognized (Cameron & Jago, 2008). Only a Arthritis and Rheumatism, 34, 505–514.
few intervention studies have been published up to now (e.g., Appelt, C. J., Burant, C. J., Siminoff, L. A., Kwoh, C. K., & Ibrahim, S. A.
Foster et al., 2008; Furze et al., 2005; Goodman, Morrissey, (2007). Arthritis-specific health beliefs related to aging among older
Graham, & Bossingham, 2005; Hall, Weinman, & Marteau, male patients with knee and/or hip osteoarthritis. Journals of Gerontol-
2004; Humphries & Ozakinci, 2008; Karamanidou, Weinman, & ogy A, Biological Science and Medical Sciences, 62, 184 –190.
Horne, 2008; McAndrew et al., 2008; Petrie, Cameron, Ellis, Ballantyne, P. J., Gignac, M. A. M., & Hawker, G. A. (2007). A patient-
centered perspective on surgery avoidance for hip or knee arthritis:
Buick, & Weinman, 2002). Theoretical and conceptual issues in
Lessons for the future. Arthritis and Rheumatism, 57, 27–34.
designing interventions in the context of the CSM are discussed Bellamy, N., Buchanan, W. W., Goldsmith, C. H., Campbell, J., & Stitt,
by Deary (2008) and by Wearden and Peters (2008). The L. W. (1988). Validation study of the WOMAC: A health status instru-
present study suggests that interventions that increase patients’ ment for measuring clinically important patient relevant outcomes to
pattern of positive beliefs, especially the control components in antirheumatic drug therapy in patients with osteoarthritis of the hip or
illness perceptions—that is, increase perceived ability to con- knee. Rheumatology, 15, 1833–1840.
 EXAMINING OSTEOARTHRITIS CHANGE 63

Bellamy, N., Campbell, J., Haroui, B., Buchbinder, R., Hobby, K., Roth, Kreder, H. J., & Wright, J. G. (2002). “You’re perfect for the procedure!
J. H., & MacDermid, J. C. (2002). Dimensionality and clinical impor- Why don’t you want it?” Elderly arthritis patients’ unwillingness to
tance of pain and disability in hand osteoarthritis: Development of the consider total joint arthroplasty surgery: A qualitative study. Medical
Australian/Canadian (AUSCAN) Osteoarthritis Hand Index. Osteoar- Decision Making, 22, 272–278.
thritis and Cartilage, 10, 855– 862. Humphries, G., & Ozakinci, G. (2008). The AFTER intervention: A
Botha-Scheepers, S., Riyazi, N., Kroon, H. M., Scharloo, M., Houwing- structured psychological approach to reduce fears of recurrence in pa-
Duistermaat, J. J., Slagboom, E., Rosendaal, F. R., et al. (2006). Activity tients with head and neck cancer. British Journal of Health Psychology,
limitations in the lower extremities in patients with osteoarthritis: The 13, 223–230.
modifying effects of illness perceptions and mental health. Osteoarthri- Kaptein, A. A., van Korlaar, I. M., Cameron, L. D., Vossen, C. Y., van der
tis and Cartilage, 14, 1104 –1110. Meer, F. J. M., & Rosendaal, F. R. (2007). Using the common-sense
Cameron, L. D., & Jago, L. (2008). Emotion regulation interventions: A model to predict risk perception and disease-related worry in individuals
common-sense approach. British Journal of Health Psychology, 13, at increased risk for venous thrombosis. Health Psychology, 26, 807–
215–221. 812.
Clatworthy, J., Hankins, M., Buick, D., Weinman, J., & Horne, R. (2007). Karamanidou, C., Weinman, J., & Horne, R. (2008). Improving haemodi-
Cluster analysis in illness perception research: A Monte Carlo study to alysis patients’ understanding of phosphate-binding medication: A pilot
identify the most appropriate method. Psychology & Health, 22, 123– study of a psycho-educational intervention designed to change patients’
142. perceptions of the problem and treatment. British Journal of Health
Deary, V. (2008). A precarious balance: Using a self-regulation model to Psychology, 13, 205–214.
conceptualize and treat chronic fatigue syndrome. British Journal of Kellgren, J. H. (1963). Atlas of standard radiographs of arthritis. In J. H.
Health Psychology, 13, 231–236. Kellgren (Ed.), The epidemiology of chronic rheumatism (Vol. 2, pp.
Dieppe, P. A., & Lohmander, L. S. (2005). Pathogenesis and management 1–13). Oxford, England: Blackwell Scientific.
of pain in osteoarthritis. Lancet, 365, 965–973. Kellgren, J. H., & Moore, R. (1952). Generalized osteoarthritis and Heb-
Doyle, D. V., Dieppe, P. A., Scott, J., & Huskisson, E. C. (1981). An erden’s nodes. British Medical Journal, 1952, 1, 181–187.
articular index for the assessment of osteoarthritis. Annals of Rheumatic Kratz, A. L., Davis, M. C., & Zautra, A. J. (2007). Pain acceptance
Diseases, 40, 75–78. moderates the relation between pain and negative affect in female
Elder, R. G. (1973). Social class and lay explanations of the etiology of
osteoarthritis and fibromyalgic patients. Annals of Behavioral Medicine,
arthritis. Journal of Health & Social Behavior, 14, 28 –38.
33, 291–301.
Ferreira, V. M., & Sherman, A. M. (2007). The relationship of optimism,
Lawrence, J. S. (1963). Generalized osteoarthritis in a population sample.
pain and social support to well-being in older adults with osteoarthritis.
American Journal of Epidemiology, 90, 381–389.
Aging and Mental Health, 11, 89 –98.
Leventhal, H., Brissette, I., & Leventhal, E. A. (2003). The common-sense
Foster, N. E., Bishop, A., Thomas, E., Main, C., Horne, R., Weinman, J.,
model of self-regulation of health and illness. In L. D. Cameron & H.
& Hay, E. (2008). Illness perceptions of low back pain patients in
Leventhal (Eds.), The self-regulation of health and illness behaviour (pp.
primary care: What are they, do they change and are they associated with
42– 65). London: Routledge.
outcome? Pain, 136, 177–187.
Leventhal, H., Weinman, J., Leventhal, E. A., & Phillips, L. A. (2008).
Furze, G., Lewin, R. J. P., Murberg, T., Bull, P., & Thompson, D. R.
Health psychology: The search for pathways between behavior and
(2005). Does it matter what patients think? The relationship between
health. Annual Review of Psychology, 59, 1–29.
changes in patients’ beliefs about angina and their psychological and
McAndrew, L. M., Musumeci-Szabó, T. J., Mora, P. A., Vileikyte, L.,
functional status. Journal of Psychosomatic Research, 59, 323–329.
Burns, E., Halm, E. A., Leventhal, E. A., et al. (2008). Using the
Gignac, M. A. M., Cott, C., & Badley, E. M. (2002). Adaptation to
disability: Applying selective optimization with compensation to the common sense model to design interventions for the prevention and
behaviors of older adults with osteoarthritis. Psychology and Aging, 17, management of chronic illness threats: From description to process.
520 –524. British Journal of Health Psychology, 13, 195–204.
Goodman, D., Morrissey, S., Graham, D., & Bossingham, D. (2005). The Moss-Morris, R., Weinman, J., Petrie, K. J., Horne, R., Cameron, L. D., &
application of cognitive-behaviour therapy in altering illness represen- Buick, D. (2002). The Revised Illness Perception Questionnaire (IPQ-
tations of systemic lupus erythematosus. Behaviour Change, 22, 156 – R). Psychology & Health, 17, 1–16.
171. Newman, S., Steed, L., & Mulligan, K. (2004). Self-management inter-
Hagger, M. S., & Orbell, S. (2003). A meta-analytic review of the ventions for chronic illness. Lancet, 364, 1523–1537.
common-sense model of illness representations. Psychology & Health, Orbell, S., Johnston, M., Rowley, D., Espley, A., & Davey, P. (1998).
18, 141–184. Cognitive representations of illness and functional and affective adjust-
Hall, S., Weinman, J., & Marteau, T. M. (2004). The motivating impact of ment following surgery for osteoarthritis. Social Science & Medicine,
informing women smokers of a link between smoking and cervical 47, 93–102.
cancer: The role of coherence. Health Psychology, 23, 419 – 424. Peat, G., Greig, J., Wood, L., Wilkie, R., Thomas, E., & Croft, P. (2005).
Hampson, S. E., & Glasgow, R. E. (1996). Dimensional complexity of Diagnostic discordance: We cannot agree when to call knee pain “os-
older patients’ illness representations of arthritis and diabetes. Basic and teoarthritis.” Family Practice, 22, 96 –102.
Applied Social Psychology, 18, 45–59. Petrie, K. P., Cameron, L. D., Ellis, C. J., Buick, D., & Weinman, J. (2002).
Hampson, S. E., Glasgow, R. E., & Zeiss, A. M. (1994). Personal models Changing illness perceptions after myocardial infarction: An early in-
of osteoarthritis and their relation to self-management activities and tervention randomized controlled trial. Psychosomatic Medicine, 64,
quality of life. Journal of Behavioral Medicine, 17, 143–158. 580 –586.
Hill, S., Dziedzic, K., Thomas, E., Baker, S. R., & Croft, P. (2007). The Riyazi, N., Meulenbelt, I., Kroon, H. M., Ronday, K. H., Hellio le Grav-
illness perceptions associated with health and behavioural outcomes in erand, M. P., Rosendaal, F. R., Breedveld, F. C., et al. (2005). Evidence
people with musculoskeletal hand problems: Findings from the North for familial aggregation of hand, hip, and spine but not knee osteoar-
Staffordshire Osteoarthritis Project (NorStOP). Rheumatology, 46, 944 – thritis in siblings with multiple joint involvement: The GARP study.
951. Annals of Rheumatic Diseases, 64, 438 – 443.
Hudak, P. L., Clark, J. P., Hawker, G. A., Coyte, P. C., Mahomed, N. N., Riyazi, N., Rosendaal, F. R., Slagboom, E., Kroon, H. M., Breedveld,
64 KAPTEIN ET AL.

F. C., & Kloppenburg, M. (2008). Risk factors in familial osteoarthritis: and impact are greater among U.S. women than men: Intervention
The GARP sibling study. Osteoarthritis and Cartilage, 16, 654 – 659. opportunities. Journal of Women’s Health, 16, 441– 453.
Skinner, T. C., Carey, M. E., Cradock, S., Daly, H., Davies, M. J., Toye, F. M., Barlow, J., Wright, C., & Lamb, S. E. (2006). Personal
Doherty, Y., Heller, K., et al. (2006). Diabetes education and self- meanings in the construction of need for total knee replacement surgery.
management for ongoing and newly diagnosed (DESMOND): Pro- Social Science & Medicine, 63, 43–53.
cess modeling of pilot study. Patient Education and Counseling, 64, Wearden, A., & Peters, S. (2008). Therapeutic techniques for interventions
369 –377. based on Leventhal’s Common Sense Model. British Journal of Health
Theis, K. A., Helmick, C. G., & Hootman, J. M. (2007). Arthritis burden Psychology, 13, 189 –193.

Call for Nominations:

Sport, Exercise, and Performance Psychology
The Publications and Communications (P&C) Board of the American Psychological Association
and Division 47 (Exercise and Sport Psychology) of the APA have opened nominations for the
editorship of Sport, Exercise, and Performance Psychology for the years 2011–2016. The editor
search committee is co-chaired by Ed Acevedo, PhD, and Robert Frank, PhD.
Sport, Exercise, and Performance Psychology, to begin publishing in 2011, will publishes
papers in all areas of sport, exercise, and performance psychology for applied scientists and
practitioners. This journal is committed to publishing evidence that supports the application of
psychological principals to facilitate peak sport performance, enhance physical activity partici-
pation, and achieve optimal human performance. Published papers include experimental studies,
qualitative research, correlational studies, and evaluation studies. In addition, historical papers,
critical reviews, case studies, brief reports, critical evaluations of policies and procedures, and
position statements will be considered for publication.
Editorial candidates should be available to start receiving manuscripts in July 2010 to prepare for
issues published in 2011. Please note that the P&C Board encourages participation by members of
underrepresented groups in the publication process and would particularly welcome such nominees.
Self-nominations are also encouraged.
Candidates should be nominated by accessing APA’s EditorQuest site on the Web. Using your
Web browser, go to http://editorquest.apa.org. On the Home menu on the left, find “Guests.” Next,
click on the link “Submit a Nomination,” enter your nominee’s information, and click “Submit.”
Prepared statements of one page or less in support of a nominee can also be submitted by e-mail to
MollyDouglas-Fujimoto,ManagingDirector,EducationalPublishingFoundation,atmdouglas-fujimoto@
apa.org.
The deadline for accepting nominations is January 31, 2010, when reviews will begin.