Journal of Psychosomatic Research 67 (2009) 417–423

Illness beliefs predict disability in rheumatoid arthritis

Helen Gravesa,⁎, David L. Scotta, Heidi Lemppb, John Weinmanc

aAcademic Department of Rheumatology, King's College London School of Medicine at Guy's, King's College, St Thomas Hospitals, SE5 9RJ London, UKbNIHR BRC Guy's and St. Thomas' Hospital Foundation Trust, Academic Rheumatology, King's College London School of Medicine, SE5 9RJ London, UK

cHealth Psychology Section, Psychology Department at Guy's, Institute of Psychiatry, London Bridge, SE1 9RT London, UK

Received 12 March 2008; received in revised form 5 January 2009; accepted 8 January 2009

Abstract

Objective: This study examined the relationships betweenpatients' beliefs about rheumatoid arthritis (using Leventhal'sCommon Sense Model) and their levels of disability, health-relatedquality of life, and disease activity. A proposed illness beliefsconstruct of “seriousness” was also investigated, combining beliefsabout illness identity (symptoms), consequences, and timeline.Method: A cross-sectional study evaluated 125 patients withrheumatoid arthritis from two South East London hospitals.Questionnaires assessed their illness beliefs, disability, and qualityof life. An objective measure of disease activity was also obtained.Results: Higher disability scores were associated with beliefs aboutidentity (r=.31, Pb.01) and consequences (r=.28, Pb.01). Stronger

⁎ Corresponding author. Academic Department of Rheumatology,Weston Education Centre, Cutcombe Road, SE5 9RJ London, UK.Tel.: +44 0207 848 5604; fax: +44 0207 848 5202.

E-mail address: helen.2.graves@kcl.ac.uk (H. Graves).

0022-3999/09/$ – see front matter © 2009 Elsevier Inc. All rights reserved.doi:10.1016/j.jpsychores.2009.01.006

control beliefs were associated with lower disability (r=−.40, Pb.01)and better physical (r=.20, Pb.05) quality of life. Disease activityscores, although positively related to disability scores (r=.39,Pb.01), showed no associations with illness beliefs. Multivariateanalysis resulted inmodels accounting for 45.5%, 27.3%, and 19.3%variance in disability, “physical quality of life” scores, and “mentalquality of life” scores, respectively. The hypothesis for a proposed“seriousness” construct was not supported. Conclusion: Patients'beliefs about their rheumatoid arthritis are associated with disabilityand quality of life and cannot be explained by disease status.Longitudinal research is needed to assess the stability of beliefs.© 2009 Elsevier Inc. All rights reserved.

Keywords: Disability; Illness beliefs; Quality of life; Rheumatoid arthritis; Self-regulation; Common Sense Model

Introduction

A purely medical model is insufficient to explainadjustment to a chronic illness such as rheumatoid arthritis(RA) and a biopsychosocial framework, incorporatingpatients' beliefs, may be more appropriate [1–3]. There isno cure for RA and treatments are only partially effective [4],so exploring individual differences that mediate betweenRA, adjustment and subsequent quality of life (QoL) is apriority. The uncertain course of the illness involves loss offunction, considerable discomfort, and possible disabilityand deformity, placing a significant burden upon theindividual and their family [5]. This, in turn may impact

on perceived QoL, commonly used as a measure ofadjustment to chronic disease.

During recent years, it has become clearer that patients'response to illness and their QoL are strongly influenced bythe way in which they perceive their condition [6,7]. Illnessperceptions are a core part of Leventhal's Common SenseModel (CSM) of illness behavior [8,9], which postulates thatthese beliefs lie along five dimensions: identity (label of thedisease and symptoms), timeline (perceived illness duration—acute chronic or cyclical), cure/control (beliefs abouttreatment or potential for cure), consequences (likelyrepercussions of living with the illness), and cause (etiology).These illness beliefs are a framework for interpretation of theillness threat; they guide future illness-related behaviors anddetermine coping, appraisal of outcomes, and adjustment.

Illness beliefs have been associated with physical andpsychological functioning in a range of chronic conditions[10–16]. There is a scarcity, however, of studies directly

418 H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

applying the CSM to the measurement of beliefs in RApatients. Most research has focused on patients withosteoarthritis, and this, together with existing data from RAsamples, shows that beliefs (particularly perceived conse-quences) play a significant role in adjustment. Researchsuggests that perceiving serious negative consequences iscommon among patients [17–19], and studies show associa-tions between these beliefs and aspects of well-beingincluding life satisfaction, physical symptoms, and depres-sion [20]. Strong beliefs in negative consequences have alsoemerged to predict poor physical functioning [21] and lowmood [22,23].

Patients' perceptions of personal control have also shownassociations with physical and psychological functioning[23–27], which suggests that perceiving more control maybe more adaptive. Stronger control beliefs correlate nega-tively with objective measures of disease severity such aserythrocyte sedimentation rate (ESR), grip strength, andlength of morning stiffness [25] and explain substantialamounts of variance in physical functioning, over and abovethat accounted for by medical variables [28]. In addition,patients' beliefs about their RA have been related to outcomewhen clinical measures have not. For example, one study[29] found that patients with stronger beliefs in a chronictimeline and negative consequences (perceiving RA as moreserious) experienced more disability, psychiatric morbidity,and pain and those with lower perceived control experiencedless. No relationships were found between these outcomesand objective disease measures such as ESR, surgery, ornumber of present second-line medications.

Although previous studies have highlighted the impor-tance of investigating beliefs of patients living with arthriticconditions, most have measured beliefs of patients withosteoarthritis; a common, degenerative condition of later life,which does not usually require specialist care [30]. RA onthe other hand, is a severe, prolonged, systemic illnesscausing significant disability, with intensive treatments [31].Individuals living with RA are therefore likely to hold verydifferent illness beliefs.

More data about these beliefs are needed in order todevelop a broader, biopsychosocial approach to meetcomplex, long-term needs [32]. It has been suggested thatbeliefs about consequences, together with symptom attribu-tions (number of symptoms ascribed to the illness) andperceived timeline, may form an intensity, or “seriousness”construct, which may play an important role in self-manage-ment [18,33–36]. Of the studies that have focused on beliefsamong RA patients [19–21,23,28], none have empiricallyinvestigated the proposed construct of seriousness.

We aimed to empirically examine the importance ofperceived seriousness as a higher-order construct (not areduction in the number of existing dimensions) and toconfirm the importance of beliefs about consequences andcontrol in subsequent adjustment. Studies focusing on otherchronic illness groups have been used to inform interven-tions, which have shown promising results [37,38].

We therefore predicted that disability (ability to performaspects of daily living) would be positively related to beliefsin the seriousness of RA (higher identity, timeline, andconsequences scores) and negatively related to control beliefs(personal and treatment control). We predicted that QoL(physical and psychological functioning) would be nega-tively associated with beliefs about the seriousness of RA andpositively associated with beliefs about control. Finally, wepredicted that illness beliefs would add significantly to theproportion of variance in disability and QoL explained bydemographic and clinical factors (these include age, diseaseduration, and disease activity, which may affect disabilitylevels and QoL). As the person ages, experiences longerdisease duration, or greater disease activity, there will be anexpected decline in level of functioning).

Method

Design

This was a cross-sectional study for which patients wererecruited consecutively during attendance at outpatientclinics in two South East London hospitals. Potentialparticipants were referred to the researcher by clinic doctors,specialist nurses, or approached directly during waiting time.Ethical approval was obtained from the St. Thomas' HospitalResearch Ethics Committee.

Participants

We recruited 125 participants (King's College Hospital,n=101; University Hospital Lewisham, n=24) from 143approaches, with a participation rate of approximately 87%.Patients were eligible to take part if they had a confirmeddiagnosis of RA, were aged over 18 years, able tounderstand and speak English, and were currently prescribedmedication for RA. Patients were excluded if they had anyserious communication, learning or hearing difficulties, anysignificant comorbidity (e.g., advanced cancer), seriousmental health problems, or if they would be in anyanticipated discomfort if required to fill out a questionnaire(e.g., breathless). Clinic doctors and specialist nurses madethis judgment.

Demographic information is shown in Table 1. Themajority of the sample was female (73%), reflecting thegender distribution of RA in the general population [31].Participants varied widely in the duration of their illnessfrom b1 year to 65 years with a median of 10 years. Aroundhalf (51%) of participants were registered as disabled due tothe severity of their illness.

Method

The details of the study were explained to each patient in aprivate room, where signed consent was obtained. Participants

Table 1Demographic information

Gender Gender (female) 91Age (years) Median (IQR) 60 (47–67)Ethnic Origin (%) White British 68

White Other 12Black Caribbean 8Indian 6Black African 2Black British 2Asian other 1Declined to answer 1

Marital Status (%) Married/living as married 67Single 28Divorced/separated 18Widower/widowed 12

Work Status (%) Employed 33Retired 38Unable to work due to RA 22Other 7Registered disabled (%) 51%

Time Since Diagnosis (years) Median (IQR) 10 (4–20)

IQR, interquartile range.

419H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

then completed validated questionnaires assessing illnessrepresentations, perceived QoL, and disability. A recentobjective measure of disease severity [Disease ActivityScore 28 (DAS28)] was obtained from clinic nurses.

The Illness Perception Questionnaire—Revised, RA-specificversion [39]

The revised version of the Illness Perception Ques-tionnaire (IPQ-R) comprises nine scales designed to assessthe dimensions underlying patients' perceptions of theirillness as outlined by Leventhal [7,8]. The identity constructis assessed by asking participants which of 14 symptomsthey have experienced since their RA and asking them toidentify the symptoms they feel are related to their RA.Beliefs about timeline (two scales: acute/chronic andcyclical, scores range from 6 to 30 and 4 to 20,respectively), consequences (scores range from 6 to 30),personal control (amount of control the person feels theyhave over their illness, scores range from 6 to 30), andtreatment control (the amount of control the person feelsthat treatment will have over their illness, scores range from5 to 25) were also assessed. In addition, research hasidentified a further two constructs [39], illness coherence(the extent to which the individual understands their illness,scores range from 5 to 25) and emotional representations(the emotions associated with the illness, scores range from6 to 30) which are assessed by 38 statements using five-point, Likert-type scales (“strongly agree” to “stronglydisagree”). A summary score is obtained for each dimen-sion. A score above the midpoint on any of these scales isconsidered to represent particularly strong beliefs for thatdimension [33]. Causal beliefs are assessed by the extent towhich patients agree with each of 18 causal items. Strength

of agreement is assessed using a five-point Likert-typescale. Examples of possible causes listed are “stress orworry,” “ageing,” and “altered immunity.”

Short-Form Health Survey Questionnaire [40,41]

The 36-item Short-Form Health Survey Questionnaire(SF-36) is a widely used instrument for assessing QoLacross eight dimensions: limitations in physical activitiesdue to health, social limitations due to physical andemotional problems, limitations in usual role activities dueto physical problems, role limitations due to emotionalproblems, bodily pain, general mental health, vitality, andgeneral health perception. Subscales are summed scoresfrom two-, three-, five-, or six-point scales which areconverted to percentages; a higher score indicating betterfunctioning. In addition to summary scores for eachsubscale, two overall summary scores were obtained: onefor the physical component (assessing role functioning dueto physical problems, general physical function, vitality,pain, and general health perception) and one for the mentalcomponent (assessing social and role functioning, vitality,mental health, and general health perception). Summaryscores range from 0 to 100, with higher scores representingbetter functioning.

The Health Assessment Questionnaire [42]

The “short” Health Assessment Questionnaire (HAQ) isdesigned to measure disability in rheumatological diseasesand consists of eight categories, each containing two tothree items. Patients rate the difficulty they experienceperforming activities of daily living on a four-point scaleranging from 0 to 3. Scores are automatically amended to“2” should the patient use aids or specially adapted devices.Scores for each category are summed to provide an overall“disability score” ranging from 0 to 24, which is thenconverted to a score ranging from 0 to 3; a higher numberrepresents greater disability.

Disease Activity Score [43]

The DAS 28 involves an assessment of 28 joints and wasdeveloped specifically as a measure of disease activity inRA. It is used to monitor the disease activity of RA patientsin everyday clinical practice, is able to discriminate betweenclinically relevant levels of disease activity [44], and hasbeen extensively validated [45]. The measure involvescounting the number of swollen and tender joints, obtainingthe ESR (an inflammatory marker) and calculating asummary score. The patient may also be asked for asubjective judgment of disease activity, providing a furtherfigure between 0 and 10. Final DAS scores range from 0 to10. Threshold scores of 2.6, 3.2, and 5.1 have beenidentified, indicating disease remission, low disease activity,and high disease activity, respectively.

Table 3Univariate analysis of associations between demographic variables, clinicalfactors, illness beliefs, and disability (HAQ)

Variables Coefficient (95%CI) P value

GenderFemale 0Male −0.31 (−0.64 to 0.03) .08EthnicityWhite 0Black 0.23 (−0.22 to 0.69) .32Asian −0.28 (−0.79 to 0.22) .26Age 0.02 (0.01 to 0.03) b.01DAS28 0.21 (0.12 to 0.29) b.01Time since diagnosis 0.03 (0.01 to 0.04) b.01Illness beliefsIdentity 0.10 (0.05 to 0.16) b.01Consequences 0.06 (0.02 to 0.09) b.01Personal control −0.08 (−0.11 to −0.05) b.01Treatment control −0.06 (−0.10 to −0.02) .01Illness coherence −0.01 (−0.04 to 0.02) .38Timeline cyclical 0.00 (−0.05 to 0.04) .93Emotional representations 0.02 (−0.01 to 0.04) .27

420 H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

Results

Statistical procedures

Data were analyzed using SPSS Version 14.0 forWindows. Summary statistics of median and interquartileranges are presented for continuous variables, and thepercentages for categorical variables (Table 1). Cronbach'sα values for IPQ-R scales are shown in Table 2. Exploratorylinear regression models (Tables 3 and 4) with robuststandard error were constructed to assess the relationshipsbetween outcome variables (illness beliefs, disability, andQoL scores) and potential covariates (demographic andclinical factors, and disease activity). The validity of themodels was assessed initially by plotting the fitted valuesagainst the “Studentized” residuals. Variables that retainedindependent associations with outcome at univariate level(Pb.02) were considered for multivariate analysis, and theadjusted P value which is less than .05 was retained. AllP values are two-sided. The multivariate results arepresented in Table 5.

Disability

Age and longer disease duration were positively related todisability, as were identity (r=.31, Pb.01) and consequences(r=.28, Pb.01) scores. Patients who attributed more symp-toms to their illness and held stronger beliefs that it wouldhave negative consequences also reported greater disability.Both personal and treatment control beliefs were inverselyrelated to disability; stronger personal control beliefs wereassociated with lower levels of disability (r=−.40, Pb.01) aswere stronger beliefs in the ability of treatment to control RA(r=−.24, Pb.01).

Disease activity scores were associated with disability(r=.39, Pb.01) and physical function summary (r=−.38,Pb.01) as expected. However, no significant associationswere found between disease activity and any dimensions ofillness beliefs that were significantly associated with HAQfunctioning scores.

Table 2IPQ scale descriptives

ScaleItemsin scale Cronbach's α Range Mean S.D.

Identity 14 0.76 1–14 6.46 2.53Timeline (acute/chronic) 6 0.77 9–30 24.57 4.30Consequences 6 0.68 10–30 21.94 4.32Personal control 5 0.76 7–30 19.58 4.42Treatment control 4 0.72 5–24 16.84 3.47Illness coherence 5 0.87 5–27 17.22 4.85Timeline (cyclical) 4 0.86 4–21 14.83 3.55Emotional representations 6 0.88 7–30 18.75 5.59

It is not appropriate to compute an internal reliability statistic for the causaldimension as participants simply indicate which causal attributionsthey endorse.

Physical and Mental QoL

The IPQ-R dimension of identity was inversely associatedwith physical (r=−.23, Pb.05) and mental functioning (r=−.40, Pb.01). Patients who attributed more symptoms totheir RA reported poorer physical and psychologicalfunctioning. The consequences dimension was negativelyassociated with the physical functioning (r=−.22, Pb.05),social functioning (r=−.19, Pb.05), and mental healthsubscales (r=−.22, Pb.05). Patients with stronger beliefsthat RA will have negative consequences reported worsephysical, social, and mental functioning. There were nosignificant correlations between beliefs about acute/chronictimeline and QoL, although cyclical timeline scores wereassociated with limitation due to emotional problems (r=−.20, Pb.05), pain (r=−.19, Pb.05), and energy/vitality (r=−.23, Pb.01).

Personal control was associated with physical (r=.20,Pb.05) summary scores and with the subscales of physicalfunctioning (r=.30, Pb.01) and energy/vitality (r=.19,Pb.05). Patients with stronger beliefs in their own personalability to control their RA reported better physical function-ing. Treatment control was associated with mental function-ing summary scores (r=.26, Pb.01) and the subscales ofhealth change (r=.21, Pb.05), physical function (r=.25,Pb.01), role limitation due to emotional problems (r=.22,Pb.05), mental health (r=.29, Pb.01), and energy/vitality(r=.24, Pb.01). Patients with stronger beliefs in the ability oftreatment to control their illness reported better physicalfunctioning, better health change over the past year, less rolelimitation due to emotional problems, and better mentalhealth and energy levels.

Disease activity was inversely associated with a numberof QoL subscales. Patients with higher disease activity werealso more likely to report worse perception of their health in

Table 5Multivariate analysis of variables associated with disability and QoL scores

Variable Coefficient (95% CI) P value R2

HAQ disability Age 0.01 (0.00 to 0.02) .02Time sincediagnosis

0.01 (0.00 to 0.03) .01

DAS28 0.16 (0.09 to 0.22) b.01Identity 0.07 (0.01 to 0.13) .02Consequences 0.04 (0.01 to 0.07) b.01Treatment control 0.00 (−0.04 to 0.04) .88Personal control −0.05 (−0.08 to −0.02) b.01 .45

SF-36 physicalsummary

Age −0.13 (−0.24 to −0.02) .020DAS score −1.98 (−2.80 to −1.17) b.01Identity −0.75 (−1.36 to −0.14) .02 .19

SF-36 mentalsummary

Identity −1.17 (−1.20 to −0.35) b.01Treatment control 0.50 (−0.11 to 1.11) .11Emotionalrepresentations

−0.64 (−1.04 to −0.23) b.01 .27

421H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

general (r=−.21, Pb.05), a more negative change in healthover the past year (r=−.21, Pb.05), worse physical function-ing (subscale) (r=−.25, Pb.01), more role limitation due tophysical problems (r=−.19, Pb.05), less energy/vitality (r=−.20, Pb.05), and worse physical functioning (summaryscale) (r=−.38, Pb.01). Disease activity was positivelyassociated with pain. Those patients with worse diseaseactivity also reported experiencing more pain (r=.43, Pb.01).

Linear regression models for disability and QoL

DisabilityExploratory univariate analysis (Table 3) revealed that the

variables age, DAS28, time since diagnosis, and the illnessbeliefs dimensions of identity, consequences, personalcontrol, and treatment control were significant (Pb.02) andwere therefore included in the multivariate analysis(Table 5). The R2 value for the final multivariate modelwas 45.5%.

QoL (SF-36 physical component summary score)Exploratory univariate analysis (Table 4) revealed that the

variables age, DAS score, and identity were significant(Pb.02) and were therefore put forward for multivariateanalysis (Table 5). The R2 value for the final modelwas 19.3%.

QoL (SF-36 mental component summary score)Exploratory univariate analysis (Table 4) revealed that the

variables identity, treatment control, and emotional repre-sentations were significant (Pb.02) and were therefore putforward for multivariate analysis (Table 5), although only

Table 4Univariate analysis of associations between demographic variables, clinicalfactors, illness beliefs, and QoL scores

Variables Coefficient (95%CI) P value

Physicalsummary

Age −0.14 (−0.26 to −0.02) .02Time since diagnosis −0.09 (−0.23 to 0.05) .19DAS score −2.01 (−2.79 to −1.23) b.01Identity −0.78 (−1.39 to −0.16) .01Timeline 0.24 (−0.08 to 0.57) .14Consequences −0.27 (−0.67 to 0.13) .19Personal control 0.40 (0.02 to 0.78) .04Treatment control 0.24 (−0.17 to 0.64) .26Illness coherence 0.11 (−0.23 to 0.45) .53Timeline cyclical −0.31 (−0.67 to 0.06) .10Emotional representations −0.11 (−0.43 to 0.22) .52

Mentalsummary

Age −0.12 (−0.25 to 0.02) .08Time since diagnosis −0.07 (−0.23 to 0.09) .41DAS score −0.21 (−1.49 to 1.06) .74Identity −1.84 (−2.52 to −1.16) b.01Timeline −0.14 (−0.61 to 0.32) .54Consequences −0.44 (−0.92 to 0.04) .07Personal control 0.33 (−0.13 to 0.79) .16Treatment control 0.86 (0.35 to 1.37) b.01Illness coherence 0.31 (−0.11 to 0.73) .15Timeline cyclical −0.55 (−1.17 to 0.06) .08Emotional representations −0.89 (−1.22 to −0.55) b.01

identity and emotional representations remained significantin the final model. The R2 value for the final modelwas 27.3%.

Perceived seriousness construct

This study did not find support for a higher-orderconstruct of seriousness in this sample of RA patients. Ifthis were a valid construct, then we would expect thecomponent variables of identity, consequences, and timelinedimensions to be strongly correlated with each other. Thiswas not the case and so this construct was not subjected tofurther empirical analysis.

Discussion

The present study examined the relationship of illnessbeliefs to disability and QoL in RA. The hypothesis for aproposed seriousness construct combining identity, conse-quences, and timeline dimensions among RA patients wasnot supported. Although these variables did show associa-tions with each other, they did not relate sufficiently tosuggest conceptual overlap. It may therefore be that thesedimensions are demonstrating logical interrelationships (e.g.,those patients with a long disease duration may anticipatemore negative consequences) but are not worthy of furthertheoretical investigation as a higher-order construct. How-ever, correlations were in the expected direction for theidentity and consequences dimensions, in line with previousresearch (e.g., Refs. [19,20]).

Timeline beliefs (acute/chronic and cyclical) were notsignificantly associated with the primary outcome measures(HAQ disability, SF-36 physical component, and SF-36mental component scores). However, patients with strongerbeliefs in a cyclical timeline for RAwere also more likely toreport less energy/vitality, perhaps due to experiencing anexhausting cycle of symptom flares. The systemic nature of

422 H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

the illness also means that patients are rarely free ofsymptoms and are therefore likely to view their RA aschronic (mean=24.57, S.D.=4.30). This is in contrast to otherlong-term conditions, for example, hypertension, wherepatients may develop acute timeline beliefs due to theabsence of symptoms (mean=3.55, S.D.=0.44) [46].

As expected, stronger beliefs in personal control overRA were related to reduced disability and better physicaland psychological functioning, in line with previousfindings (e.g., Refs. [20,21]). Treatment control was alsoinversely related to disability and positively associatedwith mental function. Although control beliefs areassociated with outcome in RA, it may be that thisrelationship is complex. It is possible that the cyclical andunpredictable disease course fosters a complicated set ofbeliefs, requiring longitudinal analysis, to establish thedirectionality of these effects.

A greater number of perceived symptoms were related topsychological functioning, as was greater confidence in theability of treatment to control the condition. However,disease activity was not related to psychological function-ing or to any of the dimensions of illness beliefs.Perceiving more symptoms and more control was relatedto physical QoL, as was DAS28; but again, disease activitywas not associated with either identity or control dimen-sions. This suggests that patients' perceptions of theirillness may not be a function of disease severity. When theillness is worse, we may expect corresponding beliefs, butthe absence of a relationship suggests this is not the case;there is wide variation among individuals, and illnessbeliefs cannot be explained by disease status. Patients withvery active disease do not necessarily view their illness asworse than others. This is in line with the results of Carlisleet al. [29].

There are a number of limitations with the present study.Although the associations between perceived consequencesand outcomes is in line with previous research, the α valuefor this subscale was just below the minimum acceptablelevel and the result may therefore be an underestimate of the“true” relationship between beliefs about consequences andoutcomes [47]. We also acknowledge that the small samplesize does limit statistical power and, therefore, the resultsshould be interpreted with some degree of caution. Thecross-sectional design of the study also means that nodirection of causality can be inferred; disability or QoL mayinfluence beliefs rather than vice versa. It is possible that theobserved associations between beliefs and outcome are aresult of patients having accurate representations of theirillness. Those experiencing more negative consequences,more symptoms, and less control would be expected to havegreater disability and poorer QoL, although the fact thatbeliefs showed no relationships with disease activitysuggests this is not the case.

This study provides support for the utility of Leventhal'sCSM in rheumatological disease and highlights theconsiderable variability of illness beliefs between indivi-

duals. The fact that objective measures of disease activity areunrelated to beliefs about illness is important as it suggeststhat illness beliefs are sufficient to influence disability andfunctioning despite severity of illness. Further studies shouldbe longitudinal and consider any change in representationsover the unpredictable disease course. Findings may informinterventions, which should target beliefs considered“maladaptive” or “adaptive” in achieving improved out-comes, as has proved effective in other long-term conditions(e.g., Ref. [37]).

Acknowledgments

The authors are grateful to Fowzia Ibrahim for herstatistical advice and support. We also thank the outpatientstaff that facilitated access to patients and the patients whogave their time to take part. Funding for this study wasprovided by a grant from the Arthritis Research Campaign.

References

[1] Fries JF, Bellamy N. Introduction. In: Bellamy N, editor. Prognosis inthe rheumatic diseases. London: Kluwer, 1991.

[2] Shipley M, Newman SP. Psychological aspects of rheumatic diseases.Baillieres Clin Rheumatol 1993;7:215–9.

[3] Pimm T. Self-regulation and psycho-educational interventions forrheumatic disease. In: Petrie KJ, Weinman JA, editors. Perceptions ofhealth and illness: current research and applications. Amsterdam:Harwood Academic Publishers, 1997.

[4] Hirano PC, Laurent DD, Lorig K. Arthritis patient education studies,1987–1991: a review of the literature. Patient Education andCounseling 1994;24:9–54.

[5] Simpson C. The patient's journey: rheumatoid arthritis. BMJ 2005;331:887–9.

[6] Petrie KJ, Weinman J. Why illness perceptions matter. Clin Med 2006;6:536–9.

[7] Petrie KJ, Jago LA, Devcich DA. The role of illness perceptions inpatients with medical conditions. Curr Opin Psychiatry 2007;20:163–7.

[8] Leventhal H, Meyer D, Nerenz D. The commonsense representation ofillness danger. In: Rachman S, editor. Contributions to medicalpsychology, Vol. 2. New York: Pergamon Press, 1980. pp. 17–30.

[9] Leventhal H. Illness representations and coping with health threats.In: Baum S, Taylor SE, Singer JE, editors. Handbook of psychologyand health, Vol 4. Hillsdale, NJ: Lawrence Erlbaum Associates, 1984.pp. 219–52.

[10] Scharloo M, Kaptein AA, Schlsser M, Pouwels H, Bel EH, Rabe KF,et al. Illness perceptions and quality of life in patients with chronicobstructive pulmonary disease. Journal of Asthma 2007;44:575–81.

[11] Rutter CL, Rutter DR. Illness representation, coping and outcome inirritable bowel syndrome (IBS). Br J Health Psychol 2002;7(Part 4):377–91.

[12] Jopson NM, Moss-Morris R. The role of illness severity and illnessrepresentations in adjusting to multiple sclerosis. J Psychosom Res2003;54:503–11.

[13] Scharloo M, Baatenburg de Jong RJ, Langeveld TP, van Velzen-Verkaik E, Doorn-op den Akker MM, Kaptein AA. Quality of life andillness perceptions in patients with recently diagnosed head and neckcancer. Head Neck 2005;27:857–63.

[14] Kaptein AA, Helder DI, Scharloo M, Van Kempen GMJ, Weinman J,Van Houwelingen HJC, et al. Illness perceptions and coping explainwell-being in patients with Huntington's disease. Psychology andHealth 2006;21:431–46.

423H. Graves et al. / Journal of Psychosomatic Research 67 (2009) 417–423

[15] French DP, Lewin RJP, Watson N, Thompson DR. Do illnessperceptions predict attendance at cardiac rehabilitation and quality oflife following myocardial infarction? Journal of PsychosomaticResearch 2005;59:315–22.

[16] Hagger MS, Orbell S. A meta-analytic review of the common-sensemodel of illness representations. Psychology and Health 2003;18:141–84.

[17] Lorig KR, Cox T, Cuevas Y, Kraines RG, Britton MC. Converging anddiverging beliefs about arthritis: Caucasian patients, Spanish speakingpatients, and physicians. Journal of Rheumatology 1984;11:76–9.

[18] Hampson SE, Glasgow RE, Zeiss AM. Personal models ofosteoarthritis and their relation to self-management activities andquality of life. Journal of Behavioural Medicine 1994;17:143–58.

[19] Van Lankveld W, Bosch PVP, Van De Putte L, Narbng G, Van DerStaak C. Disease-specific stressors in rheumatoid arthritis: coping andwell-being. Rheumatology 1994;33:1067–73.

[20] Treharne GJ, Kitas GD, Lyons AC, Booth DA. Well-being inrheumatoid arthritis: the effects of disease duration and psychosocialfactors. J Health Psychol 2005;10:457–74.

[21] Groarke A, Curtis R, Coughlan R, Gsel A. The impact of illnessrepresentations and disease activity on adjustment in women withrheumatoid arthritis: a longitudinal study. Psychology and Health2005;20:597–613.

[22] Schiaffino KM, Shawaryn MA, Blum D. Examining the impact ofillness representations on psychological adjustment to chronic ill-nesses. Health Psychology 1998;17:262–8 American PsychologicalAssociation.

[23] Sharpe L, Sensky T, Allard S. The course of depression in recent onsetrheumatoid arthritis: the predictive role of disability, illness percep-tions, pain and coping. Journal of Psychosomatic Research 2001;51:713–9.

[24] Affleck G, Tennen H, Pfeiffer C, Fifield J. Appraisals of control andpredictability in adapting to a chronic disease. J Personality and SocialPsychology 1987;53:273–9.

[25] Fitzpatrick R, Newman S, Lamb R, Shipley M. Helplessness andcontrol in rheumatoid arthritis. International Journal of Health Science1990;1:17–23.

[26] Tennen H, Affleck G, Urrows S, Higgins P, Mendola R. Perceivingcontrol, construing benefits and daily processes in rheumatoid arthritis.Journal of Behavioural Science 1992;24:186–203.

[27] Skevington SM. Depression and causal attributions in the early stagesof a chronic painful disease: a longitudinal study of early synovitis.Psychology and Health 1993;8:51–64.

[28] Scharloo M, kaptein A, Weinman J, Hazes J, Willems L, Bergman,et al. Illness perceptions, coping and functioning in patients withRheumatoid Arthritis, chronic obstructive pulmonary disease andpsoriasis. Journal of Psychosomatic Research 1998;44:573–85.

[29] Carlisle ACS, John AMH, Fife-Schaw C, LloydM. The self-regulatorymodel in women with rheumatoid arthritis: relationships betweenillness representations, coping strategies, and illness outcome. BritishJournal of Health Psychology 2005;10:571–87.

[30] Jones A, Doherty M. ABC of Rheumatology: OSTEOARTHRITIS.BMJ 1995;310:457–60.

[31] Akil M, Amos RS. ABC of Rheumatology: RHEUMATOIDARTHRITIS–I: CLINICAL FEATURES AND DIAGNOSIS. BMJ1995;310:587–90.

[32] Hale ED, Treharne GJ, Kitas GD. The common-sense model of self-regulation of health and illness: how can we use it to understand andrespond to our patients' needs? Rheumatology (Oxford) 2007;46:904–6.

[33] Horne R, Weinman J. Self-regulation and self-management in asthma:exploring the role of illness perceptions and treatment beliefs inexplaining non-adherence to preventer medication. Psychology andHealth 2002;17:17–32.

[34] Heijmans M, de RD. Assessing illness representations of chronicillness: explorations of their disease-specific nature. J Behav Med1998;21:485–503.

[35] Meyer D, Leventhal H, Gutmann M. Common-sense models ofillness: the example of hypertension. Health Psychol 1985;4:115–35.

[36] Frostholm L, Fink P, Christensen KS, Toft T, Oernboel E, Olesen F,et al. The patients' illness perceptions and the use of primary healthcare. Psychosom Med 2005;67:997–1005.

[37] Petrie KJ, Cameron LD, Ellis CJ, Buick D, Weinman J.Changing illness perceptions after myocardial infarction: anearly intervention randomized controlled trial. Psychosom Med2002;64:580–6.

[38] Frostholm L, Fink P, Oernboel E, Christensen KS, Toft T, Olesen F,et al. The uncertain consultation and patient satisfaction: the impact ofpatients' illness perceptions and a randomized controlled trial on thetraining of physicians' communication skills. Psychosom Med 2005;67:897–905.

[39] Moss-Morris R, Weinman J, Petrie KJ, Horne R, Cameron LD, BuickD. The Revised Illness Perception Questionnaire (IPQ-R). Psychologyand Health 2002;17:1–16.

[40] Ware JE, Sherbourne CD. The MOS 36-item short-form health survey(SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473–83.

[41] Ware JE, Snow KK, Ksinski MA, Gandek B. SF-36 Health Survey:manual and interpretation guide. The Health Institute, 1993.

[42] Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patientoutcome in arthritis. Arthritis Rheum 1980;23:137–45.

[43] Prevoo ML, van 't Hof MA, Kuper HH, van Leeuwen MA, van dePutte LB, van Riel PL. Modified disease activity scores that includetwenty-eight-joint counts. Development and validation in a prospectivelongitudinal study of patients with rheumatoid arthritis. ArthritisRheum 1995;38:44–8.

[44] Vander Cruyssen B, Van Looy S, Wyns B, Westhovens R, Durez P, Vanden Bosch F, et al. DAS28 best reflects the physician's clinicaljudgment of response to infliximab therapy in rheumatoid arthritispatients: validation of the DAS28 score in patients under infliximabtreatment. Arthritis Research & Therapy 2005;7.

[45] Fransen J, Stucki G, van Riel LCM. Rheumatoid arthritis measures:Disease Activity Score (DAS), Disease Activity Score-28 (DAS28),Rapid Assessment of Disease Activity in Rheumatology (RADAR),and Rheumatoid Arthritis Disease Activity Index (RADAI). ArthritisCare and Research 2003;49:214S–24S.

[46] Ross S, Walker A, MacLeod MJ. Patient compliance in hypertension:role of illness perceptions and treatment beliefs. J Hum Hypertens2004;18:607–13.

[47] Cronbach LJ. Coefficient alpha and the internal structure of tests.Psychometrika 1951;16:297–334.