Objective: Critical appraisal of instruments to assess illness representations in rheumatology.
Methods: A first search (MEDLINE, EMBASE, PsycINFO) identified articles describing development of instruments assessing illness representations. A second search identified articles applying them in rheumatology. Appraisal was performed using checklists.
Results: Five instruments were identified: the Illness Perception Questionnaire (IPQ), the Revised Illness Perception Questionnaire (IPQ-R), the Illness Cognition Questionnaire (ICQ), the Implicit Models of Illness Questionnaire (IMIQ) and the Meaning of Illness Questionnaire (MIQ). The number of items varied from 18 to 70. Internal consistency was good. Construct validity was moderate to good for all instruments, but was not tested for the IMIQ. Predictive validity was assessed for the ICQ and IPQ and was low to moderate. Sensitivity to change was tested for ICQ and was good. Applications in rheumatology are increasing. Significant relationships with different outcomes and additional evidence for predictive validity were found.
Conclusions: Five instruments that assess illness representations can be used in rheumatology. The number of subscales and items vary. The ICQ and IPQ are most extensively validated and have been most frequently applied in rheumatology. Illness representations have relevant associations with self-report and objective outcomes. They should be considered when exploring health in rheumatology.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The biopsychosocial model of disease is increasingly accepted in rheumatology as a valuable framework to understand impact of rheumatic diseases.1 The World Health Organization (WHO) approved this disease model by endorsing the International Classification of Functioning, Disability and Health (ICF) as the model for functioning and health. In this model, personal factors are recognised to have an important contribution to health-related functioning and participation.2 Personal factors are particularly important to understand adjustments individuals make when they are confronted with a (acute or chronic) disease. A variety of personal factors that might be relevant have been proposed, including gender, ethnicity, educational level and personality traits.1 It is accepted that an individual’s personality can be either protective (facilitator) or a risk factor (barrier) when adjusting to an illness. Personality traits that have been studied include coping strategies, optimism, expectancies, sense of purpose in life and illness representations.3 Illness representations, cognitions or perceptions refer to the beliefs subjects create about illness when confronted with it in order to make sense of and deal with it.4 In chronic disease, current literature suggests the value of illness representations in understanding and predicting mental and physical outcome.3 Moreover, cognitive therapies are partly directed at influencing illness beliefs and representations. Insight in these representations can be helpful to predict and monitor the effect of such treatments.3
The classic concept of illness representations was realised in the 1980s. The initial models distinguished four common dimensions of beliefs that people create about their disease: identity or the type of symptoms and knowledge about the symptoms that belong to the disease; time-line or the perception of the course the disease has and will have; cause of the illness; and the beliefs on severity of the consequence or the impact of the illness on physical, mental and social functioning.5 6 Later, other dimensions were proposed such as personal responsibility, disruptiveness7 and curability and controllability.8
To better explore the relevance of illness representations in rheumatology, it is clear that one needs instruments with a sound theoretical background and correct clinimetric properties. Therefore, we searched for the available instruments and appraised their development critically. In addition, we reviewed articles that applied the instruments after the development phase in rheumatology.
First, MEDLINE, EMBASE and PsycINFO were searched up to January 2008 using the terms “illness cognition or illness perception or illness representation” and “questionnaire” in order to identify articles describing the development of questionnaires on illness representations. The references of the retrieved articles were checked. Experts were contacted and asked for additional instruments. Further inclusion criteria were: adequate description of the instrument development; generic or specific for rheumatic diseases; language English, German, Dutch or French; availability by Internet or Inter-University Librarian Loaning. Instruments that assess helplessness and self-efficacy were excluded as these were recently reviewed.9 An appraisal checklist, that was previously used in a project reviewing outcome instruments in rheumatology, was used after adaptation and can be seen online (supplemental file appendix 1).10 Criterion validity was not included because a gold standard is lacking.
MEDLINE, EMBASE and PsycINFO were searched again to identify studies that applied the instruments in rheumatology. Search terms were limited to the instrument names. All abstracts were checked and included for appraisal if they were original studies in English, German, Dutch or French and if results were reported for patients with a rheumatological condition. All diseases of the musculoskeletal system were considered except for traumatic or surgical conditions. All instruments and studies were reviewed for study population, aim, design and results by MM and checked by AB.
Search of PsycINFO provided 31 citations. Three articles cited by PsycINFO fulfilled the inclusion criteria. The Illness Cognition Questionnaire (ICQ),11 the Illness Perception Questionnaire (IPQ)12 and the Revised Illness Perception Questionnaire (IPQ-R).13 MEDLINE and EMBASE search did not add additional instruments. The excluded citations described the development of questionnaires that were disease-specific for non-rheumatological conditions or expert opinion articles. Contact with experts revealed the Implicit Models of Illness Questionnaire (IMIQ)14 and its further validation.15 References of the articles revealed the Meaning of Illness Questionnaire (MIQ)16 and its further validation.17 Therefore, five questionnaires were included. The results of the search for application in rheumatology revealed three articles for the ICQ,18–20 nine for the IPQ21–29 (one describing a follow-up),29 four for the IPQ-R,30–33 one for the IMIQ34 and four for the MIQ35–38 (one studying different aspects in the same population).36
Appraisal of the instruments
All questionnaires are self-report, generic and self-administered. All questionnaires are readily available through the original article (IPQ), authors (ICQ, MIQ) or on internet (IPQ, IPQ-R)39 except for the IMIQ and the initial MIQ. The IMIQ, however, can be reconstructed from the tables in the original article.14 The initial MIQ was slightly adapted concerning names and numbers of (sub)scales. The authors advised to present the characteristics of the latest version (MIQ-’92).40 The IPQ-R aimed to be an improvement of the IPQ, but in the further literature both instruments are used and therefore both were separately appraised. The IPQ and IPQ-R can be adapted for the disease of interest and for different languages. The ICQ is available in Dutch and English, all other questionnaires are available in English only. The instruments characteristics and clinimetrics are presented in table 1. More detailed description of the application in rheumatology can be seen online in (supplemental file, appendix 2).
When confronted with a disease, subjects are challenged to adjust to multiple aspects of illness. Illness representations are one of the personal factors that help to understand heterogeneity in adjustment. Five questionnaires assessing illness representations that can be used in patients with rheumatological conditions were found.11–14 16 Advising an instrument for application and further research in rheumatology is difficult, as direct comparison was never performed and because there was substantial heterogeneity in the definition of studied outcome. Overall, the ICQ, IPQ and IPQ-R were best validated in the developmental phase and most widely applied in rheumatology.
Several differences in the content of the questionnaires were noted and is reflected in the subscales of the questionnaires. These differences can be explained partly through the different theoretical concepts underlying illness representations. The MIQ, for example, emphasises the reappraisal of illness in the light of previous commitments and assesses aspects of stress and impact on daily life.16 The other instruments assess ideas about the illness itself independently of life commitments.11–14 In the absence of a gold standard and direct comparison, the relevance of conceptual differences between the questionnaires are difficult to judge. It is possible that different questionnaires have advantages in specific diseases or study questions.
Construct validity was evaluated against different health status outcomes. Most outcomes in the validation studies assessed patients’ perceived physical, social and psychological outcome and showed moderate to good associations. However, the ICQ and MIQ also showed significant relationships with direct clinical (joint score, inflammation) and behavioural outcomes (healthcare utilisation),11 16 17 emphasising the relevance of illness representations. More importantly predictive validity is also increasingly evident. For the ICQ and IPQ this was already shown in the developmental and for the other instruments this became evident on application in cohort studies.
Interesting is whether the concept “illness representations” measures different aspects of personality traits than other concepts such as coping or self-efficacy. The ICQ found weak to moderate correlations between the acceptance and benefits subscale and “active” coping and between helplessness and “passive” coping.11 The IPQ predicted quality of life better than coping24 and the effect of IPQ was partially mediated by coping.21 Although it seems self-evident that subscales Controllability (IPQ, IPQ-R, MIQ and IMIQ) or Helplessness (ICQ) are closely related to instruments assessing perceived self-efficacy,9 only the IPQ reported low correlation between Control/Cure subscale and Recovery Self-Efficacy Scale (R2: 0.38).12
It is discussed whether illness representations differ across disease.12 13 16 It is difficult to know to what extent these can be explained by true differences between diseases and, therefore, reflect correct patient disease knowledge. It seems, however, that subscales of illness representations contribute equally to patient reported outcome, independent of the type of disease.11 16 This suggests that the influence of illness representations on outcome is universal.
There is limited evidence (in rheumatoid arthritis) that illness representations can be changed by treatment, more specifically cognitive behavioural therapy.18 It will be a further challenge to explore this issue as well as the role of illness representations to predict response to interventions.
Feasibility was not assessed as a separate issue of our appraisal. However, the large difference between the numbers of items in the questionnaires should be noted. The ICQ has the smallest number of items, and the IPQ-R the largest number.
Five instruments are available to assess illness representations in rheumatology. In support of the biopsychosocial model of disease, illness representations help explain patient-reported outcomes but also objective clinical variables and healthcare utilisation. The ICQ, IPQ and IPQ-R were best validated in the developmental phase and were most frequently applied in rheumatology. Although the relationship with other personal factors and the independent role of the subscales and their performance in different rheumatological disease have not been fully elucidated, the review emphasises the importance of illness representations in outcome research.
Competing interests: None declared.
▸ Additional appendices 1 and 2 are published online only at http://ard.bmj.com/content/vol68/issue3