Illness Representations
Michael A. Diefenbach

3 Measures and Measurements

a) Early assessments and ad hoc measurement of illness representations

Early measurements of illness representations consisted of a mixture of closed and open ended questions. Open ended questions are often times used by researchers at first to assess the content of an illness belief. Thus, the use of open ended questions informs scale development for closed ended questions and allows the adaptation of the illness representation construct to a new area of illness beliefs.

• Identity was assessed with a question, such as "Do you have a name for the condition" (yes/no), and a follow-up, open-ended question "Can you tell me what it is?" "How can you tell?" (Diefenbach & Leventhal, 1996 xClose
  Diefenbach, M.A., & Leventhal, H. (1996). The common-sense model of illness representation: Theoretical and practical considerations. The Journal of Social Distress and the Homeless, 5, 11-38. ). In other instances, when the patient does not know the identity of a condition, the occurrence of commonly experienced symptoms (i.e., pain, headaches, breathlessness, and sleep difficulties) has been assessed. Patients answer items on a 5-point rating scale (1 = not at all - 5 = very) or a yes/no dichotomous scale (Weinman, Petrie, Moss-Morris, & Horne, 1996 xClose
  Weinman, J., Petrie, K., Moss-Morris, R, & Horne, R. (1996). The Illness Perception Questionnaire: A new method for assessing the cognitive representation of illness. Psychology and Health, 11 431-445. ).
• Timeline can be assessed with the following question: "Do you think your condition is something that will go away after a couple of days, such as a cold; will come and go, such as an allergy; or you will have forever, such as diabetes?" Patients select the option that best fits their beliefs.
• The Consequence attribute can be assessed with the following question: "Do you think your symptom/condition is minor, severe, or life threatening?" Patients are asked to select the option that best fits their beliefs.
• Causal attributes have been assessed with threat-specific causes, often derived from patient interviews or focus groups. For example, "Please indicate how much you agree or disagree with the following reasons as causes for your symptom/condition." (Format e.g., 1 = completely disagree; 5 = completely agree).
• The control/cure attribute can be assessed with a question such as, "Something can be done to cure the symptom/condition (1 = completely disagree - 5 = completely agree).

b) The Illness Perception Questionnaire (IPQ)

The popularity of the self-regulation approach among researchers prompted the development of a measure to systematically assess the five illness representation attributes with 38 items (Illness Perception Questionnaire, IPQ; Weinman et al, 1996 xClose
Weinman, J., Petrie, K., Moss-Morris, R, & Horne, R. (1996). The Illness Perception Questionnaire: A new method for assessing the cognitive representation of illness. Psychology and Health, 11 431-445. ). Illness Identity is assessed with a 12-item core symptom list that is answered on a four-point Likert scale (1 = never; 2 = occasionally; 3 = frequently; 4 = all of the time). Core symptoms focus on general symptomatology, such as pain, nausea, breathlessness, weight loss, fatigue, stiff joints, sore eyes, headaches, upset stomach, sleep difficulties, dizziness, and loss of strength. The remaining four illness representation attributes, cause, timeline, consequences and control/cure are rated on a 5-point Likert scale (1 = strongly disagree; 3 = neither agree or disagree 5 = strongly agree). The cause attribute is assessed with 10 items, representing common causal beliefs. The timeline attribute is assessed with three items. The consequence attribute is assessed with seven statements. Finally, the control/cure attribute is assessed with six statements.

The IPQ had been used with diverse patient populations suffering from conditions, such as heart disease (Cooper et. al., 1999 xClose
Cooper, A., Lloyd, GS., Weinman, J. and Jackson, G. (1999). Why patients do not attend cardiac rehabilitation: Role of intentions and illness beliefs. Heart, 82 234-236. ,) cancer (Buick, 1997 xClose
Buick, DL. (1997). Illness representations and breast cancer: Coping with radiation and chemotherapy. In: Petrie, KJ and Weinman, J (Eds.), Perceptions of health and illness. Current research and applications pp. 379-409. Amsterdam; Harwood Academic Publishers. ), arthritis (Murphy et. al., 1999 xClose
Murphy, H, Dickens, C, Creed, F and Bernstein, R (1999). Depression, illness perception and coping in rheumatoid arthritis. Journal of Psychosomatic Research, 46, 155-164. ), diabetes (Griva et. al., 2000 xClose
Griva, K, Myers, LB, and Newman, S (2000). Illness perceptions and self-efficacy beliefs in adolescents and young adults with insulin dependent diabetes mellitus. Psychology and Health 15, 733-750. ), and chronic fatigue syndrome (Heijmans, 1998 xClose
Heijmans, M (1998). Coping and adaptive outcome in chronic fatigue syndrome: importance of illness cognitions. Journal of Psychosomatic Research, 45, 39-51. ). These studies found strong support for the existence of the six illness representation attributes. The IPQ can be tailored for use across many different illnesses by modifying the question stem and by altering the instructions to focus on the specific illness under investigation.

c) The Illness Perception Questionnaire-Revised (IPQ-R)

The IPQ-R was developed to improve the measurement properties of two of the subscales (i.e., cure/control and timeline) and to broaden the scope or the original IPQ (Moss-Morris, Weinman, Petrie, Horne, Cameron, and Buick; 2002 xClose
Moss-Morris, R., Weinman, J., Petrie, K.J., Horne, R., Cameron, L.D., & Buick D. (2002). The revised illness perception questionnaire (IPQ-R). Psychology and Health, 17, 1-16. ). The cure/control subscale now distinguishes between personal control and self-efficacy beliefs (six items) and treatment control and outcome expectations (five items). For example, new items that were added to the personal control scale are "The course of my illness depends on me;" and "Nothing I do will affect my illness." New items added to the treatment control subscale include "My treatment can control my illness" and "There is nothing which can help my condition." The second subscale that was improved was the timeline subscale that now includes 10 items, including those that assess cyclical beliefs (e.g., "My symptoms come and go in cycles;" My illness is very unpredictable.").

Another aspect of the IPQ-R that was neglected in the original conceptualization is the assessment of emotional representations. Leventhal's self-regulation model specifies that individuals have emotional representations and reactions to the health threat, which might lead to emotion-based coping behavior. Consequently, six items assessing emotional representations have been included, such as "I get depressed when I think about my illness;" "Having this illness makes me feel anxious." The final addition was a subscale measuring a person's overall coherent understanding of an illness. This illness-related "meta-cognition" is measured with five items, such as "My illness is a mystery to me;" and "I have a clear picture or understanding of my condition." In total, the IPQ-R consists of 71 items divided into twelve subscales. Psychometric analyses were performed on several patient samples, totaling a combined population of N = 711 patients.

Moss-Morris, Weinman, Petrie, Horne, Cameron, and Buick (2002) xClose
Moss-Morris, R., Weinman, J., Petrie, K.J., Horne, R., Cameron, L.D., & Buick D. (2002). The revised illness perception questionnaire (IPQ-R). Psychology and Health, 17, 1-16. used eight different illness groups for the validation of the IPQ-R, including HIV, multiple sclerosis, asthma and type II diabetes patients. Psychometric analysis revealed that all of the IPR-R subscales demonstrate good internal reliability. Cronbach alpha's range from 0.79 to 0.89 for the timeline cyclical dimension and the timeline acute/chronic dimension respectively. (Moss-Morris et al., 2002 xClose
Moss-Morris, R., Weinman, J., Petrie, K.J., Horne, R., Cameron, L.D., & Buick D. (2002). The revised illness perception questionnaire (IPQ-R). Psychology and Health, 17, 1-16. ). Test-Retest reliability was investigated at two time points over a three-week period. Pearson's correlations ranged from .46 for personal control beliefs to an average of .85 for different risk factor attributions, demonstrating adequate stability for a short time frame.

To determine discriminant validity and to ensure that the IPQ-R dimensions are not just reflective of affective dispositions, Pearson's correlations were computed between the subscales of the Positive and Negative Affect Scale (PANAS). Correlations between the two PANAS subscales and the all of the subscales of the IPQ were generally small to moderate in size, with the most significant relationship between emotional representations and negative affect (NA) (r =. 54). This correlation suggested that approximately 29% of the variance of the emotional reaction to illness was accounted for by trait negative affect.

Predictive stability of the IPQ was evaluated with regard to adjustment to illness among MS patients. Illness representations predicted 15% of the variance of adjustment to disease controlling for illness severity. Among the six illness representations, the identity representation was the strongest predictor.

The large number of items of the IPQ-R and the associated time commitment to complete the measure might be perceived as an impediment to its widespread use. For that reason, Cameron and colleagues have used a shortened version of the IPQ-R (Cameron, Booth, Schlatter, Ziginskas, Harman, & Benson; 2005 xClose
Cameron, L.D., Booth R.J., Schlatter, M., Ziginskas D., Harman, J.E., Benson, S.R.C. (2005). Cognitive and affective determinants of decision to attend a group psychosocial support program for women with breast cancer. Psychosomatic Medicine, 67, 584-589. ) consisting of a shortened version of the personal control and emotional representation subscales. The personal control scale uses three of the six items and the emotional representation subscale uses four out of six items. The authors report that analyses with data from the original validation study demonstrated that the two subscales exhibited comparable discriminate and convergent validity with questionnaires of other illness beliefs and negative affectivity.