Individualizing the Measure of Functional Impairment in Rheumatoid Arthritis (RA): Development of Patient-Specific Forms of the Health Assessment Questionnaire (HAQ)

PURPOSE. Inclusion of patients’ perspectives in outcome assessment in RA had been recently pointed out by the OMERACT. HAQ is the most widely used measure of functional disability in RA patients. We therefore conducted this study to develop and validate patient-specific (i.e. individualized) form of the HAQ.
PATIENTS AND METHODS. HAQ data were prospectively obtained from 370 outpatients with active RA, treated with leflunomide over a 6-months period in a previous open-label, non-controlled study¹. Based on the 20 activities addressed by the HAQ items, patients had to rate the importance of each activity, and to select and order the 5 activities they considered the most important in their daily life. Different methods of individualization were used: 2 individualized forms of the HAQ (score of each item is multiplied, or added, by its importance), and 2 HAQ top 5 (involving only the 5 most important activities to each patient, weighted or not). Psychometric properties of these patient-specific scales were compared to those of the HAQ, and changes in patients’ priorities were assessed.
RESULTS. All patient-specific scales were highly convergent with the HAQ (rho>0.75). Internal consistency was good (Cronbach’s alpha: 0.87 to 0.88). Sensitivity to change, evaluated by standardized response mean (SRM) was lower for all individualized scales compared to HAQ (SRM: 0.64 to 0.69 vs 0.74). Severity and importance scores of each item were weakly or not correlated (rho: 0.11 to 0.29). Patients’ priorities however seemed to be rather consistent over the 6-month period, for each item 63% to 92% of patients made the same choice at both initial and final visits.
CONCLUSION. Evaluation of patient’s priorities showed that the measure of importance of an activity gave complementary information to the measure of disability that could be relevant in clinical practice. However, individualized scales did not have better properties than the HAQ. This could be explained by the fact that the HAQ, taking only the most disabled activity of each domain is already somewhat individualized. Therefore for RA patients, the recommendation is to use the parent form.
¹ Nguyen M., Clin Drug Invest 2004; 24(2):103-112
Psychometric properties of the parent and individualized scales

Psychometric properties of the parent and individualized scales

	HAQ	HAQ individualized with multiplicative method	HAQ Individualized with additive method	HAQ top5 (5 items)	Weighted HAQ top5 (5 items)
Construct validity Spearman's correlations with HAQ	1 p<0.0001	0.88 p<0.0001	0.77 p<0.0001	0.77 p<0.0001	0.76 p<0.0001
Internal consistency Cronbach's alpha	0.87	0.88 p=0.19	0.88 p=0.21	NE	NE
Sensitivity to change SRM	0.74	0.69 p=0.05	0.68 p=0.02	0.65 p=0.053	0.64 p=0.02

NE: not estimable; SRM: standardized response mean. Psychometric properties of each individualized scale were compared to those of the HAQ.

 R. Seror, None; F. Tubach, None; G. Baron, None; P. Ravaud, None.