Which HAQ is best? A comparison of the HAQ, MHAQ and RA-HAQ, a difficult 8 item HAQ (DHAQ), and a rescored 20 item HAQ (HAQ20): analyses in 2,491 rheumatoid arthritis patients following leflunomide initiation.

Abstract

OBJECTIVE: To determine whether the full Health Assessment Questionnaire (HAQ), the shortened modified HAQ (MHAQ), or the new shortened RA-HAQ, developed on the basis of Rasch item response theory (IRT), performs best in terms of distributional characteristics, detection of functional loss, and identification of change in functional status in patients with active rheumatoid arthritis (RA). METHODS: A total of 2,491 clinic patients with RA with active disease from the practices of 519 US rheumatologists were assessed by questionnaire at the time leflunomide was started and at subsequent followup when there had been sufficient time for response. RESULTS: The HAQ scores were almost normally distributed along the 0-3 scale, but 95% of MHAQ and RA-HAQ values were clustered between 0 and 1.5. Normal or minimally abnormal scores (0 or 0.125) were noted in 6.6% of HAQ but in 21-22% of MHAQ/RA-HAQ. Mild functional loss (< or =0.375) was found in 12.7, 39.1, and 36.1% of patients by the HAQ, MHAQ, and RA-HAQ, respectively. This indicates that the MHAQ and RA-HAQ generally fail to identify appropriately the extent of functional loss in RA. The HAQ was significantly better at detecting changes than the MHAQ or RA-HAQ, with relative efficiencies of 1.28 and 1.37 compared to the MHAQ and RA-HAQ, respectively. This results in roughly a 20-26% reduction in sample size requirements. Two additional HAQ were identified that performed better than the HAQ itself, a 20 item HAQ without the use of aids and devices and an 8 item HAQ composed of the most difficult item in each of the 8 HAQ subscale categories. CONCLUSION: The HAQ is better (more efficient) than the MHAQ or RA-HAQ at detecting treatment change, and identifies the extent of functional disability better than the shortened questionnaires. The 3 questionnaires have different means, sensitivities, and distributional properties and cannot be thought of as simply different versions of the same questionnaire. The benefits of the MHAQ and RA-HAQ are that they are short and easier to score. But these benefits come at the price of loss of sensitivity and loss of sensitivity to change. The 20 item HAQ and the difficult 8 item HAQ are intriguing additional choices that are worthy of further study.