Background The health assessment questionaire (HAQ) in rheumatoid arthritis (RA) has been widely validated as a patient reported outcome measure (PROM). In the United Kingdom regulatory bodies such as the national institute of healthcare and clinical excellence (NICE) have been using improvement in HAQ as a surrogate for efficacy of drugs in RA and has informed their calculation of quality adjusted life year, and calculating health costs using incremental costs effectiveness ratios. Most clinical trials in RA report their primary outcome as improvement in clinical parameteers such as swollen and tender joints, inflammatory markers and patient and physician global assessment of disease. What is not clear how the two sets of parametes interact. Some reports¹ indicate that there is a strong correlation in early disease, but this has not been validated.

Objectives We set out to determine the relationship bewteen HAQ scores and clinical paremeters of the Disaese activity score (DAS28) in addition to physician global.

Methods Patients were recruited from a single centre from the RAMS study in the North west of England. This is a study of patients with early newly diagnosed RA commencing methotrexate A subset of patients filled in the HAQ questionanire and this as used as an outcome variable using linear regression and the swollen joints, tender joints, patient global assessment of disease as well as physician assessment of disease in addition to inflammatory markers were used as explanatory variables. These were then adjusted for age.

Results 81 patients were included in the analysis. median age was 63.1 years (IQR 52.9,72.5). 50 (61.7%) were female. the median HAQ score at baseline was 1 (IQR 0.5,1.5) the median DAS28 score 5.3 (IQR 4.5,6.3). Tender joints at baseline correlated well with HAQ score Beta=0.058 95% CI, 0.04,0.08 (p<0.01). Swollen joints did not correlate with the HAQ beta=0.000 (95%CI −0.3,0.3). Physician global correlated well with disease beta=0.014 (95%CI 0.005,0.022). Patient global assessment also correlated well with HAQ (beta 0.014 95% CI 0.008,0.020). CRP did not correlate with HAQ (beta 0.00261 95% CI −0.002,0.007)

Conclusions In this small study, patient and physician related outcome measures correlate with HAQ scores at baseline more than measures of joint swelling and inflammatory markers. This indicates that using HAQ as an outcome measure underestimates the effect of treatment. When assessing the efficacy of drugs using HAQ this should be taken into account. Validity of the approach needs to be reviewed.

Reference [1] Ann Rheum Dis. 2000Mar;59(3):223–6.

Disclosure of Interest None declared