Meta-analysis compares HAQ-DI, SF36-PF, and SF36-PCS for physical function in psoriatic arthritis trials

OBJECTIVES: To empirically compare the discriminant capacities of three outcome measurement instruments for assessment of physical functioning for psoriatic arthritis (PsA): HAQ-DI, SF36-PF and SF36-PCS. METHODS: We applied a network meta-analysis technique in a sample of randomized trials (RCTs) for PsA. For randomized comparison, we calculated net effect size estimates for each outcome measurement instrument using standardized mean differences (SMDs); positive values indicated a beneficial effect of the intervention compared to the control groups. We analyzed the differences between outcome measurement instruments at the trial level by applying a multiple-treatment meta-analysis to compare the SMDs within and across randomized comparisons for each outcome measurement instrument. RESULTS: From 42 articles (31 RCTs), 57, 18, and 18 randomized comparisons enabled a direct comparison between HAQ-DI and SF36-PCS (difference in SMDs: 0.057, 95 % confidence interval, CI: 0.003 to 0.110), SF36-PF and SF36-PCS (difference in SMDs: 0.101, 95 % CI: 0.018 to 0.184); and HAQ-DI and SF36-PF (difference in SMDs:0.059, 95 % CI:0.142 to 0.024), respectively. The network meta-analysis technique confirmed that both HAQ-DI and SF36-PF were more responsive to change than SF36-PCS, with differences between SMDs of 0.057 (95 % CI: 0.003 to 0.110) and 0.109 (95 % CI: 0.032 to 0.185), respectively. No difference in discriminatory capacity between HAQ-DI and SF36-PF was noted. CONCLUSIONS: HAQ-DI and SF-36-PF were equally responsive to change and superior to SF36-PCS in PsA RCTs. We illustrated a new method for quantitative comparison of the performance of different outcome measurement instruments for a particular domain.