Testing scaling assumptions, reliability and validity of medical outcomes study short-form 36 health survey in psoriatic arthritis.

OBJECTIVES The Medical Outcome Survey short-form 36 (SF-36) is widely used in studies of PsA. We test the fundamental scaling assumptions for the validity of SF-36 eight scales and two summary scores in PsA. METHODS We tested the five scaling criteria according to the International Quality of Life Assessment project and the validity of the physical component summary (PCS) and mental component summary (MCS) scores. The construct validity for SF-36 in PsA was evaluated. RESULTS A total of 168 PsA subjects according to CASPAR criteria (46.4% females) with SF-36 data were analysed. The mean (S.D.) age and duration of illness were 47.7 (11.9) and 8.4 (7.3) years. PsA subjects had poorer quality of life as compared with normal population. The norm-based mean (S.D.) PCS and MCS were 31.6 (14.2) and 45.2 (12.7), respectively. The Pearson's correlations between an item and its hypothesized scale were all >0.4, indicating item internal consistency. All item-own scale correlations exceed item-other scale correlations, except minor scaling failure in general health. The Cronbach's alpha-coefficients of internal reliability (Cronbach's alpha) were all above the standard value of 0.7, indicating a unique concept in each scale. Principal component analysis explained 69.4 and 73.7-99.9% of the total variance and total reliable variance in all SF-36 scales. The two-factor model is supported. CONCLUSIONS PsA had high impact on quality of life. All eight scales and two summary scores of SF-36 satisfied criteria for scaling assumption. SF-36 is a valid measurement for quality of life in PsA. The reporting of PCS and MCS in PsA is supported.