Psychometric validation of the SF-36® Health Survey in ulcerative colitis: results from a systematic literature review

Abstract

Purpose: To conduct a systematic literature review of the reliability, construct validity, and responsiveness of the SF-36® Health Survey (SF-36) in patients with ulcerative colitis (UC). Methods: We performed a systematic search of electronic medical databases to identify published peer-reviewed studies which reported scores from the eight scales and/or two summary measures of the SF-36 collected from adult patients with UC. Study findings relevant to reliability, construct validity, and responsiveness were reviewed. Results: Data were extracted and summarized from 43 articles meeting inclusion criteria. Convergent validity was supported by findings that 83% (197/236) of correlations between SF-36 scales and measures of disease symptoms, disease activity, and functioning exceeded the prespecified threshold (r ≥ |0.40|). Known-groups validity was supported by findings of clinically meaningful differences in SF-36 scores between subgroups of patients when classified by disease activity (i.e., active versus inactive), symptom status, and comorbidity status. Responsiveness was supported by findings of clinically meaningful changes in SF-36 scores following treatment in non-comparative trials, and by meaningfully larger improvements in SF-36 scores in treatment arms relative to controls in randomized controlled trials. The sole study of SF-36 reliability found evidence supporting internal consistency (Cronbach’s α ≥ 0.70) for all SF-36 scales and test–retest reliability (intraclass correlation coefficient ≥0.70) for six of eight scales. Conclusions: Evidence from this systematic literature review indicates that the SF-36 is reliable, valid, and responsive when used with UC patients, supporting the inclusion of the SF-36 as an endpoint in clinical trials for this patient population.