The SF-36 as a health status measure for epilepsy: a psychometric assessment.

Abstract

A considerable literature is now available on the applications and psychometric properties of the MOS SF-36 Health Survey. In epilepsy, the SF-36 has been used as a health status measure in its own right and as the stem for two condition-specific measures. This paper replicates for epilepsy previous work to support use of the SF-36 across a range of clinical conditions. Data were obtained from a European-wide descriptive study of quality of life of adults with epilepsy; analysis of responses on the SF-36 is based on 4,929 subjects in eight countries. Missing value rates for all SF-36 items were low; percentages for whom complete information available for subscales ranged from 95.7% to 98.6%. All subscales passed tests for item-internal consistency and item-discriminant validity. Reliability coefficients exceeded the standard recommended for group comparisons across all subscales. Floor effects were negligible for all but the two role disability subscales; there were substantial ceiling effects for five of the SF-36 subscales. We conclude that the SF-36 is a valid and reliable health status measure for descriptive studies of people with epilepsy, but ceiling effects may limit its usefulness as an outcome measure in the assessment of new treatments.