Use of the short form health survey (SF-36) in patients with amyotrophic lateral sclerosis: Tests of data quality, score reliability, response rate and scaling assumptions

Abstract

Objectives: To evaluate response rate, data quality, score reliability and scaling assumptions of the 36 item Short Form Health Survey (SF-36) in a large scale pan European survey of amyotrophic lateral sclerosis (ALS) patients. Design: A questionnaire based survey of patients diagnosed with ALS across 15 European countries. Sample: Patients presenting at neurological clinics for treatment of their condition were asked to partake in the survey. Results: 948 patients have been recruited into the survey, from whom responses have been gained in 754 (79.5%). Scores on the eight dimensions of the SF-36 were found to manifest high internal consistency reliability. Items were, in most instances, found to be most highly correlated with their own (corrected) scale score than with other scale scores. However, on two dimensions (role-physical and role-emotional) there was high levels of missing data, together with substantial floor and ceiling effects. The two factor model (of underlying constructs of physical and emotional health) for the SF-36 suggested by the developers was not supported in this patient group. Conclusion: The SF-36 appears to provide reliable information for this patient group, and for the most part there are high levels of item completeness and good spread of scores. This is not, however, true for the role functioning dimensions. Furthermore, the underlying two factor model for the SF-36 was not supported. The implications for measuring health status in this patient group are discussed.