The SF-36 walk-wheel: a simple modification of the SF-36 physical domain improves its responsiveness for measuring health status change in spinal cord injury.

OBJECTIVE

To evaluate the validity and responsiveness of a modified SF-36 within a spinal cord-injured (SCI) population.

STUDY DESIGN

SF-36 scores collected at baseline and on completion of a randomized controlled trial in 305 patients with SCI and neuropathic bladder.

SETTING

New South Wales, Australia.

METHODS

Subjects were administered the standard SF-36 plus three additional questions, in which 'walk' was replaced with 'wheel' for three of the physical function (PF) questions. Discriminant validity was determined by comparing participants with paraplegia and tetraplegia using the effect size (ES). Responsiveness was assessed in the subset of patients who developed a urinary tract infection (UTI) during the trial using the standardized response mean (SRM).

RESULTS

Compared with the standard SF-36, the SF-36 walk-wheel modification (SF-36ww) increased the mean PF score from 18 to 39 (P<0.001) and the physical composite score from 33 to 37 (P<0.001). Discriminant validity was similar for both versions (PF paraplegia/tetraplegia: ES 1.09(SF-36) vs 1.08(SF-36ww), n=305). Among 138 SCI patients who developed a UTI, the SF-36ww almost doubled PF responsiveness for all neurological levels (SRM increased from 0.36 to 0.68), more so in tetraplegic (SRM, 0.11 vs 0.58; n=77) than paraplegic groups (SRM, 0.77 vs 0.86; n=61).

CONCLUSION

The SF-36ww is a simple, pragmatic modification of the SF-36 PF items, which addresses some problems of content validity and floor effect for SCI subjects and greatly improves responsiveness, particularly for those with tetraplegia. Because it comprises a simple addition to the standard SF-36, external comparisons are preserved.