Reproducibility of performance measures of physical function based on the BASFI, in ankylosing spondylitis.

OBJECTIVE To establish the test-retest reproducibility of performance measures of physical function based on the Bath Ankylosing Spondylitis Functional Index (BASFI) questionnaire in patients with AS. METHODS Data were obtained from 65 AS patients. They were tested on two occasions by one assessor with a 1-week interval. Physical function was assessed via eight performance measures based on items used in the BASFI questionnaire, representing activities of daily life, which AS patients frequently report to be problematic. For each activity, a performance score was determined. Pain and exertion were measured using a 10-cm horizontal visual analogue scale (VAS) and Borg's modified scale, respectively. Test-retest reproducibility was assessed for all measurements using intraclass correlation coefficients (ICCs) and by calculating the standard error of measurement (SEM). RESULTS Adequate intrarater reliability was found. For performance scores, ICCs ranged from 0.73 to 0.96. Measurements of exertion and pain also showed adequate intrarater reliability, with the exception of one performance measure, namely the test for the ability to look over one's shoulder. For this test, the ICCs were 0.66 and 0.69 for exertion and pain, respectively. The remaining ICCs for exertion ranged from 0.71 to 0.88 and for pain from 0.74 to 0.83. The SEM for performance scores ranged from 4 to 9% of the observed score. The SEM for exertion ranged from 8 to 11% and for pain from 10 to 15%. CONCLUSIONS Performance measures of physical function based on the BASFI questionnaire have adequate to excellent test-retest reproducibility. Due to the presence of measurement error, measurements are accurate for group assessment; repeated measurements are advised for an adequate assessment of individual patients.