Test-retest reliability and responsiveness of the whiplash disability questionnaire
Objective The aim of this dissertation is to provide a literature review on the key features of whiplash injuries, and review assessment tools which have been used to measure disability within whiplash populations. This dissertation includes a pilot study, which was completed to further investigate the test-retest reliability and responsiveness of the Whiplash Disability Questionnaire (WDQ), with the aim of providing evidence as to whether the WDQ is a useful measurement tool for clinical practice. Study Design The pilot study was a questionnaire based, cross-sectional, longitudinal study of patients with whiplash associated disorders. Background The Whiplash Disability Questionnaire was developed specifically for use with subjects suffering from whiplash associated disorders (WAD). Analysis of published investigations to date has provided evidence that the WDQ has good internal consistency, content, construct and face validity (Pinfold et al., 2004), excellent reliability, but questionable responsiveness (Willis et al., 2004). Methodology Subjects presenting for treatment of WAD symptoms (n = 19) were recruited from a selection of private practices, private hospitals and public hospitals in London, United Kingdom. The subjects were asked to fill out three WDQ’s. The first (WDQ1) was completed prior to the first treatment and the second questionnaire (WDQ2) 24 hours later. The third questionnaire (WDQ3) was completed at the cessation of the course of treatment. At this time the subject and therapist were also asked to fill out a Global Discharge Questionnaire, which rated the level of perceived change in symptoms following treatment. Reliability was analysed by calculating an intraclass correlation coefficient (ICC), typical error, Bland and Altman plots and limits of agreement. A kappa statistic was used to assess the level of agreement between therapist and subject on the Global Discharge Questionnaire. To measure responsiveness effect size, standardized response mean (SRM) were calculated. Results Results showed excellent reliability (ICC = 0.96, typical error = 6.4 points, Bland and Altman bias of 0.63, limits of agreement = -17 to 18 points). Responsiveness was also shown to be very strong (effect size = 1.2, SRM = 1.9). Conclusion These results support previous reports of strong reliability and enhance previously reported levels of responsiveness, deeming the WDQ to be a sensitive, reliable tool for clinicians to use as a measure of disability in subjects suffering from WAD.