Inter-rater reliability of the Frenchay Activities Index in patients with stroke and their carers

Publisher Website
Google Scholar
Abstract
Objective: To measure the inter-rater reliability of the interview-administered version of the Frenchay Activities Index (FAI). Design: Comparison of FAI score on the same person when administered by two raters (mean time between interviews 15.2 days). Subjects: Fifty-nine Oxfordshire residents who either had had a stroke (n = 35) or were the main carer (n = 24). Results: The 95% limits of agreement for the FAI totals were –9.9 to +8.4. The kappa statistic for nine of the 15 items showed a good level of agreement between the two research interviews (0.64–0.80). The other six items showed fair or moderate strength of agreement (0.26–0.52). Three items showed significant differences between the two raters p < 0.05 (Wilcoxon's sign paired rank sum test). The mean difference between the total scores was –0.76 (95% confidence interval from –1.98 to 0.46). Spearman's rho correlation coefficient for FAI totals of rater B against A was r(59) = 0.93 (p < 0.001). Conclusion: The FAI is a reliable tool for measuring outcome following stroke. Suggestions are made to strengthen the reliability, and consequently the validity of the measure.