Good inter-rater reliability of the Frenchay Activities Index in stroke patients.
OBJECTIVE: To assess inter-rater reliability of the Frenchay Activities Index (FAI) when used by occupational therapists in stroke patients. DESIGN: Independent administration of the FAI by occupational therapists in 45 stroke outpatients. SETTING: Outpatient departments of two Dutch rehabilitation centres. STATISTICS: Agreement between pairs of raters was assessed at item level using weighted Kappa and the percentage absolute agreement and for the total score using intra-class correlations (ICC), Bland-Altman plots and computation of the smallest real difference (SMR). RESULTS: ICC of the total score was good (0.90; 95% confidence interval (95% CI): 0.82-0.94). The Bland-Altman plot did not reveal relationships between agreement and height of scores. Reliability at item level was good (kappa >0.60) in 11 out of 15 items. The item 'gainful work' showed high agreement but a low kappa due to an extremely skewed score distribution. Reliability was insufficient in three items: 'local shopping', 'social occasions' and 'actively pursuing hobby'. Clarification of scoring instructions of these items might further improve the inter-rater reliability of the FAI. CONCLUSION: The FAI is a reliable instrument to measure outcomes of outpatient rehabilitation in patients with stroke.