Background and aims: The aim of this study was to develop and evaluate the UK version of the Functional Assessment Measure (UK FIM+FAM).
Design: Before and after evaluation of inter-rater reliability.
Development: Ten 'troublesome' items in the original FIM+FAM were identified as being particularly difficult to score reliably. Revised decision trees were developed and tested for these items over a period of two years to produce the UK FIM+FAM.
Evaluation: A multicentre study was undertaken to test agreement between raters for the UK FIM+FAM, in comparison with the original version, by assessing accuracy of scoring for standard vignettes.
Methods: Baseline testing of the original FIM+FAM was undertaken at the start of the project in 1995. Thirty-seven rehabilitation professionals (11 teams) each rated the same three sets of vignettes - first individually and then as part of a multidisciplinary team. Accuracy was assessed in relation to the agreed 'correct' answers, both for individual and for team scores. Following development of the UK version, the same vignettes (with minimal adaptation to place them in context with the revised version) were rated by 28 individuals (nine teams).
Results: Taking all 30 items together, the accuracy for scoring by individuals improved from 74.7% to 77.1% with the UK version, and team scores improved from 83.7% to 86.5%. When the 10 troublesome items were taken together, accuracy of individual raters improved from 69.5% to 74.6% with the UK version (p <0.001), and team scores improved from 78.2% to 84.1% (N/S). For both versions, team ratings were significantly more accurate than individual ratings (p <0.01). Kappa values for team scoring of the troublesome items were all above 0.65 in the UK version.
Conclusion: The UK FIM+FAM compares favourably with the original version for scoring accuracy and ease of use, and is now sufficiently well-developed for wider dissemination.