Classifying post-stroke fatigue: Optimal cut-off on the Fatigue Assessment Scale

Objective Post-stroke fatigue is common and has debilitating effects on independence and quality of life. The Fatigue Assessment Scale (FAS) is a valid screening tool for fatigue after stroke, but there is no established cut-off. We sought to identify the optimal cut-off for classifying post-stroke fatigue on the FAS. Methods In retrospective analysis of two independent datasets (the ‘2015’ and ‘2007’ studies), we evaluated the predictive validity of FAS score against a case definition of fatigue (the criterion standard). Area under the curve (AUC) and sensitivity and specificity at the optimal cut-off were established in the larger 2015 dataset (n = 126), and then independently validated in the 2007 dataset (n = 52). Results In the 2015 dataset, AUC was 0.78 (95% CI 0.70–0.86), with the optimal ≥ 24 cut-off giving a sensitivity of 0.82 and specificity of 0.66. The 2007 dataset had an AUC of 0.83 (95% CI 0.71–0.94), and applying the ≥ 24 cut-off gave a sensitivity of 0.84 and specificity of 0.67. Post-hoc analysis of the 2015 dataset revealed that using only the 3 most predictive FAS items together (‘FAS-3’) also yielded good validity: AUC 0.81 (95% CI 0.73–0.89), with sensitivity of 0.83 and specificity of 0.75 at the optimal ≥ 8 cut-off. Conclusion We propose ≥ 24 as a cut-off for classifying post-stroke fatigue on the FAS. While further validation work is needed, this is a positive step towards a coherent approach to reporting fatigue prevalence using the FAS.