Determining the minimal clinically important difference of the hand function sort questionnaire in vocational rehabilitation.

To estimate the Minimal Clinically Important Difference (MCID) of the French version of the Hand Function Sort questionnaire (HFS-F). As a comparison, the MCID of the Disabilities of the Arm, Shoulder, and Hand (DASH) was also estimated. We included French-speaking patients hospitalized in a multidisciplinary rehabilitation program for chronic pain of the upper limb after an accident. HFS-F and DASH scores were collected at admission and discharge; the Patient Global Impression of Change measure (PGIC; 7 levels) was collected at discharge. The MCID was estimated by 2 methods: the anchor-based method (receiver operating characteristic [ROC], delta (Δ) mean of scores) and the objective method based on the distribution of scores (standard error of measurement, SEM). We included 225 patients. By the anchor-based method, the MCID for the HFS-F and DASH was +26 (SD 35) (P<10 <sup>-4</sup> ) and -13 (SD 13) (P<10 <sup>-4</sup> ), respectively, and by the ROC curve, it was +10 to +12 for the Δ-HFS-F and -7.5 to -5 for the Δ-DASH. The area under the ROC curve (AUC) was 0.726 [0.638-0.781] for Δ-HFS-F and 0.768 [0.701-0.83] for Δ-DASH. The correlations between the anchor and delta scores were>0.38 (P<10 <sup>-4</sup> ). The SEM was 16.2 for the HFS-F and -4.3 for the DASH. Values below the SEM must be rejected. Our anchor was significantly correlated with the outcome. Therefore, we propose an MCID for the HFS-F of 26, corresponding to approximately 10% progression of the score