Is it possible to determine the minimal clinically important difference (MCID) of the French version of the hand function sort (HFS-F) for patients hospitalized in musculoskeletal rehabilitation?

OBJECTIVE: The HFS is a pictorial questionnaire with 62 items; it is a self-report functional capacity evaluation of the upper limb [1]. The MCID is important in assessing the effectiveness of a therapy. It has not been estimated for HFS-F [2]. The aim of this study was to estimate the MCID of the HFS-F for patients hospitalized in musculoskeletal rehabilitation for chronic pain of the upper limb. As a comparison, the MCID of the DASH (disabilities of the arm, shoulder and hand) was also estimated. MATERIAL/PATIENTS AND METHODS: French speaking patients (18-65 years), hospitalized from January 1, 2012 to June 30, 2015, various pathology of upper limb in the aftermath of an accident. The pain has at least lasted three months. REPORTS: of HFS-F scores and DASH at the entrance and exit, of the global scale of change (Likert 7 levels) at the exit. The MCID was estimated using two methods: the subjective feeling of patient (ANOVA-ROC) and the objective method based on the distribution of scores (standard error of measurement: SEM). RESULTS: Two hundred and twenty five patients were enrolled, 82% men, age 43±12 years, 65% proximal damage (shoulder, elbow), 35% distal damage (hand-wrist). The difference of the scores in subjectively improved patients was 26/248 (ANOVA, P<10(-4)), the values of sensitivity/specificity were 0.51-0.81 for the threshold values of MCID between 25/248 and 30/248 (area under the ROC curve (AUC) =0.72 [0.65-0.78]). The SEM gave a value of 28/248. The difference in the DASH scores in subjectively improved patients was-12/100 (P<10(-4)), corresponding to the MCID commonly accepted for this questionnaire [3], the values of sensitivity/specificity were 0.25-0.54 for DASH values of-13/100 at-11/100 (AUC=0.31 [0.24-0.37]). DISCUSSION - CONCLUSION: Both used methods are consistent to propose a MCID forHFS-Fbetween 25/248 and 30/248, corresponding to 11% improvement of the score. This estimate is useful in clinical practice. In this sample, the DASH seems less relevant to determine patients subjectively improved