Hand rehabilitation needs valid evaluation tools; the 400-point Hand Assessment (HA) is an exhaustive but not standardised tool. The aim of this study was to validate a standardised version of this test.
A modified version and a standardised prototype was made for this prospective validation study (four centres, three countries). Psychometric properties studied: reliability (intra-rater and inter-rater, standard error of measurement [SEM], minimum detectable change [MDC],internal consistency); content validity, construct validity with Jebsen Taylor hand function test, QuickDASH, MOS-SF 36 and pain; responsiveness, using an anchor-based approach (ROC curve with area under curve, mean response change) with calculation of MCID. For SEM, MDC and responsiveness, QuickDASH was used for comparison.
One hundred and seventy-six patients with hand/wrist injuries were included between May 2013 and February 2015. One hundred and seventy were available for final analysis: 67% men; mean age 43.4 ± 13.2 years; both manual and office workers (46, 5% of each); 37% had a hand or wrist fracture. Reliability: ICC intra-rater = 0.967 [0.938-0.982]; inter-rater = 0.868 [0.754-0.932]. Distribution-based approach: for 400-point HA/QuickDASH: SEM = 3.48/4.52, MDC = 9.065/12.53, internal consistency of 400-point HA: Cronbach α = 0.886.
Content validity was good according to COSMIN guidelines. Construct validity: correlation coefficient: Jebsen-Taylor hand function test = - 0.573 [- 0.666-0.464], QuickDASH = - 0.432 at T0 [- 0.545-0.303], - 0.551 at T3 [- 0.648-0.436]; MOS-SF 36 physical component = 0.395 [0.263-0.513]; no correlation with MOS-SF 36 mental component = 0.142 [- 0.009 + 0.286] and pain = - 0.166 [- 0.306 + 0.018]. Responsiveness: Anchor-based approach: AUC Δ400-point HA = 0.666 [0.583-0.749], AUC ΔQuickDASH = 0.556 [0.466-0.646]. MCID (optimal ROC curve cut-off): 6.07 for 400-point HA, - 2.27 for QuickDASH. MCID with mean response change + 12.034 ± 9.067 for 400-point HA and - 8.03 ± -9.7 for QuickDASH. The patient's global impression of change was only correlated with the Δ400-point HA.
The 400-point HA standardised version has good psychometric properties. For responsiveness, we propose an MCID of at least 12.3/100. However, these results must be confirmed in other populations and pathologies.
This study was retrospectively registered into ISCTRN registry (Number ISRCTN25874481) the 07/02/2019.