Determining the Michigan Hand Outcomes Questionnaire minimal clinically important difference by means of three methods
London, D. A.; Stepan, J. G.; Calfee, R. P.
Plastic & Reconstructive Surgery
BACKGROUND: To interpret patient-rated outcome measures, clinicians rely on the minimal clinically important difference. The authors studied the range of minimal clinically important difference scores for the overall Michigan Hand Outcomes Questionnaire score its and subscales in a population with various diagnoses and treatments.; METHODS: Patients with a single, unilateral, atraumatic hand/forearm diagnosis completed the Michigan Hand Outcomes Questionnaire before treatment and at 4 +/- 1 weeks and 12 +/- 2 weeks after treatment. Three methods were used to calculate the minimal clinically important difference: two anchor question methods based on satisfaction (mean change and receiver operating characteristic) and a statistical distribution method.; RESULTS: One hundred eighty-six patients were enrolled, with a baseline median overall questionnaire score of 60.7. Using the mean change method, a minimal clinically important difference of 13 was calculated for the overall questionnaire score. Using a receiver operating characteristic curve (0.8 effect size), the minimal clinically important difference was 9, with an area under the curve of 0.92. When receiver operating characteristic analysis was based on dichotomization of reported satisfaction, the minimal clinically important difference was 12, with an area under the curve of 0.85. Calculating the difference by statistical distribution gave a value of 8. For the subscales, by all methods, triangulated minimal clinically important difference estimates ranged from 10.9 to 14.4.; CONCLUSIONS: The minimal clinically important difference for the overall Michigan Hand Outcomes Questionnaire score in atraumatic hand/forearm conditions falls between 8 and 13. Multiple analytic methods produce nonidentical but similar minimal clinically important differences. The authors recommend using difference estimates in these ranges when planning a clinical trial to investigate hand/forearm function across a range of diagnoses and treatments.