Using Rasch Analysis to Evaluate the Psychometric Properties of the Brief MHQ in Patients with Traumatic Finger Amputations

Abstract

Background: Digit amputations affect 45,000 Americans each year and are associated with substantial healthcare expenditures and loss of wages. Few patient-reported outcome measures (PROMs) are validated in patients with digit amputations. The brief Michigan Hand Outcomes Questionnaire (bMHQ) is a 12-item PROM used in several hand conditions. However, its psychometric properties have not been investigated in patients with digit amputations. Methods: The reliability and validity of the bMHQ was investigated using Rasch analysis. Data were collected from the Finger Replantation and Amputation Challenges in Assessing Impairment, Satisfaction, and Effectiveness (FRANCHISE) study. Participants were divided into replantation and revision amputation cohorts and then further separated into single-digit amputation (excluding thumb), thumb-only amputation and multiple-digit amputation (excluding thumb) subgroups. Each of the six subgroups were analysed for item fit, threshold ordering, targeting, differential item functioning (DIF), unidimensionality and internal consistency. Results: All treatment groups demonstrated high unidimensionality (Martin-Löf test = 1) and internal consistency (Cronbach's α > 0.85). The bMHQ is not a reliable PROM in individuals with single-digit or multiple-digit amputations. The aesthetics, satisfaction and two-handed activities of daily living (ADLs) items had the poorest fit to the Rasch model across all categories. Conclusions: The bMHQ is not well-suited for measuring outcomes in patients with digit amputations. We recommend clinicians use more comprehensive assessment tools, such as the complete MHQ, to measure outcomes in these complex patient populations. Level of Evidence: Level III (Diagnostic)