ABSTRACT
BACKGROUND: When the Michigan Hand Questionnaire (MHQ) was originally developed, an exploratory factor analysis (EFA) was used to reduce the originally large number of generated items to the 63 items currently present on the questionnaire. Confirmation of the implied factor model of the existing MHQ has never been performed. The objective of this study was to confirm the factor model used to create the existing MHQ, and to possibly shorten the existing MHQ using factor analysis. METHODS: Patients attending the Plastic Surgery Clinic at the QEII Health Sciences Centre with a hand complaint were asked to complete the MHQ. Confirmatory factor analysis was performed to explore the implied factor structure of the original EFA and to examine the interplay between the MHQ subscales. Further item-reduction was performed using clinically guided decisions as well as factor analysis-guided statistics. RESULTS: Initial confirmatory factor analysis showed that original EFA model does not optimally explain the relationships between items in the existing MHQ and their corresponding factors. Our abbreviated model of the MHQ consists of 23 items, and performed more favorably in all goodness-of-fit parameters than the original 63-item questionnaire. CONCLUSIONS: The factor model of the existing MHQ does not fully take advantage of the relationship between items in the MHQ and the proposed factors. This study proposes a shortened version of the MHQ that more accurately reflects hand health as well as a factor-based interpretation of the subscales that takes interdependent relationships into account.
Subject(s)
Disability Evaluation , Hand Injuries/diagnosis , Health Status Indicators , Models, Theoretical , Surveys and Questionnaires , Adult , Factor Analysis, Statistical , Female , Hand Injuries/physiopathology , Hand Injuries/psychology , Humans , Male , Middle Aged , PsychometricsABSTRACT
BACKGROUND: The Michigan Hand Outcomes Questionnaire (MHQ) remains one of very few to be based on validated and systematic methods in its creation. However, test-retest reliability and internal validity have not been appropriately investigated in any other English-speaking population outside the original development sample. The objective of this study is to examine the reliability and internal validity of the MHQ in a Canadian population. METHODS: All patients with a clinical hand problem but no hand immobilization who attended our Plastic Surgery Clinic were invited to participate in this study. Patients were asked to complete the MHQ, and then contacted to arrange for completion of the MHQ a second time. Internal validity was estimated using Cronbach alpha. Reliability was estimated using the method of Bland and Altman, and intraclass correlation coefficients. Correlation analysis between score differences and time between tests was performed. RESULTS: Full versions of the MHQ were completed by 116 participants on their initial visit, while only 77 participants completed a second MHQ. Cronbach alpha ranged from 0.84 to 0.95 for the MHQ subscales. Intraclass-correlation coefficients ranged from 0.71 to 0.84 for the subscales. The magnitude of the limits of agreement for the subscales ranged from 13.8 to 26.2. The difference in scores between the first and second tests was not related to the time between tests. CONCLUSION: The high values of Cronbach alpha indicate high internal validity, but that there may be redundancy between items in the MHQ, which could be eliminated without losing information. Reliability of the MHQ is considered moderate. The limits of agreement are considered wide, indicating potential issues with respect to interpretation of a change in score between 2 administrations of the MHQ.
