Development of the Participation Measure--3 Domains, 4 Dimensions (PM-3D4D): A New Outcome Measure for Rehabilitation

Abstract Objectives To describe the development of a participation measure that assesses 3 domains (productivity, social, and community) and 4 dimensions (frequency, diversity, desire for change, and perceived difficulty) of participation and to evaluate the initial psychometric properties in rehabilitation outpatients. Design A mixed-method approach included a literature review, item selection, expert reviews, cognitive interviews, and field testing with rehabilitation outpatients. Confirmatory factor analysis (CFA) and Rasch analysis were used to validate the construct validity of the difficulty dimension of the instrument. Setting Outpatient rehabilitation programs. Participants An expert panel consisting of 12 rehabilitation and measurement experts contributed to measurement development; 20 rehabilitation outpatients participated in cognitive interviews; and a sample of rehabilitation outpatients (N=556) (average age, 61.36±23.62y; 53% men) participated in field testing. Interventions Not applicable. Main Outcome Measure The Participation Measure--3 Domains, 4 Dimensions (PM-3D4D). Results A scoring method for each dimension of the PM-3D4D was established. The instrument displayed good overall model fit in the CFA and unidimensionality across 3 domains after removing and collapsing locally dependent items identified from a principal component analysis. However, considering the poor personal reliability of the social subscale and its high correlation with the community subscale, we decided to merge the 2 subscales into 1. The combined subscale showed improved reliability and good construct validity by demonstrating a good model fit (comparative fit index, .985; Tucker-Lewis Index, .982, root mean square error of approximation, .061) and item fit. Conclusions The PM-3D4D is a newly developed participation measure designed to assess multiple domains and dimensions of participation by rehabilitation patients. The psychometric analysis results supported the construct of the instrument and helped item revision. Further examination of the validity and reliability of the PM-3D4D will be conducted.