A Psychometric Study of the Model of Human Occupation Screening Tool (MOHOST)
Abstract
Objective: This study examined the psychometric properties of the Model of Human Occupation Screening Tool (MOHOST) using item response theory and classical test theory approaches for clients with psychiatric disorders. Methods: Data, including demographic variables and scores on the MOHOST and a version of the Health of the Nation Outcomes Scale, were retrieved from case records of 1039 adult psychiatric service users. Results: Participants ranged in age from 18 to 102 and 57% were female and 43% were male. Most (94%) were unemployed, retired, or receiving other education or training. The items that make up each of the MOHOST subscales demonstrated good discriminant validity and excellent goodness of fit showing that the items measured the MOHO constructs unidimensionally. All subscales were able to distinguish clients into at least three statistically distinct strata and showed convergence with an independent measure of functioning. Conclusion: Findings from this study must take into account implicit limitations associated with the use of Rasch analysis and classical test theory. At the same time, results did support use of the MOHOST for research and clinical purposes. The MOHOST demonstrated good construct validity, item separation reliability, and concurrent validity. As a measure of occupational participation, the MOHOST offers practitioners and researchers a valid and reliable measure of volition, habituation, communication/interaction skills, process skills, motor skills, and environmental influences on participation.