Staff ratings of occupational engagement among people with severe mental illness – psychometric properties of a screening tool in the day center context

Abstract

BackgroundStaff who plan and organize day center activities may need to observe the attendees’ performance and progression. This led us to develop a tool for that purpose, termed General Occupational Engagement in people with Severe mental illness (GOES). The aim was to investigate its psychometric properties in terms of factor structure, internal consistency, corrected item-total correlations (CITC), convergent and discriminant validity, and test-retest stability.MethodsNinety-three day center attendees were assessed by the GOES and instruments addressing constructs hypothesized to be either similar to (activity level, motivation for day center attendance, perceptions of the worker role, hours spent in the day center) or divergent from the GOES (attendees’ ratings of engagement in specified occupations, self-rated health, psychosocial functioning, psychiatric symptoms). A second sample of 41 attendees were included for the test-retest analysis. Exploratory factor analysis, Cronbach’s alpha analysis, Pearson correlations and paired-samples t-tests were performed.ResultsExploratory factor analysis indicated one factor, which was in line with the intentions of the scale. The alpha value was 0.85 and all CITC were above 0.30. The tests for convergent validity resulted in correlations ranging between 0.23 and 0.47, most of which were moderately strong and mainly confirmed the hypotheses. Discriminant validity was clearly indicated, since all correlations with the selected constructs were <0.20. GOES also showed preliminary test-retest stability (r = 0.32).ConclusionsThe GOES is ready for use in rehabilitation services and research where productive and other types of activities are of interest. It may serve as an important supplement to attendees’ self-reported occupational engagement.

Functional recovery of individuals with serious mental illnesses: Development and testing of a new short instrument for routine outcome monitoring.

“…Convergent validity is assumed if the expected association is found between two measures that are expected to point in the same direction (Westen & Rosenthal, 2003). Discriminant validity is assumed if there is a weak correlation between two measures that are expected to be dissimilar due to differences in the constructs measured or in the measuring procedures—for instance, if the measure is a self-report instrument or is based on an observer’s rating (Eklund & Leufstadius, 2007; Eklund & Bejerholm, 2017).…”