Dimensions of treatment engagement among youth and caregivers: Structural validity of the REACH framework.

Abstract

Although the literature on treatment engagement varies in its characterization and enumeration of the relevant dimensions, the dimensionality of treatment engagement has yet to be tested empirically using a uniform measurement approach. We therefore examined the structural validity of a hypothesized five-factor model, using a confirmatory factor-analytic approach applied to youth and caregiver reports of their own perceived level of engagement. Data were obtained from 1,807 primarily Hispanic American (56.0%) and African American/Black (26.3%) youth (Mage = 12.7; 46.8% female) and/or their caregivers participating in school mental health services in Los Angeles, California, and rural South Carolina. Participants (N youth records = 1,415; N caregiver records = 1,361) rated 35 self-report indicators of treatment engagement, hypothesized to represent five REACH dimensions (Relationship, Expectancy, Attendance, Clarity, and Homework), approximately 4 weeks following an intake assessment. Results uniformly supported the hypothesized five-factor models relative to one-factor, youth χdiff(10)² = 2,092.96, p < .001; caregiver χdiff(10)² = 4,570.93, p < .001, and four-factor, youth χdiff(4)² = 225.15, p < .001; caregiver χdiff(4)² = 843.06, p < .001, alternative models. Modification indices and expected change coefficients did not indicate substantive points of strain in the five-factor models, and tests of model invariance uniformly supported the REACH structure across youth age, youth race, region, and caregiver language. Findings supported a five-factor structure that appears to generalize well across multiple groups, and they set the stage for advances in measurement and improved conceptualization of treatment engagement in research and clinical care. (PsycInfo Database Record (c) 2022 APA, all rights reserved).