Predictors of Patient Engagement in a Care Coordination Program

To identify predictors of patient engagement in a care coordination program for individuals with work disabilities, using a continuum-based definition of engagement and the PROGRESS-Plus framework.

This observational analysis used deidentified data from the Vermont RETAIN Phase II program, a cluster-randomized controlled trial testing the impact of early delivery of evidence-based stay-at-work (SAW) and return-to-work (RTW) strategies on employment outcomes of individuals with work-limiting physical or mental health conditions recruited from the primary care setting. We included 429 intervention participants with complete data on patient engagement and relevant predictors in this analysis. Based on care manager assessment, we categorized patient engagement into four ordinal levels (none, low, medium, and high). We conducted bivariable analyses to identify potential predictors of engagement, followed by multivariable ordinal logistic regression. Variables were defined according to the PROGRESS-Plus framework, encompassing sociodemographic, clinical, and work-related characteristics.

Having an injury or illness-related absence from work significantly predicted higher odds of engagement (OR = 2.87, 95% CI: 1.81-4.56), while unemployment predicted lower engagement (OR = 0.43, 95% CI: 0.27-0.63) compared to employer-based job. Participants with a high school education or less (OR = 0.40, 95% Cl: 0.23-0.69), followed by college education (OR = 0.52, 95% Cl: 0.31-0.88) engaged less than those with post-graduate education. Also, poor or fair self-rated health (OR = 1.87, 95% Cl: 1.31-2.68) and age (OR = 1.02 per year, 95% CI: 1.01-1.04) significantly predicted higher engagement odds. Multivariate model diagnostics supported the proportional odds assumption (χ²(14) = 18.5, p = 0.183).

These findings highlight the importance of tailored work disability prevention strategies based on health, work, and educational profiles to optimize patient engagement.