Patient Activation Changes as a Potential Signal for Changes in Health Care Costs: Cohort Study of US High-Cost...

Programs to improve quality of care and lower costs for the highest utilizers of health services are proliferating, yet such programs have difficulty demonstrating cost savings.

In this study, we explore the degree to which changes in Patient Activation Measure (PAM) levels predict health care costs among high-risk patients.

De-identified claims, demographic data, and serial PAM scores were analyzed on 2155 patients from multiple medical groups engaged in an existing Center for Medicare and Medicaid Innovation-funded intervention over 3 years designed to activate and improve care coordination for high-risk patients.

In this prospective cohort study, four levels of PAM (from low to high) were used as the main predictor variable. We fit mixed linear models for log₁₀ of allowed charges in follow-up periods in relation to change in PAM, controlling for baseline PAM, baseline costs, age, sex, income, and baseline risk score.

Total allowed charges were derived from claims data for the cohort. PAM scores were from a separate database managed by the local practices.

A single PAM level increase was associated with 8.3% lower follow-up costs (95% confidence interval 2.5-13.2%).

These findings contribute to a growing evidence base that the change in PAM score could serve as an early signal indicating the impact of interventions designed for high-cost, high-needs patients.