There are a large variety of care management approaches designed to control utilization and spending for higher complexity patients, particularly those with chronic diseases, while maintaining quality outcomes. A study in the journal Health Services and Outcome Research Methodology deals both with approaches to analyzing the results of such care management programs, as well as an actual evaluation of certain programs. (HSORM Article) One fundamental difference in structure is between programs that are driven by payers and implemented through a payer’s staff or contracted vendors and those that are implemented through a provider. For this study, the researchers used a large health plan that had its own internal care management programs, but also had initiated a pilot program in which a subset of the plan’s network providers could do the care management. These providers tended to adopt one of two models for their care management. The first was use of a centralized staff which interacted with the selected patients, primarily by phone. The second involved care managers who were embedded as part of the care teams and worked directly with patients and the care team. The researchers analyzed the utilization and cost effects of these programs.
There were around 500 patients in the centralized provider-based model, 750 in the embedded care manager provider-based model, and 3565 in the health plan care management model. Claims data, risk scores and certain provider practice indicators were used as data sources. Quarterly per member per month costs were the primary outcome and were examined from one and a half years before the patient’s engagement in the program to up to 4 years after such engagement. At a high-level, the results suggest that an provider embedded care manager model had lower spending and the savings increased over time, while the centralized provider model had initial savings but those savings declined over time, and the health plan care management effort showed lower savings than the provider-based models. The primary reason for the better performance of the embedded care manager model is hypothesized to be more patient engagement resulting in greater behavior change. This seems consistent with common sense and the likely preferences of many patients.