New Study Finds that CareFirst’s Medical Home Model Needs Refinement Before Expansion to Medicare Beneficiaries

Aug 02, 2017

In a new edition of JAMA Internal Medicine, Mathematica’s Greg Peterson and co-authors found that extending CareFirst’s medical home model to Medicare beneficiaries did not reduce hospitalizations or Medicare spending as intended. The authors conclude that the model needs further refinement and testing before being scaled to Medicare beneficiaries more broadly.

CareFirst, the largest health insurer in the Mid-Atlantic region of the United States, runs a medical home program that provides financial incentives to primary care practices and care coordination for high-risk patients. From 2013 to 2015, CareFirst extended the program to Medicare fee-for-service beneficiaries in participating practices. Nurses hired by CareFirst worked with primary care providers to develop individual care plans and coordinate care for 3,500 high-risk Medicare beneficiaries. CareFirst paid incentives to practices that met quality and cost targets for their Medicare patients and provided technical assistance to help practices meet those aims.

The Mathematica study, an independent evaluation of the model funded by the Centers for Medicare and Medicaid Services, addressed the question of whether or not the program improved quality of care and reduced hospitalizations, emergency department visits, and spending for Medicare patients. In a difference-in-differences analysis with intervention practices and matched comparison practices, the program was not associated with outcome improvements for Medicare patients. All-cause hospitalizations did decline by 10 percent, but this was matched by similar changes in the comparison group, suggesting outside market factors drove the decline.  As a result, the program did not produce Medicare savings to offset the cost of the program. These results contrast with other recent studies of the commercial—not Medicare—populations that found that CareFirst program reduced spending at least enough to offset the cost of the interventions.

Mathematica's Greg Peterson, Ph.D.
Mathematica's Greg Peterson, Ph.D.

“The CareFirst model would need further refinement, and testing of these refinements, before being rolled out more broadly to Medicare beneficiaries,” Peterson said. “The needs and opportunities for Medicare populations differ from those for the commercial populations, and further adaptions of the model to the Medicare population might make it more effective.” For example, CareFirst could adapt its strategy for targeting practices’ Medicare patients to more precisely identify those at high risk of avoidable hospitalizations or emergency room visits.

Peterson and co-authors conclude that, although the findings do not support initiating a broad rollout of CareFirst’s medical home program to Medicare patients, program refinements like better targeting within the Medicare population may produce more favorable results.

Read more about the study in JAMA Internal Medicine's July 31, 2017 Editor's Note "The Importance of Independent Evaluation" and in the abstract. The full journal article is available to JAMA subscribers.

Related Publications