Michigan also outperformed other states in reducing hospitalizations and emergency department use.


May 22—An emphasis on care management features was credited with helping the only state to achieve significant savings in a recently concluded multi-payer medical home pilot.

Michigan was among the six of eight states that achieved any Medicare savings in the Multi-Payer Advanced Primary Care Practice (MAPCP) pilot, which ran from 2011 to 2016. The model tested whether medical homes that receive monthly care management fees for most of their patients across multiple insurers—including Medicare—perform better on quality and spending measures. They were compared with medical homes that lacked payer alignment and with non-medical home practices.

Michigan was the only state to obtain significant Medicare savings through the first two years, with $336 million. Third-year financial results are expected later this year.

The program initially was designed as a three-year pilot but was extended to 2016 for five of the states. The MAPCP demonstration was one of three medical-home pilots initiated by the Centers for Medicare & Medicaid Services (CMS).

Michigan’s results towered over those of other states, where outcomes ranged from Medicare losses of nearly $40 million in Maine to savings of $27 million in Vermont.

Jean Malouin, MD, medical director of Value Partnerships for Blue Cross Blue Shield of Michigan (BCBSM), noted that “many of these programs that have been evaluated and tested have not been successful.”

In leading BCBSM’s participation in the MAPCP pilot, Malouin focused on care management features previously associated with positive outcomes.

Care management features most associated with positive outcomes in the pilot included care delivery by multidisciplinary teams, care delivery in collaboration with the primary care provider (PCP), and attention to care transitions.

Especially important was patient selection.

“It’s pretty easy to identify last year’s high-cost patient,” Malouin said in a presentation on the pilot. “But what’s really challenging is identifying next year’s high-cost patient.”

Other Results

Those approaches produced clinical improvements as well as financial improvements. For instance, BCBSM studied its MAPCP practices and found 37.9 fewer emergency department visits and 40 fewer inpatient admissions per 1,000 members than in a comparison group.

In contrast, little evidence of utilization reduction was found in the overall program, according to a contractor’s analysis.

“For preventable hospitalizations, we found no significant differences in all eight states in terms of avoidable catastrophic events,” according to the report by RTI International. 

In the BCBSM results, the input of primary care physicians—in conjunction with a risk-stratification system—was most helpful in identifying high-risk patients, according to Malouin.

Providers were given patient lists that included risk scores and utilization information, and patients identified as part of the “very high” risk category and dual-eligible enrollees were flagged for appropriate care management. Beyond these groups, the program also assigned care management to patients flagged by the PCPs.

Patient engagement was most successful, Malouin said, if PCPs introduced flagged patients directly to their care manager.

About 850 care managers were trained through the program, with special emphasis on integrating behavioral health care and palliative care and on addressing the social determinants of health.

The state also collected data from multiple payers and aggregated it in one database, aiming to provide a comprehensive picture of a patient’s information. Specifically included were data on when patients received benefits from multiple insurers and when they visited physicians from different practices, physician organizations, or hospitals.

“This information has been very helpful,” Malouin said.

Other Differences

Additional differences in the Michigan MAPCP program, as noted in the RTI report, included the state’s use of the demonstration to combine separate efforts into a new collaborative while the other states continued to use existing collaboratives.

Michigan’s MAPCP initiative also had the most participants, with 1.2 million total and 268,000 Medicare fee-for-service beneficiaries. The entire MAPCP pilot had more than 900,000 Medicare beneficiaries. Michigan’s initiative included 314 practices, 1,600 providers, and five payers.

Michigan was one of four states that required practices to participate in activities designed to help them transform their practices. These efforts included learning collaboratives, practice coaches, webinars, and phone calls.

Michigan and North Carolina paid support teams to help practices change how they delivered care and to conduct quality improvement activities.

The Michigan Data Collaborative added commercial data to its provider dashboards, which allowed provider organizations to assess their relative performance against other organizations and against performance benchmarks. The dashboards also allowed organizations to drill down to the individual-provider and individual-patient levels to help improve performance

Michigan’s positive results came despite some challenges specific to that state. For example, RTI noted that Michigan practices and physician organizations replaced individual patient lists from each payer with an all-payer patient list. But practices and physician organizations “reported frustration with the timeliness and accuracy of the all-payer patient list and continued having difficulty billing the G-codes for care management services,” the RTI report noted.

When the MAPCP program concluded at the end of 2016, the 350 participating practices were then absorbed into the state’s ongoing provider-delivered care management program. 


Rich Daly is a senior writer/editor in HFMA’s Washington, D.C., office. Follow Rich on Twitter: @rdalyhealthcare

Publication Date: Monday, May 22, 2017