Embargoed Until:
June 26, 2007
12:01 a.m. Eastern Time



Christopher Fleming

Pay-For-Performance In Medicaid Managed Care: Money Talks, But Only If There’s Enough Of It, And Only If Plans Talk To Providers As Well

Mathematica Researchers Describe Lessons From A California Initiative To Improve The Timeliness Of Well-Baby Care

Bethesda, MD -- Pay-for-performance can work in a Medicaid managed care setting, but only if plans place enough dollars at stake and communicate well with providers, researchers from Mathematica Policy Research say in a Health Affairs Web Exclusive published today. http://content.healthaffairs.org/cgi/content/abstract/hlthaff.26.4.w516

In work supported by the California HealthCare Foundation (CHCF), the Mathematica authors describe the results of a 2003-05 P4P demonstration aimed at improving the timeliness of well-baby care. The demo, the Local Initiative Rewarding Results (LIRR) Collaborative Demonstration, funded by the CHCF and the Robert Wood Johnson Foundation, involved seven California Medicaid managed care plans operating in different regions.

Five of the plans implemented various new incentives for physicians and other primary care providers, while two did not. The specific goal of all five incentive programs was to ensure that providers saw infants for six well-baby check-ups in the infants’ first fifteen months of life. The plans, which requested anonymity in reporting results, varied greatly in size. The smallest had 72 primary care providers serving infants, while the largest had more than 2,000 such providers.

LIRR Demo: Modest Results,
But A Guide For The Future

The overall results of the LIRR demonstration were mixed at best: In four out of five plans, the incentives did not produce significant effects. In only one plan, referred to as “Plan D,” might the new incentives (in concert with existing incentives) have had a dramatic effect, although data constraints prevented the Mathematica researchers from confirming a causal relationship.

Despite these results, the LIRR demo was a pioneering effort that will enable P4P programs in the Medicaid managed-care arena to be better designed in the future, say Mathematica senior health researcher Suzanne Felt-Lisk and coauthors. Had the plans in the LIRR “known what we know now -- the critical importance of up-front strong provider communications plans and the importance of clear and timely support and feedback to providers on their performance -- and had they designed their incentives to better take into account the amount of change they were asking of providers, we suspect that their programs might have been more effective,” say Felt-Lisk and her coauthors, Mathematica health researcher Gilbert Gimm and research analyst Stephanie Peterson. “Plans were generally optimistic about building on what they had learned and were planning to retain the incentives, at least for the short term, at the close of the demonstration.”

Successful Plan Had Good Communication With Providers,
Provided Adequate Monetary Incentives

Plan D was the one plan where provider incentives appear to have yielded significant results. According to provider interviews conducted by Felt-Lisk’s team, Plan D conducted good communications with providers, both around the well-baby incentives and in general.

In addition, Plan D provided “stair-step” well-baby incentives, meaning that providers achieved bonuses for each additional well-baby visit, with a maximum award to each provider of $470 per baby on top of the provider’s regular capitated payment. Moreover, before entering the LIRR demonstration, Plan D was already offering providers an incentive for conducting well-baby visits, so many providers were in the habit of reporting administrative data on such visits to the plan.

In contrast, Plan B, in which the LIRR demo incentives had no effect, offered physicians $100 for reaching six well-baby visits for each infant, less than one-fifth of Plan D’s maximum award. Further, Plan B offered providers no money for any well-baby visits short of the ultimate goal of six for each infant. Finally, Plan B had no existing incentives for well-baby visits; as a result, going into the LIRR demo, providers reported administrative data on such visits to Plan B at only half the rate of such reports to Plan D.

The LIRR demonstration also revealed several other factors crucial to the success of P4P in the Medicaid managed-care context:

Patient Characteristics. Across all plans, providers reported that the characteristics of low-income parents often made timely well-baby visits difficult to achieve. Providers noted that low-income parents tended to be focused on economic subsistence and often lacked adequate transportation. Moreover, these parents tended to be more mobile and less likely to see the same provider for the fifteen-month period in which the six well-baby visits should take place.

Provider Characteristics. Patient outreach is the major activity necessary to improve well-baby visit rates, but many providers lacked the staff, funds, and technology necessary to perform such outreach. Many provider offices relied on monthly lists from plans identifying patients due for well-baby visits, even though these lists were sometimes out-of-date.

Plan Coordination. “The incentives would have been much stronger if competing plans in the same market area had jointly implemented them, to touch a higher proportion of each physician’s patient caseload. This approach was deemed not feasible by the LIRR plans during the demonstration period because of the history of some plans’ fierce rivalries,” Felt-Lisk and coauthors say.

Combined Approach. The Mathematica researchers suggest that “the next logical step may be to abandon the idea that any single approach alone will be enough to dramatically improve health care for low-income populations and instead to test approaches that bring provider incentives, . . . member incentives, and technical assistance [to providers] to bear simultaneously on areas of critical importance to the population.”

Christianson: Mathematica Study Offers Chance To Compare
Different P4P Programs Using The Same Measure

“Because they were able to compare the experiences of different P4P programs that use the same performance measure for similar populations of patients, Felt-Lisk and colleagues’ study generates findings that are instructive to policymakers,” Jon Christianson, the James A. Hamilton Chair in Health Policy and Management at the University of Minnesota, says in a Perspective on the Mathematica paper. “They underscore again that effective communication with physicians is essential for the success of P4P programs and that policymakers need to build enough resources and time into the implementation process to accomplish this.” http://content.healthaffairs.org/cgi/content/abstract/hlthaff.26.4.w528

The Mathematica findings also “highlight the need to consider the ‘starting point’ for physicians when designing rewards,” Christianson notes, referring to the fact that providers were already in the habit of reporting data on well-baby visits to Plan D before the LIRR incentives were put in place. “Finally, they at least suggest that physicians in resource-constrained practices serving low-income patients may have difficulty responding to financial incentives and improving their performance even when payment is based on simple, widely accepted performance measures.”


Health Affairs, published by Project HOPE, is the leading journal of health policy. The peer-reviewed journal appears bimonthly in print with additional online-only papers published weekly as Health Affairs Web Exclusives at www.healthaffairs.org.


©2007 Project HOPE–The People-to-People Health Foundation, Inc.