The 10 participants in a CMS physician group practice pay-for-performance demonstration project achieved impressive performance targets for diabetes care, but only two qualified for financial rewards resulting from savings to the Medicare program.
All 10 practices met at least seven of 10 quality benchmarks for diabetes care in the first year of the project, but only the Marshfield (Wis.) Clinic and the University of Michigan Faculty Group Practice, Ann Arbor, achieved savings of 2% based on projections from the baseline year (April 1, 2003, to March 31, 2004). While some participants voiced expectations that more savings will be generated once programs related to the demonstration become more fully developed, there was also some discussion that the CMS does not pay for the case-management and care-coordination services needed to achieve the quality targets.
Among the 10 practices, which include some 5,000 physicians providing care for more than 220,000 Medicare fee-for-service beneficiaries, it was said that the total cost of care was $21 million below projectionswith Marshfield and Michigan accounting for $9.5 million of that amount. Under a formula that provides for an 80-20 split between the practices and the Medicare program, Marshfield collected more than $4.5 million and Michigan received more than $2.7 million.
Acting Deputy CMS Administrator Herb Kuhn attributed at least some of the lack of savings to project startup costs. The project itself is actually now early into its third year, but first-year results were only released last week.
While the first year looked at diabetes care for Medicare fee-for-service beneficiaries, quality measures for congestive heart failure and coronary artery disease were added in the second, and measures for hypertension and breast and colon cancer screening have been added in the third. The measures were drawn from the CMS Doctors Office Quality, or DOQ, project. The demonstration was ordered by Congress under the Benefits Improvement and Protection Act of 2000.
James Lee, an internist and quality management expert at one of the participating practices, Everett (Wash.) Clinic, said that the overall impact of the program was much greater than just the improvements measured under the Medicare lens. And, in an American Medical Group Association-sponsored news conference, representatives of the participating groups attributed many of the benefits seen during the project to information technologymainly to the creation of diabetes patient registries.
Participants credited the use of these registries with helping to reduce hospitalizations, shorten hospital stays and avoid emergency department visits. Barbara Walters, senior medical director at Dartmouth-Hitchcock Clinic, Bedford, N.H., said the demonstration project has created an innovation think tank.
We share best practices, what works andperhaps more importantlywhat doesnt work, Walters said.
David Abelson, senior vice president and chief medical information officer for Park Nicollet Health Services, St. Louis Park, Minn., said that many of the benchmarks were being achieved by having nurses act as case managers who help coordinate care among the many providers a typical Medicare beneficiary sees. He said despite the effectiveness of this approach, Medicare doesnt provide reimbursement for care coordination and case management. Because of this, Abelson said that pay-for-performance is a back-end reconciliation for the fact that the CMS does not provide reimbursement for cheap and effective services.