WASHINGTON – Knowing which Medicare managed care plans provide high-quality care is a good idea, but Medicare beneficiaries will need more information to figure out which providers to see within a particular plan, Glenn M. Hackbarth, J.D., said at a meeting of the Medicare Payment Advisory Commission.
In Boston, for instance, “If you're a Medicare beneficiary trying to get good health care, knowing that Harvard Pilgrim Health Care is ranked number one doesn't tell you at all where to go within the Harvard Pilgrim network to get really outstanding care,” said Mr. Hackbarth, a Bend, Ore., health care consultant who is chairman of MedPAC. “That's something that needs to be understood about plan rankings.”
The commission staff ranked about 150 Medicare Advantage plans, which are managed care plans offered to Medicare beneficiaries. Senior analyst Niall Brennan presented their work.
To assess the plans, staff members looked at plan scores on nine different measures within the Health Plan Employer Data and Information Set (HEDIS) database. The researchers looked only at HMO plans; PPO and fee-for-service programs within Medicare Advantage were excluded because they weren't required to report on all HEDIS measures.
Measures included breast cancer screening, ?-blocker treatment after a heart attack, antidepressant medication management, hypertension management, osteoporosis management, cholesterol management, follow-up after mental illness, and colorectal cancer screening. One additional measure, diabetes care, counted double in the scoring, Mr. Brennan said.
Scores were calculated based on how well each plan did compared with other plans–for instance, a plan that scored above the 90th percentile on a particular measure received 1 point, while a plan that scored between the 75th and 90th percentile received a 0.88. Ten points constituted a perfect score.
Mr. Brennan noted that the total scores of the plans the staff evaluated ranged from 1.2 to 9.8. And different measures had very different levels of result. For example, “giving ?-blockers after a heart attack is a well-established quality measure, and plans tend to score very highly on it,” he said. “Other measures [such as osteoporosis management] are a little lower and more widely dispersed.”
In general, bigger plans seemed to perform better, with plans that had more than 50,000 enrollees having an average HEDIS score of 7.5 compared with an average of 5.7 for plans with fewer than 10,000 enrollees. Not-for-profit plans had slightly higher scores than did for-profit plans. And while there was not a lot of geographic variation, Boston area plans did score “significantly higher” in terms of quality, he said.
Scores also appeared to remain stable over time; for example, 74% of the plans in the lowest quartile in 2003 remained in the lowest quartile in 2004. “The results of our analysis are quite interesting, and highlight the need for a pay-for-performance program in Medicare Advantage, and the feasibility of using HEDIS data for differentiating among Medicare Advantage plans,” Mr. Brennan said.
MedPAC commissioners, however, were not sure how useful the scores would be, especially if the quality data were supposed to be used as the basis for a pay-for-performance plan.
“If this looks like a good measure for implementing a pay-for-performance program, I'd like to step back and say that it isn't,” said Commission Vice-Chair Robert D. Reischauer, Ph.D., president of the Urban Institute.
Mr. Hackbarth noted that beneficiaries needed to look at more than just managed care plans. For plan data to be really useful, “beneficiaries ought to be able to compare plan offerings and choices to the ambient level of quality of fee-for-service Medicare in that same community,” he said. “We're moving in the right direction, but there are a host of questions about how you do these things.”
Commission member John M. Bertko, vice president and chief actuary for Humana in Louisville, Ky., said more attention should be given to data collection systems other than HEDIS, which requires use of an electronic health record. “In some markets, we literally need to go and knock on doctors' office [doors] to collect these data,” he said.