Cutting Costs by Profiling Physicians

Think back to Atul Gawande's influential article about healthcare costs in the New Yorker earlier this year and the debate it sparked about overutilization. The article was forwarded throughout the industry, became mandatory reading at the White House, and may even shape future healthcare policy. All of this from a relatively simple comparison of Medicare costs in two Texas towns.

Kind of broad brush, isn't it? There are a lot of good physicians who don't overutilize—even in McAllen, TX, where Medicare spends $15,000 per enrollee, almost twice the national average. And there are plenty of physicians who overutilize in regions that look below average on paper, at least according to the Dartmouth Atlas and other studies that measure healthcare spending geographically.

The problem is that policymakers are trying to address how physicians use resources, yet most of the data that drives policy decisions and healthcare research measures how regions use resources.

CMS is trying to develop more granular data at the physician level. The Government Accountability Office, MedPAC, and the Congressional Budget Office have all recommended in recent years that CMS profile physician resource use and provide feedback as a step toward improving Medicare efficiency.

But how?

MedPAC has explored episode-based profiling, which measures the resources used for treating a particular episode of care or a specific illness. But we're still locked into a fee-for-service system, so other options have to be considered.

This week, the GAO released a report evaluating per capita profiling, a method that measures a patient's resource use over a fixed period of time and connects that resource use to physicians.

The report focused on cardiologists, radiologists, internists, and orthopedic surgeons in Miami, Phoenix, Pittsburgh, and Sacramento, and risk adjusted data to account for patients' health conditions.

In each of the four metropolitan areas, physicians were fairly stable in their resource use, even when their patients' resource use varied. And patients seen by "high-resource use" physicians generally were heavier users of institutional services (such as hospital services) than those seen by lower resource use physicians.

Those findings aren't all that remarkable on their own. The purpose of the report was really to determine whether per capita profiling could be used to measure physician resource use, and the results suggest it can. And the entire point of figuring that out was to go one step further and provide physicians' feedback on their resource use.

Will that have any effect? Researchers asked private insurers who have tried similar programs that question and concluded that feedback has no more than a moderate influence on physician behavior. Moderate could mean many things, and the GAO seems to be hoping that feedback from CMS will be more of a motivator than feedback from a private payer "because of the relatively large share of physicians' practice revenues that Medicare typically represents."

Most physicians like data and enjoy receiving feedback on their own performance. Their competitive nature kicks on and they will often improve in areas they didn't even realize they were lagging behind, even resource use.

Even if the impact is only "moderate," that could add up to a lot of dollars.