Trendspotter: Cost Profiling Used to Build Narrow Networks

July 21, 2010

Health plans’ latest effort to offer narrow networks to their customers, or this “tiering” of physicians and hospitals, is of course related to their cost of care. And the insurers will use cost profiling based on claims data to help determine which providers are in or out of their favored networks.

By a strange coincidence, the AMA and medical societies in 47 states sent a letter to health insurers protesting physician cost profiling just as The New York Times ran a front page story about health plans’ latest effort to offer narrow networks to their customers.

This “tiering” of physicians and hospitals is, of course, related to their cost of care. And the insurers will use cost profiling based on claims data to help determine which providers are in or out of their favored networks.

The controversy over physician profiling has been going on for more than a decade. But it got new legs recently when RAND Corp. researchers published a study on it in the New England Journal of Medicine. The study found that, because health plans used imprecise measures in their cost profiling, nearly a quarter of all physicians were misclassified by their use of resources. In some specialties, physicians were mistyped two-thirds of the time. Citing this and other studies, the medical societies’ letter asked the major carriers to work with them to develop more reliable rating methods and to allow impartial outside bodies to review their ratings.

What harm is done when health plans profile physicians erroneously? Here’s what the letter says:

"Patients are being encouraged, and often incentivized, to leave longstanding relationships with physicians they trust, or see certain physicians and physician groups, based on information that RAND has shown to be incorrect 25 percent or more of the time. Physicians’ reputations are being unfairly tarnished using unscientific methodologies and calculations. Some physicians have even been placed in three different efficiency tiers by three different insurers based on calculations using the same data."

What the AMA and the other societies are referring to is tiering, which has suddenly become au courant among insurers. According to the Times, “the country’s biggest insurers are promoting affordable plans with reduced premiums that require participants to use a narrower selection of doctors or hospitals.”

Among the plans that are starting to investigate or push these tiered plans are WellPoint, United, Aetna, and Cigna. In one plan that Aetna is selling in New York, the narrow network includes only half of the doctors and two-thirds of the hospitals that the carrier normally offers in its plans.

It doesn’t take a brain surgeon to figure out that the doctors in these networks are either being paid less per unit of work or are able to deliver care more efficiently than other physicians. And how does an insurer calculate which doctors are more efficient? By using claims-based cost profiling, of course.

One interesting aspect of the RAND study is that it used claims data to construct what the researchers considered to be fair profiles of physician resource use in “episodes of care” for diabetes, heart attack, or urinary tract infection. The researchers then compared those profiles with the cost profiles developed by the four health plans from which they’d obtained their data. Their conclusion was that typical insurance company profiles might produce misleading results.

An uninformed observer might draw the inference that insurers are just using the wrong methodology. Perhaps if they employed the RAND methods or some other advanced approach, they could deliver consistently accurate cost profiles. But there are a couple of problems with that argument: For one thing, claims data is inherently erroneous, because it doesn’t mirror the clinical details of a case, because a practice or a hospital might be trying to maximize reimbursement, and/or because billing clerks make mistakes. So, until every physician has an electronic health record that’s hooked up to insurance companies, cost profiling is likely to misclassify some doctors.

Second, if more employers start offering limited-choice plans, an increasing number of patients will choose them because of the cost. The doctors in those plans, however, will not be able to take all of those patients. So either the other physicians will have to become more efficient, or what’s more likely, they’ll be paid less.

From a societal point of view, this might not be a bad thing. But a lot of physicians are going to be very unhappy if this is what profiling leads to. And I suspect that the hidden agenda of the medical societies’ letter is to kill profiling altogether.