Physician cost profiling varies widely, depending on formula used

Whether a physician is assigned to a high, average or low cost category by a health plan may have to do with the way various expenses are attributed by the insurer when patients are cared for by more than one doctor.

That is the conclusion reached by a study that appeared in the May 18 Annals of Internal Medicine (link).

"This study provides more evidence that efforts to create physician cost profiles are still a work in progress," said Ateev Mehrotra, MD, MPH, lead author and professor at the University of Pittsburgh School of Medicine, and researcher with the RAND Corp. "More work is needed in order to make cost profiling a meaningful tool to curb health costs."

Insurers' increasing use of cost profiling to attempt to contain expenses has long chafed physicians, and previous studies have suggested these measures can be inaccurate. For example, another paper by the same RAND researchers, in the March 18 New England Journal of Medicine, found that 22% of physicians would be misclassified in a two-tier system (link).

In the most recent paper, researchers studied the effect of 12 rules used to attribute costs when patients visited more than one physician. In an analysis of 1.1 million patients participating in one of four health plans in Massachusetts over a two-year period -- the same plans and time frame as the March study -- 91% saw more than two physicians. About 61% saw more than five.

A rule attributing the total cost of an episode of care to the physician responsible for at least 30% of related expenses and visits -- the most common formula currently used by insurers -- served as the baseline. When a different rule was used for comparison, between 17% and 61% of physicians were assigned to different cost categories.

The authors did not find that a particular formula was better than any other. Nor have they come out against this strategy for controlling health care costs. Rather, they are advocating transparency so physicians know how they are being evaluated.

The American Medical Association strongly opposes the use of tiered and narrow physician networks that deny patient access to or attempt to steer patients toward certain physicians based primarily on cost of care.

Some forms of profiling, however, are acceptable if they follow certain principles. These include disclosing the methods for collecting and analyzing data to relevant organizations and the physicians under review. Profiles should also be based on valid, accurate, and objective data and used primarily for educational purposes.

The Massachusetts Medical Society has sued the state's Group Insurance Commission, which purchases health coverage on behalf of many government employees, over its physician profiling program.