Pay for Performance

One of the ways proposed to limit costs in health care is to switch from the current "pay for service" to "pay for performance." The idea is that physicians are compensated for doing things and that this is a perverse incentive to do unnecessary testing and procedures. The theory is that by paying for outcome, it eliminates this disincentive.

This concept is flawed for several reasons. In the first place, physicians rarely are compensated by the amount of testing they do. The equipment for MRI's, CT's, blood work, etc., by and large is owned by hospitals or other entities. The "Stark laws" actually prohibit physicians, and even their families, from owning the entities at which most of the testing is done. Surgeons are compensated by the procedures they perform, but our long experience with managed care has shown that the number of denied procedures is small. Too small, in fact, to justify the administrative expense of going through the approval process.

A more serious flaw is in the statistical parameters needed to reach probability. Most primary care physicians will have about 3500-4000 patient visits per year, and of course, these are in multiple categories. To reach a statistically significant number of visits even for a common illness such as diabetes, would be difficult for an individual physician. And of course, each of those patients comes with a pre-existing set of risk factors that needs to be "risk adjusted" to get a valid conclusion. Assuming that all physicians have a certain amount of difficult patients that will "even out" over time is simply not valid.

If there were a pay for performance rule, the incentive would be to avoid taking care of the sickest or most uncooperative patients. Not exactly a very desirable situation.

Also, the amount of money in the physician pool is much smaller than most people realize. About 20% of the health premium dollar goes to "physician and clinical" services. That includes the amount for most outpatient testing. And physicians have to pay overhead expenses, which are frequently more than 50% of payments received. So, even in the most optimistic scenario, a 10% savings due to pay for performance would lead to a 20% or more reduction in physician income. If penalized to this extent for something essentially beyond one's control, most physicians would not cooperate or simply quit altogether.

Last but not least, the idea that physicians somehow do unnecessary things for the sake of maximizing income is insulting. Granted that there are 600,000 physicians in the US and there are bound to be some who are dishonest, but the number of physicians involved in fraudulent activity is very small. To base a reimbursement scheme on this premise is foolish.