A slightly puzzling result . . .

. . . followed by "Aha!"

In 1999, the last legitimately elected president of the United States directed that the Federal Employees Health Benefits program provide parity in coverage of behavioral health services -- mental health and substance abuse treatment -- with coverage of other health care services. This has long been a major bone in the throat of people in the behavioral health field -- that most health plans sharply limit benefits.

The main justification for relegating behavioral health to second class status is the argument that the need for such services is not sharply defined. The insurance companies are afraid that if they offered open-ended benefits, half of their customers would end up like Woody Allen, in psychoanalysis for life. So, Clinton also directed that there be a comprehensive evaluation of the policy to find out how it affected cost, access and quality of behavioral health care for federal employees.

The results are reported by Howard Goldman et al in the New England Journal of Medicine (abstract here, full text available only to the elect). Federal employees have a buffet of plans to choose from, so the evaluators could compare plans that introduced full parity with others that did not. They found that consumption of behavioral health services went up over the years in both groups of plans, but by similar amounts. The difference, obviously, was that people who lacked parity in coverage paid more out of pocket.

My first reaction was that it seems very surprising that having to pay out of pocket did not reduce people's use of these services. On grounds of equity, this still seems to support providing parity in benefits -- insurance is supposed to spread risk and obviously it does that successfully in this case -- but it doesn't seem to do anything to increase access and appropriate utilization.

Then it occurs to me that 100% of the sample are people with good, steady, fairly secure jobs that provide comprehensive benefits. If these folks need mental health or substance abuse services, they can afford to pay something out of pocket. Furthermore, they are unlikely to be among the most severely mentally ill or they wouldn't have those good steady jobs in the first place.

So what this tells us is that extending parity for behavioral health services doesn't have to lead to abuse, waste, overutilization, or a nation of Woody Allens permanently preoccupied with their neuroses and pipelining money to Madison Avenue shrinks. For needier and more vulnerable populations, it is still likely that it will lead to better access and more appropriate utilization. So let's do it.