Mental Health Parity and Insurance Costs

Parity in mental health and substance abuse coverage is a goal for many advocates for the mentally ill. Normally, insurance coverage for treatment of so-called "mental illnesses" (I'll deal with that term on another day) and addictions has it's own sets of limits, in some cases draconian. Parity in mental health and substance abuse coverage means that people have the same co-pay, deductible, and visit limitations as they do for traditional medical care.

The argument against parity has been that it would increase insurance costs significantly. The fear has been that people would sail into long-term therapy or stay in the hospital indefinitely, leaving insurance companies to pick up the tab. This article, by Goldman, et al., (there are 14 authors!) in The New England Journal of Medicine challenges that assertion.

The background is simple. In 1999, President Clinton directed the Office of Personnel Management to ensure parity for federal employees. Parity was established on January 1, 2001 for all 8.5 million enrollees in the federal system. This figure includes retirees and dependents along with the federal employees.

The authors matched 9 federal insurance plans with parity to 9 similar non-federal insurance plans without parity. They matched them on features of the plans and geographic location. They then compared the costs of the plans over the years 2000 (prior to parity) and 2001 (after parity).

From 2000 to 2001, costs of both plans rose the same amount. Had parity increased costs, the plans with parity would have increased more than the plans without parity. Because they did not, Goldman, et al., concluded that parity did not increase insurance costs.

I have a couple of of quibbles with this study. First, I'm not convinced that they matched the federal and non federal plans very well. I'd feel a lot better if they matched the plans based on the demographics of the enrollees. We know that variables such as socioeconomic class can influence rates of mental illness and substance abuse.

Second, I'm not sure that one year is enough to see a trend in costs. The major costs come from treatment of chronically mentally ill, especially in inpatient settings. That might take several years to emerge.

Finally, I'm uncomfortable with the statistical analysis. Part of it is my problem. They used a technique called "difference-in-differences" analysis, and I'm unfamiliar with it. Googling indicates that it's frequently used in epidemiological research, but I'd feel better with a parametric test, such as ANOVA. Their willingness to base conclusions on negative results furthers my uneasiness. However, they do have this to say about it:

A finding of negative results always raises the question of whether the effect of parity on use and spending was really limited or whether the evaluation lacked the necessary power to detect an effect. Two factors lead us to believe that the effect really was limited: the estimated differences between the results for enrollees in the FEHB Program and comparison enrollees were relatively small in magnitude. The sample sizes used in the analysis were large and were sufficient to show significant effects of similar policy measures, such as the effect of carving out mental health and substance-abuse care while holding the benefit design constant. (p.1384)

It's not a bad argument, but it's still somewhat shaky. On the other hand, there is an argument supporting parity which has been well documented. Providing mental health and substance abuse treatment costs brings down the costs of other medical services, especially in primary care.

Overall, this study throws the ball back into the anti-parity court. On a logical basis, it's really up to them to show that costs do increase. Goldman, et al., have shown, if nothing else, that the data is out there, just begging for analysis.

Unfortunately, I doubt we'll see any more data analysis. The belief that we need to save money on mental health treatment is very entrenched, and it seems to fall into the don't-bother-me-with-the-data-when-my mind-is-made-up category. It's like the belief that raising the minimum wage eliminates jobs. Every time we do it, the economy booms, but people still repeat that objection over and over. Still, this study is a start in a long overdue discussion.