Monday, May 6th 2013

Like All Insurance, Medicaid’s Role In Health Debated

Last week you could watch liberals scramble in response to a New England Journal of Medicine published study.

In 2008 Oregon found resources within the state budget to expand Medicaid. There were far more eligible for the expansion than funds available. Oregon held a lottery to enroll 10,000 new people onto the Medicaid rolls. Harvard, MIT and RAND have a group who saw that lottery as a great playground for studies. The first of these studies has now published its results. The study followed approximately 6,000 people who got Medicaid in 2008 and 6,000 who didn’t and looked at how their blood pressure, cholesterol, diabetes and depression did post randomization. Here’s the important quote from the conclusion that has so many talking,

This randomized, controlled study showed that Medicaid coverage generated no significant improvements in measured physical health outcomes in the first 2 years

As the Affordable Care Act promises the greatest increase in covering the uninsured by Medicaid expansion, if states take it, the study is important.

I linked to Mother Jones above. Here’s what Kevin Drum had to say,

It’s more likely that access to Medicaid did improve health outcomes than that it had zero or negative effects. It’s just that the study was too small to say that with certainty. For laymen, as opposed to stat geeks, the headline result of the Oregon study was “Possibly positive but inconclusive,” not “Had no effect.”

Here’s Jonathan Cohn at The New Republic,

Of course, even if Medicaid isn’t improving health, it’s certainly not making health worse, as some critics have claimed. Meanwhile, it’s improving mental health and providing economic security to some of the most economically vulnerable people in the country.

Justin Wolfers resorted to hypotheticals,

And, as a final example, here’s Matt Yglesias,

There are many nuances that need to be realized about the this study. The measures in the Oregon study where all surrogates for long term health. Admittedly hardly definitive quantities. And it “trended” towards significance in those measures. And we don’t know what the study’s power was.

I concede that significance is a relative crude tool in research. See here for a good explanation of such. But I think that fact plays poorly for this study. The reality is that asking to find a significance defined as a p <0.05, depending on pre-test probabilities, is not asking a lot.

Even without knowing the power with specificity, if this study couldn’t find signifigance at that level then to raise policy questions based on this study is not unreasonable. The beneficial effects of Medicaid on these quantities must be small if they’re real.

And so, only Kevin Drum in hindsight realized the real question this study raises,

Even if they’re real, are these results worth the money spent? That’s a different question, and there’s just no way to answer it with this study. That would require a much larger, longer-term research project.

I strongly disagree with the assertion this study doesn’t help answer that question.

The fact that most proponents of the Affordable Care Act are loathe to discuss, in my limited experience, is the fact that insurance status is not a terribly strong influencer of health. Sure, it influences health utilization and some surrogate markers. But I’m with the studies that imply its effect on major things like mortality is negligible conceding that such is debated.

The Institute of Medicine’s estimate that lack of insurance leads to 18,000 excess deaths each year is almost certainly incorrect. It is not possible to draw firm causal inferences from the results of observational analyses, but there is little evidence to suggest that extending insurance coverage to all adults would have a large effect on the number of deaths in the United States.

The link in the quote is my own.

There are two associations that stunningly are a matter of debate.

The first is the association between insurance status and health care access. Especially with the Medicaid population there are other impressive obstacles to utilizing health care. Those include education and understanding of health issues, time constraints, literacy rates, angst at interacting with providers of different socioeconomic status amongst others. No matter their insurance status those of lower socioeconomic status utilize less health care and have poorer health.

The second is even more stunning, the association between health care access and health. It is true access to care is sometimes associated with secondary measures of health – some of the same measures in the Oregon study and others – but its association with the big quantities such as life expectancy is hardly unequivocal.

I find that personally a little bit shocking.

No doubt health care saves lives and improves health. Our own anecdotes tell us that. But it may very well be doing that on the margins. On a population level how much it effects such, especially as compared to things we might label under the umbrella ‘public health’ is questionable. And so we’re expanding Medicaid at considerable costs and for, perhaps, limited benefit. To what such limited benefit this study may help elucidate. And that is certainly something worthy of policy debate.

The issue is complex. There are cost savings and cost shifting in expansion and there are benefits to consider other than large population based health measures. But to pretend that such doesn’t deserve debate and that this Oregon study doesn’t inform that debate is silly.

The Oregon study is a blow to the logic behind Medicaid expansion.