found that, in the decade following the introduction of Medicare in 1965, the establishment of a universal health insurance program for the elderly had "no discernable impact on their overall mortality." A large observational study of more than 600,000 respondents conducted from 1986 through 2000 found that, after adjusting for demographics, health, and behavior "the risk of subsequent mortality is no different for uninsured respondents than for those covered by employer-sponsored group insurance." Last year, a rigorous, randomly controlled experiment by a team of top-flight health researchers working in Oregon found that there was no statistically significant difference in measured physical health outcomes between people awarded Medicaid coverage in a lottery and a control group of people who were not. Initial findings from the same study also reported no mortality effect.Does health insurance save lives, or make people healthier? The answer to this question might seem obvious: Of course people who are insured are healthier and likely to live longer. But the evidence has not all been so straightforward. A 2005 study published by the National Bureau of Economic Research
In other words, it’s not at all obvious whether health insurance makes people healthier or increases longevity. Which is why the result of a new study on RomneyCare, published in the Annals of Internal Medicine, is practically counterintuitive. The study looks at Massachusetts' experience in establishing the state-based, near-universal health insurance program in 2006. A trio of researchers from Harvard and the Urban Institute compared Massachusetts counties in 2007-2010 with similar control counties in other states, and found a corresponding 2.9 percent decrease in mortality. More specifically, they found a 4.5 reduction in "mortality amenable to health care"—basically, deaths that we know how to prevent using medical services—following the introduction of the Bay State’s health reform. No similar mortality reduction was found in the control counties.
This is a big study with a really large sample size, which means we ought to give it more weight. It also looks at a somewhat longer time than the Oregon experiment—four years, rather than Oregon’s two. In other words, it’s a good solid study that you should not dismiss.
So does this mean that RomneyCare saved lives? Well…maybe. It’s entirely possible, and plausible. But the study can’t quite answer that question one way or another. It’s not randomly controlled, like the Oregon experiment was, which means it’s subject to problems with confounding variables and selection effects. It’s a study that does a pretty good job of proving correlation, but doesn’t quite get us to causation.
To which a frustrated liberal health wonk might demand to know: Well, what else could have caused the same effect? There’s a measurable drop in mortality in the four years following the implementation of a state-wide, near-universal health insurance scheme, and that’s just a coincidence?
It may not be a complete coincidence; indeed, I suspect, though I cannot prove, that if we could determine causality with anything approaching real certainty, we’d find out that it’s not. But there may be other, additional factors at play as well.
Massachusetts is a wealthy, highly educated state that already had one of the highest insured rates in the nation. It’s home to a slew of elite hospitals and health care systems. Judging by the unemployment rate, which topped out at 8.7 percent, the state weathered the recession a good bit better than the United States as a whole. The authors control for the difference in unemployment rates, of course, but wealth, employment, and health are correlated in all sorts of complex ways that we don’t fully understand; the recession’s lighter toll may be expressing itself in other, subtler ways.
Massachusetts is different from other places in ways that are obvious and ways that are not so obvious. The implementation environment in Massachusetts, for example, was radically different than the one Obamacare faces nationally: broad public approval, bipartisan political support, a booming economy. Which means that even if you assume that RomneyCare’s increase in coverage was the sole cause of the drop in mortality, enacting a similar program nationwide wouldn’t necessarily produce the same results. Government-run health care pilot programs are notoriously difficult to scale up. Just because something works in one place, even if it works really well, does not mean it will work as well or at all in another.
But let’s assume, for a moment, that it does. The result of a nationwide 3 percent drop in mortality would be about 17,000 fewer deaths per year. But as with anything, there’s a cost. To put it another way, the study suggests that in order to save one life each year, about 830 people would need to enroll in health coverage. That’s an expensive proposition. As Michael Cannon, director of health policy at the Cato Institute, estimates in a post at Forbes, "If we assume the per-person cost of covering those 830 adults is roughly the per-person premium for employer-sponsored coverage in Massachusetts in 2010 (about $5,000), then a back-of-the-envelope calculation suggests that RomneyCare spent $4 million or more per life saved," perhaps more when the law's other costs are included. It’s not enough to simply ask whether a policy has any effect at all. It’s also necessary to ask whether there are different, more effective sets of policies that could be achieved for a similar price tag.
And as Cannon points out, his estimated cost for RomneyCare probably doesn't pass a basic cost-effectiveness test:
The World Health Organization considers a medical intervention to be "not cost-effective" if it costs more than three times a nation’s per-capita GDP per year of life saved. This in turn suggests that RomneyCare would have to give every person it saves an average of nearly 30 additional years of life to meet the World Health Organization’s criteria for cost-effectiveness. Given that the mortality gains were concentrated in the 35-64 group, that seems like a stretch.
On the same note, it’s worth remembering that for the past several years, the price tag in Massachusetts has looked unaffordably expensive. In a 2011 review of the state’s health reform published in Health Affairs, a team of researchers looked at the results of the program over the same time frame measured by the new study. What they found was more coverage, more utilization of care—and costs that could not be supported over time. Not in Massachusetts. Not anywhere. "The pre-2010 status quo is not a sustainable option for Massachusetts or the nation," the report said. Around the same time, state health officials were also describing the program’s costs as unsustainable, warning that, if left unchecked, they will crowd out everything else the government needs to do. Some reforms have been put in place by then, but even still, cost-containment is a challenge—in part because more coverage has led to greater use of care.
The new Massachusetts study sheds some genuinely useful, though not perfectly conclusive, light on the question of how health insurance effects health. But it offers less help with the public policy questions that come next: What should we do, and how much will we have to pay for it? Health insurance programs like RomneyCare won't help anyone if their costs turn out to be unsustainable.