This week a team of researchers released a study on the change in mortality rates during health care reform enacted in Massachusetts under Governor Mitt Romney, which eventually served as a template for the Affordable Care Act. They estimate that between 2007 and 2010 the overall rate declined significantly, and that the portion considered “amenable to health care” amounting to about 1 saved life per 830 that had gained insurance coverage. This estimated effect was particularly prevalent in low-income counties with higher rates of uninsured.
There’s been a lot written so far about this study. I would recommend Adrianna McIntyre’s summary and Austin Frakt’s editorial to catch up. In another excellent post Harold Pollack roughly extends the math to imagine similar consequences under the ACA, which “implies that ACA will prevent something like 24,000 deaths every year. That’s almost the number of Americans who die in auto crashes. It’s more than the number who die of AIDS or the number who are murdered every year.” As Pollack goes on to write, though, it’s a crude assumption and one the study authors don’t even attempt to guess.
Yet while I think it’s sensible to believe that expanding affordable access to health care would save some lives, it’s often folded into the broader and infinitely more debatable questions like “does health insurance improve health” or “is health insurance worth it?” Of course while this question of mortality fulfills the former (living, after all, is a health improvement over dying), it’s not enough to answer the former. As several folks have mentioned health insurance provides many essential benefits beyond giving people a greater chance of living. But whenever the application of that inquiry applies to public funding of coverage for those that can’t afford it conservatives often reply “No,” whether it’s mortality, financial security, mental health, etc. That was largely the response to one of the Oregon Medicaid studies, for instance, though some of the response to the MA results are more tempered.
I won’t attempt to litigate that perennial response here. Instead I’ll point to Michael Cannon of the libertarian Cato Institute, who writes that the implicated per-person cost of increasing coverage — between three and four million dollars — technically fails the World Health Organization’s definition of “cost-effective,” and that there might very well be better ways to increase coverage and save lives.
It’s a good question and response. How much is too much for reducing mortality with health care reform? Is there a better way?
These are entirely legitimate inquests, taken seriously by health care researchers and various wonks, and shouldn’t be shrugged away. Given resource constraints to provide health care to its citizens (whether real, imaginary, or artificially constructed), what is the most cost-effective way for a country to provide access? What is the most cost-effective avenue for delivering that care afterwards? This is a subject that various public and private organizations routinely attempt to answer (reference Cannon’s usage of the WHO), and for which a part of the ACA (the Patient Centered Outcomes Research Institute, or PCORI) was established to assist in that effort.
Of course these types of questions have evoked talk about rationing, death panels, and visions of indifferent government bureaucrats deciding whether you live or die depending on how much it costs. Nevermind, as the great Uwe Reinhardt noted, that they implicitly do just that in numerous other areas — most notably in defense. Nor that many of the ACA’s provisions were implemented to curb more egregious acts of similar decision-making by private, indifferent, bureaucrats. As a result politicians and legislators, at all levels of government, at the very least prefer to pretend these considerations are off the table, sort to speak.
Except when it comes to the poor. Enter Bill Gardner, challenging ACA critics to consider, among other things, the proper counter-factual here:
First, Cannon believes that there are policies that would deliver more benefit than Romney- or ObamaCare. If you are a critic of the ACA and this is what you believe, Cannon’s argument obligates you to do that better thing with the money. This poses an acute test for leaders in the 24 states that refused the ACA’s expansion of Medicaid. How many of those states refused to expand Medicaid, but then did nothing else for the health of their uninsured? If politicians in those states just refused the money and let the poor die, they do not have standing to make Cannon’s criticism about paths not taken.
There’s another consideration raised by Bill, which you should click-thru to read, and a similar take from Andrew Sprung here. While the Massachusetts study doesn’t conform neatly to teasing out the implications for different states, given that researchers observed greater effects in low-income counties, it’s not unreasonable to assume expanding Medicaid would produce a not-insignificant amount of mortality decrease. Moreover, using the MA study ratio of lives saved per insurance coverage-gained and combined with other research, the Advisory Board’s Tom Liu posits that the existence of the Medicaid gap would fall into fourth place in the CDC’s leading causes of preventable deaths in the United States.
So this is a valid inquiry for policy-making, and leaves us with the more narrow question for legislators and governors in states that have not expanded Medicaid to their low-income residents, doing “nothing else for the health of their uninsured,” and who are responsible solely for their respective states budgets. The federal government actually pays 100 percent of the bill for every new enrollee covered by the expansion for the first three years. After that the reimbursement rate will gradually fall but never below 90 percent, which is still a much more advantageous deal for states than pre-ACA arrangements. How much is too much for saving lives with health care reform? If the continued refusal to expand Medicaid coverage is the response then their answer right now is effectively zero.