A study in this week’s New England Journal of Medicine finds that when three states expanded their Medicaid programs, mortality rates fell 6 percent relative to four neighboring states. The study found evidence that the mortality gains were concentrated in poorer counties — i.e., where people were most likely to become eligible for Medicaid.

As always, the study comes with caveats. The results “may not be generalizable to other states,” may have been driven by unobservable confounding factors, et cetera. Speaking only for myself, I hope these results are accurate. I hope Medicaid does save lives. That program spends nearly half a trillion dollars per year. It damn well better save lives.

Even so, that does not mean politicians should expand Medicaid. If saving lives is the goal, then politicians should instead find the lowest-cost way of doing so, because that enables the greatest number of lives to be saved with the available resources. It is generally accepted among health economists that other strategies (e.g., discrete health programs targeted at hypertension or diabetes) could save more lives per dollar spent than expanding health insurance. This study says nothing about how much it costs to save lives through Medicaid, much less whether alternative uses of those resources could save even more lives. It could be that other uses of the money would save — I don’t know — twice as many lives.

Absent evidence that Medicaid saves the most lives per dollar spent, expanding Medicaid does not show how much politicians care about saving lives. It shows how little they care about saving lives, because they are willing to forgo additional reductions in mortality for the sake of…whatever else expanding Medicaid gives them.

[Editor’s Note: See this Cato essay for more on Medicaid reforms.]