By Saurabh Jha, MD
“Extraordinary claims require extraordinary evidence,” said Carl Sagan.
The claim that health insurance improves health outcomes is hardly ground breaking. Studying whether insurance affects health status is like wondering whether three meals a day lead to a higher muscle mass than total starvation.
Well that’s what I thought. Until I read the study on Oregon’s Medicaid program by Baicker and colleagues in the NEJM earlier this year and, more recently, Avik Roy’s short treatise “How Medicaid Fails the Poor”.
Baicker et al found that Medicaid enrollees fared no better in terms of health outcomes than those without insurance. That is, no insurance no difference.
The study is an exemplar of policy research laced with regression equations, control of known confounders and clear separation of variables. There is only so much rigor social science can achieve compared to the physical sciences. Yet this is about as good a study as is possible.
The one thing the study did not lack was sample size. It’s useful to bear in mind sample size. Large effects do not need a large sample size to show statistical significance. Conversely, if study with a large sample size does not show even a modest effect, it means that the effect probably does not exist.
There are several interpretations of the Medicaid study, interpretations inevitably shaped by one’s political inclination. The ever consistent Paul Krugman, consistent in his Samsonian defense of government programs against philistines and pagans, extolled critics of Medicaid as “nuts” and asked, presumably rhetorically, “Medicaid is cheaper than private insurance. So where is the downside?”
Unlike Krugman I am not a Nobel laureate and am about as likely to win a Nobel Prize as I am of playing the next James Bond, so it’s possible that I am missing something blatantly obvious. Could the downside of a government program paying physicians, on average 52 cents, and as low as 29 cents, for every dollar paid by private insurance in a multiple payer system be access?
Indeed, it’s darn impossible for patients on Medicaid to see a new physician. As Avik Roy explains “…massive fallacy at the heart of Medicaid….It’s the idea that health insurance equals healthcare”.
But wait. It gets better.
I am accustomed to US healthcare throwing more plot twisters than Hercule Poirot’s sleuth work. But one I least expected was that patients on Medicaid do worse than patients with no insurance (risk-adjusted, almost). I am not going to be that remorseless logician, which John Maynard Keynes warned us about, who starting with one mistake can end up in Bedlam, and argue that if you are for Medicaid that is morally equivalent to sanctioning mass murder. Rather, I ask how it is possible that possessing Medicaid makes you worse off than no insurance whatsoever.
To some extent this may artifactually appear so because poverty correlates with ill health, and studies that show Medicaid patients faring worse than uninsured, cannot totally control for social determinants of health.