This year, the Milbank Memorial Fund has released three issue briefs meant to explain important topics in health care to policymakers. Two pertain to transforming payment and integrating providers in a multi-payer setting and one other I will return to below. What's interesting and laudable is that none of the studies covered in the issue briefs appeared in Milbank's journal, The Milbank Quarterly. Milbank is playing a valuable translational role, explaining research and not merely promoting that which appears under its own name. (Milbank is not unique in this regard.)
Christopher Koller is president of the Milbank Memorial Fund. "We really try to go where the interest is, and rely heavily on policymaker input," he said. "We try to start with what the research shows, and then add experience to it as appropriate."
I found the July 2014 Milbank Memorial Fund Issue Brief particularly interesting. It examined three erroneous claims about Medicaid and three papers that some use to support those claims:
- Claim: Medicaid coverage is worse for health than being uninsured. Paper: Primary payer status affects mortality for major surgical operations, by LaPar et al.
- Claim: Oregon's Medicaid expansion has not improved the health of patients receiving coverage. Paper: The Oregon Experiment — Effects of Medicaid on Clinical Outcomes, by Baicker et al.
- Claim: Medicaid coverage increases emergency department use. Paper: Medicaid Increases Emergency-Department Use: Evidence from Oregon's Health Insurance Experiment, by Taubman et al.
Though Aaron and I have written about all of these claims and studies before (links at the end of this post), I want to return to the first one.
The Issue Brief makes one important point I've made about the LaPar study previously. Its results are confounded by selection so outcomes cannot be said to have been caused by insurance status.
Although the study finds associations between Medicaid patients and poor health outcome, the study design prevents us from concluding that Medicaid is the cause of these differences. The study design does not account for many confounding variables, such as the Medicaid patients’ increased incidence of acquired immunodeficiency syndrome, depression, liver disease, neurologic disorders, and psychoses. The Medicaid patients also had the highest incidence of metastatic cancer among the groups studied.
The Issue Brief makes other points I do not recall making, one of which is very important: not all outcomes were worse for Medicaid patients.
[A] closer look at the study’s data shows a much more nuanced set of associations. First, the differences in health outcomes among the insurers were very small, and second, the outcomes were mixed. For example, fewer Medicaid patients died when having lung resections, pancreatectomy, and aortic aneurysm operations, while uninsured patients had better mortality outcomes in esophagectomy, colectomy, gastrectomy, hip replacement, and coronary artery bypass grafting (CABG).
Therefore, to cite this study as evidence that Medicaid causes worse outcomes than being uninsured is both confusing correlation with causation and also cherry picking types of surgery.
Another finding seldom mentioned by those citing this study is that after the outcomes were risk adjusted, privately insured patients had higher pulmonary, urinary, gastrointestinal, systemic, and procedure-related complications than did uninsured patients. But these findings are seldom used to argue that being uninsured is superior to being privately insured.
It's really not surprising that there are so many seemingly counterintuitive results from this study, as correlations can often suggest the opposite of the truth when important factors are not controlled for. I think the Milbank Issue Brief does a very nice job of highlighting just how troublesome it is to rely on this study to bolster the claim that Medicaid causes worse outcomes. The most logical conclusion, in light of results from stronger studies, is that this claim is false.
Additional, Relevant Reading
- All Milbank Memorial Fund Issue Briefs are here.
- On the paper listed in claim #3, above, look at the posts found here.
- Many posts about the paper and claim #2 are here. (Perhaps this one is a good summary.)
- About the study by LaPar et al. and claim #1 look here and here.
Austin B. Frakt, PhD, is a health economist with the Department of Veterans Affairs and an associate professor at Boston University’s School of Medicine and School of Public Health. He blogs about health economics and policy at The Incidental Economist and tweets at @afrakt. The views expressed in this post are that of the author and do not necessarily reflect the position of the Department of Veterans Affairs or Boston University.