Re (3): UVa's Surgical Outcomes Study

Cross-posted from The Agenda on National Review Online.

Austin Frakt is back for more, with two more posts on the issue of Medicaid’s poor performance in outcome studies vs. the uninsured. I am afraid that Austin makes a couple of analytical errors. These will take some time to go through, so I forgive readers who don’t want to wade in the tall weeds with me. In summary, the evidence Austin cites to argue that Medicaid might be just fine actually demonstrates the opposite: that Medicaid patients get worse medical care than do the uninsured, and that PPACA’s massive expansion of Medicaid will have tragic consequences.

To begin with, Austin is re-energized by some useful commentary from Aaron Carroll. Writes Carroll:

- Insurance doesn’t equal care.  Insurance can affect how likely you are to get care and how quickly you might get it.  But any study that looks at insurance has to adjust for many, many other variables in order to get the true effect of insurance.

- There is a large body of literature out there on insurance.  A lot of it shows that people with private insurance do better than those with public insurance or those without insurance; that should not be a surprise.  Most people (and most of your docs) would rather have private insurance than Medicaid.  But would you really rather have no insurance than Medicaid?  If so, that is everyone’s right.  Don’t get the Medicaid.  I wager few would make that choice.

- I find it interesting that most of the literature that Avik cites is about surgery.  Surgery is different than other types of care (like emergency care) in that it is harder to refuse.  So it may be that the uninsured are getting care on a compassionate basis.  Few would provide a screening mammogram or yearly colonoscopy to someone uninsured, however, and you would get that with Medicaid.

I entirely agree that insurance does not equal care: indeed, this is the point I have been trying to make all along. Medicaid, in particular, because it is heavily underfunded and mismanaged, is insurance of a kind, but not care. And as I pointed out in my previous post, arguments that the studies don’t adjust for enough variables miss the point. We are trying to distinguish between cause and effect: what elements of Medicaid’s poor outcomes are caused by Medicaid, and what elements are inherent to the Medicaid population? Clearly, things like one’s race or gender or age are not caused by Medicaid. But the chronically poor care that Medicaid patients receive, because they have poor access to physicians, is directly responsible for many of their comorbidities. Therefore, correcting for comorbidities, while a useful tool, treats Medicaid overly generously.

Medicaid, due to its extreme underpayment and other problems, gives its beneficiaries very poor access to health care—things like annual checkups, screenings, etc. In Texas, for example, only 32 percent of all physicians are willing to see Medicaid patients. A nationwide survey of major metropolitan areas found that 55% of physicians accepted Medicaid, with particularly low numbers for specialists.

Skeptics of the Virginia study want us to say: “The studies are flawed! They don’t consider all of subtle ways in which Medicaid patients are worse off!” Well, what if the main reasons that Medicaid patients are worse off has something to do with…Medicaid?

As to Carroll’s point about the fact that the uninsured aren’t getting mammograms and colonoscopies, but Medicaid patients are: that isn’t true, because Medicaid patients can’t get appointments to see doctors (see above). If Carroll was right, it would only render even more striking the fact that the uninsured have their cancers detected earlier than do Medicaid patients. With breast cancer specifically (since he brings up mammograms), Medicaid patients were 31% more likely to have late-stage breast cancer than the uninsured. With surgical resections for colon cancer (since he brings up colonoscopies), in a separate study, Medicaid patients had a 27% higher risk of mortality than the uninsured, and a 9% higher risk of surgical complications.

Carroll makes another minor point, which Frakt echoes:

I’m going to take a tiny issue with Avik’s first post.  It was based on (as far as I can tell) a meeting abstract…I’m not saying the results of Avik’s discussed study aren’t valid.  I’m saying I can’t tell.  And neither can you, without more information.  The peer review for a meeting just isn’t the same as for full publication.  You have less time, different criteria, and almost nothing by which to judge the work.

As I have commented elsewhere, the full study is in press in the Annals of Surgery. It should be out in a few weeks, relieving Carroll's concerns. I certainly agree that the detailed study results, published in a peer-reviewed journal, will be useful. Skeptics of the study are setting themselves up for disappointment if they hope that the detailed results will exonerate Medicaid. Indeed, the detailed results are likely to show that social factors (e.g. alcohol and drug abuse) were counted as comorbidities, and were not meaningfully different between Medicaid and the uninsured anyway.

The Virginia study is based on the largest publicly available inpatient database in the U.S., that of the Nationwide Inpatient Sample. The NIS represents approximately 20% of all hospital discharges in the country. The UVa study is the most comprehensive analysis, with the largest sample size, of any published study comparing surgical outcomes on the basis of insurance status.

So, back to Austin Frakt. Austin implies that the Virginia study, and the other studies I cite, are “cherry-picked.” I can assure him that the four studies I described are representative. He is more than welcome to slog through the medical literature, as I did, and see if he comes up with a different conclusion. He won’t.

Austin then goes and picks out a single study, the thing he had earlier criticized me for, and uses that study to suggest that it is premature to conclude that Medicaid patients fare worse than the uninsured. Even worse, Austin actually misinterprets the results of the study he cites! Let us quote its authors (emphasis mine):

Most of the reasons for insurance-related disparities noted above for the uninsured are also applicable to Medicaid patients. Differences in the intensity of inpatient care, limited access to health care services, unmet health needs, and suboptimal management of chronic medical conditions were also reported for Medicaid patients in prior research. These factors likely contributed to the higher in-hospital mortality in this patient population, evidenced by the sequential decrease in odds after adjusting for comorbidities and disease severity.

Translation: Medicaid patients have poorer access to physicians. As a result, their diseases are managed very poorly. This leads them to die in the hospital at higher rates than the uninsured (in Austin’s study, Medicaid patients had a higher mortality than the uninsured with heart attacks and pneumonia, but did better than the uninsured with strokes). As the authors note, additional adjustments for risk factors increase the evidence that Medicaid is the problem, because those risk factors are the ones that Medicaid exacerbates.

I am glad that Austin has taken on this challenging subject. But I encourage him to take the effort he spends attempting to dispute that Medicaid has problems with access to care and medical outcomes, and apply it to actually solving those problems. (Suggesting that we throw an extra $200 billion a year into Medicaid is not viable.) I hope I am wrong, but I fear that he is less than enthusiastic about addressing the problems, because that would require him to acknowledge that PPACA’s massive expansion of Medicaid was a catastrophic mistake.