Roy vs. Pollack on health outcomes, Medicaid, and welfare dependency
by Harold Pollack
I appreciate that Avik Roy took the time to respond to my last post. There needs to be more talk across the ideological divide. Read what he wrote, to make sure that I fairly describe it. Here's my take.
In brief, I stand by what I wrote. I'm also unimpressed by Roy's bromides regarding Medicaid's furtherance of "welfare dependency, that leads to family breakdown and social disrepair."
I should say at the outset that I didn't know that Roy is an investment professional who evaluates clinical trials. That certainly gives him an interesting perspective on many subjects. I still believe he is foolish to mix things up with Austin Frakt and Aaron Carroll on Medicaid policy.
There is no evidence that Medicaid is causally harmful or worse than being uninsured. Certainly that assertion is not supported by the UVa study that Roy cites. He way over-interprets the findings. He never satisfactorily addresses the specific critiques that Austin Frakt or I raised in response to his columns.
There are many good reasons to believe that basic health coverage improves population health. (For more, see here.) PPACA does some important things to strengthen these links and to raise Medicaid reimbursement rates. I would have liked to see the legislation do more to maximize the population health benefits.
There is also much evidence that Medicaid is much less effective than it needs to be in improving recipient health. Low provider reimbursement rates and poor administration are the critical defects.
As the guardian of an adult intellectually disabled man who is dually-eligible for Medicare and Medicaid, I am well aware of Medicaid's shortcomings. He moved into our house weighing more than 340 pounds. He had multiple medical and social challenges, not to mention an improperly diagnosed genetic disorder. He had many issues that would not be controlled in the UVa study.
I've written many, many columns and op-eds about the need to raise Medicaid reimbursement rates (e.g. here and here and here and here) and the need to address Medicaid's other shortcomings. At the same time, I am grateful for what Medicaid has done to help my extended family under difficult circumstances. For all its shortcomings, Illinois Medicaid has paid huge amounts of money to finance my brother-in-law Vincent's essential care at the University of Chicago Medical Center to address some serious medical concerns.
Roy identifies genuine issues that require serious analysis. We do need to be careful about policies that may lead people—particularly sick people—to move from private insurance into Medicaid. We don't really know if this is helpful or harmful. I take Roy's point that there are reasons to be concerned.
I myself would modify the Affordable Care Act to give Medicaid recipients greater freedom to participate in health insurance exchanges. Governors have been pushing for this, since this shifts some of their Medicaid costs onto the federal government. I'm fine with that. I would offer this option to the healthy low-income and TANF groups. I would not try this with the elderly or the disabled, whose specialized issues may be poorly-handled in the new exchanges. I hope to say more about this on another occasion.
Roy also raises methodological issues regarding instrumental variables analysis. This issue can't be answered with generalities. It is much more context-specific. I agree that IV analyses are rarely equivalent to a true randomized trial. The question really is: what do we do when randomized evaluations aren't/can't be done. I'll let Jon Gruber and Austin Frakt tackle that question.
There is one final point, which gets to a difference in mentality and, well, respect for men and women who require government help. Responding to my empirical claims, Roy writes:
Let’s go through these one-by-one. Medicaid patients face greater socioeconomic obstacles: something that can be exacerbated, if not caused, by welfare dependency. Medicaid patients use different hospitals: something that is in many cases a direct result of their insurance status. Some receive Medicaid because they have specific disabilities: which is controlled for in the study. Privately insured patients may have better access to surgery: a direct result of insurance status.
Do you see a pattern here? Many of the factors Harold raises as flaws of the study are actually flaws of Medicaid. It’s Medicaid that restricts access to the best hospitals and the best doctors and the best treatments. It’s Medicaid, i.e., welfare dependency, that leads to family breakdown and social disrepair. (For those who seek a more extensive discussion of this problem, read Charles Murray’s landmark book, Losing Ground: American Social Policy 1950-1980.)
Consider his comment: "It's Medicaid, i.e. welfare dependency, that leads to family breakdown and social disrepair." Roy twice laments "welfare dependency." I'm not exactly sure what he means here. My in-laws cared for Vincent in their family home for almost 40 years, until the day his mother died. There was no family breakdown or social disrepair to fix, only the need to provide care for a complicated and tragic disability. If it weren't for Medicaid and accompanying programs, my in-laws might well have needed to institutionalize Vincent in a forbidding public institution.
Roy tries to clarify things by writing: "For those who seek a more extensive discussion of this problem, read Charles Murray’s landmark book, Losing Ground: American Social Policy 1950-1980."
This doesn't help. Murray's (in my view deeply flawed) book appeared a quarter-century ago. It was written to critique an AFDC system that doesn't exist anymore. In 1996, about 12.3 million single mothers and children received traditional welfare. Today, 4.3 million do. The typical Medicaid surgical patient is elderly or disabled, or resides in a household headed by low-income adults who are not welfare recipients.
Roy should be much more respectful towards traditional welfare recipients, too. I've done a fair amount of research on this population, including various surveys of low-income, predominantly African-American single mothers. (Kristin Seefeldt's work covers some of the pertinent issues for this post better than I do.) For many of these women and their children, Medicaid is an essential resource in staying healthy and becoming economically self-sufficient.
Medicaid must do more to give patients access to good primary and specialty care, and to address social determinants of health. Roy remains off-base in attacking the program, and, at times, in his disparagement of so many Americans who rely on Medicaid for help.