• What’s the value of Medicaid? Read Chris Conover

    I recently discussed how the expansion of Medicaid through the ACA may benefit people living with HIV/AIDS. Specifically, getting insurance may encourage previously uninsured people at risk to get tested. Some of them will discover that they are HIV+, enabling them to benefit from early combined antiretroviral treatment (cART). If so, they will not only greatly extend their own lives, but also prevent significant numbers of fresh HIV infections, because cART radically reduces the probability of sexual transmission of HIV.

    This is a plausible argument for one benefit of expanding Medicaid. But plausible arguments are one thing, confirmatory data are another. A critic can and should ask: If expanding Medicaid really saves lives, why don’t we have data supporting a strong mortality benefit for people receiving Medicaid versus being uninsured?

    Chris Conover has a thoughtful post on just this question. He makes two valuable points. First, he questions some previous research that purported to show a large mortality cost of being uninsured. Second, he poses a nifty thought experiment designed to raise questions about whether expanding Medicaid is a cost-effective way to increase life expectancy.

    Conover begins by reviewing some claims for dramatic mortality benefits associated with becoming insured, based on observational studies (such as this one). He demolishes these claims, for reasons that are worth quoting.

    All are so-called observational studies meaning that the two groups being compared (uninsured and privately insured) each self-selected itself into the group. Many uninsured admittedly are in that group because they lack the means to pay for coverage, but the key point here is that unlike a randomized controlled trial, in which people are randomly assigned to be in either the “treatment” group or “control” group, there is no reason to suppose that the characteristics of the two groups will be comparable.

    This argument also demolishes the commonly heard claim that Medicaid is harmful compared to being uninsured. This claim is similarly based on observational studies that are methodologically weak for causal inference purposes, studies that Conover sensibly ignores.

    Unfortunately, Conover does not discuss the most relevant scientific studies on the effects of getting health insurance (like this one). This is the instrumental variables literature, summarized by Austin here. In these studies, researchers look at ‘experiments of nature,’ where large groups of people received or did not receive Medicaid for arbitrary reasons that mimic random assignment. Austin:

    My take-away from the Medicaid-[Instrumental Variable] literature review is: there is no credible evidence that Medicaid results in worse or equivalent health outcomes as being uninsured. […] [It] strongly suggests Medicaid is better for health than no insurance at all.

    This conclusion was reprised in the NEJM by Austin, Aaron, Harold Pollack, and Uwe Reinhardt.

    So what’s my bottom line on the mortality benefit of Medicaid? Getting Medicaid will likely give the average uninsured person at best a small mortality benefit, albeit of uncertain amount. It likely gives the small group of uninsured people with dire health conditions — e.g., people living with HIV/AIDS — a larger benefit. (And, although this is a different question, we could build a better insurance program for the poor than Medicaid, and it would likely have a better effect.)

    Conover, however, then poses a challenge to people who hold views like mine. He notes that there are preventive health interventions targeting smokers that would save lives with greater certainty than expanding Medicaid. So if we are interested in reducing mortality — and if the health benefit of Medicaid is likely small — then is increasing Medicaid the most cost-effective way to improve longevity?

    It would be uncharitable to read Conover as literally proposing to repeal the Medicaid expansion and instead invest billions of dollars in preventive health care. This is, instead, a worthwhile thought experiment. Conover wants people who defend the Medicaid expansion to explain why given the small (possibly negligible in his view, but not mine) and uncertain (his view and mine) mortality benefit.

    I have no trouble defending the expansion, in part because I have no need to do so on mortality benefits alone. Insurance has financial and quality of life benefits over and above its effect on mortality. And even though it may have only a small mortality benefit for the average uninsured person, it will mean life or death for others, and those people matter. For these reasons, people across the world view access to health care for all citizens as a matter of justice.

    Moreover, I do not attribute the small benefit of Medicaid primarily to the deficiencies of that scheme for financing health care. Remember, what insurance does for health is to provide access to health care. My view is that the effect of insurance is much smaller than it should be, given the money that we spend on it, because the effect of accessing health care is much smaller than it should be, given the money that we spend on it. Universal access to health care is a solvable problem: check almost any developed country. Conover’s thought experiment ought to shock us into asking the harder and deeper question: “Why doesn’t our health care system give the poor and rich alike better value for our health care dollars?”

    That question has many answers. Here are a few: We don’t pay enough attention to the social determinants of health. We don’t work hard enough to fix the manifold deficiencies in the quality of health care. We miss opportunities for cost-effective prevention. We don’t follow patients effectively, to support them in adhering to their treatments. And we don’t invest enough in medical research to achieve fundamental advances both in treatments and in how to deliver them.

    Perhaps Conover sees the matter differently. Nevertheless, I urge you to think through the point highlighted by his thought experiment, that extending health insurance has only a small and uncertain effect on mortality. One thing that I hope we can all agree on is that our reaction to the literature on the effect of Medicaid on health should be: We can do better.


  • Health insurance and outcomes

    This is a TIE-U post associated with Karoline Mortensen’s Introduction to Health Systems (UMD’s  HLSA 601, Fall 2012). For other posts in this series, see the course intro.

    As a health outcome, mortality has its advantages: apart from zombies, it’s a relatively unambiguous condition, and it’s unambiguously bad. In contrast, hospitalization, even emergency department use, can be good or bad, depending on the circumstances. Mortality has its disadvantages too: it’s a relatively rare event, requiring a long follow-up to observe in large numbers. And, it isn’t representative of the full spectrum of risks we want health care to mitigate. Still, it’s worth knowing the extent to which health insurance affects mortality. As recent events from the presidential campaign trail illustrate, there is some controversy in this area.

    It’s hard to know the effect of health insurance on mortality with great precision. One reason is that other things, like poor health and low income, lead to both increased mortality and uninsurance. Worse, not all of those other things are observable, so it’s hard to control for them. One often sees estimates of associations coupled with strained language that is all too easy to interpret as causation. That’s not to say that insurance doesn’t have a causal effect on financial and health outcomes, including mortality. It’s just very hard to estimate the sizes of the causal effects. All this could be addressed with a randomized trial in which people are randomly assigned to insured and uninsured states, but such an experiment is unethical. Clever researchers find other ways around the problem, e.g., the Oregon Health Study.

    This week, Mortensen assigned two readings that relate insurance to financial and health outcomes, including mortality: a primer on the uninsured by the Kaiser Family Foundation and a paper on health insurance and mortality by Andrew Wilper and colleagues (PDF). Both are ungated. Apart from my broad comments about association and causation above, instead of discussing either of these documents, I am going to summarize a relatively recent NBER paper that uses a clever technique.

    In an August 2012 NBER paper, Bruce Meyer and Laura Wherry applied a powerful statistical approach to the issue of whether Medicaid saves lives.

    This paper provides new evidence on the mortality effects of expansions in public health insurance eligibility for children in the mid 1980s to early 1990s. Our identification strategy exploits a unique feature of several early Medicaid expansions that extended eligibility only to children born after September 30, 1983. This feature led to a large cumulative difference in public health insurance eligibility for children born on either side of this birthdate cutoff. Children in families with incomes at or just below the poverty line gained close to five additional years of eligibility if they were born in October 1983 rather than just one month before. Black children were more likely to be affected by the Medicaid expansions and gained twice the years of eligibility on average as white children at this birthdate cutoff. The average gain in eligibility also varied by state of residence due to differences in state Medicaid eligibility rules prior to and during the expansions, as well as differences in the demographic characteristics of state residents.

    The authors are exploiting a discontinuity in children’s Medicaid eligibility. The intuition is that there is no reason to believe that kids born just before the September 30, 1983 cutoff are different from those born just after: there are no relevant, unobservable factors. A before/after cutoff date indicator is akin to a random flip of a coin. Broadly, your understanding of randomized controlled trials applies. The following chart shows the discontinuity in action.

    The authors argue that this regression discontinuity design is stronger than relying on state variation in Medicaid eligibility, the main technique in many of the studies on Medicaid outcomes I’ve reviewed on this blog.

    We should emphasize that this source of variation in Medicaid eligibility is not as clearly exogenous as birthdate. States with improving child mortality may be less likely to expand coverage, for example.

    The study results are nuanced.

    The results provide strong evidence of an improvement in mortality for black children under the Medicaid expansions. Although there is some evidence of a decline in mortality rates during the period of coverage (ages 8-14), we find a substantial decline in the later life mortality of black children at ages 15-18. The regression estimates indicate a 13-18 percent decrease in the internal mortality rate of black teens born after September 30, 1983. Because deaths due to internal causes are more likely to be influenced by access to medical care, this result supports an improvement in the underlying health of black children under the eligibility expansions. Early evidence indicates that this gain in health is not reversed during the early adult years. We find no evidence of an improvement in the mortality of white children under the expansions. Furthermore, we are unable to detect mortality improvements among children residing in those states with the largest gains in public health insurance eligibility under the expansions.

    Why wouldn’t mortality improvements be larger in states with larger gains in public health insurance eligibility? The authors explanation boils down to, eligibility isn’t everything:

    One potential explanation for this result is that larger expansions in eligibility do not directly translate into higher levels of enrollment or improvements in access to medical care for children. Increases in enrollment require outreach by states and the dissemination of eligibility information to parents. In addition, larger numbers of new beneficiaries may strain limited supply-side resources, particularly if physicians opt not to accept Medicaid clients due to low reimbursement rates or slow payment.

    Readers particularly interested in this subject should also be familiar with the recent article by Sommers, Baicker, and Epstein. See also the FAQ on the effects of health insurance on health.

    UPDATE: Added explanation for why there were no observed mortality improvements in states with the largest health insurance eligibility.


  • Oregon Health Study results, part 1

    Officially I’m on vacation, but I’m taking a break to write briefly about the the NBER-published first set of results from the Oregon Health Study, which are out today. The study has a randomized design and examines the difference in health care utilization and some self-reported outcomes for those on Medicaid vs. the uninsured. It is authored Katherine Baicker, Amy Finkelstein, Jonathan Gruber, Joseph Newhouse, Sarah Taubman, Bill Wright, Mira Bernstein, Heidi Allen and members of the Oregon Health Study Group.

    Just to refresh your memory, here’s what I wrote about that study in August 2010,

    Not since the RAND Health Insurance Experiment (HIE) has there been a randomized controlled experiment of the effect of insurance on health outcomes. Finally, a second one is underway, the Oregon Health Study (OHS). It’s being conducted by Heidi Allen, Katherine Baicker, Amy Finkelstein, Sarah Taubman, Bill J. Wright, and the Oregon Health Study Group who report on the study design in the most recent edition of Health Affairs.

    [T]he Oregon Health Study [is] a randomized controlled trial that will be able to shed some light on the likely effects of [Medicaid] expansions. In 2008, Oregon randomly drew names from a waiting list for its previously closed public insurance program. Our analysis of enrollment into this program found that people who signed up for the waiting list and enrolled in the Oregon Medicaid program were likely to have worse health than those who did not. However, actual enrollment was fairly low, partly because many applicants did not meet eligibility standards.

    Get excited! But don’t get too excited. The study runs through 2010 and no outcome results are available yet.

    Some results are available now. They’ve been reported by Ezra Klein, David Leonhardt, and Gina Kolata. Compared to the uninsured group, those in the Medicaid group:

    • received 30% more hospital care,
    • received 35% more outpatient care,
    • were 15% more like to use prescription drugs,
    • received 60% more mammograms,
    • received 20% more cholesterol checks,
    • were 15% more likely to have had a blood tested for high blood sugar or diabetes,
    • were 45% more likely to have had a pap test within the last year (for women),
    • had lower out-of-pocket medical expenditures and medical debt,
    • had a 40% lower probability of needing to borrow money or skip payment on other bills because of medical expenses,
    • incurred $778 more in spending on health care in one year, a 25% increase over the uninsured mean spending level,
    • were 25% percent more likely to report themselves in “good” or “excellent” health,
    • were 70% more likely to have a usual source of care,
    • were 55% more likely to see the same doctor over time,
    • reported better physical and mental health,
    • were 10% percent less likely to screen positive for depression.
    Klein mentions that “there was no evidence of “crowd-out”: Medicaid coverage didn’t make someone more or less likely to purchase private insurance.” One would hope that all of this would also lead to other objective measures of better health. However, a fuller examination of health outcomes are left for part 2 of the study, for which there are no results yet.

    I expect the research team will find that Medicaid does lead to better health. Such a finding would be consistent with some of the results above (better self-reported physical and mental health, less likely to be depressed, the myriad of higher propensity for preventative tests and treatment). It would also be consistent with a body of other evidence summarized on this blog (see the Medicaid-IV tagged posts) and in a NEJM paper by me, Aaron Carroll, Harold Pollack, and Uwe Reinhardt.

    In looking at the NBER paper, I see that the team used an IV method that addresses any potential bias due to unobservable differences between treatment and control groups. They used the selection by lottery as an instrument. It is clearly unrelated to outcomes but tightly related to placement in treatment and control groups. It’s a perfect instrument and an accepted, proven way to handle contamination or other unobservable differences between treatment and control groups in a ranomized design (see Angrist and Pischke).

    I look forward to reading subsequent work from the Oregon Health Study.

    UPDATE: Removed a paragraph on potential treatment-control bias based on feedback from study authors.

  • Medicaid and health outcomes (again)

    Avik Roy has read and posted about the papers I reviewed as part of my Medicaid-IV series. If you’ve forgotten, the purpose of that series of posts was to examine studies that use proven, sound methods to infer the causal effect of (as opposed to a correlation between) Medicaid enrollment on health outcomes. From that series, I concluded that there is no credible evidence that Medicaid is worse for health than being uninsured. Considering only studies that show correlations (not causation), Avik disagrees.

    Avik’s post is long, but you can save yourself some trouble by skipping the gratuitous attack on economists in general, and Jon Gruber in particular, as well as the troubled description of instrumental variables (IV).* About halfway down is his actual review of the papers; look for the bold text.

    The point I want to drive home in this post is why an IV approach is necessary in studying Medicaid outcomes. People enrolling in Medicaid differ from those who don’t. They differ for reasons we can observe and for those we can’t. An ideal study would be a randomized controlled trial (RCT) that randomizes people into Medicaid and uninsured status. Thats neither practical nor ethical. So we’re stuck, unless we can be more clever.

    The next best thing we can do is look for natural experiments. That’s what IV exploits. In this case, the studies I examined use the state-level variation in Medicaid eligibility (and related programs). That variation obviously affects enrollment into Medicaid (you can’t enroll unless you’re eligible), though it is not determinative. Importantly, state-level variation in Medicaid eligibility rules does not itself affect individual-level health. Other than figuratively, do you suddenly take ill when a law is passed or a regulation is changed? Do you see how Medicaid eligibility rules are somewhat like the randomization that governs an RCT, affecting “treatment” (Medicaid enrollment) but not outcomes directly? (If this is unclear, go here.)

    Note that IV studies can, and should in some cases, control for observable factors. (The studies I reviewed use quite sophisticated controls, including fixed effects and interactions, that greatly reduce the ambitiousness of the assumptions required to obtain causal estimates. In contrast, assumptions for inference of causality in the studies Avik prefers are far greater.) But controlling for observable factors alone is insufficient. That brings me to a study that Avik has cited many times as evidence that Medicaid produces worse health than no insurance at all. Tyler Cowen referenced the same study in his book, about which I wrote earlier. It’s the UVa surgical outcomes study, formerly known as: Primary Payer Status Affects Mortality for Major Surgical Operations, by LaPar and colleagues.

    Avik has summarized this study, so I’ll skip that. It examines 11 surgical outcomes by insurance status, adjusting for many observable factors, but, crucially, with no controls for unobservable factors that affect selection. All adjusted outcomes for Medicaid enrollees are worse than for the uninsured. With only one exception, adjusted outcomes for Medicare beneficiaries are worse than for the uninsured too. Got that? Not just Medicaid enrollees, but Medicare beneficiaries too, fare worse than the uninsured. Any theory to explain what’s going on in Medicaid had better explain Medicare too. It cannot be just that Medicaid enrollees see lower quality providers.

    You know what theory is consistent with these results? It’s a pretty famous one? I just described it above: selection (or omitted variable) bias. It is well known that studies that do not exploit purposeful (i.e., an RCT) or natural (i.e., natural experiment or instrumental variables) randomness can suffer from selection bias. Even controlling for observable characteristics is not enough in the field of health care. This is well known. I’ve explained it before, even in a diagram.

    The authors of the UVa surgical outcomes study acknowledge the possible presence of selection bias in trying to explain their results. They say as much in many places in the text of their paper,  writing,

    Another possible explanation for the differences we observed among payer groups is the possibility of incomplete risk adjustment due to the presence of comorbidities that are either partially or unaccounted for in our analyses [sic]. […]

    Several explanations for inherent differences in payer populations have been suggested. Factors including decreased access to health care, language barriers, level of education, poor nutrition, and compromised health maintenance have all been suggested. […]

    There are several noteworthy limitations to this study. First, inherent selection bias is associated with any retrospective study. […]

    For example, the proportion of Medicaid patients may be artificially inflated due to the fact that normally Uninsured patients may garner Medicaid coverage during a given hospital admission. […] [I]n our data analyses and statistical adjustments there exists a potential for an unmeasured confounder. Due to the constraints of NIS data points, we are unable to include adjustments for other well-established surgical risk factors such as low preoperative albumin levels or poor nutrition status.

    Kudos to the authors for acknowledging the limitations of their study. That the results have been repeated elsewhere without such disclaimers is a disservice to science.

    Moving on, to Avik’s great credit, he unearthed a Medicaid-IV study I had overlooked: The Link Between Public and Private Insurance and HIV-Related Mortality, by Bhattacharya, Goldman, and Sood (ungated PDF available). It examines mortality outcomes in an HIV population using IV methods to control for selection into insurance category (uninsured, public, and private). Table 5 is the key table. It confused me at first, as it has Avik. Just reading the table, it looks as if the “best” model produces the results in the bottom row, which suggest private insurance decreases mortality by 50% and public insurance increases it by 8%, relative to no insurance.

    But, reading the text, it is clear that the results in that bottom row are based on a faulty model, which the authors explain. (I will too, below.) The model based on sound methodology produces results in the second to last row of Table 5, a 79% and 66% reduction in mortality for the privately and publicly insured, respectively, relative to the uninsured. Table 6 also reports the results of the preferred model, though there is a typographical error on the mortality results: they’re missing minus signs in the first two rows. (I confirmed this with the authors.)

    The results of this study are stated very clearly by the authors, “both private and public insurance decrease the likelihood of death.”

    Now, what’s wrong with the model that shows Medicaid killing people, the one Avik thinks is best? It includes AZT and HAART** treatment indicators on the right hand side. That’s a problem because AZT and HAART treatment are more likely for those with insurance and HAART is indicative of poor health. Essentially, they’re “caused” by insurance and highly predictive of the outcome of interest, mortality. This is an example of “bad control,” i.e. controlling for an outcome. It should be clear that having the outcome — or something very close to it — on both the left and right hand sides is a problem. It soaks up too much of the effect of insurance but, being an outcome, it isn’t a proper control. About this, the authors write, “Of course, there is concern that HAART itself may be endogenous, since receipt of therapy almost certainly reflects disease severity and ability to adhere to the complicated regimen.” (Why was this faulty model even included in the paper? My guess has been confirmed by the authors via email: reviewers requested the authors include it. It was not included in their NBER working paper that predates the peer reviewed one. It really is too bad this model was inserted into the paper because it seems to have tricked some readers. However the authors very clearly indicate which model is most sound. Anyone appreciating the essence of good research design will understand it, as I explained above.)

    Bottom line: once again, we find that Medicaid is shown not to be bad for health, but only if proper econometric techniques are employed. Sadly, it is easier to ignore the need for such techniques and to misunderstand them than to do the work to educate oneself in their use. The real tragedy is that it leads to an unwarranted conclusion that Medicaid is harming people. We can certainly craft a better Medicaid program, and we should. But we should always use proper science in considering any program. If we don’t, we may mistake improvements to Medicaid as harmful. I’m sure advocates for change, myself included, would not welcome such an outcome.

    * Avik dismisses IV as a “fudge factor,” casually and erroneously discrediting a vast amount of mainstream work by economists and several entire sub-disciplines. Since IV is a generalization of the concepts that underlie randomized controlled trials (differing in degree, but not in spirit, from purposeful randomization), and can be used to rehabilitate a trial with contaminated groups — a not infrequent occurrence — it is unwise to trivialize IV and what it can do.

    ** HAART = highly active anti-retro-viral therapy.

    UPDATE: I fixed my explanation of the “bad control” problem in the Bhattacharya, Goldman, and Sood study.

    UPDATE 2: The authors of the Bhattacharya, Goldman, and Sood study confirmed the typos in Table 6.

    UPDATE 3: Those authors also confirmed that the faulty model was requested by reviewers, as I suspected.

  • Medicaid-IV summary

    I’ve found, read, and reviewed six papers in my Medicaid-IV series. These are ones that met my criteria for sound methodology for estimating the causal effect of Medicaid coverage on health outcomes. This post is a concluding summary.

    First let me note that just because I found six papers doesn’t mean there aren’t others that qualify. I just don’t know about them. So, point me to them, and I’ll take a look.

    Four of the six papers were either by Currie and Gruber or used methods pioneered by them. The three Currie and Gruber papers focused on maternal and child health, finding that 1980s and 1990s Medicaid expansions improved the health of children (increasing use of care, use of high-tech birth interventions, and birth weight, and decreasing infant and child mortality). Busch and Duchovny, using similar methods, found that late 1990s and early 2000s Medicaid expansions increased cancer-screening rates and decreased the likelihood of forgoing a doctor visit because of cost. A paper by Goldman and colleagues examined the effect of insurance on an HIV population using an IV technique with Medicaid-related instruments. Their methods were sound, and their results were not statistically significant.

    I’m not satisfied with the literature in this area, but that’s not the fault of the scholars who’ve worked in it. The areas of focus have been relatively narrow–prenatal and neonatal care, HIV patients–and therefore, not necessarily generalizable to the broader population. Moreover, much of the work with significant results was conducted with data nearly two decades old. Therefore, I think more research needs to be done to prepare for the 2014 expansion of Medicaid authorized by the ACA.

    The sixth paper I reviewed promises just that. The Oregon Health Study is a randomized trial of the effects of enrollment in the Medicaid program of that state. It’s the first randomized trial of health insurance since the RAND health insurance experiment. Results are not out yet but when they emerge they’re sure to be important and relevant.

    My take-away from the Medicaid-IV literature review is: there is no credible evidence that Medicaid results in worse or equivalent health outcomes as being uninsured. That is Medicaid improves health. It certainly doesn’t improve health as much as private insurance, but the credible evidence to date–that using sound techniques that can control for the self-selection into the program–strongly suggests Medicaid is better for health than no insurance at all.

    There are observational studies that purport to reveal otherwise, that Medicaid coverage is worse or no better than being uninsured. One cannot draw such conclusions from such studies if they do not control for the unobservable factors that drive Medicaid enrollment. Causal inference requires appropriate techniques. Even a regression with lots of controls, even propensity score analysis, is insufficient in this area of study.

    Finally, none of this means Medicaid is a program without flaws. It is badly in need of reform. It should be federalized or otherwise protected from state-level fiscal woes. Physicians and hospitals treating Medicaid patients should be reimbursed at rates closer to those of Medicare or private insurance. (That might mean lowering the latter, not only increasing the former.) So long as they’re evidence-based, I’m not opposed to adjustments in the design of Medicaid to increase the value of care delivered to the population that relies on it.

    However, what we should not do is fool ourselves into thinking Medicaid is not capable of improving health. Based on high-quality evidence to date, it is and it has.

  • Medicaid expansion and the technology of birth

    Remember my Medicaid-IV series? No? Guess what, I hardly recall it myself. So, let’s recap. Two months ago, in the first post of the series I wrote,

    An individual’s health status affects Medicaid enrollment (the ill are more likely to enroll). Medicaid enrollment affects an individual’s health status too (one can argue about which way, for the better or worse). The two are simultaneous. That makes inferring the causal effect of Medicaid on health outcomes difficult.

    A few weeks ago I described the right way to tease out the causal effect of Medicaid enrollment on health outcomes:

    There are undoubtedly studies that consider Medicaid vs. uninsured outcomes using the random variations provided by the natural experiment that is Medicaid. Characteristics of the program vary by state and year, making it a perfect set-up for such an analysis of this issue. […]

    [I] will […] describe the relevant literature as I read the papers. I’m not going to filter or cherry pick papers based on their findings. All that matters to me is the quality of the methods applied. Feel free to send me links to papers you think qualify (look for peer-reviewed, natural or randomized experiments and/or instrumental variables approaches; the run-of-the-mill observational study that controls for observable individual characteristics won’t do). […] When I think I’ve summarized them all, I’ll post a conclusion that reports on the full body of evidence.

    To date, readers have sent me zero papers. But I’ve found six on my own, the one summarized below being the last. In a post later this week I’ll sum up all that I’ve found.

    I’ve already summarized two 1996 papers by Currie and Gruber. They wrote a third, a 1997 NBER working paper titled “The Technology of Birth: Health Insurance, Medical Interventions, and Infant Health.” The techniques are similar to those of their other papers so you can click back to read about them. I’ll just cut right to the results. From the abstract:

    [U]sing Vital Statistics data on every birth in the U.S. over the 1987-1992 period [we study t]he effects of insurance status on treatment and outcomes […] identified using the tremendous variation in eligibility for public insurance coverage under the Medicaid program over this period. Among teen mothers and high school dropouts, who were largely uninsured before being made eligible for Medicaid, eligibility for this program was associated with significant increases in the use of a variety of obstetric procedures. On average, this more intensive treatment was associated with only marginal changes in the health of infants, as measured by neonatal mortality. But the effect of eligibility on neonatal mortality is sizeable among children born to mothers whose closest hospital had a Neonatal Intensive Care Unit, suggesting that insurance-induced increases in use of `high tech’ treatments can have real effects on outcomes. Among women with more education there is a counter- vailing effect on procedure use. Most of these women had private insurance before becoming Medicaid-eligible, and some may have been ‘crowded out’ onto the public program. These women moved from more generous to less generous insurance coverage of pregnancy and neonatal care. This movement was accompanied by reductions in procedure use without any discernable change in neonatal mortality.

    A shorter and cruder summary is that Medicaid increases utilization of high-tech interventions, relative to being uninsured, and this has measurable health effects. However, relative to private insurance, Medicaid reduces the use of high-tech procedures and has no effect on health.

  • Medicaid expansion and health care utilization

    This is the next  post in my Medicaid-IV series. I presume anyone interested in this post has read all the others in the series, so I don’t have to repeat myself. This post is about the 2005 JHE article by Busch and Duchovny, “Family coverage expansions: Impact on insurance coverage and health care utilization of parents.” Because my interest at the moment is about the “health care utilization” part and not the “insurance coverage” part of the paper, I’ll skip much of the paper (though I read it!).

    The relevant part of the abstract reads,

    With the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), Medicaid eligibility ceased to be tied to receipt of cash assistance. Since then, states have had a growing number of opportunities to expand health coverage to low-income working families beyond previous AFDC limits. As of 2001, 20 states have raised income eligibility limits for parents to or beyond 100% of the Federal Poverty Level. … [W]e use [1995-2002] Behavioral Risk Factor Surveillance System [data] to examine changes in health care utilization. We find that these expansions increased cancer-screening rates. Of previously uninsured mothers not receiving cancer screening, 29% now receive these screens. Finally, our results indicate the expansions decreased the likelihood that a parent needed to see a doctor but did not because of cost.

    Yep, that pretty much sums up the findings. Of course, methods matter. The authors use the same style of instrumental variables method as Currie and Gruber (I told you it had become standard). Since I’ve assumed you’ve read prior series posts, you know what I’m talking about: the key independent variable is Medicaid eligibility and it is instrumented with the average Medicaid eligibility rate under each of the year-state Medicaid rules where the average is computed over a fixed population.

    The authors’ conclusion has a nugget worth sharing, just because it is a segue to an important aspect of the ACA’s Medicaid expansion. In the authors’ analysis, they only considered income-based eligibility changes. But there were other changes to Medicaid eligibility during the period of study as well, including elimination or modification of asset tests and income disregards. The ACA’s Medicaid expansion is typically described in terms of income (increasing eligibility up to 133% of the federal poverty level) but it is worth emphasizing that it also eliminates the asset test and income disregards.

  • Medicaid and saving babies

    As mentioned at the end of my prior post in the Medicaid-IV series Janet Currie and Jon Gruber published a 1996 paper on the effect of Medicaid expansion on infant mortality and birth weight. Here’s the abstract:

    A key question for health care reform in the United States is whether expanded health insurance eligibility will lead to improvements in health outcomes. We address this question in the context of the dramatic changes in Medicaid eligibility for pregnant women that took place between 1979 and 1992. We build a detailed simulation model of each state’s Medicaid policy during this era and use this model to estimate (1) the effect of changes in the rules on the fraction of women eligible for Medicaid coverage in the event of pregnancy and (2) the effect of Medicaid eligibility changes on birth outcomes in aggregate Vital Statistics data. We have three main findings. First, the changes did dramatically increase the Medicaid eligibility of pregnant women, but did so at quite differential rates across the states. Second, the changes lowered the incidence of infant mortality and low birth weight; we estimate that the 30-percentage-point increase in eligibility among 15-44-year-old women was associated with a decrease in infant mortality of 8.5 percent. Third, earlier, targeted changes in Medicaid eligibility, which were restricted to specific low-income groups, had much larger effects on birth outcomes than broader expansions of eligibility to women with higher income levels. We suggest that the source of this difference is the much lower take-up of Medicaid coverage by individuals who became eligible under the broader eligibility changes. Even the targeted changes cost the Medicaid program $840,000 per infant life saved, however, raising important issues of cost effectiveness.

    This study shares the same methodological approach, and many of the strengths and weaknesses of the Currie and Gruber paper I reviewed previously. So, I’m not going to repeat myself. There is one element of this study worth emphasizing, however. As stated in the abstract, the authors examined two types Medicaid expansions in the 1980s, one targeted and one broad.

    The targeted expansions were essentially modest changes to Medicaid eligibility around the edges of the program’s ties to AFDC (I’m obviously grossly simplifying). The broad expansion began in 1987 and liberalized the income cutoffs for pregnant women. By 1990 all states were required to cover pregnant women with incomes up to 133% of poverty and had the option of extending coverage up to 185% of poverty with federal matching funds.

    Results of the study differ across the two types of expansions. The targeted expansion had much stronger effects:

    [W]e find that a 30-percentage-point increase in eligibility under targeted programs would have been associated with a highly significant 7.8 percent decline in the incidence of low birth weight; a similar increase in eligibility under the broad programs would have decreased the incidence of low birth weight by only 0.2 percent. Similarly, a 30-percentage-point increase in targeted eligibility would have been associated with an 11.5 percent decline in infant mortality, compared to a 2.9 percent decline under the broad policy changes.

    The authors attribute this difference in outcomes across type of expansion to different rates of take-up. Lower take-up under the broad expansion attenuated its effect. To the extent that these findings can be generalized, they would seem to suggest that the broad Medicaid expansion under the ACA will have relatively small effects on health. However, the ACA’s expansion comes with an individual mandate, so take-up should occur at a much higher rate than under the broad expansions in the 1980s.

  • Medicaid and child health

    Next up in my “Medicaid-IV” series–in which I’m reviewing papers that use instrumental variables techniques to estimate the effects of Medicaid on health outcomes–is the widely-cited 1996 Quarterly Journal of Economics paper by Currie and Gruber on Medicaid and child health (link to ungated version).

    Not surprisingly, the authors do a superb job of explaining their approach and interpreting their results. So, I’m going to liberally quote from the paper. Let’s start with the abstract just to get an overview, then I’ll hit some important issues not fully revealed by such a brief summary.

    We study the effect of public insurance for children on their utilization of medical care and health outcomes by exploiting recent expansions of the Medicaid program to low-income children. These expansions doubled the fraction of children eligible for Medicaid between 1984 and 1992. … [E]ligibility for Medicaid significantly increased the utilization of medical care, particularly care delivered in physicians’ offices. Increased eligibility was also associated with a sizable and significant reduction in child mortality.

    By “exploiting recent expansions of the Medicaid program” the authors mean they use state-year variations in those expansions to construct an instrument that is not correlated with individual characteristics but is correlated with Medicaid eligibility, and therefore with Medicaid enrollment. The instrument and how it works are mind-benders (I didn’t get it upon first encounter). It’s the average Medicaid eligibility rate under each of the year-state Medicaid rules where the average is computed over a year-but-not-state-varying population of kids. (I know that’s hard to grok. I could spend a whole post explaining it further, but I won’t. You’ll have to trust me that it is a valid instrument and has become standard technique for instrumenting for Medicaid status. Or you can read the paper. This is advanced material!)

    A good question is, “Why the focus on kids?” Currie and Gruber have a great answer:

    A potential problem with utilization measures, however, is that they confound access and morbidity. For example, the Medicaid expansions may have increased access to hospitals, but at the same time they could have increased the use of preventive care, improving health status and reducing the demand for hospital care. One way to surmount this problem is to focus on utilization that is explicitly preventative, and therefore unaffected by morbidity. Pediatric guidelines recommend at least one doctor’s visit per year for most children in our sample, so that the absence of a doctor’s visit in the previous year is suggestive of a true access problem, regardless of underlying morbidity.

    The results are well-summarized by the abstract quoted above, but I want to highlight a few things. The authors find that Medicaid eligibility cuts the probability in half that a child will go a year without seeing a physician in any setting. Much of this is due to increased visits to doctors’ offices. They also find that the 15.1 percentage point rise in Medicaid eligibility during the study period reduced child mortality by 5.1 percent. Their sub-analysis of mortality is sharp:

    If Medicaid eligibility reduces deaths by improving the utilization of care, then we would expect deaths due to “internal causes” (such as disease) to fall more than deaths due to “external causes” (such as accidents, homicides, suicides, and other external causes). [The results] show that this is indeed the case: increases in eligibility are correlated with a significant reduction in deaths due to internal causes, but have no significant effect on deaths due to external causes.

    I’ve saved the most puzzling finding for last, Medicaid eligibility was found to increase hospital visits. That sounds bad, and maybe it is. But the mechanism could be benign, as the authors explain.

    [H]ospitals may be better equipped to assist patients in claiming benefits. Potential eligibles for Medicaid must complete lengthy and complex application forms, provide extensive documentation (such as birth certificates, pay stubs, and confirmation of child care costs), and attend several interviews with caseworkers. … In response, many hospitals have established special offices, or contract with private companies, to assist Medicaid eligibles in completing these procedures. … The nontrivial costs of providing these services may be beyond the means of private doctors and clinics, leading them to recommend that potential eligibles seek care in a hospital setting.

    Before closing, it is worth noting two things. One, the control variables in the regressions do not include health status. That’s important since health status could be an outcome of Medicaid enrollment. (Inclusion of an outcome as a control variable leads to bias.) Second, as the authors point out, Medicaid expansions have two effects. They encourage additional Medicaid enrollment and discourage private coverage. Some new Medicaid enrollees had been privately insured, am effect known as “crowding out.” The estimates include all effects of Medicaid expansion on outcomes, including that due to crowding out, but do not distinguish among them.

    Finally, there is a question of generality of the findings. This is a study of Medicaid expansions that targeted children about 20 years ago. The ACA’s Medicaid expansion is far broader and occurring in four years from now. Can one generalize the findings of Currie and Gruber to other populations and eras? It’s hard to say. The authors published another paper that used the same techniques and focussed on the effect of Medicaid expansions for pregnant women, finding they lowered infant mortality and increased birth weight. So, the positive effects of Medicaid expansions on outcomes apply to more than one population, which strengthens claims of generality.

  • The Oregon Health Study

    Not since the RAND Health Insurance Experiment (HIE) has there been a randomized controlled experiment of the effect of insurance on health outcomes. Finally, a second one is underway, the Oregon Health Study (OHS). It’s being conducted by Heidi Allen, Katherine Baicker, Amy Finkelstein, Sarah Taubman, Bill J. Wright, and the Oregon Health Study Group who report on the study design in the most recent edition of Health Affairs.

    [T]he Oregon Health Study [is] a randomized controlled trial that will be able to shed some light on the likely effects of [Medicaid] expansions. In 2008, Oregon randomly drew names from a waiting list for its previously closed public insurance program. Our analysis of enrollment into this program found that people who signed up for the waiting list and enrolled in the Oregon Medicaid program were likely to have worse health than those who did not. However, actual enrollment was fairly low, partly because many applicants did not meet eligibility standards.

    Get excited! But don’t get too excited. The study runs through 2010 and no outcome results are available yet.

    The paper includes a nice summary of observational, quasi-experimental, and experimental studies of the effects of public insurance programs on outcomes. As I’ve written before, observational studies are not of primary interest to me. There’s only been one other experimental study (RAND HIE). About the quasi-experimental studies, the authors write,

    Some studies have found evidence that public health insurance reduces mortality among infants and children [8-10] and improves some outcomes—although not mortality—among the elderly. [11–14] Although they are much more persuasive than observational studies, quasi-experimental studies are not truly randomized. Thus, investigators must rely on the assumption that the people whose health insurance was affected by environmental or policy changes are otherwise identical to the people in the comparison group.

    The authors’ references 8-14 are listed below. References 8 and 9 pertain to Medicaid and are on my list of papers as part of the “Medicaid-IV” project.


    8. Currie J, Gruber J. Saving babies: the efficacy and cost of recent expansions of Medicaid eligibility for pregnant women. J Polit Econ. 1996;104(6):1263–96.

    9. Currie J, Gruber J. Health insurance eligibility, utilization of medical care, and child health. Q J Econ. 1996;111(2):431–66.

    10. Hanratty MJ. Canadian national health insurance and infant health. Am Econ Rev. 1996;86(1):276–84.

    11. Card D, Dobkin C, Maestas N. The impact of nearly universal health care coverage on health care utilization: evidence from Medicare. Am Econ Rev. Forthcoming.

    12. McWilliams JM, Meara E, Zaslavsky AM, Ayanian JZ. Health of previously uninsured adults after acquiring Medicare coverage. JAMA. 2007; 298(24):2886–94.

    13. McWilliams JM, Meara E, Zaslavsky A, Ayanian JZ. Use of health services by previously uninsured Medicare beneficiaries. New Engl J Med. 2007; 357(2):143–53.

    14. Finkelstein A, McKnight R.What did Medicare do? The initial impact of Medicare on mortality and out of pocket medical spending. J Public Econ. 2008;92(7):1644–69.