The Truth About Medicaid: How Paul Ryan Misconstrues the Evidence
SOURCE: AP/David Goldman
- Endnotes and citations are available in the PDF and Scribd versions.
- Download the report:
- Read it in your browser:
A recent report by House Budget Committee Chairman Paul Ryan (R-WI) includes an inaccurate and misleading portrayal of Medicaid, a joint federal-state health insurance program for children and certain low-income or disabled adults. Although Ryan’s report about the pitfalls of Medicaid relies on numerous studies to back his assertions, he misinterprets data, uses studies with weak methodology, and fails to include key results—some from the very studies he cites—that underscore the benefits of Medicaid coverage.
A more comprehensive, careful review of existing Medicaid research offers a very different picture from Ryan’s report. Benefits of Medicaid coverage include greater access to much-needed health care, improved health outcomes, and improved financial security for enrollees. This report dismantles Ryan’s claims by providing a clear, full picture of the evidence on Medicaid.
Medicaid’s effect on enrollees’ health and service utilization
Ryan makes numerous misleading claims about the effect of Medicaid coverage on beneficiaries’ health. To untangle and more accurately assess how Medicaid affects health care utilization and health outcomes, it is critical to first understand the demographics and health needs of the Medicaid population.
The demographics of the Medicaid population
Medicaid coverage is particularly important for low-income individuals, as they often have greater, more complex health needs.
Ryan’s version: Ryan misuses research to imply that Medicaid coverage leads to poorer health.
Ryan’s mistakes: The studies Ryan relies on show a correlation, not causation, between enrollees’ health and Medicaid coverage. Ryan suggests that these statistics show that regardless of age, enrollees in Medicaid or the Children’s Health Insurance Program, or CHIP, a public health insurance program for low-income children, are in poorer health than those covered by private insurance or the uninsured because of their enrollment in these programs. But Ryan’s use of these data is incredibly misleading for several reasons.
- The Medicaid population has significantly different health needs and demographics than privately insured populations. The Ryan report fails to acknowledge a number of additional—and important—factors such as lower socioeconomic status, prior difficulty accessing health care, and lower levels of health literacy that contribute to and affect the health outcomes of Medicaid enrollees. A large body of literature identifies various social determinants of health, including socioeconomic status and living and working environments, as risk factors for poor health outcomes. Medicaid enrollees therefore often have higher rates of chronic conditions, disability, and functional limitations. The greater, more complex health needs of these persons—not the fact that these enrollees are covered by Medicaid—affect service use and cost of care. One study demonstrates that service use for Medicaid beneficiaries would not differ significantly if they were instead enrolled in an employer-sponsored plan.Research also shows that even low-income, privately insured enrollees have very different socioeconomic and health characteristics when compared with Medicaid enrollees—making the comparison between the health outcomes and service utilization rates of Medicaid enrollees’ and the privately insured incorrect, as the Medicaid population not only starts out sicker, but faces more barriers to improving and maintaining good health.
The truth: Due to a number of interrelated factors, Medicaid enrollees are on average sicker than persons covered by private insurance and those who are uninsured. Thus, the data show that because Medicaid enrollees have greater health needs than other populations, coverage is absolutely essential in allowing them to access needed medical care.
Utilization of health services
Medicaid coverage allows enrollees to seek out needed health care services.
Ryan’s version: Ryan incorrectly asserts that Medicaid coverage improperly increases enrollees’ use of health care services.
Ryan’s mistakes: Ryan misuses two studies to claim that Medicaid coverage increases beneficiaries’ use of health services, including preventive care and emergency department, or ED, services.However, both studies cannot accurately isolate the effect of Medicaid coverage on service use, as they compare Medicaid enrollees’ use of these services to uninsured populations, who are less likely to use health care services due to significant financial barriers.
Additionally, one of the studies actually states that “neither theory nor existing evidence provides a definitive answer to … whether we should expect increases or decrease in emergency-department use when Medicaid expands.” Further research shows that the evidence on the effect of Medicaid coverage on ED services is indeed mixed. For example, a study of ED utilization by children after a Medicaid eligibility expansion found no statistically significant change. Studies of ED use after Massachusetts’ health insurance expansions also found either no change or reductions in utilization.
The truth: It is likely that Medicaid enrollees use more health care services than the uninsured both because the uninsured face significant financial barriers to receiving care and because many Medicaid enrollees seek out care for the first time after being uninsured. Presenting data that Medicaid enrollees use more health services than the uninsured affirms that insurance coverage allows people who need care to seek it out, and that being uninsured is a major barrier to receiving important medical care.
Health of enrollees
Medicaid coverage increases people’s ability to access valuable medical care, which improves the health outcomes of enrollees.
Ryan’s version: Ryan inaccurately claims that Medicaid coverage has little positive effect on enrollees’ health.
Ryan’s mistakes: Ryan cherry picks data points from a study on Oregon’s Medicaid program to erroneously claim that Medicaid does not improve the health of enrollees. Although the results of this study were frequently misused to decry Medicaid expansion, numerous scholars looked more closely at the data and the study’s methodology.
The study was randomized, which is often the gold standard for clinical research. Its randomized design, however, resulted in the study including far too few people with high blood pressure or high cholesterol to meaningfully measure the effects of Medicaid coverage on their health. As Ryan acknowledges, this coverage allowed enrollees to detect and begin treating serious health conditions, lowering rates of depression and increasing detection and management of diabetes.
Ryan claims that Medicaid does not improve beneficiaries’ health despite evidence that Oregon Medicaid beneficiaries received more preventive care, screenings, and treatment for conditions such as diabetes—and evidence that these services promote and improve health. It is also reasonable to expect that significant improvements in beneficiaries’ health may take more time than the study’s two-year period.
The truth: More data are needed to more accurately assess the effect of Medicaid on enrollees’ health outcomes, but the Oregon study illustrates the early benefits of Medicaid coverage. Moreover, Ryan ignored a number of additional studies that show that Medicaid coverage improves health outcomes and mortality rates.
- Medicaid coverage is associated with lower mortality rates. A 2012 study compared the health outcomes of people in states who had previously expanded Medicaid eligibility with neighboring states that did not expand eligibility. It found that expanded Medicaid coverage was associated with a 6 percent reduction in mortality when adjusted for demographic factors, local median income, and unemployment rates. (Unadjusted, Medicaid expansion was associated with an 8 percent reduction in mortality rates.) This drop in mortality applied to all residents in these states—not just to Medicaid patients. Additionally, more residents in expansion states reported “excellent” or “very good” health after their states expanded eligibility.
- Medicaid coverage improves infant and child health. A study examining the effect of both targeted and broad Medicaid eligibility expansions in the 1980s for pregnant women found that increased eligibility reduced the number of low-birth-weight deliveries and lowered the number of infant deaths. Reducing the number of low-birth-weight deliveries not only helped decrease costs of expensive neonatal hospital care, but more importantly it lowered the odds of these infants later developing disorders associated with low birth weight, including cerebral palsy, major seizure disorders, blindness, deafness, and learning disorders.
The same authors also studied the effect of Medicaid eligibility expansions for low-income children, finding that increased Medicaid eligibility was associated with a “sizable and significant” reduction—5.1 percent—in child mortality. Importantly, the authors are able to isolate the effects of Medicaid coverage by examining the rates of child deaths by cause: They found that lower child mortality rates were due to decreased deaths from “internal causes” such as disease, not “external causes” such as homicide or accidents.
- Medicaid coverage improves HIV mortality rates. A study examined the effect of Medicaid coverage on mortality for HIV-positive patients, finding that Medicaid coverage had a beneficial effect on enrollees’ health status, lowering the probability of six-month mortality by more than 70 percent.
Numerous studies also irrefutably show that any form of insurance coverage improves health outcomes—and that lack of health coverage is detrimental to the health and well-being of children and adults. For instance, those without insurance are more likely to have poor health and higher mortality rates than the insured. The uninsured are also significantly less likely to receive important preventive care and are more likely to delay or forego doctors’ visits, medicines, and other services that can lead to declines in health. The consequences of being uninsured also include receiving poorer care when hospitalized, even in trauma cases such as car accidents.
In addition to the health benefits of Medicaid coverage, researchers concluded that those covered by Oregon’s Medicaid program were 60 percent less likely to borrow money or skip payments because of medical bills. Oregon’s Medicaid program worked exactly as insurance is intended to by protecting beneficiaries from catastrophic medical costs and providing important financial security to enrollees.
Medicaid’s effect on access to care
Medicaid provides important health coverage to enrollees, but we must do more to ensure that all enrollees can access care when they need it.
Ryan’s version: Ryan states that Medicaid eligibility expansions alone may be inadequate to ensure access to health care.
Ryan’s mistakes: Ryan misses the point that investments in Medicaid are key to ensuring access for enrollees.
Ryan mentions studies showing that one-third of primary care doctors do not accept new Medicaid patients, and studies highlighting problems with low Medicaid reimbursement levels.However, Ryan’s report fails to mention that a study he relies on prominently recommends that raising reimbursement rates for these physicians—as the Affordable Care Act, or ACA, does—could increase the number of primary care physicians who see Medicaid patients.
Additionally, while Medicaid patients unfortunately can face difficulties in accessing care, access problems are far worse for the uninsured. Ryan also omits mention of research showing that some patients insured by private plans or Medicare may also have difficulty seeing particular providers. He does not mention that many studies used to illustrate poor access rates for Medicaid patients often compare the Medicaid network of providers to the most expensive—and expansive—private plans in the area. This comparison implies that Medicaid enrollees could afford more expensive private insurance coverage, when their insurance choices are often between Medicaid and nothing. Ryan’s misuse of this research unfairly stacks the cards against Medicaid and incorrectly suggests that all patients with private insurance can access any physician at any time.
The truth: The ACA, which Ryan has vocally opposed, raises Medicaid reimbursement rates for primary care to 100 percent of Medicare reimbursement rates in 2013 and 2014. As a result, doctors will receive an average Medicaid pay increase of 30 percent nationally, with physicians in certain states with particularly low Medicaid rates—such as 41 percent of Medicare rates in New Jersey—receiving more dramatic increases.States will also receive a total of $11 billion in new funds to expand community health centers that provide key services to Medicaid patients.
Instead of responding to access issues by proposing increased Medicaid reimbursement rates to ensure that more providers accept Medicaid patients, Ryan’s budgets and previous proposals would gut the Medicaid program, exacerbating these problems.
Medicaid’s effect on work incentives and welfare participation
Medicaid provides important health insurance to people of all income levels. While Medicaid may affect work incentives, this effect is minimal, and the Medicaid expansion and exchange tax credits provided by the Affordable Care Act may actually reduce any tax on work by smoothing eligibility cliffs.
Ryan’s version: Ryan claims that the availability of Medicaid discourages people—especially women—from working and that it results in increased participation in welfare programs.
Ryan’s mistakes: Not only does Ryan rely on methodologically questionable studies, but he also omits other important estimates by the nonpartisan Congressional Budget Office, or CBO, and ignores how the ACA accounted for work incentive issues. Ryan’s assertions are inaccurate and misleading for several reasons.
- He ignores how the ACA smooths eligibility cliffs. Ryan assumes people would not work in order to remain income eligible for the Medicaid program. A recent CBO analysis of Medicaid expansion projected that increased eligibility could modestly reduce work incentives. CBO estimated that while some people will work fewer hours or voluntarily leave the labor force to become or remain eligible for Medicaid, other people will have increased work incentives.In particular, prior to 2013, working parents with very low incomes could only qualify for Medicaid benefits with incomes below approximately 64 percent of the federal poverty level, or FPL, creating an eligibility cliff that would make a parent ineligible for Medicaid coverage if they earned a dollar more than the eligibility limit. Because the ACA allows states that choose to expand Medicaid eligibility to 138 percent of the FPL, people who may have otherwise limited their work hours to remain benefit eligible can now increase their work hours—and their incomes—while still remaining Medicaid eligible. Additionally, because individuals earning between 100 percent and 400 percent of the FPL in all states will be eligible for subsidies to help offset the costs of insurance through a marketplace plan, there is no longer an eligibility cliff at 138 percent of the FPL.States that opt out of Medicaid expansion and choose to retain eligibility levels below 138 percent of the FPL miss out on the full smoothing effects of the ACA. People in these states with incomes between the state’s current eligibility level—i.e., 64 percent of the FPL—and the ACA subsidies available at 100 percent of the FPL will face an eligibility hole, leaving no option of subsidized public coverage or subsidized purchase of private coverage for these low-income individuals. Thus, Medicaid expansion is crucial to smoothing steep eligibility cliffs that could create adverse work incentives. The ACA offers a graduated support system to help people of all income levels access health insurance.
- The problems with cited studies. Ryan distorts language from a 1991 study by Anne Winkler, claiming the study found that “Medicaid has a significant negative impact” on the probability that a female head of household would work. However, the study actually states that Medicaid would have a “generally [statistically] significant but small negative impact on an average female head’s probability of being employed.”
More problematically, however, to assess the effect of Medicaid availability on women’s decision to work, the author increases the income of women in the study by “cashing out” the estimated average market value of Medicaid benefits. Conducting the study using the assumption that Medicaid coverage directly translates to increased income is unrealistic as Medicaid coverage allows enrollees to access and pay for health care—and only health care. Although Medicaid coverage provides important value to beneficiaries, it is not cash. Because the value of Medicaid coverage varies by an individual’s health needs, service use, and personal preferences—and previously by benefits covered by each state’s Medicaid program—it is inaccurate to apply these findings to Medicaid expansion under the ACA.
Ryan also uses a 1990 paper by Robert Moffitt and Barbara Wolfe to claim that increased Medicaid benefits would increase the likelihood of persons receiving welfare. When contacted about Ryan’s use of her study, Wolfe stated that he misused her research, saying that her study was limited to a small percentage of Medicaid recipients and that her findings were limited to years before the welfare reform bill passed in 1996, which began requiring welfare recipients to work in exchange for assistance.It is also worth noting that Ryan does not mention that the first paper he relies on by Winkler found that Medicaid would have no effect on welfare participation, consistent with other previous research.
Research has also found that Medicaid coverage in Oregon did not discourage employment. The study found no difference in employment rates and earnings between the randomized group of individuals selected to enroll in Medicaid and the group waitlisted for Medicaid coverage.
The truth: Expanding Medicaid ensures that low-income working individuals and families have access to health insurance and does not create incentives for people to stop working. Where the ACA smooths eligibility cliffs, converting Medicaid to a block grant program, as Ryan has proposed, would significantly cut Medicaid funding and lower eligibility levels—exacerbating the very work incentive problem Ryan suggests is a significant problem.
Medicaid’s effect on private insurance enrollment
Medicaid “crowd out” refers to the possibility of people leaving private insurance plans for public Medicaid coverage. Research in this area is complex, and no previous crowd-out research can be used to accurately predict crowd out under the ACA. Regardless of crowd-out rates, Medicaid expansion offers important health coverage to millions of low-income Americans.
Ryan’s version: Ryan oversimplifies how Medicaid coverage expansions lead to crowd out, rather than covering the previously uninsured.
Ryan’s mistakes: Although the research on expansions of public insurance programs, such as Medicaid, leading to crowd out of private insurance is incredibly complicated, Ryan reduces the issue to a single study. While the complexity of the issue does not allow us to fully explore it here, Ryan misses several important points about Medicaid crowd-out research.
- Previous crowd-out estimates are inapplicable to potential crowd out under the ACA. Earlier projections of high crowd-out rates, such as in one study cited by Ryan, modeled crowd-out estimates on eligibility expansions for low-income children under CHIP. The authors of the study projected a high crowd-out rate when many states increased CHIP eligibility to up to 200 percent of the federal poverty level. This particularly affected coverage for children between 100 percent and 200 percent of the FPL.The population eligible for Medicaid expansion under the ACA differs from the CHIP populations used to model previous estimates. Adults who are newly Medicaid eligible have approximately half the rate of enrollment in private coverage compared to children between 100 percent and 200 percent of the FPL. Thus, although there may be a marginal crowd-out effect due to the ACA’s Medicaid expansion, there is a significantly smaller population from which potential crowd out can occur.Additionally, the ACA’s Medicaid expansion is unlike other previous expansions studied because of the sizable low-income uninsured population in the United States, as well as the law’s individual mandate. Without a mandate, a Medicaid coverage expansion might not attract the same number of uninsured. But with a mandate, millions more uninsured people will gain health coverage both through public and private insurance programs, meaning that the only thing getting crowded out, as one scholar writes, is uninsurance.
- Existing crowd-out research is complex and highly variable. A more thorough review of existing studies indicates that estimates of crowd-out rates range from no significant crowd out to rates as high as 60 percent, such as in the study Ryan uses. For example, one analysis finds that in two states—Arizona and New York—that expanded Medicaid coverage to low-income adults prior to the ACA, declines in private insurance coverage have occurred at approximately the same rate as in states that did not expand Medicaid eligibility. In the states with expanded eligibility, rates of Medicaid coverage increased and the number of uninsured increased less than other non-expansion states. These results suggest that Medicaid eligibility expansions did not lead to decreases in private insurance enrollment but did lower the number of uninsured. Another study found that crowd-out rates differed for children and pregnant women by poverty level, estimating an overall crowd-out rate of 14 percent.The variability of the findings of crowd-out research illustrates the complexity of the issue. Review of these studies indicate that crowd-out rates are influenced by several study design factors: the dataset used to examine or model estimates; researchers’ assumptions about Medicaid eligibility levels or differences in the eligibility expansion programs studied; the current coverage landscape—particularly the number and demographics of the uninsured; and researchers’ inclusion of indirect effects of eligibility expansions, such as how family members of newly eligible persons made coverage decisions.
- Focusing on crowd out minimizes the important coverage gains made possible through Medicaid expansion. Although some shift in coverage from private to public insurance is probable, the real purpose of Medicaid expansion is to ensure that all people have access to affordable, comprehensive health care coverage.
The truth: Determining the effect of Medicaid eligibility expansion on private insurance enrollment levels is complicated, with varying results and estimates—none of which can be easily applied to the Medicaid expansion available to states under the ACA. Ryan’s focus on Medicaid crowd out is oversimplified and misleading and detracts from the key issue: how Medicaid expansion results in important coverage gains for millions of Americans.
Contrary to the claims in Ryan’s report, Medicaid coverage provides affordable, comprehensive health insurance to children and certain low-income and disabled adults. This coverage can lead to improved health and greater financial security for enrollees. To take advantage of these benefits, states that have not yet expanded Medicaid eligibility to 138 percent of the federal poverty line should do so immediately. By expanding Medicaid, states can cover millions of additional individuals, reduce current spending on uncompensated care, and access billions of federal dollars to support newly eligible populations. To address many of the issues Ryan raises, states must expand their Medicaid programs—not strip them of needed resources, as Ryan has repeatedly proposed.
Emily Oshima Lee is a Policy Analyst with the Health Policy team at the Center for American Progress Action Fund.
To speak with our experts on this topic, please contact:
Print: Allison Preiss (economy, education, poverty)
202.478.6331 or email@example.com
Print: Tom Caiazza (foreign policy, health care, energy and environment, LGBT issues, gun-violence prevention)
202.481.7141 or firstname.lastname@example.org
Print: Elise Shulman (Oceans)
202.796.9705 or email@example.com
Print: Chelsea Kiene (women's issues, Center for American Progress' Legal Progress, Half in Ten)
202.478.5328 or firstname.lastname@example.org
Print: Tanya Arditi (Immigration, Progress 2050, race issues, demographics)
202.741.6258 or email@example.com
Spanish-language and ethnic media: Jennifer Molina
202.796.9706 or firstname.lastname@example.org
TV: Rachel Rosen
202.483.2675 or email@example.com
Radio: Chelsea Kiene
202.478.5328 or firstname.lastname@example.org