|
Assessing Access To Care Under Medicaid: Evidence For The Nation And Thirteen States
States have broad latitude in designing their Medicaid programs; this has important implications for access to care. To understand the consequences of state variation, we evaluate, for the nation and for thirteen study states, how well the program is providing access for beneficiaries, using the level of access available to low-income privately insured people in the local health care market as our benchmark. Overall, we find that Medicaid beneficiaries’ access matches that of the low-income privately insured for most of the ambulatory outcomes examined but is worse for dental services and prescription drugs. State-level analyses revealed some variation in the access gap.
Medicaid is a federal-state partnership that provides health care coverage to more than fifty-one million low-income Americans.1 The program covers a large share of the neediest, most vulnerable Americans, including low-income pregnant women, parents and children, elderly people, and people with disabilities. States, within broad federal guidelines, have controlled critical aspects of program design, including eligibility, service coverage, and provider payment, since Medicaid was established. Such flexibility has important consequences. One is that states vary in the share of their poor population they cover and how much they spend on them, which has created substantial inequity across states: Equally poor people living in different states face widely disparate eligibility criteria. Another important effect of the flexibility is that once enrolled in the program, people might have very different access to health care because of state variations in what is covered under Medicaid or in the level of provider reimbursement, among other things. Only limited information exists on how well the program provides access to care for its beneficiaries in individual states. Moreover, we have little insight into how well the program provides access relative to the access levels generally available within local health care markets. Understanding how beneficiaries’ access compares with that of insured populations in the same local market, and whether those differences vary across states, provides an important assessment of Medicaid’s success as a health insurance program. In this paper we compare Medicaid beneficiaries’ access with that of the privately insured population, particularly the low-income segment. We selected this as the comparison group because its members are likely to face a local health care market similar to that faced by Medicaid beneficiaries. We examine two questions: (1) Overall, does Medicaid provide access to care that is comparable with that of the low-income privately insured? (2) For selected states, how does beneficiaries’ access compare with that of the low-income privately insured in the state?
Data. To conduct the analysis, we used data from the 1999 and 2002 National Survey of America’s Families (NSAF), which collected data on the economic, health, and social characteristics of more than 100,000 children and nonelderly adults each year.2 Two important features of NSAF, a nationally representative survey, are that it oversamples the low-income population and that it has substantial, representative samples in thirteen states (Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, Washington, and Wisconsin). As such, the survey provides a large sample of Medicaid beneficiaries and the overall low-income population for the nation and for the thirteen study states. Although not a random sample, these states accounted for about half of the country’s population and 55 percent of the total Medicaid population in 2002. The states also represent a varied group of state Medicaid programs, in terms of eligibility standards, benefit packages, and program spending. We limited our sample to adults ages 19–64. We further restricted the sample to people who reported having the same insurance coverage (Medicaid or private) for the full year preceding the survey, to ensure that we focused on health care experiences or services received while in that insurance state. For the Medicaid analysis sample, this yielded a national sample of 7,493 Medicaid beneficiaries; state samples ranged from 236 in Texas to 795 in Massachusetts. For the low-income privately insured analysis sample, we also limited the sample to people with household incomes below 200 percent of the federal poverty level. For this group, our national sample is 12,064 people, with state samples ranging from 589 in New Jersey to 1,306 in Wisconsin. Beyond NSAF, we used several other data sources, including the Area Resource File of the Bureau of Health Professions, Health Resources and Services Administration; and data from Inter-Study. Methods. In this analysis we modeled health care access and service use as a function of a person’s predisposition to use health care services, factors that enable or impede use, and the need for health care.3 Predisposing factors included demo-graphic and social characteristics (for example, age, race/ethnicity, sex, education, and marital status). Enabling and impeding characteristics included individual and family resources (income and employment) and community health care resources (supply of providers in the individual’s county of residence) and economic circumstances (county unemployment level). A person’s need for services is measured by his or her health status.4 Finally, the model included a dummy variable for the 1999 survey year to account for any national changes that might have occurred during the 1999–2002 study period. The model also included a dummy variable for insurance status (Medicaid versus low-income privately insured), a dummy variable for the study state, and the interaction between these two dummy variables to capture differences in access to care and service use between Medicaid beneficiaries and low-income privately insured people within each state relative to the remaining forty-nine states and the District of Columbia (the "remaining states"). This framework provided an estimate of the difference in the value of the outcome measures for adults in each study state relative to the value for adults in the remaining states.5 We looked at six measures—three realized and three potential—that assess access to ambulatory care. All of the measures describe the person’s health care experiences in the past year. The measures are whether the respondent reported having at least one doctor visit; having at least one dental care visit; having a clinical breast exam and a Pap smear (women only); unmet need for medical care or surgery; unmet need for prescription drugs; and unmet need for dental care. We used the coefficient estimates obtained from our multivariate model to predict the values of each outcome measure by first assuming that everyone in the sample is a Medicaid enrollee and then assuming that everyone has private insurance. The difference between those two predictions is our measure of the access gap between Medicaid and the low-income privately insured in the state (referred to as the "access gap"). To facilitate the comparisons across outcomes, populations, and states, we estimated linear probability models. Study shortcomings. Similar to all survey-based research, our findings are based on self-reported data and thus rely on respondents’ perceptions of their health status, access to care, and the like. Another shortcoming is the limited sample sizes for some of the study states, especially Colorado and Texas. Because of this, we may have erroneously concluded that there is no difference between Medicaid beneficiaries and the low-income privately insured when in fact there is a difference. We also know very little about the type of health insurance coverage the low-income privately insured have. Relative to Medicaid, which has little if any cost sharing, private health insurance policies will likely impose premiums, deductibles, and copayments. Further, some policies may have maximum benefit levels. We expect the range of policies held by the low-income privately insured population to be quite diverse, with past studies reporting that low-wage workers tend to have less generous health benefits than their higher-income counterparts.6 In our study we took insurance status as given. Although we included a range of individual and area characteristics in our multivariate analysis, to the extent that we did not adequately control for systematic differences between people who enroll in Medicaid and in private insurance that affect their access to care, our results will be biased. Nonetheless, in recent work examining the affects of Medicaid on access to care where selection into insurance status was controlled for, limited evidence of bias in the estimates of the effects of Medicaid versus private coverage on access and use was found.7 Selection bias was a much more important issue in assessing the effects of having coverage versus being uninsured than in evaluating the effects of different types of insurance coverage. Finally, because of data constraints, we did not control for whether a person has a chronic illness, which can greatly affect access to care.
Characteristics of two study populations. Exhibit 1  lists selected personal characteristics of the Medicaid and low-income privately insured national-level study samples. As expected, the latter have higher incomes and are more likely to be in better health and to be employed than the former.
Access differences. We found that simple differences in access to care between Medicaid and the low-income privately insured to be significant across all six measures examined (data not shown). Once we controlled for individual and area characteristics, however, the access differences narrowed considerably. Exhibit 2 presents adjusted access differences between the two populations for the nation and for each of the thirteen states; Exhibit 3 presents national results in geographical form.
National access gap. As shown in Exhibit 3  (or by comparing the upper and lower panels of Exhibit 2), after accounting for individual and area differences, we found few access disparities between Medicaid beneficiaries and the low-income privately insured for the country as a whole. The two populations reported comparable levels of access to physician and preventive care (for women) and unmet need for medical care and dental care. However, we did find access differences for two measures: Medicaid beneficiaries were significantly less likely to have a dental visit and more likely to report unmet need for drugs, compared with the low-income privately insured. This access gap may reflect in part the benefit flexibility afforded under Medicaid. Dental care, for example, is an optional Medicaid service for adults, and in 2000, eighteen states (including Alabama, Mississippi, and Texas in our sample) provided no dental coverage or limited coverage to emergency treatment only. Moreover, even for states that do provide dental benefits, the scope of coverage varies widely, with many imposing substantial restrictions.8 In addition, Medicaid reimbursement for dental care has historically been very low, as has dentists’ participation in the program.9 Collectively, these factors could create substantial dental access problems for beneficiaries, even for those who live in states where dental care is a covered Medicaid benefit. The low-income privately insured also face access challenges in dental care. Recent research has shown that only about two-thirds of low-wage workers with private insurance have dental coverage.10 However, the type of dental coverage provided to this population could be more comprehensive or geared more to preventive care than what is provided under Medicaid. This could partly explain why we found that the low-income privately insured have better access to dental care than Medicaid beneficiaries have. Another possible reason for the disparity is that despite being poor, the low-income privately insured are still financially better off than Medicaid beneficiaries and thus are in better position to pay out of pocket for a dental visit, routine or otherwise. The other area where we found that Medicaid falls short is access to prescription drugs, also an optional benefit. Unlike dental care, however, all states cover drugs. Drug coverage also appears to be the norm among the low-income privately insured, with one study reporting that nearly 90 percent of low-wage workers with private insurance had such coverage.11 However, states can place a number of limitations on their Medicaid drug coverage, such as imposing copayments, restricting the number of prescriptions, developing preferred drug lists (PDLs), mandating that beneficiaries use generic rather than brand-name drugs, and requiring prior authorization for some drugs. Indeed, most states have implemented a host of strategies to control Medicaid drug spending, which have been shown to be associated with poorer access to drugs.12 In addition, even though all states provide drug coverage under Medicaid, cost containment policies may particularly affect perceptions of "unmet" drug needs: Although beneficiaries still might get their medications, they might not get the ones they think they need because, for example, the state has a PDL or mandatory generic substitution.
State access gaps.
Mirroring national results, we found, on balance, few access differences between Medicaid and the low-income privately insured at the state level (Exhibit 2 We observed more access gaps between Medicaid and the low-income privately insured in a couple states. In Michigan we found that Medicaid beneficiaries were less likely to report having a dental visit and more likely to report unmet need for dental care. In Washington we found that Medicaid beneficiaries fared worse on all three of the unmet-need access measures examined. Given that the gaps in Washington are confined to the potential access measures (rather than the realized measures), beneficiaries might be able to obtain basic access to care but appear to experience difficulties in securing all of the care they feel they need. Individual access measures. Doctor visits. We found that access to physician care under Medicaid is at least as good as access with private insurance in twelve of the thirteen states, with Medicaid adults in New York doing significantly better than low-income privately insured New Yorkers. Only in California was Medicaid beneficiaries’ access to physicians significantly lower than that of the low-income privately insured: There, 78 percent of Medicaid adults reported a physician visit, compared with 84 percent of low-income privately insured adults. One possible explanation for this is the types of delivery systems that these states use. Both states depend heavily on managed care for their Medicaid populations, but New York’s managed care program relies more on plans that serve primarily or exclusively Medicaid beneficiaries than California’s managed care program does. Such plans may be better able than commercial plans to deal with the complex needs and cultural diversity of Medicaid beneficiaries.14 Dental visits. Unlike the probability of a doctor visit, the gap in the probability of a dental visit was relatively large and statistically significant in nine of the thirteen states. Only in California, Massachusetts, Minnesota, and Washington did we observe comparable levels of dental visits between Medicaid beneficiaries and the low-income privately insured. Access to dental care has been a chronic problem in the Medicaid program, as discussed earlier.15 Interestingly, our results indicate that the access gap was large and significant in states with no dental coverage (Alabama, Mississippi, and Texas) as well as in states with comparatively generous dental coverage (New Jersey and Wisconsin). This suggests that issues other than coverage alone are influencing access. For example, although our model controlled for the supply of dentists in the local health care market, we did not account for the share of dentists participating in Medicaid, which, as mentioned earlier, historically has been low.16 Another factor that could be contributing to the access gap is low Medicaid reimbursement for dental services, which is often well below dentists’ normal fees. For example, although New Jersey offered full dental benefits to adult Medicaid beneficiaries in 1999, the state reimbursement level for many dental procedures was 25 percent or less of the average regional dental fees.17 Although dental coverage among the low-income privately insured is not universal, as discussed above, the type of coverage may be more comprehensive or more focused on preventive treatment than Medicaid. This could increase the likelihood that the low-income privately insured would have at least an annual dental visit. Another possible explanation is that the low-income privately insured likely have more disposable income than the Medicaid population and thus can more readily pay directly for dental services that are not covered by their health plan. Comprehensive preventive care for women. We found very low levels of preventive care for low-income women, whether in Medicaid or privately insured. (Nationwide, less than half of either population received these preventive services, which is well below national health promotion goals.)18 Further, there was little variation for this measure between the populations or across the states. In twelve states the difference between the two populations was not statistically significant. The one exception was Massachusetts, where 53 percent of women in Medicaid reported receiving preventive care, compared with 43 percent of low-income privately insured women. Unmet medical care or surgical need. We found few differences across the states in the share of Medicaid adults reporting unmet need for medical care or surgery relative to the low-income privately insured population. Less than 10 percent of adults in both groups in the nation as a whole and in each of the states reported such unmet need. Only in Washington were Medicaid adults significantly more likely than their low-income privately insured peers to report unmet need (12 percent and 6 percent, respectively). Unmet need for dental care. In keeping with the access gap for use of dental care services, we found that Medicaid beneficiaries were more likely to report unmet need in all states but Mississippi and New Jersey. However, the difference in reported unmet need levels was significant in only four states (Florida, Michigan, Minnesota, and Washington). Interestingly, the access gap for Medicaid for the use of dental care services in Minnesota and Washington was not significantly different from that of the low-income privately insured, yet beneficiaries in these states perceived more unmet need for these services. Although the likelihood of any use of dental care services was comparable between the two populations, it may be that Medicaid beneficiaries in Minnesota and Washington have more difficulty obtaining follow-up care than do the low-income privately insured, which leads to higher levels of reported unmet need. Unmet need for prescription drugs. We also observed some between-state variation in reported unmet need for prescription drugs. Particularly noteworthy is the access gap in Colorado, where 22 percent of Medicaid beneficiaries reported unmet need for drugs, compared with only 8 percent of the low-income privately insured. Since access for the latter in Colorado is the same as for the population in the nation as a whole, this difference is driven by much higher unmet need for prescription drugs under Colorado’s Medicaid program. As do many states, Colorado’s Medicaid program requires prior approval of drugs and has a closed drug formulary in some cases. Additional analyses are needed to determine whether these policies are causing the large disparity we found.
In comparing Medicaid beneficiaries with the low-income privately insured, we found that Medicaid beneficiaries are poorer, less healthy, and less likely to be employed than the low-income privately insured. When comparing simple differences in access to care, we found that Medicaid beneficiaries have significantly worse access. Once we controlled for individual and health status characteristics and local market factors, however, access disparities between the two groups all but disappeared. Medicaid is often criticized for being a costly, low-quality health insurance program that pays providers poorly and suffers from low provider participation. It is also widely believed, however, that Medicaid is a "Cadillac" health insurance program that offers a rich benefit package at little to no cost to beneficiaries. Given that we found comparable access for people whether they were covered by Medicaid or private insurance, our results help dispel both of these myths. At the same time, though, our benchmark is the low-income privately insured population, who tend not to have the most comprehensive coverage (indeed, studies report just the opposite).19 Thus, policymakers should exercise caution when considering Medicaid cutbacks—which, owing to the current fiscal situation, are actively being debated by state and federal decisionmakers. A likely outcome of benefit reductions is that beneficiaries’ access will fall below that of the low-income privately insured. Although access was generally equivalent in both groups, Medicaid enrollees fared worse with regard to dental care and prescription drugs, which are both optional services under Medicaid and which many states have opted to provide on a limited basis or to forgo altogether (in the case of dental care). These results highlight the fact that Medicaid benefit changes should be made with caution. This is especially true for prescription drugs: Although they are vital to providing high-quality health care, their rapidly rising costs have made them a conspicuous budget target in several states.20 Also, when we looked at the access gap at the state level, we observed some variation. For the majority of states, however, we either found no access gap or found that beneficiaries’ access was lower on only one measure—primarily, access to dental care. Thus, our results indicate that Medicaid is providing access to most ambulatory care services at a level comparable to what is available to the low-income privately insured in the thirteen study states. Our study focused on access to ambulatory care. To obtain a broader perspective on how well Medicaid is working, future research should look at access to specialized ambulatory care services (for example, mental health and substance abuse), as well as emergency room and inpatient use. Examining measures that are based on health care outcomes also would contribute to a fuller understanding of Medicaid. Finally, future work should examine how access for Medicaid beneficiaries compares with that of other populations, such as low-income uninsured or higher-income privately insured populations. Such information will provide important insights into what works in Medicaid and what needs to be improved.
Teresa Coughlin (tcoughli{at}ui.urban.org) is a principal research associate at the Urban Institute in Washington, D.C. Sharon Long also is a principal research associate there. Yu-Chu Shen is an assistant professor at the Graduate School of Business and Public Policy, Naval Postgraduate School, in Monterey, California. This paper was funded by the Henry J. Kaiser Family Foundation. Opinions expressed are those of the authors and do not necessarily reflect the positions of the Urban Institute, its board, or its sponsors. The authors thank Liliane Ndong and Allison Cook for their diligent research assistance; John Holahan and two anonymous reviewers for their many helpful comments on an earlier draft; and Barbara Lyons, David Rousseau, and Julia Paradise of the Kaiser Commission on Medicaid and the Uninsured for their thoughtful comments and support of this project.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Data Sources And Methods
Study Results
