Many analysts have shown that children with Medicaid coverage tend to have higher rates of utilization of care than uninsured children (c.f. Weisman and Epstein, 1990), though it is difficult to determine whether this relationship is causal. Currie and Thomas (1995) use data from the National Longitudinal Survey of Youth (NLSY), which has followed the children of the initial female respondents since 1986. The longitudinal nature of the data allows them to include a fixed effect for each child, which controls for any unobserved, constant, characteristics of the home environment and of the child that might be correlated with Medicaid status. Their estimates indicate that both private insurance coverage and Medicaid coverage are associated with a higher number of visits for illness among white children, while for African-American children, insurance coverage has no significant effect on the number of sick child visits. Both white and African-American children receive more preventive care on Medicaid than with private health insurance.
However, there is no guarantee that increases in eligibility for health insurance will be translated into increases in coverage. Studies of the first years of these expansions of the income cutoffs show that despite the high fraction of births that are being paid for by the Medicaid program, many newly eligible, uninsured, pregnant women did not take up coverage in time to benefit from improved prenatal care. For example, Currie and Gruber (1996a) suggest that as many as half of newly eligible women did not take up coverage in time. These rates of non-participation are higher than those that have been estimated for AFDC and Food Stamps (Blank and Ruggles, 1996), or unemployment insurance (Blank and Card, 1991). Moreover, non-participation was concentrated among women who were not income-eligible for AFDC, suggesting that simply increasing the income eligibility cutoff did not break the link between receipt of cash welfare and Medicaid coverage. However, Currie and Gruber (2000) find that increases in Medicaid eligibility were associated with increases in the utilization of obstetric procedures, which is consistent with the view that when women did finally become covered (sometimes when they arrived at the hospital to deliver) they received more services.
Similarly, Currie and Gruber (1996b) analyze data from the Current Population Surveys and National Health Interview Surveys and find that about half of newly eligible children took up their Medicaid benefits, and that increases in eligibility were associated with increases in the utilization of care. For example, the probability that a child did not receive a doctor's visit in the past year fell 10 percentage points from a baseline level of 19 percent. That is, becoming eligible for Medicaid was estimated to reduce the probability of going without a doctor's visit by more than half.
Currie (1999) reports evidence that for immigrant children, eligibility may be associated with increased utilization of care even when coverage does not rise. This result seems to be attributable to the fact that in some cases providers can obtain reimbursement ex poste for treating Medicaid eligibles who are not covered at the time of service. Also, parents of eligible children may minimize the transactions costs associated with becoming covered by obtaining coverage only when they need services. Currie (1999) reports that Medicaid enrollments follow a seasonal pattern, rising in the summer (when children are presumably preparing for school) and falling to their lowest level in the winter when parents would be required to recertify them.
Most states have tried to encourage takeup by adopting administrative measures designed to simplify the Medicaid application process, especially for pregnant women. Common reforms include: presuming that pregnant women are eligible for Medicaid while their applications are being processed and/or expediting the processing of applications for pregnant women; dramatically shortening and simplifying application forms; and eliminating the requirement for face-to-face interviews by allowing mail-in applications from pregnant women. At the same time, recent declines in welfare caseloads may have caused many pregnant women and children to lose their Medicaid coverage. Many poor women obtained Medicaid coverage "automatically" when they enrolled in the Aid to Families with Dependent Children (AFDC) program. Thus, the loss of AFDC (now the Temporary Assistance for Needy Families or TANF program) effectively raised the administrative bar for women seeking Medicaid coverage, by requiring them to go through a separate and unfamiliar application process (Ellwood and Kenney, 1995)(3).
Currie and Grogger (2000) conduct a comparative evaluation of the effects of three types of policies (changes in income eligibility, administrative reforms, and changes in welfare caseloads) on the use of prenatal care and infant health using data from birth certificates covering all U.S. births between 1990 and 1996. They find that increases in income cutoffs were associated with increased use of prenatal care, while decreases in welfare caseloads were associated with reduced use of prenatal care, especially among African-Americans. The administrative reforms they considered had little effect. The fact that welfare caseloads continue to be linked to Medicaid takeup suggests that transactions costs or informational problems remain an important barrier to Medicaid coverage, despite the administrative reforms that have been undertaken. Low takeup of the Medicaid expansions have inspired other types of state efforts. For example, many states have conducted outreach programs designed to get women into prenatal care (Utah ran one of the earliest campaigns, called "Baby Your Baby"), and the new SCHIP program requires states to submit an outreach plan to the Health Care Financing Administration in order to receive federal matching funds. However, there has been little systematic attempt to evaluate these outreach efforts, or even to determine whether lack of information is the main reason that Medicaid eligibles remain unenrolled. In one of the few studies to examine characteristics of state outreach programs, Aizer (2001) reports that states that contracted out outreach efforts had higher enrollments in public health insurance than states that did not.
Even those who are covered by Medicaid may have difficulty obtaining preventive care, since Medicaid typically pays about half of what private health insurance would pay. One study of new mothers who had arrived in emergency rooms to deliver with "no physician of record" found that 64 percent of the women cited their inability to find a doctor willing to accept them as the largest barrier to obtaining prenatal care (Aved et al., 1993). Baker and Royalty (1996) use data from a longitudinal survey of California physicians observed in 1987 and 1991 and find that expansions of Medicaid eligibility to previously uninsured women and children increased the utilization of care provided by public clinics and hospitals but had little effect on visits to office based physicians. This is consistent with much previous evidence that many providers either do not accept Medicaid payments, or limit the amount of time that they spend with Medicaid patients (Sloan, Mitchell, and Cromwell, 1978; Decker, 1992). These problems may be even more severe for minority mothers. American cities are highly segregated by race and income (Massey and Denton, 1993). Urban minorities often live in parts of the city that are shunned by physicians in private practice and hence are more likely to be served by large urban teaching hospitals (Fossett et al., 1992).
It is worth considering the one instance of great success in terms of takeup: the fact that most eligible pregnant women now have their deliveries paid for by Medicaid. While it may be quite difficult for individuals to overcome barriers to obtaining coverage, hospitals have both the incentive and the means to help women gain coverage since they must provide care to women in labor (Saywell, 1989). Many hospitals have established Medicaid enrollment offices on site. These offices assist people in completing applications and tell them how to obtain necessary documentation (GAO, 1994). Hospitals in at least 32 states and the District of Columbia also employ private firms to help them enroll eligible patients in the Medicaid program. Thus, it is not surprising that many births are covered by Medicaid even when prenatal care is not. This example indicates that takeup is likely to be higher when: a) the service is one that everyone wants; b) providers have incentives to facilitate takeup; and c) individual transactions costs are minimized.