Responsible researcher: Bruno Benevit
Authors: Janet Currie and Jonathan Gruber
Intervention Location: United States
Sample Size: 526,830 individuals
Sector: Healthcare
Variable of Main Interest: Infant Mortality, Underweight Newborns
Type of Intervention: Eligibility for health insurance coverage
Methodology: OLS and IV
Summary
The expansion of health services is a topic for several countries, prompting analysis of the priority demands of different population groups. A critical aspect for the healthcare system concerns the expansion of public health insurance eligibility in the United States. This article verifies whether the changes in the eligibility criteria for the Medicaid that occurred between 1979 and 1992 in the United States resulted in improvements in neonatal and prenatal health markers. Using OLS and IV methodologies, the results demonstrated that changes to Medicaid dramatically increased the eligibility of pregnant women for the program, that there was a decrease in rates of underweight newborns and infant mortality, and that the groups affected by the targeted changes in program were more positively impacted. The study also estimates that the cost per life saved was between US$840,000 and US$4.2 million.
The infant mortality rate and the incidence of underweight newborns are two of the main indicators of child health (Currie and Gruber, 1996). The United States had one of the highest infant mortality rates in the industrialized world in the 1990s, with 9 infant deaths per 1,000 births ( US House of Representatives 1992 , pp. 1116-17). This rate may reflect a large number of unhealthy newborns.Such a rate may reflect a large number of unhealthy newborns. Therefore, it is essential to evaluate the determinants related to this situation, observing how policies to expand access to health service coverage affect newborns, what are the costs related to such policies and how priority groups adopt such policies.
Between the 1980s and 1990s there was a rapid expansion in the eligibility of pregnant women for Medicaid with the aim of increasing the use of prenatal care. Until the early 1980s, eligibility for Medicaid was tied to receipt of social security payments under the Aid to Families with Dependent Children (AFDC) program. This linkage had the effect of limiting eligibility to very low-income women in single-parent families, making access to the program very restrictive. Recent eligibility extensions to other groups provide a case study in whether changes to health insurance eligibility can actually improve children's health.
Medicaid is a federal program in partnership with states that provides health insurance for low-income populations . Eligibility for health coverage through this program for women and children has historically been associated with AFDC participation. This linkage caused restrictions on access to Medicaid and restricted access to the program. This was due to several aspects of the design of the AFDC program, whether due to restrictions on social benefits for only families headed by women, such as the AFDC – Unemployed Parents , or due to the very low and heterogeneous income range parameters between the states.
Because of these restrictions, states had the option of extending Medicaid to some groups of pregnant women who were not in AFDC throughout the period from 1979 to 1992. Such eligibility changes during this era can be divided into two types. The first type was categorized as “targeted eligibility” changes, and considers the expansion of coverage to groups covered by AFDC in a less restrictive way with respect to the family structure of beneficiaries and to individuals who had large medical expenses. The second type was categorized as “broad eligibility” changes, and concerns changes that allowed states to expand access to all women above the previous minimum income threshold. The target audiences for both changes clearly differed in their socioeconomic contexts.
The article contains several analyzes regarding the expansion of eligibility for the Medicaid program. The database used for the main analysis consisted of aggregate newborn data from all US states available in Vital Statistics Medicaid eligibility index as an intervention variable, representing the fraction of women aged 15 to 44 in each state and the year in which they would be eligible for Medicaid coverage in the event of pregnancy.
To control the possible bias arising from omitted characteristics specific to each state or year, an instrumental variable was adopted for the real eligible fraction. This instrument was created from a sample of 3,000 women from the Current Population Survey (CPS) for each year by simulating the eligibility of each woman in each state to control for differences in each state's legislative environments. In addition to the instrumental variable, the study also employs the ordinary least squares (OLS) method.
Analyzes were disaggregated by type of eligibility change. The first type consists of targeted changes, covering specific groups that gained access to eligibility through more flexible access to Medicaid linked to other social programs. The second type consists of broad changes, which expanded access to Medicaid for all women with income equivalent to up to 185% of the federal poverty level.
In the first analysis, the outcome variables of incidence of births underweight (less than 2,500 grams) and the infant mortality rate in each state and year were considered. The estimates considered general eligibility, targeted eligibility changes, and broad eligibility changes. Additionally, these results were re-estimated using robust regression techniques to control the effects of outlier through the adoption of weights and other specifications considering additional variables.
Medicaid health coverage using linear probability models. The data set consists of 526,830 observations for the period 1979 to 1992. Observations for targeted and broad changes cover the period 1987 to 1992. All regressions include a full set of state and year dummies.
The third analysis observes the efficiency relationship of the policy's allocated resources. For this purpose, we used data on total expenditure on doctors, hospital inpatient departments and hospital outpatient departments and other clinics, for all non-disabled children and non-disabled/non-elderly adults. We normalize spending using the state's female population aged 15 to 44. Spending was normalized and deflated to thousands of 1986 dollars.
Finally, the fourth analysis checks whether eligibility had an effect on the demand for prenatal care before the third month of pregnancy due to the targeted changes. This analysis used data from 1979 to 1990 from the National Longitudinal Survey of Youth (NLSY) on the month in which prenatal care began.
The results of the analysis regarding newborn indicators indicate that the increase in Medicaid resulted in a 1.9% and 8.5% decrease in the incidence of underweight births and the infant mortality rate, respectively. Evaluating the effect of eligibility changes disaggregated, the results indicate that targeted eligibility had larger and more significant effects compared to broad eligibility. Robustness analyzes showed small variations in the magnitude of the effects, but did not change the inference regarding the impact on the infant mortality rate.
The analysis estimates regarding the predictive characteristics of Medicaid demonstrate that general eligibility had smaller effects than those identified in other social programs. The effects of targeted eligibility were again higher, while no effects of broad eligibility were identified. In general, the people most likely to be covered by the program are non-white, single, low-income women with children.
The main result regarding the effect of expanding Medicaid on health spending was the existence of a significant positive effect on spending considering broad eligibility. Individuals with broad eligibility had larger effects compared to the effects for targeted eligibility, an unexpected result for the authors.
Cost-benefit analyzes indicate that the cost of saving a life through targeted eligibility changes was $840,000, while the cost of saving a life through broad eligibility changes was $4.2 million. In general, the authors identify that the costs in targeted eligibility were efficient compared to the estimates of the value per life saved presented in the literature.
The results of the analysis on the probability of delay in prenatal care indicate that targeted Medicaid decreases the probability of delay by almost half when considering the instrumental outcome variable.
This article evaluated how the expansion of eligibility for the Medicaid that occurred between 1979 and 1992 in the USA impacted prenatal care and the health indices of newborns.
Through OLS methods and instrumental variables, three main conclusions were identified. First, the changes dramatically increased Medicaid eligibility for pregnant women, but they occurred heterogeneously across states. Second, the changes reduced the incidence of infant mortality and underweight children. Medicaid eligibility have had much larger effects on birth outcomes than broader eligibility expansions, a phenomenon explained by the income gap between the groups affected by such changes. Ultimately, the targeted changes were estimated to cost the Medicaid $840,000 per life saved. Various methods for measuring this value present in the literature allow different interpretations regarding the efficiency of the policy.
References
Currie, J. and Gruber, J. (1996), “Saving Babies: The Efficacy and Cost of Recent Changes in the Medicaid Eligibility of Pregnant Women”, Journal of Political Economy , Vol. 104 No. 6, pp. 1263–1296.
Check the Institution's registration in the e-MEC System here
© 2025 All Rights Reserved
