Abstract
Background:
Medicaid provides insurance for 40% of US births. Patterns of preconception Medicaid enrollment are not well-described.
Methods:
Using Medicaid Analytic Extract files, this retrospective cohort study of individuals with a 2014 Medicaid-insured birth examined months of Medicaid enrollment and changes in enrollment status during 84 months prior to birth. We used linear regression to assess the association between enrollment months and 2014 age, race and ethnicity, rural residence, any Medicaid eligibility due to disability, foster care involvement, or pregnancy, chronic health conditions, and state of residence. We examined variation across states in the relationship between enrollment months and changes in enrollment status.
Results:
We identified 944,068 individuals in 30 states. Individuals had a median of 40.6 (SD 27.3) enrollment months of Medicaid and 3.2 (SD 2.3) changes in enrollment status. Twenty-eight percent were enrolled in Medicaid for ≥ 63 months. In regression analysis, age, race and ethnicity, and chronic health conditions were associated with enrollment months. Each year of age was associated with −0.93 enrollment months (95% CI −1.27, −0.60). For chronic conditions, hypertension was associated with an additional 3.12 enrollment months (95% CI 2.53, 3.71), diabetes with 3.28 (95% CI 2.24, 4.32), and mental health with 6.27 (95% CI 5.55, 6.98). The relationship between enrollment months and changes in enrollment status varied across states.
Conclusions:
Medicaid plays a substantial role in preconception health insurance for individuals with Medicaid-insured births. This is particularly true for younger individuals and those with comorbidities associated with adverse birth outcomes.
Keywords: Medicaid, preconception, birth outcomes, insurance churn
INTRODUCTION
Medicaid provides insurance for 40% of births in the United States and plays a central role in initiatives to improve pregnancy outcomes.1 Health before conception is one predictor of pregnancy outcomes. Preconception health care can improve pre-pregnancy health and pregnancy outcomes by addressing chronic disease management, pregnancy-specific topics such as folate-containing vitamin supplementation or family planning, and early entry into prenatal care.2–6 High-quality preconception care may also improve health activation, trust, and communication with clinical teams, enabling timely diagnosis of problems and disease self-management, including during pregnancies. 7,8 However, receipt of preconception care is low.9,10
One reason for the low receipt of preconception care may be that, in the US, childbearing coincides with periods of high uninsurance during early adulthood.11 Health insurance is a key facilitator of health care access.12–14 However, individuals who rely on Medicaid during pregnancy may experience periods of Medicaid eligibility and ineligibility, and consequently enrollment and disenrollment, throughout their reproductive years. Given that the median age of first birth in the US is 27 years, for some the prime preconception window may overlap with pediatric Medicaid eligibility.15 Changes in insurance enrollment status, termed churn, may interfere with health care continuity, and churn is associated with worse control of chronic diseases and lower self-reported health.16–18 Specific to pregnancy, continuous Medicaid coverage before and after pregnancy is associated with early entry into prenatal care, receipt of postpartum care, and quality of postpartum care.19–22
Little is known about the role of Medicaid as a provider of preconception insurance for the pregnancies it ultimately covers.23–25 Though churn in the prenatal and postpartum periods is well-established, churn in the preconception period has not been described. Therefore, the goal of this study was to define Medicaid enrollment over an extended preconception period for a cohort of Medicaid-insured births and identify factors associated with greater Medicaid enrollment.
METHODS
Data and cohort
This retrospective cohort study used Medicaid Analytic Extract (MAX) data from 2007 – 2014. Thirty-two states reported data to MAX through 2014. We excluded 2 states with 10% or greater missing data on Hispanic ethnicity. The 30 remaining states included Arkansas, Arizona, California, Connecticut, Georgia, Hawaii, Iowa, Idaho, Indiana, Kentucky, Michigan, Minnesota, Missouri, Mississippi, New Jersey, New York, Ohio, Oklahoma, Oregon, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Utah, Virginia, Vermont, Washington, West Virginia, and Wyoming. We analyzed all 2014 births to individuals aged 14 – 45 in these states. If 2 births were observed in 2014 for the same individual, we included only the one that occurred earlier in the year. This dataset pre-dates the transition from MAX to T-MSIS Medicaid data system which occurred from 2014 – 2016, with 11 states on T-MSIS in 2014 and 30 in 2015. The COVID public health emergency from 2020 – 2023 led to temporary changes in Medicaid eligibility that are not consistent with routine Medicaid rules. Therefore, this dataset provides the most recent opportunity for an extended lookback across a majority of states that is not interrupted by either data transition or the COVID public health emergency. This study was reviewed and deemed exempt by the Institutional Review Board at our institution. Analyses were conducted using R and Stata, Version 18 (College Station, TX).26,27
Medicaid enrollment
We used enrollment files from 2007 – 2014 to classify the number of Medicaid enrollment months for each person during the 84 months prior to the 2014 birth. This study period included both enrollment during pregnancy for the 2014 birth and preconception enrollment. We classified enrollment months as both a continuous and a categorical variable. The categorical variable classified enrollment as <42 months (<50% of the 84-month study period), 42 – 62 months (50 – 74%), and ≥63 months of enrollment (≥75%). We counted the number of times Medicaid enrollment status changed, including changes from enrolled to unenrolled or from unenrolled to enrolled. Separately, we count the number of churn cycles, defined as becoming unenrolled and then becoming enrolled again. We used Medicaid enrollment files to create 5 dichotomous variables related to reasons for Medicaid eligibility (not mutually exclusive) during the study period. These variables assessed whether individuals were ever enrolled in Medicaid because of (1) disability status, (2) foster care placement, or (3) pregnancy eligibility.28 All included individuals had Medicaid-insured pregnancies, but some were enrolled under non-pregnancy eligibility codes during all pregnancy months. Individuals with any enrollment under age 19 years were classified as having (4) pediatric eligibility. Finally, we classified codes for adults that were not linked to unemployment, disability, or status as parents as (5) traditional adult eligibility.
Other measures
We examined other measures associated with health care access and health based on prior literature.29 Maternal age was classified in years at the time of 2014 birth. Given that income tends to increase throughout the reproductive years, we hypothesized that younger individuals would have more months of preconception Medicaid enrollment. We created a dichotomous variable indicating any other observed birth during the study period prior to the 2014 birth. Race and ethnicity are linked to health care access and health through mechanisms linked to interpersonal and structural racism.30–32 Medicaid enrollment data record race and ethnicity based on individual’s self-report. Most individuals who were missing race reported Hispanic ethnicity. We therefore aggregated self-reported race and ethnicity into the following 8 categories: non-Hispanic American Indian or Alaska Native, non-Hispanic Asian, non-Hispanic Black, non-Hispanic Pacific Islander, non-Hispanic White, non-Hispanic multiple races, Hispanic, and no race or ethnicity reported. We classified rural residence as a dichotomous variable based on zip codes eligible for rural health funding through HRSA.
Health conditions
We classified individuals with common chronic conditions using claims related to their 2014 pregnancy, including any claims from the start of pregnancy to 3 months after birth. We took this approach to classifying common chronic conditions because health care use and receipt of recommended screenings is more consistent during pregnancy than during other life stages. Thus, classifying conditions based on claims from this period may reduce confounding by indication. Consistent with prior studies, we classified hypertension and diabetes using ICD codes and medications included in treatment guidelines.33–36 To classify hypertension or diabetes, we required only 1 ICD code, but in the absence of relevant ICD codes we required 2 instances of medication. We classified pregnancy-induced hypertension or gestational diabetes if there were no medications prior to 2nd trimester, or if no ICD codes indicated pre-pregnancy or chronic hypertension or diabetes. We created a dichotomous variable indicating common mental health conditions, identified by ICD codes for both depression and anxiety.
Analysis
We examined characteristics of the cohort across categories of enrollment months (<42 months, 43 – 62 months, and ≥63 month), including demographics, characteristics of enrollment, and health conditions. To compared across categories, we used chi-squared tests for categorical variables and oneway ANOVA tests for continuous variables. We also examined how enrollment differed across age groups and between those with and without prior observed pregnancies, again using chi-squared and oneway ANOVA tests.
We examined the association between individual characteristics and enrollment months using multivariable linear regression with enrollment months as the dependent variable. Given the potential role of Medicaid in providing preconception health care access, we were interested in the association between enrollment and individual characteristics known to be associated with adverse birth outcomes, including race and ethnicity, age, and chronic health conditions. We selected non-Hispanic Black as the reference group for race and ethnicity in this regression because this was one of the largest groups and was the group with the highest enrollment months. We were also interested in preconception Medicaid enrollment for populations who may face particular barriers to care such as those in rural areas or with a history of foster care involvement. We separately created a multivariable linear regression model with changes in enrollment status as the dependent variable and the same independent variables.
We examined the association of enrollment months and changes in enrollment status across the 30 included states. We conducted a stratified analysis, with a regression for each state with enrollment months as the dependent variable and changes in enrollment status as the independent variable. We were interested in understanding the range of the regression coefficients and whether they tended to be negative or positive. We did not adjust for other individual level characteristics in these regressions to best understand the unadjusted variation across states.
RESULTS
We identified 944,068 individuals with 2014 births in 30 states (Table 1). Over the 84-month study period, these individuals had a mean of 40.6 (SD 27.3) months of Medicaid enrollment, and a mean of 3.2 (SD 2.3) changes in enrollment status. Forty-six percent of the cohort were enrolled in Medicaid for at least half of the 84-month study period, and 28% for at least three-quarters of the period. Younger individuals and those with Black race had greater enrollment months.
Table 1:
Characteristics of cohort stratified by total enrollment months over 84-month look-back period
| <42 months N=506,654 |
42 – 62 months N=173,600 |
>=63 months N=263,814 |
Total N=944,068 |
||
|---|---|---|---|---|---|
| Demographics | |||||
| Age | 26.9 (5.6) | 25.6 (5.1) | 24.8 (5.8) | 26.1 (5.6) | <0.001 |
| Race and ethnicity | <0.001 | ||||
| Non-Hispanic American Indian / Alaska Native | 4,539 (1%) | 2569 (1%) | 4,012 (2%) | 11,120 (1%) | |
| Non-Hispanic Asian | 16,783 (3%) | 3,035 (2%) | 3,737 (1%) | 23,555 (3%) | |
| Non-Hispanic Black | 81,121 (16%) | 39,793 (23%) | 77,131 (29%) | 198,045 (29%) | |
| Non-Hispanic Pacific Islander | 5,732 (1%) | 998 (1%) | 1,968 (1%) | 8,698 (1%) | |
| Non-Hispanic White | 213,361 (42%) | 73,576 (42%) | 97,342 (37%) | 384,279 (41%) | |
| Non-Hispanic more than one race | 9,514 (2%) | 4,903 (3%) | 7,963 (3%) | 23,389 (2%) | |
| Hispanic, any race | 160,888 (32%) | 47,637 (27%) | 70,353 (27%) | 278,878 (30%) | |
| No race or ethnicity reported | 14,716 (3%) | 1,089 (1%) | 1,308 (1%) | 17,113 (2%) | |
| Rural residential zip code | 99,849 (20%) | 41,094 (24%) | 55,577 (21%) | 196,520 (21%) | <0.001 |
| Medicaid enrollment characteristics | |||||
| Months enrolled | 18.2 (11.7) | 51.9 (6.1) | 76.4 (7.0) | 40.6 (27.3) | <0.001 |
| Changes in enrollment | 2.9 (2.0) | 4.7 (2.5) | 2.7 (2.4) | 3.2 (2.3) | <0.001 |
| ≥ 3 churn cycles | 43,723 (9%) | 58,019 (33%) | 36,982 (14%) | 138,724 (15%) | <0.001 |
| Reason for eligibility | |||||
| Pediatric | 87,707 (17%) | 74,055 (43%) | 128,818 (49%) | 290,580 (31%) | <0.001 |
| Disability | 6,580 (1%) | 6,224 (4%) | 25,515 (10%) | 38,319 (4%) | <0.001 |
| Foster care | 2,863 (1%) | 4,665 (3%) | 13,655 (5%) | 21,183 (2%) | <0.001 |
| Pregnancy | 139,548 (28%) | 43,708 (25%) | 31,507 (12%) | 214,763 (23%) | <0.001 |
| Traditional adult | 488,161 (96%) | 166,022 (96%) | 227,396 (86%) | 881,579 (93%) | <0.001 |
| Health characteristics | |||||
| Prior birth | 223,485 (44%) | 114,902 (66%) | 167,296 (63%) | 505,683 (54%) | <0.001 |
| Hypertension | 55,626 (11%) | 19,603 (11%) | 30,966 (12%) | 106,195 (11%) | <0.001 |
| Diabetes | 65,186 (13%) | 21,133 (12%) | 31,255 (12%) | 117,574 (12%) | <0.001 |
| Depression or anxiety | 37,059 (7%) | 20,167 (12%) | 37,561 (14%) | 94,787 (10%) | <0.001 |
Stratifying by age (Table 2, Figure 1a) younger women were more likely to have greater months of enrollment and fewer changes in enrollment status. They were more likely to have enrollment related to foster care or pediatric eligibility, and less likely to have enrollment related to pregnancy eligibility or traditional adult eligibility. Disability eligibility was stable across ages, at approximately 4%. We found that 35% of those under 18 years of age had codes indicating traditional adult eligibility, suggesting coding errors. Stratifying by prior Medicaid birth (Table 2), those with a prior birth had more months of enrollment over the study period as well as more changes in enrollment status. They were also more likely to have pregnancy eligibility (25% for those with prior birth versus 20% for those with no prior birth).
Table 2:
Differences by age at 2014 birth and by observation of birth prior to 2014
| Age at 2014 birth | Prior Medicaid birth* | Total N=944,068 |
||||||
|---|---|---|---|---|---|---|---|---|
| < 18 N=27,216 |
18 – 19 N=70,855 |
20 – 24 N=322,163 |
25 – 29 N=279,459 |
>=30 N=244,375 |
No | Yes | ||
| Prior Medicaid birth (2007 – 2013) | 2,553 (9.4%) | 17,276 (24.4%) | 174,881 (54.3%) | 177,048 (63.4%) | 133,925 (54.8%) | -- | 505,683 (100%) | 505,683 (53.5%) |
| Age | 16.4 (0.8) | 18.6 (0.5) | 22.1 (1.4) | 26.8 (1.4) | 33 (3.3) | 25.4 (6.2) | 26.7 (5.0) | 26.1 (5.6) |
| Medicaid enrollment characteristics | ||||||||
| Months enrolled | 57.4 (27.9) | 51.0 (29.3) | 43.4 (26.2) | 37.8 (26.0) | 35.5 (27.6) | 33.3 (27.9) | 47.0 (25.1) | 40.6 (27.3) |
| Changes in enrollment status | 2.6 (2.4) | 2.7 (2.3) | 3.4 (2.3) | 3.3 (2.3) | 2.9 (2.2) | 2.4 (2.0) | 3.8 (2.4) | 3.2 (2.3) |
| ≥ 75% coverage | 15,167 (56%) | 32,994 (47%) | 96,245 (30%) | 63,744 (23%) | 55,664 (23%) | 96,518 (22.0%) | 167,296 (33.1%) | 263,814 (28%) |
| ≥ 3 churn cycles | 3,079 (11%) | 8142 (11%) | 56,026 (17%) | 44,429 (16%) | 27,048 (11%) | 35,394 (8.0%) | 103,330 (20.4%) | 138,724 (15%) |
| Reasons for eligibility | ||||||||
| Pediatric | 100% | 66,409 (94%) | 224,171 (70%) | -- | -- | 133,807 (30.5%) | 156,773 (31%) | 290,580 (31%) |
| Disability | 1188 (4%) | 3018 (4%) | 13268 (4%) | 10,475 (4%) | 10,370 (4%) | 18,437 (4.1%) | 19,882 (3.9%) | 38,319 (4%) |
| Foster care | 1883 (7%) | 4768 (7%) | 13,256 (4%) | 1266 (<1%) | 10 (<1%) | 11,075 (2.5%) | 10,108 (2.0%) | 21,183 (2%) |
| Pregnancy | 967 (4%) | 11,823 (17%) | 79,486 (25%) | 69,371 (25%) | 53,116 (22%) | 88,355 (20%) | 126,408 (25%) | 214,763 (23%) |
| Traditional adult | 9,439 (35%) | 50,680 (72%) | 308,899 (96%) | 273, 751 (98%) | 238,810 (98%) | 390,899 (89%) | 490,680 (97%) | 881,579 (93%) |
Note: p-values comparing characteristics across categories by age and, separately, by prior Medicaid birth (using one-way ANOVA for continuous variables and chi-squared tests for categorical variables) are < 0.001 for all comparisons.
Indicates Medicaid birth observed in 2007 – 2013 data.
Figure 1a:
Mean number of Medicaid enrollment months and mean number of changes in Medicaid enrollment status by age
After adjustment in regression analysis (Table 3) we found that younger age and race and ethnicity were both associated with changes in enrollment months. Each additional year of age was associated with −0.93 (95% CI −1.27, −0.60) fewer enrollment months over study period. Individuals who reported non-Hispanic Black race and ethnicity had the highest enrollment months. Compared to this reference group, non-Hispanic White race and ethnicity was associated with −10.05 (95% CI −11.97, −8.12) enrollment months and Hispanic ethnicity with −11.79 (−14.76, −8.82) enrollment months. Residence in a rural zip code was associated with 2.80 (95% CI 1.85, 3.75) additional enrollment months. Ever noting disability or foster care as a reason for enrollment was associated with increased enrollment months (disability 16.86, 95% CI 12.58, 21.13; foster care 15.51, 95% CI 14.26, 16.75) while traditional adult eligibility was associated with decreased enrollment months (−9.61, 95% CI −12.44, −6.78) and pregnancy eligibility was not associated with any change in enrollment months. Hypertension, diabetes, and depression or anxiety were all associated with greater enrollment months.
Table 3:
Association between individual characteristics and preconception enrollment months
| Number of months enrolled in Medicaid beta coefficient (95% CI)* |
Number of changes in Medicaid enrollment status beta coefficient (95% CI)* |
|
|---|---|---|
| Age (years) | −0.93 (−1.27, −0.60) | −0.02 (−0.04, −0.01) |
| Race and ethnicity | ||
| Non-Hispanic American Indian / Alaska Native | −4.98 (−8.98, −0.98) | 0.20 (−0.07, 0.48) |
| Non-Hispanic Asian | −16.70 (−19.89, −13.47) | −0.59 (−0.76, −0.43) |
| Non-Hispanic Black | REF | REF |
| Non-Hispanic Pacific Islander | −18.40 (−21.00, −15.81) | −0.54 (−0.73, −0.34) |
| Non-Hispanic White | −10.05 (−11.97, −8.12) | −0.20 (−0.41, 0.01) |
| Non-Hispanic multiple races | −6.60 (−9.47, −3.74) | 0.21 (0.02, 0.40) |
| Hispanic | −11.79 (−14.76, −8.82) | 0.02 (−0.18, 0.23) |
| No race or ethnicity reported | −24.49 (−28.09, −20.89) | −0.98 (−1.24, −0.73) |
| Rural zip code | 2.80 (1.85, 3.75) | 0.12 (0.07, 0.18) |
| Disability eligibility | 16.86 (12.58, 21.13) | −0.84 (−1.17, −0.50) |
| Foster care eligibility | 15.51 (14.26, 16.75) | 0.22 (−0.05, 0.48) |
| Pregnancy eligibility | −4.44 (−13.13, 4.25) | 0.58 (0.36, 0.80) |
| Traditional adult eligibility | −9.61 (−12.44, −6.78) | 0.84 (0.56, 1.11) |
| Prior Medicaid birth during study period | 15.96 (8.70, 23.21) | 1.11 (0.82, 1.40) |
| Chronic hypertension | 3.12 (2.53, 3.71) | −0.05 (−0.10, −0.01) |
| Pregnancy-induced hypertension | 1.10 (0.71, 1.48) | 0.00 (−0.03, 0.03) |
| Chronic diabetes | 3.28 (2.24, 4.32) | −0.05 (−0.10, 0.01) |
| Gestational diabetes | 0.83 (0.38, 1.29) | −0.04 (−0.06, −0.02) |
| Depression or anxiety during pregnancy | 6.27 (5.55, 6.98) | −0.09 (0.02, 0.16) |
State of residence is included as a fixed effect and standard errors are adjusted to account for clustering within states.
Italics: p-value < 0.05; Bold italics: p-value < 0.001
In a multivariable regression predicting changes in enrollment status, the factors associated with greater enrollment months described above tended to be associated with changes in enrollment status, and in the same direction (Table 3). The main exceptions to this pattern were disability eligibility which was associated with −0.84 (95% CI −1.17, −0.50) fewer changes in enrollment status and traditional adult eligibility, which was associated with 0.84 (95% CI 0.56, 1.11) additional changes in enrollment status. Chronic hypertension, gestational diabetes, and depression or anxiety were also exceptions, but with very small effect sizes (differences of < 0.1 changes in enrollment status).
At the state level, enrollment months ranged from 19.9 to 55.8 months (Figure 1b) and mean changes in enrollment status ranged from 2.1 to 4.6. We found substantial variation in the relationship between enrollment months and changes in enrollment status by state. For the full cohort, in bivariate regression, each change in enrollment status was associated with an additional 0.84 (95% CI 0.82, 0.86) enrollment months. In bivariate regressions stratified by state, each change in enrollment status was associated with a change in enrollment months ranging from −3.2 (−3.7, −2.7) to 4.9 (4.7, 5.2).
Figure 1b:
Mean number of Medicaid enrollment months and mean number of changes in Medicaid enrollment status by state
DISCUSSION
In this 30-state cohort of Medicaid-insured 2014 births, Medicaid was a substantial source of preconception health insurance during the 84 months before birth. For the 46% of individuals enrolled in Medicaid during half or more of the total period, Medicaid was clearly the primary provider of insurance during this time. Medicaid was particularly important for younger individuals and for certain groups known to have high rates of adverse birth outcomes, including individuals of non-Hispanic Black race or non-Hispanic American Indian or Alaska Native race, those with disabilities, and those with chronic conditions including hypertension, diabetes, and depression or anxiety. Medicaid was also particularly important for those who had a prior birth during the study period. This suggests that Medicaid played an even larger role in interconception insurance than in preconception insurance for those who ultimately had Medicaid insured births.
Understanding access to preconception health care is increasingly important as more young adults experience chronic diseases, such as hypertension, diabetes, and depression during their reproductive years.37–39 There is an 8-year gap between the time that many individuals age out of pediatric Medicaid coverage (age 19) and the current mean age of first birth in the US (age 27). These younger individuals have relatively high levels of Medicaid insurance during the study period with a median of 42 months of coverage. People in this age range also have high rates of uninsurance. Considering our finding on relatively high durations of Medicaid coverage, and prior reports of frequent uninsurance in this period, we suspect that many of these younger individuals who go on to have Medicaid-insured births lack other significant sources of insurance during early adulthood.
Optimizing preconception coverage for all individuals may be one component of addressing racial inequities in birth outcomes.40–42 Multiple overlapping factors, often related to structural factors that influence exposure to environmental toxins, investment in health institutions, and implicit or explicit clinician bias, drive these racial inequities.30,43,44 However, Black women are more likely than White women to have chronic conditions that influence birth outcomes before pregnancy, and have worse outcomes than White women with similar conditions.31,45 Preconception care can ensure management of these conditions before pregnancy, and improve access to early prenatal care for women with known risk factors for adverse birth outcomes. Black individuals in our sample had greater months of Medicaid enrollment, though also more frequent transitions in coverage, compared to non-Black individuals. These findings reinforce that health insurance coverage is necessary but not sufficient to ensure health and that Medicaid has an opportunity to improve health promotion for young adult enrollees as a strategy to address birth inequities.
This study had several limitations. Though our data provided comprehensive information on preconception Medicaid enrollment, we are unable to assess whether individuals in this study had other sources of insurance coverage during the study period. We are also unable to assess whether individuals remained eligible for Medicaid during the times that they were not enrolled. In addition, our estimates of Medicaid enrollment may underestimate current preconception Medicaid enrollment for at least 3 reasons. First, our data only accounted for Medicaid enrollment that occurred in the state where the 2014 birth occurred. Some individuals may have had Medicaid coverage in other states during the study period. Second, some individuals may have had a new Medicaid identifier in the same state, due to administrative errors. Third, our data predate both the Medicaid expansion, which began to provide Medicaid to a broader range of low-income adults in some states starting in 2014, and the postpartum Medicaid extension, which provided up to a year of postpartum coverage in most states by 2023. Medicaid expansion reduced uninsurance among women of reproductive age and reduced pregnancy-related churn.46–48 Because of these policy changes, finding may not be generalizable to the current era. In particular, we suspect that more recent data would show greater Medicaid enrollment in the years preceding birth. However, it is possible that associations related to age, race, ethnicity, and health status may remain relevant.
Despite these limitations, our findings indicate that Medicaid should consider itself a major provider of preconception health insurance for Medicaid births. Initiatives to improve preventive care delivery and to consider transitions from primary care to pregnancy care and from pregnancy care to primary care may be aligned with Medicaid efforts to improve birth outcomes and reduce costs associated with adverse birth outcomes. Medicaid has a particular opportunity to improve preconception health among individuals with higher rates of adverse birth outcomes, including individuals with Black race and those with disability eligibility or certain chronic diseases.
Our findings also have implications for Medicaid policies related to eligibility and enrollment that vary between states. Across states, we observed a wide range in eligibility months, changes in enrollment status and, most strikingly, in the relationship between changes in enrollment status and enrollment months. There is a substantial literature suggesting churn is a barrier to health insurance and to health care.16–22 We were not able to observe how many individuals remained eligible for Medicaid in the months that they were not enrolled. However, the range we observed in the relationship between enrollment months and changes in enrollment status suggests that, during our study period, states may not have consistently adopted enrollment policies that align with the goal of supporting health care access and health for the Medicaid-eligible population. A negative correlation between months of coverage and frequency of coverage changes could have reflected that individuals were disenrolled from Medicaid because of changes in eligibility (e.g., increases in income) or that administrative barriers prevented enrollment of people with ongoing eligibility. Meanwhile a positive correlation between months of coverage and frequency of coverage changes could be interpreted as moderate administrative barriers for individuals with ongoing eligibility leading to gaps in coverage but not outright loss of coverage or populations that underwent multiple intermittent phases of eligibility. Further understanding of these dynamics could help reduce unnecessary administrative burdens which create costs for both enrollees and for states.
In conclusion, Medicaid plays a substantial role in preconception health insurance for individuals who ultimately have Medicaid-insured births, particularly for individuals with comorbidities associated with adverse birth outcomes. These findings suggest that Medicaid has an opportunity to improve birth outcomes through preconception health promotion during early adulthood. States seeking to improve birth outcomes may consider improving continuity of Medicaid enrollment during early adulthood or partnering with health systems to support continuity of care for those transitioning in and out of Medicaid. Medicaid should be considered a core driver of preconception health care access in the United States.
Funding:
Data were provided via an agreement between the Centers for Medicare and Medicaid Services, a component of the U.S. Department of Health and Human Services, and the Children’s Hospital of Philadelphia (CHOP) under project “Variations in Health and Health Outcomes for Medicaid Insured Patients” (DUA RSCH-2017-51234). Funding support for the data agreement was provided by the Harriet and Ronald Lassin Endowed Chair in Pediatric Neonatology at CHOP. Dr. Gregory was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under Award Number K23HD102560.
Footnotes
Conflicts of interest: The authors report no competing interests or conflicts of interest.
REFERENCES
- 1.Dubay L, Hill I, Garrett B, et al. Improving Birth Outcomes And Lowering Costs For Women On Medicaid: Impacts Of “Strong Start For Mothers And Newborns.” Health Aff (Millwood). 2020;39(6):1042–1050. [DOI] [PubMed] [Google Scholar]
- 2.Dean SV, Lassi ZS, Imam AM, Bhutta ZA. Preconception care: closing the gap in the continuum of care to accelerate improvements in maternal, newborn and child health. Reprod Health. 2014;11(Suppl 3):S1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Osterman MJK, Hamilton BE, Martin JA, Driscoll AK, Valenzuela CP. Births: Final Data for 2020. National Vital Statistics Reports. 2022;70(17). [PubMed] [Google Scholar]
- 4.Petersen EE, Davis NL, Goodman D, et al. Racial/Ethnic Disparities in Pregnancy-Related Deaths — United States, 2007–2016. MMWR Morb Mortal Wkly Rep. 2019;68(35):762–765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ogunwole SM, Chen X, Mitta S, et al. Interconception Care for Primary Care Providers: Consensus Recommendations on Preconception and Postpartum Management of Reproductive-Age Patients With Medical Comorbidities. Mayo Clin Proc Innov Qual Outcomes. 2021;5(5):872–890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gregory EF, Passarella M, Levine LD, Lorch SA. Interconception preventive care and recurrence of pregnancy complications for Medicaid-insured women. J Womens Health. Published online February 28, 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Becker ER, Roblin DW. Translating primary care practice climate into patient activation: the role of patient trust in physician. Med Care. 2008;46(8):795–805. [DOI] [PubMed] [Google Scholar]
- 8.Mancuso JM. Impact of health literacy and patient trust on glycemic control in an urban USA population. Nurs Health Sci. 2010;12(1):94–104. [DOI] [PubMed] [Google Scholar]
- 9.Poels M, Koster MPH, Boeije HR, Franx A, van Stel HF. Why Do Women Not Use Preconception Care? A Systematic Review On Barriers And Facilitators. Obstet Gynecol Surv. 2016;71(10):603–612. [DOI] [PubMed] [Google Scholar]
- 10.Oza Frank R, Gilson E, Keim SA, Lynch CD, Klebanoff MA. Trends and Factors Associated with Self Reported Receipt of Preconception Care: PRAMS, 2004–2010. Birth. 2014;41(4):367–373. [DOI] [PubMed] [Google Scholar]
- 11.Uninsured Rates Highest For Young Adults Aged 19 to 34.
- 12.Finkelstein A, Taubman S, Wright B, et al. The Oregon Health Insurance Experiment: Evidence from the first year. Q J Econ. 2012;127(3):1057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Aron-Dine A, Einav L, Finkelstein A. The RAND Health Insurance Experiment, three decades later. J Econ Perspect. 2013;27(1):197–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Schoen C, Lyons B, Rowland D, Davis K, Puleo E. Insurance matters for low-income adults: results from a five-state survey. Health Aff (Millwood). 1997;16(5):163–171. [DOI] [PubMed] [Google Scholar]
- 15.Osterman MJK, Hamilton BE, Martin JA, Driscoll AK, Valenzuela CP. Births: Final Data for 2022. Natl Vital Stat Rep. 2024;73(2):1–56. [PubMed] [Google Scholar]
- 16.Edmonds A, Ludema C, Eron JJ, et al. Effects of Health Insurance Interruption on Loss of Hypertension Control in Women With and Women Without HIV. J Womens Health (Larchmt). 2017;26(12):1292–1301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Yabroff KR, Han X, Zhao J, Kirby J, Nogueira L, Zheng Z. Association of Health Insurance Coverage Disruptions With Mortality Risk Among US Working-Age Adults. JAMA Health Forum. 2022;3(11):E224258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Sommers BD, Gourevitch R, Maylone B, Blendon RJ, Epstein AM. Insurance Churning Rates For Low-Income Adults Under Health Reform: Lower Than Expected But Still Harmful For Many. Health Aff (Millwood). 2016;35(10):1816–1824. [DOI] [PubMed] [Google Scholar]
- 19.Egerter S, Braveman P, Marchi K. Timing of insurance coverage and use of prenatal care among low-income women. Am J Public Health. 2002;92(3):423–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gavin NI, Adams EK, Manning WG, Raskind-Hood C, Urato M. The impact of welfare reform on insurance coverage before pregnancy and the timing of prenatal care initiation. Health Serv Res. 2007;42(4):1564–1588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.DeSisto CL, Rohan A, Handler A, Awadalla SS, Johnson T, Rankin K. The Effect of Continuous Versus Pregnancy-Only Medicaid Eligibility on Routine Postpartum Care in Wisconsin, 2011-2015. Matern Child Health J. 2020;24(9):1138–1150. [DOI] [PubMed] [Google Scholar]
- 22.Eliason EL, Gordon SH. The Association Between Postpartum Insurance Instability and Access to Postpartum Mental Health Services: Evidence From Colorado. Womens Health Issues. 2022;32(6):550–556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Witt WP, Wisk LE, Cheng ER, Hampton JM, Hagen EW. Preconception Mental Health Predicts Pregnancy Complications and Adverse Birth Outcomes: A National Population-Based Study. Matern Child Health J. 2012;16(7):1525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Haas JS, Fuentes-Afflick E, Stewart AL, et al. Prepregnancy health status and the risk of preterm delivery. Arch Pediatr Adolesc Med. 2005;159(1):58–63. [DOI] [PubMed] [Google Scholar]
- 25.Misra DP, Guyer B, Allston A. Integrated perinatal health framework: A multiple determinants model with a life span approach. Am J Prev Med. 2003;25(1):65–75. [DOI] [PubMed] [Google Scholar]
- 26.R Development Core Team. R: A language and environment for statistical computing. Published online 2018. [Google Scholar]
- 27.StataCorp. Stata Statistical Software: Release 18.0. Published online 2023. [Google Scholar]
- 28.Pullmann MD, Heflinger CA, Satterwhite Mayberry L. Patterns of medicaid disenrollment for youth with mental health problems. Med Care Res Rev. 2010;67(6):657–675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Babitsch B, Gohl D, von Lengerke T. Re-revisiting Andersen’s Behavioral Model of Health Services Use: a systematic review of studies from 1998-2011. Psychosoc Med. 2012;9:Doc11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Burris HH, Hacker MR. Birth outcome racial disparities: A result of intersecting social and environmental factors. Semin Perinatol. 2017;41(6):360–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Denny CH, Floyd RL, Green PP, Hayes DK. Racial and ethnic disparities in preconception risk factors and preconception care. J Womens Health (Larchmt). 2012;21(7):720–729. [DOI] [PubMed] [Google Scholar]
- 32.Biener AI, Zuvekas SH. Do racial and ethnic disparities in health care use vary with health? Health Serv Res. 2019;54(1):64–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lewey J, Levine LD, Yang L, Triebwasser JE, Groeneveld PW. Patterns of postpartum ambulatory care follow-up care among women with hypertensive disorders of pregnancy. J Am Heart Assoc. 2020;9(17). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: Executive Summary: A Report of the American College of Cardiology/American Heart Association Task F. J Am Coll Cardiol. 2018;71(19):2199–2269.29146533 [Google Scholar]
- 35.American Diabetes Association. Pharmacologic approaches to glycemic treatment: Standards of medical care in diabetes. Diabetes Care. 2019;42(Supplement 1):S90–S102. [DOI] [PubMed] [Google Scholar]
- 36.Bennett WL, Chang HYY, Levine DM, et al. Utilization of primary and obstetric care after medically complicated pregnancies: An analysis of medical claims data. J Gen Intern Med. 2014;29(4):636–645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Goodwin RD, Dierker LC, Wu M, Galea S, Hoven CW, Weinberger AH. Trends in U.S. Depression Prevalence From 2015 to 2020: The Widening Treatment Gap. Am J Prev Med. 2022;63(5):726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Lawrence JM, Divers J, Isom S, et al. Trends in Prevalence of Type 1 and Type 2 Diabetes in Children and Adolescents in the US, 2001-2017. JAMA. 2021;326(8):717–727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Hardy ST, Sakhuja S, Jaeger BC, et al. Trends in Blood Pressure and Hypertension Among US Children and Adolescents, 1999-2018. JAMA Netw Open. 2021;4(4):e213917–e213917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Pregnancy Mortality Surveillance System | Maternal and Infant Health | CDC. [Google Scholar]
- 41.Hoyert DL. Health E-stat, February 2022. Published online 2020.
- 42.Louis JM, Menard MK, Gee RE. Racial and ethnic disparities in maternal morbidity and mortality. Obstetrics and Gynecology. 2015;125(3):690–694. [DOI] [PubMed] [Google Scholar]
- 43.Howell EA. Reducing Disparities in Severe Maternal Morbidity and Mortality. Clin Obstet Gynecol. 2018;61(2):387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Minehart RD, Bryant AS, Jackson J, Daly JL. Racial/Ethnic Inequities in Pregnancy-Related Morbidity and Mortality. Obstet Gynecol Clin North Am. 2021;48(1):31–51. [DOI] [PubMed] [Google Scholar]
- 45.Bornstein E, Eliner Y, Chervenak FA, Grünebaum A. Concerning trends in maternal risk factors in the United States: 1989-2018. EClinicalMedicine. 2020;29–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Bellerose M, Collin L, Daw JR. The ACA Medicaid Expansion And Perinatal Insurance, Health Care Use, And Health Outcomes: A Systematic Review. Health Aff (Millwood). 2022;41(1):60–68. [DOI] [PubMed] [Google Scholar]
- 47.Lee LK, Chien A, Stewart A, et al. Women’s coverage, utilization, affordability, and health after the aca: A review of the literature. Health Aff. 2020;39(3):387–394. [DOI] [PubMed] [Google Scholar]
- 48.Chen J, Ouyang L, Goodman DA, et al. Association of Medicaid Expansion Under the Affordable Care Act With Medicaid Coverage in the Prepregnancy, Prenatal, and Postpartum Periods. Women’s Health Issues. 2023;33(6):582–591. [DOI] [PMC free article] [PubMed] [Google Scholar]