Teen Birth Trends: Addressing Intergenerational Effects
Introduction
In 2023 alone, 140,977 births were recorded among mothers younger than 20 years old, a rate more than three times higher than that observed in most other high-income nations (Osterman et al., 2025). Despite the decreasing trend over the past two decades, teen births remain a public health challenge. However, this burden is not evenly distributed; it disproportionately affects rural counties and regions in the Southern United States, which report the highest rates of teen births.
The implications of teen childbearing extend beyond immediate maternal outcomes, significantly influencing a wide array of adverse outcomes for their offspring. Extensive empirical research has documented associations between teen parenthood and increased risks of low birth weight, premature birth, and elevated infant mortality. Furthermore, children born to teenage parents exhibit lower kindergarten-readiness scores and higher rates of dropping out of high school. In the long term, teen births correlate with reduced lifetime earnings, increased reliance on social safety-net programs, and heightened prevalence of chronic health issues such as obesity, depression, and chronic diseases during young adulthood (Congress.gov | Library of Congress, 2025). These associations suggest that teen parenthood may perpetuate intergenerational cycles of socioeconomic disadvantage, highlighting its importance as a critical target for public health policy intervention.
Despite these observed correlations, it remains uncertain whether poor child outcomes are direct consequences of early parenthood or result from underlying systemic factors such as poverty, limited educational opportunities, or inadequate access to healthcare, which predispose adolescents to early childbearing. Not only does teen birth result in poorer outcomes for offspring, but also for teen mothers who face greater economic vulnerability due to reduced participation in higher education and the labor market.
Several factors have contributed to the decline in U.S. teen birth rates (TBR) over recent decades. The most significant contributors are improved contraceptive use—including greater reliance on effective, long-acting methods—and, to a lesser extent, a reduction in sexual activity among younger teens (Geiger, 2019). Broader social and economic factors, such as increased access to education and job opportunities, may also play a role.
Given these complexities, our study aims to identify counties and populations at the highest risk of teen birth to more accurately target resources and policy interventions. Specifically, we seek actionable insights into the policy debate surrounding disparities in reproductive healthcare access and early family planning, which have been exacerbated by varying state-level policies enacted in recent years. By examining county-level data, we aim to clarify whether teen parenthood itself is a determinant of negative child outcomes or a consequence of broader structural factors, informing more effective and targeted intervention strategies.
Data
This analysis uses the 2025 National County Health Release data, which provides county-level information on key socioeconomic and health-related factors, accessed from the County Health Rankings & Roadmaps website. Missing data points were imputed using mean replacement.
Feature selection was performed using machine learning algorithms to identify the most predictive variables while reducing overfitting and noise.
We also incorporated Kaiser Family Foundation survey data on state-level health insurance coverage for children (ages 0–18), which reports the proportion insured by employer plans, Medicaid, other public sources, non-group plans, and those uninsured. The most recent available data was used, which was from 2023.
Outcome Variable:
- Teen births: Number of births per 1,000 female population ages 15–19
Top Four Predictors:
- High school completion: Percentage of adults (25+) with a high school diploma or equivalent
- Uninsured adults: Percentage of adults under 65 without health insurance
- Food insecurity: Percentage of the population lacking adequate access to food
- Some college: Percentage of adults ages 25–44 with some post-secondary education
Methods
We first utilized PCA followed by Gaussian Mixture Modeling (GMM) to cluster counties based on similarities in predictor variables andTBR. The GMM approach allowed us to assign counties probabilistically, overcoming limitations of rigid clustering methods such as k-means. Three clusters emerged: counties with high TBR and poor socioeconomic conditions, counties with intermediate characteristics, and counties with low TBR and favorable socioeconomic factors.
Next, we conducted feature selection to identify the strongest predictors of cluster assignment using two separate methods. Initially, we created a decision tree with GMM-derived clusters as outcomes and ranked predictor importance using metrics such as the Gini index. Separately, counties were divided into tertiles based on their raw TBR, and a random forest model was trained to verify feature importance rankings.
Having identified the most influential variables using the decision tree, we proceeded with multilinear regression analysis, removing previously planned regularization methods (Ridge and LASSO) as they were no longer necessary.
Finally, we conducted a landscape assessment of teen-birth intervention strategies. We analyzed state-level payor mixes—calculating the proportions of children covered by employer-sponsored insurance, Medicaid/other public plans, and those uninsured—and explored additional contextual predictors, such as food insecurity and educational attainment, to further understand factors influencing teen pregnancy risk.
Results
Where one lives is a strong predictor for teen birth risk. The geographic regions predominantly in the Southern United States have worse outcomes: the highest TBR, shorter life expectancy, worse access to resources, etc. Contrastingly, the Northeast and northern Midwest tend to be assigned to the lowest-risk cluster, with the best health and socioeconomic outcomes. The counties in the West tend to fall into the intermediate risk cluster (Fig. 1).
Where are the highest risk counties?
This study identified four key variables that determined risk-cluster assignment for TBR: adult uninsured status, food insecurity, some college completion, and high school completion rates. The strong correlation between lower educational attainment and higher teen birth rates is well-supported—teen childbearing often disrupts schooling, as physical demands of pregnancy and the responsibilities of early parenthood can make it difficult for teens to remain in or complete school.
Identifying the top four socioeconomic predictors
The food insecurity and adult uninsured status also emerged as influential predictors. While there is limited prior research on the connection between food insecurity and TBR, our findings suggest that poorer food insecurity may signal broader socioeconomic disadvantage. Areas with higher food insecurity scores are likely to experience not just limited access to healthy foods, but also to other critical resources, which may contribute to higher teen birth rates. Food insecurity may also be a reflection of household circumstances where a child may have to prioritize participating in the labor force or spending time outside of school, obtaining food or other essentials for themselves and their family. This relationship warrants further investigation to understand better the pathways linking food insecurity, resource availability, and teen pregnancy.
Interestingly, the random forest model identified the same top four features as the GMM combined with the single decision tree. Furthermore, the XGBoost model identified the same top features, ensuring confidence that we have found the most relevant features within the dataset. The shapes of the relationships between the top-identified predictors and TBR were confirmed using general additive models.
Visualization of predictor relationships using GAM
Visualization of the relationships of these predictors against TBR was confirmed through general additive models. The relationship between food insecurity and TBR was strongly positive and linear, with TBR increasing steeply at higher levels of food insecurity (Fig. 3, top left). Higher rates of adults with “some college” experience were associated with a linear decrease in TBR, suggesting educational attainment is protective (Fig. 3, top right). The effect of high school completion on TBR showed a mild nonlinear pattern: TBR decreased with increasing high school completion rates, with diminishing returns at the highest levels (Fig. 3, bottom left). Higher rates of uninsured adults were associated with modest increases in TBR, but the effect was less pronounced than for the other predictors (Fig. 3, bottom right). Analysis of the KFF data also highlighted differences in insurance coverage profiles between states. States with the highest TBRs tended to have higher proportions of residents who were uninsured or who relied on Medicaid or other public insurance programs. In contrast, states with the lowest TBRs were characterized by higher rates of employer-sponsored insurance coverage. These findings point to the importance of insurance coverage—and, more broadly, access to affordable health care—as a potential lever for reducing teen pregnancies and improving overall child and adolescent outcomes.
State TBRs against insurance status of children
States in the highest teen birth rate quartile had a significantly higher proportion of children with non-employer-based insurance coverage (mean = 56.8%) compared to those in the lowest quartile (mean = 45.8%; t = 5.74, p < 0.001; 95% CI for the difference: 7.0% to 15.0%). Linear regression was used to create a policy simulation, which revealed that even a modest improvement in several key variables can lead to a significant reduction in TBR. Specifically, the analysis suggests that a 2% decrease in food insecurity and a 3% reduction in the number of uninsured adults, along with a 2% increase in high school graduation rates and a 1% increase in some college attendance, could result in four fewer teen births per 1,000 females.
Small changes in policy yield noticable decrease in teen births
Using linear regression, we suggested some small percentage improvements in the top predictor variables (Fig. 5 top right). From these changes, there is an estimated reduction in about 4 births per 1,000 teenage females in the population. The bar chart shows the decrease in TBR after the conditional changes (Fig. 5 left), and the correlation coefficients between each predictor and TBR is shown for this model (Fig. 5 bottom right).
Recommendations
Based on our findings, we recommend that UnitedHealth Group and its partners prioritize programs that:
These recommendations are aimed at breaking cycles of disadvantage and ensuring healthier futures for adolescents and their children.
Discussion
From our analysis, food insecurity appears to be a robust proxy for broader socioeconomic deprivation and limited resource availability. Jurisdictions with higher levels of food insecurity tend to have elevated teen birth rates, likely because chronic food stress forces adolescents to prioritize work or caretaking over school, reducing educational engagement and increasing pregnancy risk. Supporting this, research on voting restrictions has shown that areas with more barriers to civic participation—often mirroring impoverished regions with weaker healthcare infrastructure—also experience higher teen birth rates (Pabayo et al., 2023).
We also observed that states with higher adult uninsurance rates have higher teen birth rates, reflecting diminished access to health education and reproductive services among families. Importantly, educational attainment variables further reinforce this picture: counties with lower high-school graduation rates and smaller proportions of adults with any college experience exhibit the highest teen birth rates. This inverse relationship suggests that sustained engagement in secondary and post-secondary education provides protective effects against early childbearing, whether through increased access to resources, education, stronger support networks, or delayed entry into parenting.
A study of girls along the U.S.-Mexico border found that circumstances of food insecurity promoted financial reliance for basic needs on male partners (Reed et al., 2022). This subsequently compromised these girls’ control of making their own reproductive health decisions including contraceptive use and pregnancy planning. The same study also found that economic vulnerability may decrease girls’ perceptions of future opportunities available to them, and this dynamic decreases intentions and behaviors to delay pregnancy.This supports our findings relating TBR to food insecurity and educational attainment, demonstrating the importance in empowering girls to have control over their health decisions.
Together, these findings highlight the dynamic between material hardship, healthcare access, and educational opportunity in shaping teen pregnancy risk. Interventions that simultaneously improve food security, expand insurance coverage, and bolster high-school completion and college enrollment are therefore critical to reducing teen birth rates and mitigating their intergenerational consequences.
Limitations:
Teen birth rates are a powerful indicator for child and community well-being because they are closely linked to long-term health, educational, and economic outcomes—both for young mothers and their children. Numerous studies demonstrate that children born to teen mothers face a higher risk of adverse outcomes ranging from health status to economic status. These effects often extend into adolescence and adulthood, perpetuating cycles of disadvantage across generations.
For healthcare organizations and policymakers, monitoring teen birth rates as a public-health metric is critical for assessing how social determinants and barriers to care contribute to intergenerational health disparities. Teen pregnancy is strongly associated with socioeconomic factors such as income, education, and access to reproductive health services—issues that lie at the heart of public health and health equity efforts (Congress.gov | Library of Congress, 2025). Interventions targeting the child-adolescent population are particularly impactful, as supporting young people during this critical developmental period can break cycles of poor outcomes and drive meaningful change for entire communities.
Additionally, the topic of teen birth rates is especially timely given the current national debate around reproductive health access. As states revisit policies affecting contraception, abortion, and sexual health education, understanding the patterns and predictors of teen pregnancy is essential for designing effective, equitable, and evidence-based interventions. For UnitedHealth Group and others working at the intersection of health analytics and strategy, focusing on teen birth rates can inform smarter resource allocation, identify populations most in need of support, and help drive improvements in child and maternal health outcomes.