Spatial-Temporal Trajectories of County-Level Low Birthweight Rates and Contextual Determinants in the United States

This longitudinal study of over 18 million births across 3,108 U.S. counties from 2016 to 2021 reveals a significant increase in low birthweight rates driven by county-level socioeconomic and maternal health factors rather than metropolitan status, highlighting the need for targeted, place-based interventions to address structural vulnerabilities.

The Gist

Imagine the United States as a massive, sprawling garden with over 3,000 different plots (counties). In this garden, the health of the "seedlings" (newborn babies) is a top priority. Specifically, the gardeners are worried about "low birthweight" seedlings—those born too small (under 5.5 pounds). These tiny seedlings are more fragile and face a harder path to a healthy, long life.

For a long time, people assumed that the biggest problem was simply where you lived. They thought, "If you live in the big, busy city (metropolitan), you're safe. If you live in the quiet, open country (non-metropolitan), you're at risk."

This study, however, decided to take a closer look at the soil, the weather, and the gardeners themselves across the entire country from 2016 to 2021. Here is what they found, explained simply:

1. The "City vs. Country" Myth

The researchers expected to find a clear line: cities doing well, rural areas struggling. But the map looked different. It was like looking at a garden where the "city" plots and "country" plots were both full of healthy and sick seedlings mixed together.

The Big Surprise: Once they adjusted for what was actually happening in the soil (like poverty, lack of insurance, or racism), it didn't matter if the plot was in a city or the country. Location alone wasn't the villain. A struggling neighborhood in a big city faced the same risks as a struggling neighborhood in a small town.

2. The Real Culprits: The "Soil Conditions"

Instead of the address, the study found that the conditions of the soil were what really determined if a seedling would grow strong. Think of these as the "ingredients" in the county's recipe for baby health:

• The "Black Soil" Factor: Counties with higher populations of Black women had significantly higher rates of low birthweight. The study suggests this isn't about biology, but about the "toxic soil" of structural racism. It's like trying to grow a plant in soil that has been polluted by decades of unfair treatment, segregation, and lack of resources. The stress of this unfairness physically affects the mother and the baby.
• The "No Umbrella" Factor (Insurance): In counties where many people didn't have health insurance, babies were smaller. Without an umbrella (insurance), mothers couldn't get the prenatal care they needed to stay dry and healthy during the storm of pregnancy.
• The "Empty Stomach" Factor (Underweight): If the mothers in a county were often underweight, the babies were more likely to be underweight too. You can't build a strong house on a weak foundation.
• The "Smoky Air" Factor: Counties where more mothers smoked during pregnancy saw more low birthweight babies. It's like trying to grow a plant in a room filled with smoke; the seedling just can't breathe right.
• The "No Doctor" Factor: If mothers didn't see a doctor for prenatal care, the risk skyrocketed.

3. The "Storm" of 2020-2021

The study noticed something scary: the garden got much worse in the last two years (2020–2021). The rate of small babies jumped up significantly everywhere.

The Metaphor: Imagine a sudden, massive storm (the COVID-19 pandemic). This storm knocked down fences, closed the garden gates (hospitals), scared the gardeners, and made it hard to get water (healthcare). This storm hit the most vulnerable plots the hardest, making the existing problems much worse.

4. The Takeaway: Fix the Soil, Not the Signpost

The main lesson of this paper is that we can't just look at a map and say, "Help the rural areas." That's like trying to fix a garden by only watering the country plots and ignoring the dry patches in the city.

The Solution: We need to fix the soil conditions everywhere.

• We need to ensure every mother has health insurance (an umbrella).
• We need to fight racism and improve the "soil quality" in communities of color.
• We need to make sure mothers can eat well and get to the doctor, whether they live in a skyscraper or a farmhouse.

In short: A baby's health isn't determined by the zip code on the mailbox; it's determined by the quality of the neighborhood, the fairness of the system, and the support the mother receives. To grow strong seedlings, we have to fix the garden, not just the address.

Technical Summary

1. Problem Statement

Low birthweight (LBW; <2,500 grams) is a critical public health indicator associated with increased risks of infant mortality, neurologic impairment, and chronic diseases (e.g., cardiovascular disease, diabetes). While maternal risk factors (e.g., smoking, age, race) are well-documented, national and state-level estimates often mask significant county-level heterogeneity.

Existing literature lacks a comprehensive understanding of:

• How LBW rates evolve geographically across metropolitan vs. non-metropolitan counties over time.
• Whether rural–urban disparities persist after adjusting for county-level maternal composition, socioeconomic conditions, and healthcare access.
• The specific role of contextual determinants (structural and socioeconomic factors) in driving these patterns.

2. Methodology

Study Design and Data Sources

• Design: Longitudinal study covering the period 2016–2021.
• Sample: 18,193,145 singleton births across 3,108 U.S. counties. Multiple gestations and births with congenital anomalies were excluded.
• Data Sources:
  • Birth Data: Restricted-use National Center for Health Statistics (NCHS) natality files (maternal characteristics, birthweight, residence county).
  • Contextual Data: County Health Rankings (unemployment, insurance status, violent crime, primary care provider rates).
• Time Aggregation: Data was aggregated into three two-year periods (2016–2017, 2018–2019, 2020–2021) to stabilize rates.
• Classification: Counties were categorized using the NCHS Urban–Rural Classification Scheme into: Large Metropolitan, Small Metropolitan, and Non-Metropolitan.

Statistical Analysis

• Spatial Smoothing: To address data sparsity in small counties, Bayesian hierarchical spatial models were used to estimate smoothed Standardized Incidence Ratios (SIRs).
  • The model included a spatial random effect (conditional autoregressive Gaussian process) to capture neighborhood influence and a non-spatial random effect.
  • Inference was conducted using the R-INLA package (Integrated Nested Laplace Approximation).
• Regression Modeling: Generalized Estimating Equations (GEE) with negative binomial regression were employed to examine associations between county-level contextual variables and LBW rates.
  • Offset: Logarithm of the total births per county.
  • Predictors: County-level percentages of maternal characteristics (race, BMI, smoking, prenatal care, education, age) and contextual factors (unemployment, uninsured, crime, primary care).
  • Interaction Terms: Tested for effect modification by metropolitan status (e.g., interaction between county type and time, race, or socioeconomic indicators).
  • Standardization: All predictors were standardized (mean=0, SD=1).

3. Key Contributions

• Granular Geographic Analysis: Provides the first detailed analysis of LBW trajectories at the county level across the entire U.S., moving beyond state-level aggregates.
• Methodological Rigor: Utilizes Bayesian spatial modeling to generate reliable estimates for counties with small populations, reducing the noise associated with raw rates.
• Re-evaluation of Rural-Urban Disparities: Explicitly tests the hypothesis that rural/urban location is a primary driver of LBW, adjusting for deep-seated socioeconomic and maternal health factors.
• Temporal Insight: Captures the trajectory of LBW rates leading up to and including the 2020–2021 period, offering insights into the potential impact of the COVID-19 pandemic on birth outcomes.

4. Key Results

Descriptive Trends

• Overall Increase: LBW rates increased significantly over the study period.
  • 2016–2017: 6.42%
  • 2018–2019: 6.56%
  • 2020–2021: 9.10% (a sharp rise).
• Demographic Disparities:
  • Race: Infants born to Black mothers consistently had higher LBW rates compared to other racial groups.
  • Prenatal Care: Mothers with no prenatal care had significantly higher LBW rates (17–19%) compared to those with care (6%).
  • Smoking & BMI: Smoking during pregnancy and maternal underweight status were strongly associated with LBW.
  • Age: Extremes of maternal age (<20 and >40) showed higher risks.

Spatial Distribution

• Elevated LBW risk (SIR > 1 with >95% exceedance probability) was persistent in specific clusters, including parts of the South, Midwest, Northeast, and specific Western states (CA, OR, WA, AZ, NM).
• These spatial patterns remained consistent across all three time periods.

Multivariate Regression Findings

• Metropolitan Status: After adjusting for covariates, metropolitan status (large metro, small metro, non-metro) was NOT significantly associated with LBW rates. No significant interactions were found between metropolitan status and time or other maternal characteristics.
• Significant Predictors of Increased LBW:
  • % Black Women: IRR = 1.14 (95% CI: 1.13–1.15).
  • % Underweight Women: IRR = 1.11 (95% CI: 1.10–1.12).
  • % No Prenatal Care: IRR = 1.03 (95% CI: 1.02–1.03).
  • % Smoking During Pregnancy: IRR = 1.05 (95% CI: 1.04–1.06).
  • % Uninsured Population: IRR = 1.04 (95% CI: 1.03–1.05).
• Temporal Effect: Compared to 2016–2017, LBW rates increased significantly in 2018–2019 (IRR = 1.09) and 2020–2021 (IRR = 1.37).

5. Significance and Implications

Interpretation of Findings
The study challenges the assumption that rural location is the primary driver of poor birth outcomes. Instead, it suggests that structural and socioeconomic conditions (specifically racial composition, insurance coverage, and access to prenatal care) are the salient drivers of LBW risk, regardless of whether a county is urban or rural. The sharp increase in 2020–2021 likely reflects pandemic-related disruptions (stress, reduced care access, economic instability).

Policy and Public Health Implications

• Targeted Interventions: Interventions should not be based solely on rural/urban classifications. Instead, resources must be directed toward counties with high proportions of Black populations, uninsured residents, and women lacking prenatal care.
• Structural Focus: Addressing LBW requires tackling structural racism, expanding insurance coverage, and improving access to high-quality prenatal and nutritional support.
• Future Research: The authors recommend moving to sub-county scales (census tracts) to better capture neighborhood deprivation and integrating measures of chronic stress and environmental hazards.

Conclusion
The paper concludes that while LBW rates are rising across the U.S., the disparities are driven by place-based socioeconomic vulnerabilities rather than metropolitan status alone. Effective reduction of LBW requires geographically targeted strategies that address the specific structural and maternal health deficits identified in high-risk counties.