Geographic gaps in U.S. breastfeeding rates tied to race, income, disability support

Across the United States, a mother's decision to breastfeed—or her ability to do so—depends less on her own choice than on where she lives. In some counties, nine out of ten newborns receive breast milk before leaving the hospital. In others, barely more than one in five do. This geographic chasm, researchers have discovered, is not random. It traces itself along lines of race, income, disability, and the availability of support that varies wildly from place to place.

A comprehensive analysis of 3,011 counties across the continental United States, published in PLOS Global Public Health, mapped these disparities with precision. The data came from birth certificates recorded between 2018 and 2019—contemporaneous records that avoid the distortions of memory—combined with census information about the communities where those births occurred. The researchers used a statistical approach called multiscale geographically weighted regression, which allows different factors to operate at different scales. Some influences shape breastfeeding patterns across entire regions. Others matter only in specific clusters of counties. This flexibility revealed something that simpler national statistics had obscured: breastfeeding initiation is fundamentally a local problem, requiring local solutions.

The geography is stark. The Northeast, parts of the West Coast, and Central Texas show the highest initiation rates. The Gulf Coast states—Louisiana, Mississippi, Alabama—and much of Appalachia lag substantially behind. But the reasons for these patterns are not uniform. Education matters everywhere. A one-standard-deviation increase in the share of adults with a bachelor's degree correlates with a 1.62 percentage-point increase in breastfeeding initiation across all counties. Race also operates globally: counties with larger Black populations show consistently lower initiation rates, a difference of 2.56 to 2.62 percentage points for each standard-deviation increase in Black population share. Yet other factors tell more complicated stories. Hispanic population share, for instance, shows a strong positive association with breastfeeding initiation—but only in the Northeast, Mid-Atlantic, and parts of the Southeast. In the West and Southwest, where Hispanic populations are actually larger, the relationship disappears. The researchers hypothesize that regional differences in which Hispanic subgroups live in each area, and varying levels of acculturation, may explain this east-west divide.

Disability presents perhaps the most striking local variation. In three distinct clusters—one spanning Arkansas, Tennessee, Missouri, Illinois, and Kentucky; another in Virginia; and a third in Oklahoma and South Texas—higher disability prevalence correlates with substantially lower breastfeeding initiation rates, with declines as steep as 10 percentage points. Yet in two smaller clusters along the Alabama-Georgia border and in eastern Arkansas and northeastern Louisiana, the opposite occurs: higher disability prevalence associates with increases in breastfeeding initiation of up to 8.2 percentage points. These communities appear to be doing something different—perhaps maintaining stronger support networks or community resources—but the research cannot yet say what. Female-headed households with children show similarly localized effects, with strong negative associations in the Upper Midwest, Mississippi Delta, Appalachia, and Eastern North Carolina, but no significant relationship in much of the West and Northeast.

Income growth, residential stability, and access to public assistance all show regional patterns. Income growth correlates with lower breastfeeding initiation in the Southeast—a counterintuitive finding that the researchers attribute to the region's demographic composition and historical patterns. Residential stability matters most in the Southeast as well, where population turnover associates with lower initiation rates. Public assistance shows its strongest negative association in the Northern Plains. Female labor force participation, by contrast, shows a strong positive relationship with breastfeeding initiation in the Southeast and Mid-Atlantic, suggesting that in these regions, women's workforce engagement may correlate with greater awareness or institutional support for breastfeeding.

The statistical model itself improved dramatically when accounting for geography. A traditional regression model explained 64.1 percent of the variation in breastfeeding initiation rates and violated key statistical assumptions. The geographically weighted model explained 85.2 percent of the variation and satisfied those assumptions, suggesting it captured something real about how place shapes maternal behavior. The local intercepts—the baseline breastfeeding rates after accounting for measured factors—reveal persistent regional differences that point to unmeasured influences: cultural norms, community trust in healthcare, institutional infrastructure, and historical patterns of health behavior that vary geographically and remain invisible in the data.

The researchers identify three cost-effective interventions with proven track records. Baby-Friendly Hospital designation, which trains staff in lactation support and implements evidence-based protocols, could expand access to professional guidance. Peer-to-peer support networks—whether in-person groups like La Leche League or online forums for working and rural mothers—address isolation and provide practical knowledge. Expanding WIC eligibility and removing income restrictions would increase access to breastfeeding support among the lowest-income families. All three strategies could be particularly transformative in the Gulf Coast and Appalachia, where initiation rates lag furthest behind national targets. The challenge ahead is not discovering what works, but recognizing that what works in one place may need adaptation for another, and that sustainable improvement requires understanding the specific barriers and assets of each community.