Modeling the Likelihood of Low Birth Weight: Findings from a Chicago-Area Health System

Objective: This study presents a statistical model of the incidence of low birth weight (LBW) births in a large, Chicago-area hospital system. The study was undertaken to provide a strategic framework for future health system interventions. Methods: Administrative and electronic health records were matched to census Zip Code Tabulation Area (ZCTA) household poverty data for 42,681 births in 2016–2019 at seven system hospitals, serving a diverse patient population. A logistic regression model of LBW incidence was estimated to test the independent significance of maternal sociodemographic characteristics after controlling for clinical risk factors. Results: The incidence of LBW was 6.3% overall but 11.3% among non-Hispanic Black patients as compared to 5.1% among non-Hispanic White patients. LBW incidence ranged from 9.2% for patients from the poorest ZCTA (20% + poor households) compared to 5.6% of patients from the most affluent (< 5% poor) ZCTA. Nulliparous patients, patients with pre-existing chronic conditions, and patients with hypertensive disorders of pregnancy were significantly more likely to have LBW births. After controlling for clinical risk factors and poverty level, non-Hispanic Black patients were still over 80% more likely and to have a LBW birth. Discussion: Study findings reveal the joint effects of social and clinical risk factors. Findings profile our highest-risk populations for targeted interventions. Promising prenatal care redesign programs include pregnancy patient navigators, home and group visits, eHealth telemonitoring, improved mental health screening, and diversification of the maternity care workforce. Decreasing LBW births should be a national public health policy priority and will require major investments in the most impacted communities.