Double Disadvantage in Delivery Hospital for Black and Hispanic Women and High-Risk Infants

Abstract

Objective: To determine whether delivery hospitals that perform poorly for women also perform poorly for high-risk infants and to what extent Black and Hispanic women receive care at hospitals that perform poorly for both women and infants.

Methods: We examined the correlation between hospital rankings for severe maternal morbidity and very preterm morbidity and mortality in New York City Hospitals using linked birth certificate and state discharge data for 2010-2014. We used mixed-effects logistic regression with a random hospital-specific intercept to generate risk standardized severe maternal morbidity rates and very preterm birth neonatal morbidity and mortality rates for each hospital. We ranked hospitals separately by these risk-standardized rates. We used k-means cluster analysis to categorize hospitals based on their performance on both metrics and risk-adjusted multinomial logistic regression to estimate adjusted probabilities of delivering in each hospital-quality cluster by race/ethnicity.

Results: Hospital rankings for severe maternal morbidity and very preterm neonatal morbidity-mortality were moderately correlated (r = .32; p = .05). A 5-cluster solution best fit the data and yielded the categories for hospital performance for women and infants: excellent, good, fair, fair to poor, poor. Black and Hispanic versus White women were less likely to deliver in an excellent quality cluster (adjusted percent of 11%, 18% vs 28%, respectively, p < .001) and more likely to deliver in a poor quality cluster (adjusted percent of 28%, 20%, vs. 4%, respectively, p < .001).

Conclusions for practise: Hospital performance for maternal and high-risk infant outcomes is only moderately correlated but Black and Hispanic women deliver at hospitals with worse outcomes for both women and very preterm infants.

Keywords: Disparities; Hospital; Maternal and child health; Quality; Very preterm.

Figures

Figure I.

Scatterplot of hospital rankings for severe maternal morbidity and very preterm birth morbidity/mortality and quality clusters based on k-means analysis, New York City, 2010-2014

Figure 2.

Adjusted* percent of deliveries by race/ethnicity and quality cluster, New York City hospitals, 2010-2014. *Adjusted percent is predicted probability estimated from multinomial logistic regression model adjusting for race-ethnicity, maternal age, parity, multiple birth, maternal education, Medicaid insurance, nativity, asthma, prepregnancy diabetes, prepregnancy hypertension, gestational diabetes, gestational hypertension, prepregnancy BMI, smoking during pregna