Healthcare system variability and inpatient maternal mortality in the United States

To examine the association of inpatient maternal mortality with variability in healthcare services delivery such as hospital size, urban/rural designation, teaching/non-teaching status, regional location, and insurance coverage.

This is a pooled, cross-sectional analysis of the National Inpatient Sample (2012-2014). Information on maternal demographics, clinical conditions, and birth outcomes were identified using respective ICD9-CM codes. Bivariate and multivariate analysis using logistic regression models were used to describe maternal characteristics and to calculate the risk of mortality with each independent variable.

The weighted sample included 12,409,939 hospital records (82.6% are 18-34-year-old and 49.5% are Caucasians). Maternal death during hospitalization occurred in 1310 cases (12/100,000 live birth). Women with cardiovascular disorders, hemorrhage or sepsis were 33.6, 4.7, and 5.4 times more likely to suffer inpatient maternal mortality. Compared to small-sized hospitals, delivery at medium or large size hospitals is associated with higher mortality, adjusted odds ratios (aOR) 1.8 (1.4-2.3), and 2.2 (1.8-2.8), respectively. Adjusted OR for inpatient maternal mortality in urban non-teaching or urban teaching compared to rural hospitals were 2.2 (1.7-3.0) and 2.9 (2.2-3.9), respectively. Women in the South have higher maternal mortality compared to Northeast, aOR 1.7 (1.5-2.1). Women coved with public insurance experience higher inpatient maternal mortality compared to those with private insurance, aOR: 2.6 (2.1-3.2) and 1.9 (1.6-2.1), respectively.

Factors related to variability in healthcare delivery may play a role in inpatient maternal mortality. Some could be explained by the case mix and the clinical conditions affecting birthing outcomes. A qualitative analysis is needed to explore how these factors relate to increased maternal mortality in certain hospital settings.