58: Detecting the maternal in mortality: A data fusion approach to improving pregnancy-associated mortality ascertainment


      To utilize a data fusion approach to identify cases of maternal mortality not detected with standard surveillance strategies. With rising maternal mortality rates in the United States, understanding the full spectrum of pregnancy-associated mortality has the potential to inform clinical and community-level interventions with the goal of improving the safety of women’s obstetric experiences.

      Study Design

      Mixed methods identification of a retrospective cohort from the electronic health records at a tertiary medical center from 2011-2018. The primary outcome was pregnancy-associated mortality during pregnancy or within the first postpartum year in this cohort versus the hospital-identified cohort. Secondary outcomes included causes of death and comorbidities. Cases were identified via a standard medical informatics service query and with the TriNetX cohort discovery tool as patients with a deceased vital status and evidence of antecedent pregnancy exposure. This cohort underwent chart review to confirm timing of death versus timing of last appreciable pregnancy and to characterize the details of these deaths.


      During the study period, our hospital was aware of 23 maternal deaths. In contrast, our mixed methods cohort extraction identified 18 additional patients—a 78% increase—of whom a greater proportion represented postpartum deaths (Figure 1). The majority of newly ascertained mortalities were related to cardiac causes or other medical comorbidities (Figure 2). While many hospital ascertained cases were associated with deaths following delivery of a living infant, more deaths following early pregnancy loss or termination were identified through our novel technique.