Health Equity/Community Health
Poster Session 3
Ajleeta Sangtani, MD
Fellow
University of Michigan
Ann Arbor, MI, United States
LeAnn A. Louis, MD MPH (she/her/hers)
Resident Physician
University of Michigan Hospital System
Ann Arbor, MI, United States
Ashley Hesson, MD, PhD (she/her/hers)
MFM Fellow
University of Michigan
Ann Arbor, MI, United States
Anita M. Malone, MD, MPH
University of Michigan
Ann Arbor, MI, United States
We conducted a retrospective study using data extracted from the Vital Statistics Natality Birth Data. Individuals were included if they delivered between 2003-2007 (the "early" group) or 2014-2019 (the "late" group), had a prior cesarean delivery, and were cephalic at the time of presentation. Patients delivering between 2008-2013 were excluded to allow for a washout period. Patients were also excluded if their TOLAC status was unknown. Three categories of race were analyzed: non-Hispanic White, Hispanic, and non-Hispanic Black. Statistical analysis was conducted using a Poisson regression and post-hoc testing with estimated marginal means in R.
Results:
2,134,498 births met criteria during the two study periods, 22% of whom were in the early group and 78% were in the late group. Rates of TOLAC increased in all three groups, though the effect size was largest in the Hispanic population (P=0.01). Although intrapartum cesarean rates were not significantly different between race categories in the early group, Black patients were more likely to undergo intrapartum cesareans compared to White patients in the late group (p< 0.01). Likewise, Hispanic patients were more likely to undergo intrapartum cesareans in the late group (p=0.01). Figure 1 demonstrates the trend in intended and actual mode of delivery over time.
Conclusion:
Nationally, intrapartum cesarean rates for Black and Hispanic patients significantly increased after original TOLAC calculator was released without increasing the rates of planned cesarean deliveries or changing the intrapartum cesarean rates in White patients. These findings highlight new, unintended bias associated with tools intended to enhanced clinical decision-making.