Abstract
The US infant and maternal mortality and morbidity rates are far above those for the rest of the
developed world. One potential driver of these poor outcomes is the characteristics of hospitals where
women deliver, especially the annual delivery volume. There is some evidence that the volume of deliveries
affects maternal outcomes (mortality and serious morbidity), but this evidence is not consistent for all
maternal outcomes or across all types of locations and has frequently omitted fetal deaths from the
analyses. Further, there has not been a careful examination of the effect of obstetric volume on joint
maternal-infant outcomes dyad, or how obstetric volume modifies the observed improvements in neonatal
mortality and morbidity at NICUs. Finally, many of these studies have focused only on the volume of
preterm infants, ignoring overall delivery volume, chronic medical conditions or co-existing complications,
and low-risk deliveries. Two smaller European studies have found and association between the volume of
the deliveries of term, low-risk infants and newborn outcomes, without examining maternal outcomes.
That hospital delivery volume could be a contributor to the poor US reproductive outcomes could
have significant policy implications. There could be benefits from some consolidation of obstetric services,
but there are trade-offs between consolidation and access, with no data on either the appropriate
thresholds, or how such thresholds change when routine access to medical care is limited, such as rural
areas. In the 1990s Portugal closed all deliveries services with a volume <1500 deliveries/year and
experienced a decrease in the maternal mortality rate from 9.2 to 5.3/100,000. Portugal simultaneously
closed all small NICUs, so the resulting very large decrease in neonatal (8.1 to 2.9/1000) and perinatal
(16.4 to 6.6/1000) mortality could have resulted from either changes in NICU and obstetric volume. While
such data is compelling, the larger variation in both patient risk and hospitals that deliver infants in the
United States requires innovative studies to inform US and state policy about how organize obstetric care.
The proposed study will address the following specific aims:
Aim 1: What are the relationships between the volume of obstetric services and maternal and infant
morbidity and mortality (including fetal deaths)? Do these effects very by patient risk?
Aim 2: Are there differences in the volume-outcome effects of delivery volume for rural vs. urban areas?
We will use linked vital statistics-patient discharge data from CA, MA, MO, PA, SC and WA for 1995-
2020. We will exploit the panel nature of the data (repeated observations of each hospital over time) to
control for unobserved, hospital-specific factors that affect outcomes. The objective is to identify the delivery
volumes needed to optimize the outcomes for pregnant women and their babies, allowing for evidence-
based policies at the state and national-level to guide the development of perinatal delivery systems.