Additionally, some of the point estimates in our model may
have broader confidence intervals and be less accurate than
others, as our input data were taken from numerous sources.
However, we endeavored to examine the robustness of the
model by varying such estimates broadly in the sensitivity
analyses. Interestingly, we found that small variations in our
assumptions regarding the probability of maternal mortality in
cesarean hysterectomy and probability of maternal mortality
in cesarean delivery affected maternal mortality most strongly,
followed by variations in the probability of cesarean hysterectomy
in the setting of placenta accreta. However, while the
estimates of complications and mortality varied, they consistently
rose throughout the range of the assumptions in the
rising cesarean rate models and fell in the falling cesarean rate
model. This solidifies our finding that multiple repeat
cesareans – the most common cause of cesarean hysterectomies
– would be an important driver of increasing maternal
mortality in the future if current trends continue.