The Unintended Consequences of Reducing Early Elective Deliveries

Early term births (i.e., occurring at 37-38 weeks) are associated with adverse infant health outcomes, compared to full term (39-40 weeks) deliveries. Nonetheless, many patients choose to induce labor or schedule a Caesarian section (C-section) surgery prior to the recommended full term.

In response, 21 states implemented fiscal or administrative policies to reduce early elective deliveries (EED). Little is known about the impact of these policies on EED rates. Notably, the one published study on the topic suggests that the reduction in EEDs is primarily driven by a reduction in early term [labor] inductions (ETIs) only, while C-section rates remain relatively constant. Further, unpublished evidence from Texas suggests that EED policies increased the overall rate of C-sections. Considering C-sections are costlier and more dangerous than vaginal deliveries, uncovering the mechanisms underlying this effect is important.

Previous work suggests the physician induced demand (PID) hypothesis as an explanation for this result: physicians may only be willing to change their early-term behavior for the less lucrative labor inductions. In addition, physicians may make up for lost income in the early-term phase by increasing rates of C-sections in the post-term phase. We posit a different reason for the unintended consequence of rising overall C-sections in response to EED policies: the longer babies are in the womb, the larger they become. Rising birthweights – though considered a positive health outcome – may contribute to higher C-section rates.

This paper adds to the nascent literature on EED policies by examining their impact in a new state: South Carolina (SC). We calculate the relative contribution of ETIs versus C-sections in reducing EEDs. We use birthweight data to determine how much of the change in C-section rate is attributable to larger babies versus other drivers, such as PID.

We use national vital statistics data (2006-2014), which contain almost all US births. We use gestational age, birth method, and other variables to determine EEDs. We use difference-in-differences OLS models, comparing outcomes in SC before and after the EED policy implementation, versus states with no policy change. Models controlled for parental characteristics; year, month, and state fixed effects; and state linear time trends. Standard errors were clustered at the state level.

We find an EED policy in SC reduced the probability of having an EED by 2.5 percentage points (p<.001), an effect driven entirely by a 3.1 percentage point reduction (p<.01) in probability of having an ETI. We find no effect of the policy on early elective C-section rate. We find a statistically significant increase in the overall probability of having a C-section. When examining birthweight data, it appears that full-term C-sections increased across all birthweight categories (high, medium, low), not just among heavier babies.

Policies intended to reduce all EEDs in SC only reduced early labor inductions; early C-section rates remained unchanged. Further, overall C-section rates increased. Our results suggest this is due to physician attempts to recoup losses from reducing early labor inductions, rather than from rising birth weights accompanying longer gestation times.