We will decompose variation in Cesarean section (C-section) rates across geographic areas and over time, using confidential US vital statistics data, to determine whether obesity is capable of explaining the large unexplained variation in C-Sections. Studies to date have assumed that non-clinical factors are behind the unexplained area variation. However, if there is an important omitted clinical factor in the existing models, then that will alter the perception that current C-section rates are unwarranted and affect policy goals for their reduction. Maternal obesity is a potentially important clinical factor that has bee omitted from existing studies. Even if obesity explains a sizable portion of the variance in C-sections, it does not necessarily follow that those C-sections are medically necessary. We will therefore investigate the extent to which obese patients'higher C- Section rates are the result of clinical necessity as opposed to physicians'reaction to the classification of patients as obese. We will use regression discontinuity estimation around the clinical definitions of obesity and morbid obesity (BMIs of 30 and 40, respectively) to assess the extent to which the formal classification of patients as obese directly affects C-section probabilities. The key identifying assumption is that, all else equal, a woman with a pre-pregnancy BMI of 29.9 should not have a higher C-section probability than one with a BMI of 30.1. Preliminary analysis using data on first-births in Texas finds that there is a significant difference is C- section probabilities across thi narrow difference in weights. If patients with BMIs near 30 are clinically and socio-economically indistinguishable, we can estimate the effect of a C-section for the marginal patient. Estimates of the effect of C-sections on health outcomes and costs will improve treatment decisions and thus the health of this important subpopulation of mothers and infants. The medical literature suggests obese patients may present differently in labor, be harder to monitor and have higher morbidity in C-sections. We will therefore test whether the discontinuities in C-section rates and health outcomes vary with physician experience (both overall and with delivering obese patients). Preliminary analyses on first births in Texas suggest that it does. The full US data wil enable us to more rigorously explore this by formally modeling physician learning. If some of the higher C-section rate of obese patients is due to physicians'unfamiliarity with managing their labors, some of this increase could erode as physicians gain more experience with this population. It would also suggest that physician education as well as sorting of patients to physicians more experienced with their care could be effective mechanisms to reduce C-sections.
This project will study the role of maternal obesity in U.S. Cesarean section deliveries. By determining the extent to which obesity has contributed to the variation in C-section rates, the project will contribute to the discourse in public health on reducing area variations and lowering the C-section rate and will contribute to the literature on managing labors of obese patients.
