Treatment practices vary widely across hospitals, often with little connection to the medical needs of patients. We assess impacts of these differences in childbirth, where there is broad interest in reducing cesarean deliveries. Using a distance-based design and data from half a million births, we find that infants delivered at hospitals with higher c-section rates are born in better shape, are less likely to be readmitted to the hospital, are exhibit suggestive evidence of improved survival. These benefits are driven by the avoidance of prolonged labors that pose serious risks to infant health. In contrast, we document that these infants are substantially more likely to return to the emergency department for respiratory-related problems in the year after birth, providing some of the first design-based evidence consistent with a large observational literature linking cesarean delivery to chronic reductions in respiratory health.

Motivated by wide cross-sectional variations in intensity of care that are unrelated to quality of care, researchers and policymakers commonly claim that healthcare providers waste considerable resources, engaging in so-called “flat-of-the-curve” medicine. A key yet elusive prediction of this hypothesis is that providers ought to be able to cut back on care without sacrificing quality. This article examines the effects of a particular form of provider cutbacks – those generated by physicians working in high-pressure peer group environments. Using expansive, time-stamped discharge data from 137 hospital-based emergency departments, I document that physicians systematically alter their pace and intensity of care across frequently shuffled peer groups. Peer groups that induce a physician to work faster also induce her to order fewer tests and spend less money. Contrary to the flat-of-the-curve hypothesis, these cutbacks lead to large reductions in quality of care. This evidence, paired with the fact that slower physicians do not produce better average outcomes, suggests that cross-physician differences in resource utilization reflect substantial differences in physician productivity within a hospital.

Cesarean delivery for low-risk pregnancies is generally associated with worse health outcomes for infants and mothers. The interpretation of this correlation, however, is confounded by potential selectivity in the choice of birth mode. We use birth records from California, merged with hospital and emergency department (ED) visits for infants and mothers in the year after birth, to study the causal health effects of cesarean delivery for low-risk first births. Building on McClellan, McNeil, and Newhouse (1994), we use the relative distance from a mother's home to hospitals with high and low c-section rates as an instrument for c-section. We show that relative distance is a strong predictor of c-section but is orthogonal to many observed risk factors, including birth weight and indicators of prenatal care. Our IV estimates imply that cesarean delivery causes a relatively large increase in ED visits of the infant, mainly due to acute respiratory conditions. We find no significant effects on mothers' hospitalizations or ED use after birth, or on subsequent fertility, but we find a ripple effect on second birth outcomes arising from the high likelihood of repeat c-section. Offsetting these morbidity effects, we find that delivery at a high c-section hospital leads to a significant reduction in infant mortality, driven by lower death rates for newborns with high rates of predetermined risk factors.

No abstract