An evaluation of seasonal maternal–neonatal morbidity related to trainee cycles

Abstract

The existence of the "July phenomenon" (worse outcomes related to the presence of new physician trainees in teaching hospitals) has been debated in the literature and media. Previous studies of the phenomenon in obstetrics are limited by the quality and detail of data. To evaluate whether the months of June to August, when transitions in trainees occur, are associated with increased maternal and neonatal morbidity. Secondary analysis of an observational cohort of 115,502 mother-infant pairs that delivered at 25 hospitals from March 2008 to February 2011. Inclusion criteria were an individual who had a singleton, nonanomalous live fetus at the onset of labor, and delivered at a hospital with trainees. The primary outcomes were composites of maternal and neonatal morbidity. We evaluated the outcomes by academic quarter during which the delivery occurred, beginning July 1, and by duration of the academic year as a continuous variable. To account for clustering in outcomes at a given delivery location, we applied hierarchical logistic regression with adjustment for hospital as a random effect. Of 115,502 deliveries, 99,929 met the inclusion criteria. Race and ethnicity, insurance, body mass index, drug use, and the availability of 24/7 maternal-fetal medicine, anesthesia, and neonatology varied by quarter. In adjusted analysis, the frequency of the composite maternal and neonatal morbidity did not differ by quarter. No differences in composite morbidity were observed when using day of the year as a continuous variable (maternal morbidity adjusted odds ratio, 1.00; 95% confidence interval, 0.99-1.00 and neonatal morbidity adjusted odds ratio, 1.00; 95% confidence interval, 1.00-1.01) and after adjustment for hospital as a random effect. Odds of major surgical complications in quarter 2 were twice those in quarter 1. Neonatal injury and intensive care unit were less frequent in later quarters. Maternal and neonatal morbidity in teaching hospitals was not associated with the academic quarter during which delivery occurred, and there was no evidence of a "July phenomenon".