Regional variation in late preterm births in North Carolina.

Late preterm (LPT) neonates (34 0/7th-36 6/7th weeks' gestation) account for 70% of all premature births in the United States. LPT neonates have a higher morbidity and mortality risk than term neonates. LPT birth rates vary across geographic regions. Unwarranted variation is variation in medical care that cannot be explained by sociodemographic or medical risk factors; it represents differences in health system performance, including provider practice variation. The purpose of this study is to identify regional variation in LPT births in North Carolina that cannot be explained by sociodemographic or medical/obstetric risk factors. We searched the NC State Center for Health Statistics linked birth-death certificate database for all singleton term and LPT neonates born between 1999 and 2006. We used multivariable logistic regression analysis to control for socio-demographic and medical/obstetric risk factors. The main outcome was the percent of LPT birth in each of the six perinatal regions in North Carolina. We identified 884,304 neonates; 66,218 (7.5%) were LPT. After multivariable logistic regression, regions 2 (7.0%) and 6 (6.6%) had the highest adjusted percent of LPT birth. Analysis of a statewide birth cohort demonstrates regional variation in the incidence of LPT births among NC's perinatal regions after adjustment for sociodemographic and medical risk factors. We speculate that provider practice variation might explain some of the remaining difference. This is an area where policy changes and quality improvement efforts can help reduce variation, and potentially decrease LPT births.