Calculating length of gestation from the Society for Assisted Reproductive Technology Clinic Outcome Reporting System (SART CORS) database versus vital records may alter reported rates of prematurity

ObjectiveTo compare length of gestation after assisted reproductive technology (ART) as calculated by three methods from the Society for Assisted Reproductive Technology Clinic Outcome Reporting System (SART CORS) and vital records (birth and fetal death) in the Massachusetts Pregnancy to Early Life Longitudinal Data System (PELL).DesignHistorical cohort study.SettingDatabase linkage analysis.Patient(s)Live or stillborn deliveries.Intervention(s)None.Main Outcome Measure(s)ART deliveries were linked to live birth or fetal death certificates. Length of gestation in 7,171 deliveries from fresh autologous ART cycles (2004–2008) was calculated and compared with that of SART CORS with the use of methods: M1 = outcome date − cycle start date; M2 = outcome date − transfer date + 17 days; and M3 = outcome date − transfer date + 14days + day of transfer. Generalized estimating equation models were used to compare methods.Result(s)Singleton and multiple deliveries were included. Overall prematurity (delivery <37 weeks) varied by method of calculation: M1 29.1%; M2 25.6%; M3 25.2%; and PELL 27.2%. The SART methods, M1–M3, varied from those of PELL by ≥ 3 days in >45% of deliveries and by more than 1 week in >22% of deliveries. Each method differed from each other.Conclusion(s)Estimates of preterm birth in ART vary depending on source of data and method of calculation. Some estimates may overestimate preterm birth rates for ART conceptions.