Predicting date of birth: the best time to date a pregnancy?

Abstract

Objective: To compare the estimated date of birth calculations from last menstrual period (LMP) and ultrasounds at varying gestations (<70, 70-106, 110-140, 141-196 and 200-276) against the actual date of birth (DOB). Methods: This cohort study in a single local health district, Australia included 18,708 women with spontaneous labor who gave birth to a single live born infant without major anomalies between 2007 and 2011. Data were sourced from a computerized population birth database. The outcome of interest was duration of pregnancy expressed as total days, and the difference between actual DOB and estimated date of birth by dating method. Results: Only 5% of births occurred on the estimated date of birth regardless of the timing of the estimate. Approximately 66% of births occurred +/-7 days of the estimated date of birth, and there was little difference between ultrasound gestational week bands. The 110-140 weeks of gestation ultrasound performed as well if not better than ultrasounds conducted at other gestations. Maternal factors such as ethnicity and smoking status during pregnancy influenced duration of pregnancy; however, their explanatory power was too low to support incorporating these characteristics in dating estimations. Conclusion: An early dating scan (10 weeks or earlier) is unnecessary if LMP is reliable.