Accuracy of Birth Certificate Data for Classifying Preterm Birth.

Area of Science:

• Obstetrics and Gynecology
• Public Health Surveillance
• Perinatal Epidemiology

Background:

• Accurate classification of preterm birth (PTB) as spontaneous or indicated is crucial for clinical management and research.
• The reliability of using birth certificate data for PTB subtyping remains uncertain.
• This study addresses the need to evaluate the accuracy of birth certificate data for classifying PTB subtypes.

Purpose of the Study:

• To assess the accuracy of classifying spontaneous versus indicated preterm birth using birth certificate data.
• To compare the accuracy of birth certificate classification against a gold standard derived from medical record review.
• To identify patterns of misclassification in birth certificate data.

Main Methods:

• Retrospective cohort study involving 123 preterm birth records from Missouri birth certificates.
• Established a consensus classification of spontaneous versus indicated PTB through multi-provider review of complete medical records.
• Developed and applied an a priori algorithm to classify PTB subtypes based on birth certificate clinical data.

Main Results:

• Classification accuracy using birth certificate data was 66% compared to the consensus review.
• Misclassification predominantly involved labeling indicated PTBs as spontaneous.
• A significant source of error was the incorrect checking of preterm rupture of membranes (≥12 hours) on birth certificates, leading to spontaneous classification despite maternal or fetal indications for delivery.

Conclusions:

• Birth certificate data tends to overestimate spontaneous PTB and underestimate indicated PTB.
• Current birth certificate data limitations hinder accurate population-level surveillance of PTB subtypes.
• Proposed revisions to birth certificate clinical data elements could improve the accuracy of PTB subtyping.