Presentation (CTI)

OBJECTIVES: With the increasing use of real-world data in studies involving pregnant populations, accurate identification of pregnancy dates and gestational age has become critical. In Taiwan, the Birth Reporting Database (BRD) captures nationwide records of live births and stillbirths (birth weight ≥500 g or gestational age ≥20 weeks), with gestational age at delivery enabling derivation of pregnancy start date. However, pregnancy losses before 20 weeks are not captured, limiting gestational information for these outcomes. As an alternative, government-funded prenatal care visits are scheduled at designated gestational weeks with corresponding codes, allowing estimation of the pregnancy start date. This study evaluated whether prenatal visit timing could serve as a proxy for inferring pregnancy start date.
METHODS: Pregnancies in 2023 were identified from the BRD and linked to the National Health Insurance Research Database (NHIRD), covering 99% of visits, to obtain prenatal care records. A one-year inclusion period was applied to minimize multiple pregnancies. The difference in pregnancy start date derived from BRD gestational age versus that estimated from the first prenatal visit was assessed using descriptive statistics.
RESULTS: A total of 136,436 pregnancies were identified; 99.3% resulted in live births and 96.7% had at least one prenatal visit. The majority of first visits were coded at Week 8 (64.3%) or Week 12 (31.0%), consistent with the recommended schedule. The median difference in pregnancy start date between the two sources was 4 days (IQR: −7 to 16). Differences were smallest when the first visit was coded at Week 8 (median: 0 days) and increased with later first visits, reaching 18 days at Weeks 12 and 16, but decreased again at Week 20 (median: −3 days).
CONCLUSIONS: Early prenatal visit dates can provide a reasonable approximation of pregnancy start date and serve as a practical alternative for pregnancy losses not captured in the BRD in Taiwan.