Margulis AV, Calingaert B, Kawai AT, Rivero-Ferrer E, Anthony MS. Distribution of gestational age by maternal and infant characteristics in US birth certificate data: toward informing gestational age assumptions when clinical estimates are not available. Poster to be given at the 2022 ICPE Conference; August 26, 2022. Copenhagen, Denmark.


BACKGROUND: When data sources do not have clinical or obstetric estimates for gestational age at birth (GAB) or ICD-10-CM Z3A codes for gestational age, researchers often use a fixed duration of pregnancy based on published mode or median GABs, such as 39 weeks (w) for all live births. GAB estimates for subgroups of live births can be derived using the CDC´s US birth data files, which report clinically based GAB and many data elements from birth certificates.

OBJECTIVES: To describe the distribution of GAB stratified by plurality (e.g., twin) and within plurality by other relevant subgroups.

METHODS: We used CDC´s US birth data for 2019 (the most recent pre-pandemic year) and 2020 (the most recent available year). The study population included all live births with nonmissing GAB. We stratified data based on plurality and other variables. For each stratum, we estimated mean, standard deviation, median, mode, and percentage born at each gestational week. We present results overall and for singletons and twins for 2020. Results for 2019 were similar.

RESULTS: In 2020, 3,619,826 newborns were issued birth certificates; 3,617,213 (99.9%) had nonmissing GAB. Among singletons (3,501,693; 96.8%), the median and mode GAB were 39 w overall and in most strata. Births with smaller median or mode GAB were from women with eclampsia (0.3% of 3,617,213 live births; median and mode, 37 w), women receiving intensive care (0.2%; median, 37; mode, 39 w), newborns receiving intensive care (8.0%; median, 37; mode, 39 w), low birth weight (LBW, 6.5%; median, 35; mode, 37 w), very LBW (1.0%; median and mode, 28 w), extremely LBW (0.5%; median and mode, 25 w), and newborns not discharged alive (0.2%; median, 23; mode, 22 w). Among twins (112,633, 3.1%), the median was 36, and the mode was 37 w overall and for most subgroups. Additional noteworthy subgroups were women with 7-8 prenatal visits (0.4% of 3,617,213 live births; median, 35; mode, 36 w), with 0-6 prenatal visits (0.4%; median, 34; mode, 36 w), who smoked in pregnancy (0.2%; median and mode, 36 w), or aged 15-19 years (0.1%; median; 35; mode, 37 w).

GAB distributions were narrow in pregnancies with healthy characteristics and wider in noted subgroups (e.g., 77.8% of singletons of women with body mass index 18.5-25 kg/m2 had GAB within 1 w of the mode compared with 35.9% of pregnancies in which newborns received intensive care). Means showed more variation than medians and modes.

CONCLUSIONS: Some pregnancy subgroups had distinct GAB distributions. This information can be useful in estimating GAB when clinical estimates are not available in data sources.

Share on: