Accuracy of Menstrual History for Determining Gestational Age Among Adolescents Who Underwent Abortion in England and Wales.

Abstract

To assess the accuracy of using the last menstrual period (LMP) date for determining gestational age, compared with ultrasonography, among patients aged younger than 16 years at British Pregnancy Advisory Service (BPAS), a large national abortion service in England and Wales. This evaluation was undertaken in the context of national discussion and guidance about adolescent no-test medication abortion access. We analyzed routinely collected gestational age data for all patients younger than 16 (herein referred to as adolescents) who underwent abortion at BPAS between May 1, 2021, and April 30, 2024. We compared gestational age from ultrasonography (reference) to that from LMP date, overall and by eligibility for no-test medication abortion (where ineligibility was indicated by an unknown gestational age or gestation that exceeded 69 days). We calculated diagnostic measures of accuracy that included sensitivity, specificity, and proportion of false-negative results. For a subsample, we incorporated additional menstrual history indicators into a composite screening variable and recalculated measures. During the evaluation period, 1,555 adolescents underwent abortion; 1,282 (82.4%) provided an LMP date. On average, LMP dating overestimated gestational age by 3 days compared with ultrasonography. When using the LMP date alone, the sensitivity for correctly detecting ineligible patients was 86.6% (95% CI, 82.5-90.1) and the specificity for correctly identifying eligible patients was 78.8% (95% CI, 76.4-81.1). The proportion of false-negative results (patients incorrectly classified as eligible) was 2.8% (95% CI, 2.0-3.7). With the composite measure, sensitivity rose to 93.9% (95% CI, 90.8-96.2, P <.05) but specificity declined to 54.2% (95% CI, 51.4-57.0). The false-negative proportion reduced to 1.3% (95% CI, 0.7-1.9), including 0.5% with a gestational age beyond 83 days (8/1,555) on ultrasonography. Menstrual dating provides high sensitivity for identifying adolescents as ineligible for no-test medication abortion, with very low risk of clinically important misclassification. Adding menstrual history indicators improves sensitivity but indicates unnecessary scans for many eligible individuals. Policy must balance the rare risk of misestimation against costs of restricting adolescents' access, where disproportionate care barriers already exist.