Incidence of Ectopic Pregnancy and Diagnostic Accuracy of Ectopic Pregnancy Screening Criteria Among People Seeking Versus Not Seeking Abortion: A Retrospective Case-Control Study.

Abstract

Objectives: We estimated the incidence of ectopic pregnancy, associated risk factors, and accuracy of three history-based screening criteria for medication abortion (history of ectopic pregnancy, tubal surgery, and intrauterine device [IUD] in place) in detecting ectopic pregnancies among people seeking versus not seeking abortion.

Study design: We conducted a retrospective case-control study by reviewing electronic health records of a random sample of 2,201 ectopic and 1,153 intrauterine pregnancies of people enrolled at Kaiser Permanente Northern California (2016-2021). After classifying pregnancies as those to people seeking or not seeking abortion, we used logistic regression weighted to the total sample of pregnancies (N = 385,081) to estimate the incidence of ectopic pregnancies, associated risk factors, and accuracy of three history-based screening criteria.

Results: The adjusted incidence of ectopic pregnancy was significantly lower for people seeking abortion (.39%; 95% confidence interval [CI] [.30%, .47%]) than for people not seeking abortion (1.74%; 95% CI [1.55%, 1.94%]). People seeking abortion received ectopic pregnancy treatment earlier in pregnancy than those not seeking abortion (53 vs. 55 days gestation, p = .04). Among people seeking abortion, combined history of IUD use, ectopic pregnancy, and tubal surgery had low sensitivity (17.6%; 95% CI [12.4%, 23.8%]), high specificity (93.5%; 95% CI [90.2%, 96.0%]), and a low area under the receiver operating characteristic curve value (55.5%; 95% CI [52.5%, 58.6%]).

Conclusions: The incidence of ectopic pregnancy among people seeking abortion is extremely low and lower than among people not seeking abortion. The three history-based screening criteria we evaluated had very low sensitivity and performed poorly in accurately identifying ectopic pregnancies, suggesting they may have limited clinical utility and may create barriers to access to no-test abortion. Further research is needed to refine protocols to better identify, evaluate, and monitor people for possible ectopic pregnancies in the context of no-test provision of medication abortion.