Accuracy of Survey-Based Assessment of Eligibility for Medication Abortion Compared to Clinician Assessment

Background: With increasingly restricted access to facility-based abortion in the United States, pregnant people are increasingly relying on models of care that utilize history-based or no-test approaches to eligibility assessment. Minimal research has examined the accuracy of abortion patients' self-assessment of eligibility for medication abortion using their health history, a necessary step towards ensuring optimal access to history-based or no-test models, as well as potential over-the-counter access.

Objective: To examine the accuracy of pregnant people's eligibility for medication abortion determined using their self-reported health history as compared to clinician assessment with ultrasound and other tests.

Study design: In this diagnostic accuracy study, we recruited people seeking medication or procedural abortion from 9 abortion facilities, aged > 15, English- or Spanish-speaking, and with no prior ultrasound at recruitment facility. Before ultrasound, we surveyed participants on medication abortion eligibility, including estimated gestational duration, medical history, contraindications, and ectopic risk factors such as pain and bleeding symptoms. We compared patients' eligibility based on self-reported history to subsequent clinician assessment, focusing on overall diagnostic accuracy, or area under the receiver operating characteristic curve, sensitivity, specificity, and proportion with discordant patient and clinician eligibility assessment using 77 days as the upper gestational duration limit.

Results: Overall, 22.1% of 1,386 participants were ineligible for medication abortion according to clinician assessment. Overall diagnostic accuracy of self-assessment was acceptable (area under the receiver operating characteristic curve=0.65, 95% CI: 0.63,0.67), with sensitivity of 92.2% (88.6, 94.9%) and specificity of 37.8% (34.9, 40.7%). Very few participants (n=24, 1.7%) screened themselves as eligible when the clinician deemed them ineligible; many more (n=672, 48.5%) screened themselves as ineligible when the clinician deemed them eligible. The most common patient-reported contraindications included unexplained pain (55.5%), gestational duration >77 days (36.5%), and anemia (29.0%). On its own, unexplained pain had poor sensitivity identifying those with clinician-concern for ectopic pregnancy (41.7%, 95% CI: 15.2, 72.3%). Removing unexplained pain as a screening criterion resulted in higher accuracy (0.71, 95% CI: 0.69, 0.74) (p<0.001).

Conclusions: History-based screening protocols are highly effective at ensuring few people receive medication abortion when ineligible. However, a sizeable group screens as ineligible when they are in fact eligible, suggesting a need for more specific screening questions.