Clinical and service delivery outcomes following medication abortion provided with or without pretreatment ultrasound or pelvic examination: An updated comparative analysis.

Objectives: This study aimed to compare medication abortion outcomes among people screened without or with ultrasound or pelvic examination.

Study design: We used data collected from March 24, 2020, to September 27, 2021, at five TelAbortion Project sites that provided medication abortion with mifepristone and misoprostol by mail. Using logistic regression weighted on propensity scores, we compared outcomes in participants who had neither ultrasound nor examination before treatment (No-Test group) or had such tests (Test group). We analyzed outcomes separately for participants screened early in the analysis period (before September 15, 2020) or later. Outcomes included procedural abortion completion or ongoing pregnancy, serious adverse events, and unplanned posttreatment abortion-related clinical visits.

Results: Among 416 participants in the early period, the No-Test group had a significantly higher risk than the Test group of procedural abortion completion or ongoing pregnancy (5.6% vs 0.9%, risk difference 4.6%, 95% CI 1.5%, 7.7%) and abortion-related clinical visits (13.3% vs 6.3%; risk difference 7.0%; 95% CI 1.1%, 12.8%). Among 364 participants screened later, the risk of procedural abortion completion or ongoing pregnancy did not differ by group, while unplanned abortion-related clinical visits were less common in the No-Test group (9.9% vs 20.5%; risk difference -10.6%; 95% CI -20.1%, -1.1%). The risk of serious adverse events did not differ by group in either period.

Conclusions: When providers first began omitting ultrasound or pelvic examination before medication abortion, the practice was associated with increased risks of failure of complete abortion and posttreatment clinical visits; however, these increased risks resolved over time.

Implications: Medication abortion without pretreatment ultrasound or examination is effective and safe. This model should be routinely offered to eligible patients.