ObjectivesTo assess whether planning high-sensitivity urine pregnancy tests (HSPT) rather than facility-based tests for medication abortion follow-up may increase risk of unplanned clinical visits or procedural completion of the abortion. Study designWe used data from the TelAbortion Project, a 5-year study assessing the safety and feasibility of providing mifepristone and misoprostol by telemedicine and mail in the United States. We categorized participants by whether the pretreatment follow-up plan included HSPT at home 3–5 weeks after treatment or facility-based tests (ultrasound or serum human chorionic gonadotropin) within 2 weeks after treatment. We used multivariable logistic regression to compare likelihood of post-treatment unplanned, abortion-related clinical visits and procedural intervention in these groups. ResultsOf 1324 patients who planned HSPT follow-up and 576 who planned facility-based tests, 85% and 83%, respectively, provided outcome information. Post-treatment clinical visits were less frequent in the HSPT group (19%) than in the facility-based test group (79%); most of the latter were to obtain the planned test. However, unplanned, abortion-related visits were significantly more common in the HSPT group (adjusted risk difference: 6.5%; p < 0.01). The likelihood of procedural completion did not differ by group. Planned follow-up test was not associated with delay in procedural completion or detection of ongoing pregnancy. ConclusionsFollow-up of medication abortion with home HSPT was associated with fewer post-treatment clinical visits, modestly more unplanned, abortion-related clinical visits, and no increase in the risk of procedural interventions or delayed identification or management of treatment failures. This option is an appropriate follow-up approach after medication abortion. ImplicationsUse of home high-sensitivity pregnancy tests rather than facility-based tests for outcome assessment after medication abortion is associated with a modest increase in unplanned clinical visits but does not lead to an increase in procedural interventions or delays identification and management of treatment failure.