Medical methods of early abortion differ from surgical methods in that women themselves can potentially administer the regimens. As currently researched and offered, however, the main regimen used for medical abortion, mifepristone-misoprostol, is highly medicalized, involving several clinic visits and extensive physician involvement. We re-examined the role of clinical supervision in each step of the abortion process, using data collected during a large clinical trial of mifepristone-misoprostol abortions in the US, fielded during 1994-1995. The trial was carried out in 17 geographically diverse centers, including private, public, and nongovernmental organization clinics, and enrolled 2121 women, aged 18-45 years, seeking early abortion (< or =63 days since last menstrual period). Women received 600 mg oral mifepristone, followed 48 h later by 400 microg oral misoprostol. Evidence suggests that most women can handle most steps of the medical abortion process themselves, effectively and safely. The utility of clinic visits to ingest mifepristone and misoprostol is questionable. For many women, even the follow-up visit could perhaps be replaced by telephone follow-up, combined with home pregnancy tests. Alternatives to the present protocol might allow greater control, comfort, and convenience at lower cost. Where clinician involvement might be useful, mid-level health care providers typically possess the skills necessary to offer the method safely, implying that physicians might be necessary only as complications arise. Future research useful for determining the optimal amount of medical involvement to provide mifepristone-misoprostol safely and effectively should include self-screening tests, label comprehension tests, calendars to aid in calculating gestational age, and the development of special pregnancy tests with telephone follow-up.