This randomized trial was performed to examine the clinical efficacy of, patient acceptance of, and provider resources needed for medical and surgical abortion in women with pregnancies up to 49 days' gestation. Women with no pre-treatment preference for method of abortion were randomized to medical abortion with methotrexate and misoprostol (group 1) or surgical abortion under local anesthesia using manual vacuum aspiration (group 2). Women in group 1 received methotrexate 50 mg orally followed 5 to 6 days later by misoprostol 800 microg vaginally; the misoprostol dose was repeated if the abortion did not occur. All subjects returned for a follow-up evaluation 7 and 14 days after the methotrexate or 14 days after the vacuum aspiration. The time spent by clinical staff for all interactions with participants was prospectively recorded. Enrollment of 50 subjects took 24 months; 25 women were randomized to each group. The complete abortion rates by study day 15 were 83% (95% CI 68, 98%) and 96% (95% CI 88, 100%) for groups 1 and 2, respectively. Of the women randomized to a surgical abortion, 92% (95% CI 81, 100%) stated they would choose a surgical for a next abortion, whereas only 63% (95% CI 43, 82%) of women randomized to a medical abortion would choose that option in the future. Overall, surgical abortion requires 0 to 10% more personnel cost than medical abortion using methotrexate and misoprostol. In women who did not have a strong preference between medical and surgical abortion, the side effect profile and patient acceptability was significantly better for surgical abortion compared to medical abortion.