Objective To compare the psychological impact, acceptability and clinical effectiveness of medical versus surgical termination of pregnancy (TOP) at 13-20 weeks of gestation.Design Randomised trial.Setting Large UK tertiary centre.Sample Women accepted for TOP at 13-20 weeks of gestation.Methods Medical TOP (MTOP) using mifepristone and misoprostol or surgical TOP (STOP) by vacuum aspiration at <15 weeks of gestation, and by dilatation and evacuation at 15 or more weeks of gestation.Main outcome measures Distress 2 weeks after TOP, measured by the impact of events scale (IES), and acceptability, measured by the proportion of women who would opt for the same procedure again.Results One hundred and twenty two women were randomised: 60 to the MTOP group and 62 to the STOP group. Twelve women opted to continue their pregnancy. Follow-up rates were low (n = 66/110; 60%). At 2 weeks post-procedure there was no difference in total IES score between groups. However, compared with women undergoing STOP, women undergoing MTOP had a higher score on the IES intrusion subscale (mean difference 6.6; 95% CI 1.4-11.8), and a higher score on the general health questionnaire (GHQ) (P = 0.033). Women found STOP more acceptable: compared with MTOP, more women would opt for the same procedure again (100% versus 53%, P £ 0.001), and fewer women found the experience to be worse than expected (0% versus 53%, P = 0.001). Women who had MTOP experienced more bleeding (P = 0.003), more pain on the day of the procedure (P = 0.008), and more days of pain (P = 0.020). Of the 107 women who declined to participate, 58 (67%) preferred a STOP.Conclusions Randomised trials of women requesting midtrimester TOP are challenging. Women found STOP less painful and more acceptable than MTOP.