Emergency contraception (EC) prevents pregnancy, and it has been suggested that widespread use could reduce abortion rates; however, the use is limited. Providing EC in advance of need increases use, but there is no direct evidence that it reduces unintended pregnancy. In a randomized controlled trial of 2000 women after childbirth in Shanghai, all women not wishing to use hormonal contraception or an intrauterine device (IUD) were given a supply of condoms. Those in the intervention group also received three courses of mifepristone 10 mg with instructions for use as EC. Follow-up was by telephone at 16, 32 and 52 weeks. Over 88% of women in both groups completed 1 year of follow-up. Women with a supply of EC were more than twice as likely to use it, and to use it more than once (p<.001 for both) than women without a supply. There was no difference in pregnancy rates at 1 year (38/832 vs. 32/817). EC was not used in 89% of conception cycles, as women did not recognize the need for it. Increased use of EC may not reduce abortion rates.