Measuring true contraceptive efficacy

No investigator has attempted to measure prospectively the true efficacy of a contraceptive method, compared with a control group using no method, because contraceptive trials focus on women trying to avoid pregnancy and ethical concerns do not permit the withholding of contraception. We tested the feasibility of an approach that recruited women who desired pregnancy but were willing to postpone conception by 1 month. In this protocol, we restricted frequency and timing of intercourse to one coital act on the most fertile day of the menstrual cycle, as measured by a luteinizing hormone (LH) detection kit. Participants were randomized to use either a male latex condom, spermicidal film, or no method. In this feasibility study we recruited 58 women at three sites, with one site recruiting 25 women in 5 months. Among 54 women who completed the study, we found a 12% pregnancy rate for the group using no method (2/17; 95% confidence interval [CI], 1-36%) and an 11% pregnancy rate for the group using spermicidal film (2/18; 95% CI, 1-35%). No pregnancies occurred among the 19 women using condoms (0/19; 95% CI, 0-18%). The wide confidence intervals illustrate the small sample size of this pilot study and no conclusions can be drawn about the relative efficacy of the methods. Having demonstrated the feasibility of this study design, we now urge the initiation of a large-scale study to evaluate the efficacy of barrier methods using our randomized approach, with a control arm using no method of contraception.