Measuring and reporting of the treatment effect of hormonal emergency contraceptives

To derive summary estimates of observed pregnancy rates in women who used the Yuzpe or levonorgestrel emergency contraceptive regimen and identify the various ways in which data related to regimen effectiveness were reported, to discuss the limitations of the effectiveness estimates as they are currently reported, and to propose alternative reporting methods that are less susceptible to misinterpretation.

Analysis of pooled data from 34 experimental or observational studies that had an explicit method of follow-up to systematically evaluate pregnancy outcome after a single course of emergency contraception treatment within a given menstrual cycle.

Women who took either the Yuzpe or levonorgestrel emergency contraceptive regimen to prevent pregnancy.

We searched the MEDLINE and EMBASE databases for clinical studies that systematically evaluated pregnancy outcomes of women who used the Yuzpe or levonorgestrel regimen for emergency contraception. The effectiveness of these regimens is typically reported in relative terms as the reduction of pregnancy risk from a theoretical baseline risk. We reported the effectiveness using absolute risk reductions and numbers needed to treat. We pooled relevant data by using the β-binomial method to derive observed pregnancy rates. The pooled observed pregnancy rates for studies of the Yuzpe and levonorgestrel regimens were 2.0% (95% confidence interval [CI] 1.5-2.5%) and 1.7% (95% CI 1.2-2.2%), respectively. Against expected pregnancy rates of 4-8%, relative emergency contraceptive effectiveness ranges were 50.0-75.0% and 57.5-78.8% for the Yuzpe and levonorgestrel regimens, respectively. Absolute risk reductions were 2.0-6.0% and 2.3-6.3%, respectively. This means that 17-50 women would need to have received the Yuzpe regimen and 16-43 women the levonorgestrel regimen to prevent one pregnancy.

Emergency contraception effectiveness data are susceptible to misinterpretation when data are reported in relative terms without the expected pregnancy rate (i.e., the baseline risk). We recommend that the effect of emergency contraceptives be expressed in absolute terms and accompanied by the expected pregnancy rate used in its estimation.