Content area
Abstract
Previous research on medical abortion has followed the conceptual model for surgical abortion; efficacy is determined by calculating the proportion of successful abortions. This method is appropriate for surgical abortion because surgical abortion is a discrete event. Unlike surgical abortion, medical abortion requires multiple steps to occur over time and some women will not complete the process. Because the experience of these women cannot be handled appropriately by simple division, all previously published efficacy rates are incorrect. To calculate unbiased efficacy rates, life tables must be used.
This analysis uses multi-decrement life tables to calculate the efficacy of a mifepristone/misoprostol medical abortion regimen using data from more than 6,500 women from six previously published medical abortion studies conducted in China, Cuba, France, India, the United States, and Vietnam. A successful medical abortion is defined as one in which the pregnancy is terminated without the need for surgery. By using life tables, the important element of time in successful medical abortion can be captured. More importantly, the life table allows people who are lost to follow-up to contribute what is known of their experience to the analysis.
Using life tables, both perfect use (method failure) and typical use efficacy rates were calculated, revealing a low level of non-compliance in medical abortion. A two-dose misoprostol protocol was shown to significantly improve efficacy for women with a gestational age of 50–63 days; there was no significant difference in efficacy for women with a gestational age of 49 days or less.
Using hazard models, the impact of multiple characteristics on efficacy was evaluated. Women with lower gestational ages, women under 23 years of age, women with more than 12 years of education, and women with no previous induced abortion experience were more likely to experience a successful medical abortion. The hazard analysis showed that differences in efficacy by study still exist, even after controlling for study composition.
