Using induced abortion to measure contraceptive efficacy

Abstract

Data from a 1989-1990 case-control study of contraceptive efficacy in Norway compare contraceptive use among women who requested an abortion (1,386 cases) with women who responded to a general fertility survey (2,627 controls). In a logistic regression analysis measuring contraceptive efficacy as the odds of avoiding a pregnancy that terminated in an induced abortion compared with the odds for nonuse, consistent condom use was found to lower fecundity by 88.9%, diaphragm use by 89.3%, the pill by 97.8%, the IUD by 97.6%, vasectomy by 99.5%, and female sterilization by 99.8%. The relative contraceptive efficacy of the condom, the IUD and the pill did not vary by marital status or parity but did vary with age; the proportion by which each of these methods reduced fecundity declined among successively older age-groups.

PIP: A study in Norway compared the contraceptive experience of 1386 cases who were born in 1942-68 and who terminated a pregnancy in 1989 or 1990 with that of 2627 controls who had participated in the 1988 Norwegian fertility study and met certain criteria. The outcome measure of this comparison was the relative reduction in fecundity provided by the use of a specific contraceptive method as compared with nonuse. It was found that women who had abortions were more likely to be young, single, and be childless or have only 1 child. 57% of these women were not using contraception at the time of conception (as compared to 5% of the general population in the month preceding the survey). Compared with nonuse, consistent use of the diaphragm or condom reduced fecundity by 89%. Use of oral contraceptives or the IUD led to a 98% reduction. This reduction, and that found with condoms, decreased with age. Female sterilization led to a 99.8% reduction and vasectomy to a 99.5%. Efficacy rates which depend upon abortion as a proxy for unintended pregnancy, however, ignore the fact that some women carry unintended pregnancies to term. Adjusting the data for this fact revealed that any bias would be toward underestimating reductions in fecundity. Norway's abortion rate would not be affected by a shift from less effective to more effective methods. Since 60% of all abortions occur among the 5% of the population who do not use contraception, however, the abortion rate would be affected substantially if this 5% were to use any contraception, even a less effective method. Improved contraceptive prevalence would also lead to a reduction in births among those members of the 5% who do not choose abortion.