Three methods of estimating births averted nationally by contraception.

This paper compares the performance of three methods applicable to national-level demographic data of estimating births averted as a consequence of contraceptive practice. Two are based on the relationship between the general fertility rate (GFR) or total fertility (TFR) and contraceptive prevalence, while the third uses Bongaarts' proximate determinants (PD) model. Estimates of the number of births averted and the percentage by which the number would have increased in the absence of contraception are consistent between the GFR-based and TFR-based methods, but in general lower than the estimates generated by the PD-based method, except for a few high-contraceptive-prevalence countries. For 156 countries and areas around the world the estimated number of births that would have occurred in a recent year in the absence of contraception--the average of the estimates of the three methods--is approximately 230 million, which is more than the estimated 129 million births that actually occurred.