Harmonizing Methods for Estimating the Impact of Contraceptive Use on Unintended Pregnancy, Abortion, and Maternal Health

Abstract

Estimates of the potential impacts of contraceptive use on averting unintended pregnancies, total and unsafe abortions, maternal deaths, and newborn, infant, and child deaths provide evidence of the value of investments in family planning programs and thus are critically important for policy makers, donors, and advocates alike. Several research teams have independently developed mathematical models that estimate the number of adverse health outcomes averted due to contraceptive use. However, each modeling approach was designed for different purposes, and as such the methodological assumptions, data inputs, and mathematical algorithms initially used in each model differed; consequently, the models did not produce comparable estimates for the same outcome indicators. To address this, a series of expert group meetings took place in which 5 models-Adding it Up, Impact 2, ImpactNow, Reality Check, and FamPlan/Lives Saved Tool (LiST)-were reviewed and harmonized where possible. The group identified the main reasons for the inconsistencies in the estimates generated by the models for each of the adverse health outcome indicators. The group then worked together to align the methodologies for estimating numbers of unintended pregnancies, abortions, and maternal deaths averted due to contraceptive use, and reviewed the challenges with estimating the impact of contraceptive use on newborn, infant, and child deaths, including the lack of a conceptually clear pathway and rigorous evidence. The assumption that most influenced harmonization was the comparison pregnancy rate used by the models to estimate the counterfactual scenario-that is, if women who are currently using contraception were not using a method, how many would become pregnant? All the models now base this on the number of unintended pregnancies among women with unmet contraceptive need, bringing the estimates for unintended pregnancies, total and unsafe abortion, and maternal deaths much closer together. The agreed approaches have already been adopted by the Family Planning 2020 (FP2020) initiative and Track20, a project that supports FP2020. The experts will continue to update their models collaboratively to ensure that the most current estimation methodologies and data available are used. Valid and reliable methodologies for estimating these impacts from family planning are critically important, not only for advocacy to sustain resource allocation commitments but also to enable measurement and tracking of global development indicators. Conflicting estimates can be counterproductive to generating support for family planning programs, and this harmonization process has created a more unified voice for quantifying the benefits of family planning.

FIGURE 1.

Different Estimates of the Health Impact of Contraceptive Use in Malawi From 5 Models, Before Harmonizing the Modeling Approaches

FIGURE 2.

More Uniform Estimates of the Health Impact of Contraceptive Use in Malawi From 5 Models, After Harmonizing the Modeling Approaches