Volume 44 - Article 37 | Pages 879–898

Has contraceptive use at pregnancy an effect on the odds of spontaneous termination and induced abortion? Evidence from Demographic and Health Surveys

By David Antonio Sánchez-Páez, José Antonio Ortega

Print this page  Facebook  Twitter


Date received:26 Sep 2020
Date published:21 Apr 2021
Word count:4670
Keywords:contraceptive calendar, contraceptive use, induced abortion, pregnancy termination, reproductive health


Background: Contraceptive failure increases the chances of pregnancy termination, including both induced abortions and spontaneous terminations. Proper separation requires accounting for competing risks of pregnancy outcomes.

Objective: To measure the differential risk of spontaneous termination and induced abortion according to contraceptive use prior to pregnancy based on pooled Demographic and Health Survey calendar data.

Methods: We use multinomial logistic models controlling for demographic and socioeconomic variables to estimate the differential risk of spontaneous termination and induced abortion according to contraceptive use at the time of pregnancy. We address data limitations including recall error, omission error, and possible misclassification of outcomes.

Results: We find higher risk of induced abortion (RRR = 7.18, CI = 6.38–8.09) and spontaneous termination (RRR = 1.38, CI = 1.13–1.69) after contraceptive failure, with stronger effect for women under 30. Parity, union status, education, and wealth have a strong effect on induced abortion. Regarding spontaneous termination, age mainly explains the increased risk.

Conclusions: Since pregnancies following contraceptive failure are less likely to end in a live birth, aggregate models of the impact of family planning should reflect that contraceptive use and induced abortion conform interdependent strategies and that spontaneous termination is a competing risk of induced abortion.

Contribution: This is the first study reporting differences in the risk of spontaneous termination and induced abortion according to contraceptive use prior to pregnancy. We account for competing risks using a multinomial logit model of pregnancy outcomes conditional on pregnancy, new in the literature. Data limitations are addressed in novel ways.

Author's Affiliation

David Antonio Sánchez-Páez - Universidad de Salamanca, Spain [Email]
José Antonio Ortega - Universidad de Salamanca, Spain [Email]

Other articles by the same author/authors in Demographic Research

» Adolescent contraceptive use and its effects on fertility
Volume 38 - Article 45

» Tempo-Adjusted Period Parity Progression Measures: Assessing the Implications of Delayed Childbearing for Cohort Fertility in Sweden, the Netherlands and Spain
Volume 6 - Article 7

» Tempo-Adjusted Period Parity Progression Measures, Fertility Postponement and Completed Cohort Fertility
Volume 6 - Article 6

Most recent similar articles in Demographic Research

» An investigation of Jordan’s fertility stall and resumed decline: The role of proximate determinants
Volume 45 - Article 19    | Keywords: contraceptive use

» Low perceived susceptibility to pregnancy as a reason for contraceptive nonuse among women with unintended births
Volume 44 - Article 31    | Keywords: contraceptive use

» Women's employment and fertility in a global perspective (1960–2015)
Volume 43 - Article 25    | Keywords: reproductive health

» Young women's joint relationship, sex, and contraceptive trajectories: Evidence from the United States
Volume 42 - Article 34    | Keywords: contraceptive use

» Intimate partner violence and contraceptive use in developing countries: How does the relationship depend on context?
Volume 42 - Article 10    | Keywords: contraceptive use