Answers are grounded in this paper and cite their sources.
Introduction Contraceptive continuation is central to preventing unintended pregnancy. Although method-specific discontinuation, switching, side effects and fertility intentions are well described in earlier DHS research, contemporary episode-level evidence for adolescent girls and young women in Malawi and Zambia is limited. We estimated 12-month contraceptive method discontinuation, defined as ending a method episode whether by switching to another method or stopping use altogether, and interpreted the findings as a contemporary, confirmatory comparison rather than as evidence of novel causal relationships. Methods We analysed 2024 DHS-8 reproductive calendar data from Malawi and Zambia using contraceptive-use episodes as the unit of analysis. Weighted episode-level survival methods estimated 12-month discontinuation, with 500 bootstrap replicates resampling PSUs within DHS strata for the country contrast. Adjusted associations were examined using weighted generalized estimating equations (GEE) with woman-level clustering, while gradient boosting and elastic-net logistic regression were used as confirmatory predictive analyses with woman-grouped five-fold cross-validation. The calendars overlap the pre-COVID-19, pandemic and later periods, but no causal pandemic effect was estimated. Results Twelve-month contraceptive method discontinuation, including switching and transitions to non-use, among women aged 15–24 years was 48.7% in Malawi ( n = 4,858 segments) and 55.5% in Zambia ( n = 2,695 segments), a difference of 6.8 percentage points (95% CI 4.2–9.5). Stated reasons were dominated by pregnancy-related reasons, method switching, and side effects; reported cost or supply problems were uncommon among episodes with a recorded reason. Method type was the most consistent measured predictor. Predictive discrimination was moderate: gradient boosting achieved woman-grouped cross-validated AUCs of 0.714 in Malawi and 0.677 in Zambia, compared with 0.709 and 0.674 for elastic-net logistic regression, respectively. Conclusion Twelve-month contraceptive method discontinuation is common among adolescent girls and young women in both settings. These estimates include method switching and therefore should not be interpreted as rates of contraceptive abandonment or transition to non-use. The results mainly confirm established method and reproductive-pattern gradients using the most recent DHS-8 data, while providing a current Malawi-Zambia comparison. The findings support informed method choice, side-effect management and continuity of care, but causal claims about method effects, service quality, access or COVID-19 cannot be made from these retrospective calendar data.
No takes yet. Share an insight, caveat, or question.
Ng’ambi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: