Intersectional inequalities in demand satisfied by modern contraceptive methods among women in sub-Saharan Africa, 1994 to 2024: a multilevel analysis of individual heterogeneity and discriminatory accuracy
- 1University of Zambia Ridgeway Campus: University of Zambia School of Medicine
Abstract
Background: Demand for family planning satisfied by modern methods is Sustainable Development Goal (SDG) indicator 3.7.1, and countries have committed to universal access to family planning by 2030. Inequalities are usually reported one social dimension at a time. This hides how social characteristics combine and leaves open whether programmes should target social groups or strengthen services for all women. We measured intersectional inequalities in demand satisfied by modern methods among women in sub-Saharan Africa over three decades, assessed how well social position discriminates between women, and identified groups left behind. Methods and findings: We did a preregistered cross-sectional analysis of 139 Demographic and Health Surveys from 38 countries with fieldwork from 1994 to 2024. We included 675,584 women aged 15-49 years with demand for family planning and grouped them into 144 intersectional strata by age, union status, education, household wealth and residence. We fitted multilevel logistic models with random effects for strata and surveys (multilevel analysis of individual heterogeneity and discriminatory accuracy, MAIHDA) and estimated the variance partition coefficient (VPC), the area under the curve (AUC) and the proportional change in variance (PCV). Weighted demand satisfied rose from 27.5% in 1994-99 to 54.1% in 2018-24. Predicted demand satisfied ranged from 13.2% (95% CI: 11.0 - 15.9) among married adolescents with no schooling in the poorest rural households to 68.0% (95% CI: 63.6 - 72.0) among unmarried urban women aged 25-34 years with secondary schooling in the richest households. The strata explained only 6.9% (95% CI: 5.6 - 8.7) of individual variation, with small discriminatory accuracy (AUC = 0.65; 95% CI: 0.63 - 0.66), and a Bayesian model gave a similar VPC (7.0%; 95% credible interval: 5.7 - 8.9). Additive effects explained 95.1% of the between-stratum variance. The VPC fell from 10.1% (95% CI: 8.0 - 12.8) in 1994-99 to 4.5% (95% CI: 3.6 - 5.8) in 2018-24. Five strata of married rural women with no schooling were in the lowest decile in every period. In the latest surveys, married rural women aged 15-24 years with no schooling had lower demand satisfied than other women in 29 of 30 countries (median gap 17.5 percentage points). The 15 most disadvantaged strata held a median 16.4% of unsatisfied demand, so targeting these groups alone would reach about one in six women whose demand is not met. The main limitations are the cross-sectional design, self-reported contraceptive use and the exclusion of 15 early surveys. Conclusions: Intersectional gaps in demand satisfied by modern methods have narrowed in sub-Saharan Africa, and most of the remaining gap between groups is additive. Social position alone discriminates poorly between women, so programmes targeted only by social group would miss most women with unmet need. This supports strengthening universal services. Within universal services, married adolescents and young women with no schooling in poor rural households have been left behind for three decades in almost every country and need priority, country-specific programmes if the 2030 targets are to be met.
Related
- Intersectional Inequalities in Mental Health Need and Youth Justice Contact among Pupils enrolled in Alternative Provision
- Social inequalities and the risk of multimorbidity in Mexican adults: an intersectional approach
- A lifecourse perspective on sociodemographic intersectional inequalities in physical and mental health outcomes
- How much the burden of diabetes and hypertension is attributable to social inequalities? an intersectional analysis in Brazil
- Intersectional inequalities in youth mental health measurement and outcomes: population-representative evidence from the UK Millennium Cohort Study
BibTeX
@misc{kaonga2026intersectional,
title = {{Intersectional inequalities in demand satisfied by modern contraceptive methods among women in sub-Saharan Africa, 1994 to 2024: a multilevel analysis of individual heterogeneity and discriminatory accuracy}},
author = {P. Kaonga and S. Masambi and M. Nkole and A. Hazemba},
howpublished = {medRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.12.26367922},
url = {https://www.medrxiv.org/content/10.64898/2026.11.12.26367922v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Intersectional inequalities in demand satisfied by modern contraceptive methods among women in sub-Saharan Africa, 1994 to 2024: a multilevel analysis of individual heterogeneity and discriminatory accuracy
AU - P. Kaonga
AU - S. Masambi
AU - M. Nkole
AU - A. Hazemba
PY - 2026
DA - 2026/11/12
PB - medRxiv
DO - 10.64898/2026.11.12.26367922
UR - https://www.medrxiv.org/content/10.64898/2026.11.12.26367922v1
AB - Background: Demand for family planning satisfied by modern methods is Sustainable Development Goal (SDG) indicator 3.7.1, and countries have committed to universal access to family planning by 2030. Inequalities are usually reported one social dimension at a time. This hides how social characteristics combine and leaves open whether programmes should target social groups or strengthen services for all women. We measured intersectional inequalities in demand satisfied by modern methods among women in sub-Saharan Africa over three decades, assessed how well social position discriminates between women, and identified groups left behind. Methods and findings: We did a preregistered cross-sectional analysis of 139 Demographic and Health Surveys from 38 countries with fieldwork from 1994 to 2024. We included 675,584 women aged 15-49 years with demand for family planning and grouped them into 144 intersectional strata by age, union status, education, household wealth and residence. We fitted multilevel logistic models with random effects for strata and surveys (multilevel analysis of individual heterogeneity and discriminatory accuracy, MAIHDA) and estimated the variance partition coefficient (VPC), the area under the curve (AUC) and the proportional change in variance (PCV). Weighted demand satisfied rose from 27.5% in 1994-99 to 54.1% in 2018-24. Predicted demand satisfied ranged from 13.2% (95% CI: 11.0 - 15.9) among married adolescents with no schooling in the poorest rural households to 68.0% (95% CI: 63.6 - 72.0) among unmarried urban women aged 25-34 years with secondary schooling in the richest households. The strata explained only 6.9% (95% CI: 5.6 - 8.7) of individual variation, with small discriminatory accuracy (AUC = 0.65; 95% CI: 0.63 - 0.66), and a Bayesian model gave a similar VPC (7.0%; 95% credible interval: 5.7 - 8.9). Additive effects explained 95.1% of the between-stratum variance. The VPC fell from 10.1% (95% CI: 8.0 - 12.8) in 1994-99 to 4.5% (95% CI: 3.6 - 5.8) in 2018-24. Five strata of married rural women with no schooling were in the lowest decile in every period. In the latest surveys, married rural women aged 15-24 years with no schooling had lower demand satisfied than other women in 29 of 30 countries (median gap 17.5 percentage points). The 15 most disadvantaged strata held a median 16.4% of unsatisfied demand, so targeting these groups alone would reach about one in six women whose demand is not met. The main limitations are the cross-sectional design, self-reported contraceptive use and the exclusion of 15 early surveys. Conclusions: Intersectional gaps in demand satisfied by modern methods have narrowed in sub-Saharan Africa, and most of the remaining gap between groups is additive. Social position alone discriminates poorly between women, so programmes targeted only by social group would miss most women with unmet need. This supports strengthening universal services. Within universal services, married adolescents and young women with no schooling in poor rural households have been left behind for three decades in almost every country and need priority, country-specific programmes if the 2030 targets are to be met.
ER -