Humanitarian risk and exclusive breastfeeding: an exploratory cross-national ecological study with DHS and subnational triangulation

An estimated 252 million people needed humanitarian assistance in mid-2026. Breastfeeding is lifesaving, especially when emergencies disrupt safe water, health services and supply chains. Globally, only 47% of infants under six months are exclusively breastfed; the 2026 global investment case estimates that suboptimal breastfeeding contributes to about 383,000 premature child deaths and US$339 billion in economic losses each year. The Operational Guidance on Infant and Young Child Feeding in Emergencies frames breastfeeding protection across preparedness, response and recovery. This study examined whether higher standing humanitarian risk is associated with lower exclusive breastfeeding after accounting for income. Exploratory cross-sectional analyses used 135 of 196 countries, 17,574 children in 15 Demographic and Health Surveys (DHS), and 114 administrative units in seven Sahel countries. Humanitarian risk was measured with the Index for Risk Management (INFORM) Risk score (0–10; high risk: score ≥ 5.4). The primary analysis used income-adjusted partial Spearman correlation; DHS inference used a two-stage country-level estimator. Sensitivity analyses examined timing, health-system capacity, region, fragility and acute-crisis indicators. After continuous-income adjustment, higher humanitarian risk was associated with lower exclusive breastfeeding (ρ=−0.22, 95% confidence interval [CI] − 0.39 to − 0.03; false-discovery-rate q = 0.021). Among low- and lower-middle-income countries, median exclusive breastfeeding was 44.1% in high-risk versus 53.8% in lower-risk countries (shift − 9.4% points, 95% CI − 17.0 to − 0.6). The association persisted after health-system and regional adjustment and was concentrated in sub-Saharan Africa, but was non-significant with categorical or survey-year income or standardised-survey estimates only. In exploratory decomposition, associations were largest for institutional capacity (ρ=−0.36), lack of coping capacity (ρ=−0.25) and human hazard/conflict (ρ=−0.24), all surviving multiple-comparison correction. DHS results were inverse but imprecise (two-stage odds ratio 0.718, 95% CI 0.492–1.045). No consistent inverse within-country association was detected. At similar national income, high-risk countries had lower exclusive breastfeeding, an exploratory, specification-sensitive association. Weak coping capacity and conflict-related risk showed the largest associations. Countries most likely to face emergencies may already have more infants without exclusive breastfeeding’s protection and systems less able to sustain support. Consistent with existing guidance, breastfeeding protection and support should be core lifesaving investments in emergency preparedness and response.