Fig. 5: Results of cumulative associations between child mortality and mothers’ exposure to El Niño conditions at 0–12 preconceptional and prenatal months from country-specific analyses and meta-analyses.

The country-specific sample sizes used to derive statistics are provided in Supplementary Table S2. The effect estimate, which is presented as the center of the error bars in the figure, is computed as the HR of 90th percentile level of MEI measure relative to the reference value (set at zero). The error bars represent the 95% confidence intervals of estimates. HR hazard ratio, ENSO El Niño Southern Oscillation, MEI multivariate El Niño index. The heterogeneity is measured by Cochran’s Q, which is calculated as the weighted sum of squared differences between country-specific effects and the pooled effect across studies. The Q-statistic is 254.16; with the degree of freedom being 37, the exact p value is 3.30 × 10⁻³⁴, which is way smaller than 0.01. Adjustments are not made for multiple comparisons.