Preventable drivers of acute respiratory infections among under-five children in India’s high-risk zones: application of population attributable fractions analyses

BACKGROUND: Acute respiratory infections (ARI) remain a major contributor to morbidity and mortality among under-five children in India. Although previous national studies have identified individual-level risk factors, evidence on the spatial distribution of ARI and the burden statistically attributable to selected exposures within hotspot regions remains limited. METHODS: We analysed data from the 2019-21 National Family Health Survey (NFHS-5) Children's Recode file using a two-stage analytical framework. First, district-level ARI hotspots were identified among 221,849 under-five children across 707 districts. Second, analyses were restricted to children residing in hotspot districts (weighted n = 108,470). Survey-weighted prevalence, multivariable logistic regression, and population attributable fraction (PAF) analyses accounted for sampling weights, clustering, and stratification. FINDINGS: A total of 167 districts were identified as ARI hotspots, predominantly in central and eastern India. The unrestricted national weighted prevalence of ARI in the study dataset was 2.74%, compared with 7.23% (95% CI 7.08-7.39) within hotspot districts. Recent diarrhoeal illness (AOR 4.37, 95% CI 4.01-4.76) and maternal tobacco use (AOR 1.29, 95% CI 1.06-1.57) were independently associated with ARI. The PAF was 19.6% for recent diarrhoea and 0.7% for maternal tobacco use. Rural residence, interpreted as a contextual marker rather than a directly modifiable exposure, had a model-based PAF of 10.2%. The joint counterfactual scenario yielded a PAF of 28.7% (95% CI 20.4-36.1). CONCLUSIONS: Under-five ARI is geographically concentrated in distinct hotspot districts. Integrated interventions addressing diarrhoeal disease and maternal tobacco exposure, together with improvements in rural child-health services and infrastructure, may reduce ARI burden in these districts. Because rural residence is not itself an intervention, its PAF should be interpreted as a priority-setting indicator of place-based disadvantage rather than as the effect of relocating rural populations.