New Analysis

Pollution Sensitivity: The Cluster 3 Anomaly

A uniform NAAQS standard assumes every district generates the same respiratory burden per µg/m³. The data proves otherwise. Cluster 3 districts (UP/Bihar) produce nearly 3× more cases at the same pollution level as any other cluster — and the gap is statistically significant even after controlling for population and PM2.5.

—

extra cases/month

Cluster 3 structural excess

—

significance

controlling for PM2.5 + population

—

mean excess cases/month

avg per Cluster 3 district

—

Cluster 3 vs At Risk

at comparable PM2.5 levels

Same pollution level, vastly different outcomes

Mean monthly respiratory cases per PM2.5 concentration bin, computed separately for each K-Means cluster. At PM2.5 ~60–70 µg/m³, Cluster 3 (purple) produces 1,700–1,800 cases/month while other clusters at the same level produce 400–700. The gap is a structural shift, not a dosage difference.

Key observation

Every Cluster 3 bin — at every PM2.5 level — sits above every other cluster at the same pollution level. This is not a slope effect (the curves are roughly parallel): it is a vertical shift. Cluster 3 districts start at a higher disease baseline and stay above it regardless of how much pollution they have.

Why this finding changes the policy picture

Every other analysis in this project frames the problem as: reduce PM2.5 → reduce disease. The sensitivity analysis reveals a harder constraint: Cluster 3 districts have a baseline structural deficit of ~1,735 extra cases/month that exists independently of pollution level. This means air quality improvement in UP/Bihar will produce smaller health gains per µg/m³ of PM2.5 reduced than in Delhi or Maharashtra — not because the interventions don't work, but because a substantial share of the disease burden has other structural drivers. Healthcare infrastructure, sanitation, and nutritional access must be treated as co-equal interventions, not downstream afterthoughts.