The Effects of Air Pollution on 911 Mental Health Calls

• In Event: Environmental Determinants of Health: Evidence from Air, Water, and Heat

Abstract

Overview and Research Question:
While prior research has shown that air pollution adversely affects physical health, cognitive performance, and behavior, less is known about its impact on acute mental health crises in urban settings. This study examines whether short-term exposure to fine particulate matter (PM2.5) increases the incidence of 911 calls for emotionally disturbed persons (EDP)—a proxy for acute mental distress requiring police or emergency medical response.

Data and Setting:
We use a novel, high-frequency dataset of daily 911 calls in New York City from January 2018 to September 2024, aggregated to the zip code level. The primary outcome of interest is the number of EDP-related calls, complemented by other emergency calls such as unconsciousness, cardiac events, and vehicle accidents. Pollution exposure is measured using daily PM2.5 concentrations from the U.S. Environmental Protection Agency. Weather controls are drawn from the National Climatic Data Center. Our sample includes over 460,000 zip code-day observations.

Empirical Strategy:
We estimate Poisson count models with zip code, day-of-week, and month-year fixed effects. The primary exposure variable is PM2.5 concentration, categorized into three bins: ≤5 µg/m³, 5–8.1 µg/m³, and >8.1 µg/m³. Weather variables—precipitation, maximum temperature, and windspeed—are included as controls. Standard errors are clustered at the zip code level. We also explore robustness using alternative pollution cutoffs (e.g., WHO thresholds), lag/lead structures, and a forthcoming instrumental variables (IV) strategy exploiting variation in wind direction and wild fire.

Main Results:
We find a significant increase in mental health-related 911 calls on days with elevated PM2.5 levels. Specifically, EDP calls increase by 1.24% on days with PM2.5 between 5 and 8.1 µg/m³ and by 1.06% on days above 8.1 µg/m³, relative to days with PM2.5 at or below 5 µg/m³. Similar patterns emerge for unconsciousness and vehicle accident calls, suggesting a broader impact on emergency services.

Heterogeneity:
The effects are concentrated in zip codes with below-median income and above-median baseline EDP call volumes. We also find stronger effects on weekends and during the daytime, which may reflect increased exposure or heightened vulnerability during those periods.

Robustness and Instrumental Variable Plans:
Results remain robust when excluding the COVID-19 period and when applying WHO PM2.5 thresholds. Leads of pollution suggest that air quality forecasts may influence caller behavior. Preliminary work is underway to implement an IV strategy using wind direction as a source of exogenous variation in pollution exposure.

Policy Implications:
Our findings highlight a previously underexplored consequence of air pollution: increased demand on emergency mental health services. These results underscore the broader social costs of pollution and support stricter air quality regulations as a means to alleviate mental health strain and reduce pressure on police and emergency medical systems.

Author

• Tahreen Zahra, Carleton University
  Presenting Author