Log in / My SOA

The Impact of Wildfire-Related and Environmental Air Pollution on Morbidity

soa-research-societal-purpose-logo.png

October 2025

Authors

Rong Yi, PhD
Principal
Milliman, Inc.

Meseret Woldeyes, MS
Managing Data Scientist
Milliman, Inc.

Garrett Bradford
Principal and GIS Consultant
Milliman, Inc

Executive Summary

In recent years, wildfires have grown more frequent and intense across the United States as climate conditions become hotter and drier.[1] Contrary to common perception, wildfires are not just a West Coast phenomenon; they increasingly affect many regions across the United States.[2]

Wildfire smoke contains a complex mixture of chemical components, which undergo atmospheric reactions and condensation and form particulate matter that suspends the air. Among air pollutants, fine particulate matter with diameters of less than 2.5 micrometers in size (PM2.5) is of greatest concern due to its ability to penetrate deep into the lungs and enter the bloodstream, leading to systemic health effects.[3] Wildfire smoke also remains airborne for extended periods, allowing it to travel long distances via atmospheric currents such as jet streams, impacting air quality far from the wildfire source.[4]

This study sought to quantify the impact of PM2.5 exposure during the wildfire season and its lagged effects on the prevalence of four sets of disease conditions: circulatory conditions (CIR), mental and behavioral disorders (MBD), neoplasms (NEO), and respiratory conditions (RSP). Machine learning and statistical methods were applied to a combination of health care claims data, climate data, and community-level socioeconomic data for 2017–2023. Three major health insurance coverage types were included in the data: commercial, Medicare, and Medicaid.

The analysis considered two domains of risk drivers:

  • Clinical risks and demographics, including measures of baseline health status, age, gender, and state of residence, and demographics, including measures of baseline health status, age, gender, and state of residence, and
  •  Environmental and socio-contextual factors, including measures of PM2.5 exposure during the wildfire season and its lagged effects, extreme heat and its lagged effects, seasonality, the COVID-19 pandemic, and social determinants of health (SDOH).

Key Findings

The study yields several important findings:

  • Clinical risks and demographics explain the majority of condition prevalence rates. Clinical risks and demographics factors explained approximately 68% to 92% of prevalence rates in the four sets of disease conditions studied. The remaining 8% to 32% was attributable to environmental and socio-contextual factors.
  • Within the environmental and socio-contextual domain, PM2.5 exposure is a leading driver of disease prevalence. PM2.5 exposure during wildfires season and its lagged effects accounted for 11% to 26% of the environmental and socio-contextual domain (~1% to 4% of total prevalence). For RSP and MBD, the impact of PM2.5 exposure during wildfire season and its lagged effects exceeded that of SDOH.
  • The combined effects from exposure to PM2.5 and extreme heat are significant. The combined effects of PM2.5 exposure during the wildfire season and its lagged effects, and extreme heat and its lagged effects accounted for 2.3% to 8.6% of disease prevalence, or 26% to 56% of the total environmental and socio-contextual influence. This magnitude rivaled and often exceeded that of the COVID-19 pandemic and its aftermath.
  • PM2.5 exposure is a material driver of MBD prevalence with the largest effects observed in the Medicaid population. The combined effects of exposure from PM2.5 and extreme heat accounted for 6.7% to 7.8% of the total MBD prevalence rate, or 50–56% of the prevalence contribution from environmental and socio-contextual factors.
  • Exposure impacts are persistent and long term. The effects of PM2.5 exposure during wildfire season extended for months and, in some cases, years after the exposure event, particularly for CIR and MBD.
  • There are variations in immediate versus lagged responses. RSP exhibited immediate increases in prevalence following PM2.5 exposure during wildfire season. CIR and MBD displayed delayed responses, often peaking several months after exposure.
  • Health impacts are uneven. Climate-related health effects were not experienced equally. Communities with limited health care access, higher social vulnerability, or preexisting chronic conditions tended to experience greater health burdens during wildfire and heat events, underscoring the importance of targeted, data-driven public health interventions.

Implications for the Health Care and Insurance Industries

The findings of this study have important implications for health care systems, insurers, and policymakers.

Health insurers may consider integrating climate risks into actuarial models. Traditional risk assessments may underestimate the growing health care burden of climate change. Incorporating climate-adjusted predictive modeling could improve financial projections and premium pricing strategies.

Excess morbidity related to PM2.5 and extreme heat could place additional pressure on hospitals, emergency departments, and mental health services. Health care providers may consider approaches to meet demand surges during extreme climate events, such as using telemedicine, remote patient monitoring, and mobile health units to expand access in affected areas. It is equally important to consider coverage and delivery strategies after the wildfire peak seasons to address the delayed health impacts and health care needs from the impacted populations. In this process, data-driven prediction models similar to those in this study may be leveraged to support planning and enable quick responses.

Different populations and communities may require specific, targeted mitigation efforts, including increased health care access, air filtration subsidies, and community resilience programs to reduce exposure-related health disparities.

Policymakers and regulatory agencies may also evaluate whether existing frameworks for Medicare, Medicaid, and commercial health insurance sufficiently account for climate-related health impacts. Future updates could consider supporting infrastructure and interventions that promote health care system resilience in the face of environmental stressors.

Limitations and Future Research

While this study provides empirical insights into the health impacts of wildfire-related PM2.5 exposure, the complexity of the observed relationships and data limitations warrant cautious interpretation. Key limitations—such as challenges in exposure attribution and population representativeness—are discussed in Section 4.2. Further research is needed to expand the scope of health outcomes examined and to refine analytic methods for assessing long-term and cumulative effects, as outlined in Section 4.4.

Conclusion

Wildfire-related air pollution is not only an environmental issue-it increasingly affects public health and drives long-term increases in disease burden and health care utilization. By applying advanced analytics and fostering collaboration across sectors, health care organizations can enhance preparedness and resilience in the face of evolving climate-related health challenges.

Material

Wildfire-Related Environmental Air Pollution on Morbidity

Acknowledgements

The research team’s deepest gratitude goes to those without whose efforts this project could not have come to fruition: the volunteers who generously shared their wisdom, insights, advice, guidance, and arm’s length review of this study prior to publication. Any opinion expressed may not reflect their opinions nor those of their employers. Any errors belong to the authors alone.

Project Oversight Group members:
Sam Gutterman, FSA, MAAA, FCAS, FCA, HonFIA, CERA
Patience Manhanga, FSA, FCIA
Charlie Mathews
Donna Megregian, FSA, MAAA
Rebecca Owen, FSA, MAAA, FCA
Sandra Said, CFIRM, ACII
Aadit Sheth, FSA, CERA, FCIA
Cirhan Truswell
Georgiana Willwerth-Pascutiu, MD, DBIM

At the Society of Actuaries Research Institute:
Kara Clark, FSA, MAAA, Sr. Research Actuary
Rob Montgomery, ASA, MAAA, FLMI, Consultant-Research Project Manager
Ronora Stryker, ASA, MAAA, Sr. Research Actuary
Barbara Scott, Sr. Research Administrator

At Milliman, Inc.:
Kailey Adams, GIS Analyst
Joanne Buckle, FIA, Principal and Consulting Actuary
Rich Moyer, Principal and MedInsight Chief Product Officer
Shivangi Sharma, Assistant Manager Clinical Product, MedInsight
Mia Wafer, GIS Analyst

Questions or Comments?

Give us your feedback! Take a short survey on this report. Take Survey

If you have comments or questions, please send an email to Research@soa.org

 

[1] See “Climate Change Indicators: Wildfires,” EPA, last updated May 9, 2025, https://www.epa.gov/climate-indicators/climate-change-indicators-wildfires.

[2] The National Interagency Fire Center tracks total number of wildfires and acres burned by state by year. See “National Fire News,” National Interagency Fire Center, June 21, 2025, https://www.nifc.gov/fire-information/nfn.

[3] See “Particulate Matter Basics,” EPA, last updated May 30, 2025, https://www.epa.gov/pm-pollution/particulate-matter-pm-basics.

[4] “Wildland Fire Activity and Modeled Impacts on O3 and PM2.5,” EPA, March 2022, https://www.epa.gov/system/files/documents/2022-03/epa_454_r_22_002.pdf.