Geography, rurality, and community distress: deaths due to suicide, alcohol-use, and drug-use among Colorado Veterans.

BACKGROUND: In the USA, deaths due to suicide, alcohol, or drug-related causes (e.g., alcohol-related liver disease, overdose) have doubled since 2002. Veterans appear disproportionately impacted by growing trends. Limited research has been conducted regarding the relationship between community-level factors (e.g., rurality, community distress resulting from economic conditions) and the presence of spatial clustering of suicide, alcohol-related, or drug-related deaths. We explored community-level relationships in Colorado Veterans and compared suicide, alcohol-, and drug-related death rates between the Colorado adult population and Veterans. METHODS: 2009-2020 suicide, alcohol-related, and/or drug-related deaths were identified using qualifying multiple cause-of-death International Classification of Disease (ICD)-10 codes in CDC WONDER for the general adult population and Colorado death data for Veteran populations. Age and race adjusted rates were calculated to compare risk overall and by mortality type (i.e., suicide, alcohol-related, drug-related). In Veteran decedents, age-adjusted rates were stratified by rurality and community distress, measured by the Distressed Communities Index. Standardized mortality ratios were calculated to measure spatial autocorrelation and identify clusters using global and local Moran's I, respectively. RESULTS: 6.4% of Colorado Veteran deaths (n = 6948) were identified as being related to suicide, alcohol, or drugs. Compared to rates in the general population of Colorado adults, Veterans had 1.8 times higher rates of such deaths overall (2.1 times higher for suicide, 1.8 times higher for alcohol-related, 1.3 times higher for drug-related). Among Veterans, community distress was associated with an increased risk of alcohol-related [age-adjusted rate per 100,000 (95% CI) = 129.6 (89.9-193.1)] and drug-related deaths [95.0 (48.6-172.0)]. This same significant association was not identified among those that died by suicide. Rurality was not associated with risk for any of the deaths of interest. There was significant spatial clustering for alcohol-related deaths in southeast Colorado. CONCLUSIONS: Colorado Veterans have higher rates of deaths due to suicide, alcohol-related, and drug-related causes compared to members of the general adult population. Upstream prevention efforts, such as community-based interventions targeting alcohol-use and community economic distress, are warranted. More research is also needed to understand how community distress and other social determinants of health impact the community burden of suicide, alcohol-related, and drug-related mortality.

Evaluation of Model-Based PM2.5 Estimates for Exposure Assessment during Wildfire Smoke Episodes in the Western U.S.

Investigating the health impacts of wildfire smoke requires data on people's exposure to fine particulate matter (PM2.5) across space and time. In recent years, it has become common to use machine learning models to fill gaps in monitoring data. However, it remains unclear how well these models are able to capture spikes in PM2.5 during and across wildfire events. Here, we evaluate the accuracy of two sets of high-coverage and high-resolution machine learning-derived PM2.5 data sets created by Di et al. and Reid et al. In general, the Reid estimates are more accurate than the Di estimates when compared to independent validation data from mobile smoke monitors deployed by the US Forest Service. However, both models tend to severely under-predict PM2.5 on high-pollution days. Our findings complement other recent studies calling for increased air pollution monitoring in the western US and support the inclusion of wildfire-specific monitoring observations and predictor variables in model-based estimates of PM2.5. Lastly, we call for more rigorous error quantification of machine-learning derived exposure data sets, with special attention to extreme events.

Perceptions of green space usage, abundance, and quality of green space were associated with better mental health during the COVID-19 pandemic among residents of Denver.

BACKGROUND: The COVID-19 pandemic has impacted both physical and mental health. This study aimed to understand whether exposure to green space buffered against stress and distress during the COVID-19 pandemic while taking into account significant stressors of the pandemic. METHODS: We leveraged a cross-sectional survey on green space exposure and mental health among residents of Denver, CO that ran from November 2019 through January 2021. We measured objective green space as the average NDVI (normalized difference vegetation index) from aerial imagery within 300m and 500m of the participant's residence. Perceived green space was measured through Likert scores on five questions about vegetation near the home that captured perceived abundance, visibility, access, usage, and quality of green space. We used generalized linear models to assess the relationship between each green space exposure variable and perceived stress (PSS-4), depression (CES-D-10), or anxiety (MMPI-2) adjusted for sociodemographic and COVID-19 impact variables. RESULTS: We found significantly higher depression scores for all covid periods compared to the "before covid" period, and significantly higher anxiety scores during the "fall wave" compared to earlier periods. Adjusted for sociodemographic and pandemic stressors, we found that spending a lot of time in green space (usage) was significantly associated with lower anxiety and depression. We also observed significantly lower depression scores associated with NDVI in both buffers (objective abundance) and significantly lower anxiety scores with perceived abundance of green space. There was some evidence of lower anxiety scores for people reporting having high quality green spaces near the home (quality). We did not observe significant associations for any green space metric and perceived stress after adjustment for confounding variables. CONCLUSION: Our work provides further evidence of mental health benefits associated with green space exposure during the COVID-19 pandemic even after adjustment for sociodemographic variables and significant pandemic-related stressors.

Socio-demographic and health vulnerability in prescribed-burn exposed versus unexposed counties near the National Forest System.

Prescribed fire is an increasingly important tool in restoring ecological conditions and reducing uncontrolled wildfire. Prescribed burn techniques could reduce public health impacts associated with wildfire smoke exposure. However, there have been few assessments of the health impacts of prescribed burning, and potential vulnerabilities among populations exposed to smoke from prescribed fires. Our study area focused on counties in and near U.S. National Forests - a set of lands distributed across the U.S. In county-level analyses, we compared the sociodemographic and health characteristics of areas that were exposed with those that were not exposed to prescribe burns during the years 2010-2019 on a national level and within three regions. In addition, using spatial error regression models, we looked for associations between prescribed fire exposure and health behaviors and outcomes while controlling for spatial autocorrelation. On a national level, we found disproportionate prescribed fire exposure in rural counties with higher percentage mobile home and vacant housing units, and higher percentage African-American and white populations. Regionally, we found evidence of disproportionate exposure to prescribed burns among counties with lower percentage white population, higher percentage Hispanic population and mobile homes in the southern region, and to high poverty counties with high vacancy in the western region. These findings could indicate that vulnerable populations face potential health risks from prescribed burning smoke exposure, but also that they are not missing out on the benefits of prescribed burning, which could involve considerably lower smoke exposure compared to uncontrolled wildfire. In addition, in regression analyses, we found no evidence of disproportionate health burden in exposed compared to unexposed counties. Awareness of these patterns could influence both large-scale or institutional polices about prescribed burning practice, and could be used to build decision-making factors into modeling tools and smoke management plans, as well as community-engagement around wildfire risk reduction.

Assessment of Rural-Urban and Geospatial Differences in Perceived Handgun Access and Reported Suicidality Among Youth in Colorado.

Importance: Suicide is the second leading cause of death in adolescents, with firearms the most common method, especially in rural communities. Identifying where to target lethal means safety interventions could better leverage limited resources. Objectives: To understand the associations of rurality, school-level prevalence of easy handgun access, and suicidality measures in Colorado youth, to explore spatial distribution of school-level measures, and to identify communities with high prevalence of both easy handgun access and suicidality. Design, Setting, and Participants: This cross-sectional study used data from the 2019 Healthy Kids Colorado Survey, an anonymous cross-sectional school-based survey conducted at 256 participating Colorado high schools. Participants included students from schools recruited for statewide population-based estimates and additional schools opting in. Data were analyzed from November 9, 2020, to March 13, 2021. Exposures: Urban-centric locale according to a 7-level continuum. Geocoded location of schools was used for spatial analysis. Main Outcomes and Measures: The main outcomes were weighted prevalence for easy handgun access and 4 measures of mental health and suicidality in the previous year (ie, feeling sad for 2 weeks and considering suicide, planning suicide attempt, or attempting suicide in the past year). Results: A total of 59 556 students (49.7% [95% CI, 49.3%-50.1%] male and 50.3% [95% CI, 49.9%-50.7%] female; 53.9% [95% CI, 53.5%-54.3%] in 9th and 10th grade; 36.4% [95% CI, 36.0%-36.8%] Hispanic and 50.8% [95% CI, 50.4%-51.2%] non-Hispanic White) from 256 schools participated. Most schools were rural or in small towns (56.8% [95% CI, 50.7%-62.9%]), while more students participated from urban and suburban schools (57.8% [95% CI, 57.6%-58.0%]). Prevalence of perceived easy access to handguns increased with increasing rurality, with 36.2% (95% CI, 35.2%-37.1%) of students in rural (remote) schools reporting easy access, compared with 18.2% (95% CI, 17.3%-19.1%) for city (large) schools. The spatial distribution of easy handgun access and suicidality measures had minimal overlap, but there was correlation at school-level between easy handgun access and considering suicide (ρ = 0.203 [95% CI, 0.0748-0.331]), planning suicide (ρ = 0.300 [95% CI, 0.173-0.427]), and attempting suicide (ρ = 0.218 [95% CI, 0.0869-0.350) in the previous year. The highest quartile for prevalence of both perceived easy access to handguns and planning suicide in the previous year included 21 schools (81.0% [95% CI, 64.0%-97.9%] rural [remote] or rural [distant]). Conclusions and Relevance: These findings suggest that rural-remote communities in Colorado may benefit most from interventions focused on limiting youth access to handguns when youth are in crisis, with some communities at especially high risk. Spatially referenced data may improve targeting interventions to where they are needed most.

Hospital Preparedness, Mitigation, and Response to Hurricane Harvey in Harris County, Texas.

OBJECTIVE: This case study documents Harris County hospitals' flood preparedness and mitigation efforts before Hurricane Harvey, their collective response experience during Hurricane Harvey, and their lessons learned in the storm's aftermath. METHODS: The case study was constructed using a survey of hospital emergency managers, semi-structured interviews with local agencies involved in public health, emergency management, and health care, and an analysis of news reports and other documents from a variety of government agencies, local organizations, and hospitals themselves. RESULTS: Harris County hospitals learned their most valuable lessons through their direct and repeated experience with flooding over the years, leading to improved preparedness before Hurricane Harvey. Hospital emergency response successes included infrastructure improvements, staff resilience, advanced planning, and pre-established collaboration. However, hospitals still experienced challenges with staff burnout, roadway flooding, and patient evacuation. CONCLUSIONS: Although the current state of hospital flood preparedness and mitigation is rather advanced and mature, it is advisable that Harris County takes steps to strengthen emergency management efforts in hospitals with fewer financial and staffing resources and less direct flood experience.

Daily PM2.5 concentration estimates by county, ZIP code, and census tract in 11 western states 2008-2018.

We created daily concentration estimates for fine particulate matter (PM2.5) at the centroids of each county, ZIP code, and census tract across the western US, from 2008-2018. These estimates are predictions from ensemble machine learning models trained on 24-hour PM2.5 measurements from monitoring station data across 11 states in the western US. Predictor variables were derived from satellite, land cover, chemical transport model (just for the 2008-2016 model), and meteorological data. Ten-fold spatial and random CV R2 were 0.66 and 0.73, respectively, for the 2008-2016 model and 0.58 and 0.72, respectively for the 2008-2018 model. Comparing areal predictions to nearby monitored observations demonstrated overall R2 of 0.70 for the 2008-2016 model and 0.58 for the 2008-2018 model, but we observed higher R2 (>0.80) in many urban areas. These data can be used to understand spatiotemporal patterns of, exposures to, and health impacts of PM2.5 in the western US, where PM2.5 levels have been heavily impacted by wildfire smoke over this time period.

Extreme Weather and Climate Change: Population Health and Health System Implications.

Extreme weather and climate events, such as heat waves, cyclones, and floods, are an expression of climate variability. These events and events influenced by climate change, such as wildfires, continue to cause significant human morbidity and mortality and adversely affect mental health and well-being. Although adverse health impacts from extreme events declined over the past few decades, climate change and more people moving into harm's way could alter this trend. Long-term changes to Earth's energy balance are increasing the frequency and intensity of many extreme events and the probability of compound events, with trends projected to accelerate under certain greenhouse gas emissions scenarios. While most of these events cannot be completely avoided, many of the health risks could be prevented through building climate-resilient health systems with improved risk reduction, preparation, response, and recovery. Conducting vulnerability and adaptation assessments and developing health system adaptation plans can identify priority actions to effectively reduce risks, such as disaster risk management and more resilient infrastructure. The risks are urgent, so action is needed now.

Wildland firefighter exposure to smoke and COVID-19: A new risk on the fire line.

Throughout the United States, wildland firefighters respond to wildfires, performing arduous work in remote locations. Wildfire incidents can be an ideal environment for the transmission of infectious diseases, particularly for wildland firefighters who congregate in work and living settings. In this review, we examine how exposure to wildfire smoke can contribute to an increased likelihood of SARS-CoV-2 infection and severity of coronavirus disease (COVID-19). Human exposure to particulate matter (PM), a component of wildfire smoke, has been associated with oxidative stress and inflammatory responses; increasing the likelihood for adverse respiratory symptomology and pathology. In multiple epidemiological studies, wildfire smoke exposure has been associated with acute lower respiratory infections, such as bronchitis and pneumonia. Co-occurrence of SARS-CoV-2 infection and wildfire smoke inhalation may present an increased risk for COVID-19 illness in wildland firefighters due to PM based transport of SARS CoV-2 virus and up-regulation of angiotensin-converting enzyme II (ACE-2) (i.e. ACE-2 functions as a trans-membrane receptor, allowing the SARS-CoV-2 virus to gain entry into the epithelial cell). Wildfire smoke exposure may also increase risk for more severe COVID-19 illness such as cytokine release syndrome, hypotension, and acute respiratory distress syndrome (ARDS). Current infection control measures, including social distancing, wearing cloth masks, frequent cleaning and disinfecting of surfaces, frequent hand washing, and daily screening for COVID-19 symptoms are very important measures to reduce infections and severe health outcomes. Exposure to wildfire smoke may introduce additive or even multiplicative risk for SARS-CoV-2 infection and severity of disease in wildland firefighters. Thus, additional mitigative measures may be needed to prevent the co-occurrence of wildfire smoke exposure and SARS-CoV-2 infection.

Improving accuracy of air pollution exposure measurements: Statistical correction of a municipal low-cost airborne particulate matter sensor network.

Low-cost air quality sensors can help increase spatial and temporal resolution of air pollution exposure measurements. These sensors, however, most often produce data of lower accuracy than higher-end instruments. In this study, we investigated linear and random forest models to correct PM2.5 measurements from the Denver Department of Public Health and Environment (DDPHE)'s network of low-cost sensors against measurements from co-located U.S. Environmental Protection Agency Federal Equivalence Method (FEM) monitors. Our training set included data from five DDPHE sensors from August 2018 through May 2019. Our testing set included data from two newly deployed DDPHE sensors from September 2019 through mid-December 2019. In addition to PM2.5, temperature, and relative humidity from the low-cost sensors, we explored using additional temporal and spatial variables to capture unexplained variability in sensor measurements. We evaluated results using spatial and temporal cross-validation techniques. For the long-term dataset, a random forest model with all time-varying covariates and length of arterial roads within 500 m was the most accurate (testing RMSE = 2.9 µg/m3 and R2 = 0.75; leave-one-location-out (LOLO)-validation metrics on the training set: RMSE = 2.2 µg/m3 and R2 = 0.93). For on-the-fly correction, we found that a multiple linear regression model using the past eight weeks of low-cost sensor PM2.5, temperature, and humidity data plus a near-highway indicator predicted each new week of data best (testing RMSE = 3.1 µg/m3 and R2 = 0.78; LOLO-validation metrics on the training set: RMSE = 2.3 µg/m3 and R2 = 0.90). The statistical methods detailed here will be used to correct low-cost sensor measurements to better understand PM2.5 pollution within the city of Denver. This work can also guide similar implementations in other municipalities by highlighting the improved accuracy from inclusion of variables other than temperature and relative humidity to improve accuracy of low-cost sensor PM2.5 data.

Hurricane Harvey Hospital Flood Impacts: Accuracy of Federal Emergency Management Agency Flood Hazard Areas in Harris County, Texas.

Objectives. To compare the flood impacts experienced by Harris County, Texas, hospitals with Federal Emergency Management Agency (FEMA) flood hazard areas and Hurricane Harvey's inundation boundary.Methods. One year following Hurricane Harvey, we created a novel data set of Hurricane Harvey's flood impacts in Harris County hospitals. We then mapped the hospital flood impact data in ArcGIS alongside FEMA flood hazard areas and Hurricane Harvey's inundation boundary to classify each hospital's location in high flood-risk areas and in areas purportedly affected by Hurricane Harvey.Results. Of the 66 hospitals for which flood impact information was ascertained, 16 (24%) hospitals experienced flood impacts during Hurricane Harvey. Of these 16 hospitals, 5 (31%) were located outside a FEMA flood hazard area and 8 (50%) were located outside Hurricane Harvey's inundation boundary.Conclusions. FEMA flood hazard areas did not accurately predict all areas of Harris County, Texas, that flooded during Hurricane Harvey or which hospitals experienced flood impacts.

Putting Co-Exposures on Equal Footing: An Ecological Analysis of Same-Scale Measures of Air Pollution and Social Factors on Cardiovascular Disease in New York City.

Epidemiologic evidence consistently links urban air pollution exposures to health, even after adjustment for potential spatial confounding by socioeconomic position (SEP), given concerns that air pollution sources may be clustered in and around lower-SEP communities. SEP, however, is often measured with less spatial and temporal resolution than are air pollution exposures (i.e., census-tract socio-demographics vs. fine-scale spatio-temporal air pollution models). Although many questions remain regarding the most appropriate, meaningful scales for the measurement and evaluation of each type of exposure, we aimed to compare associations for multiple air pollutants and social factors against cardiovascular disease (CVD) event rates, with each exposure measured at equal spatial and temporal resolution. We found that, in multivariable census-tract-level models including both types of exposures, most pollutant-CVD associations were non-significant, while most social factors retained significance. Similarly, the magnitude of association was higher for an IQR-range difference in the social factors than in pollutant concentrations. We found that when offered equal spatial and temporal resolution, CVD was more strongly associated with social factors than with air pollutant exposures in census-tract-level analyses in New York City.

Machine learning models accurately predict ozone exposure during wildfire events.

Epidemiologists use prediction models to downscale (i.e., interpolate) air pollution exposure where monitoring data is insufficient. This study compares machine learning prediction models for ground-level ozone during wildfires, evaluating the predictive accuracy of ten algorithms on the daily 8-hour maximum average ozone during a 2008 wildfire event in northern California. Models were evaluated using a leave-one-location-out cross-validation (LOLO CV) procedure to account for the spatial and temporal dependence of the data and produce more realistic estimates of prediction error. LOLO CV avoids both the well-known overly optimistic bias of k-fold cross-validation on dependent data and the conservative bias of evaluating prediction error over a coarser spatial resolution via leave-k-locations-out CV. Gradient boosting was the most accurate of the ten machine learning algorithms with the lowest LOLO CV estimated root mean square error (0.228) and the highest LOLO CV Rˆ2 (0.677). Random forest was the second best performing algorithm with an LOLO CV Rˆ2 of 0.661. The LOLO CV estimates of predictive accuracy were less optimistic than 10-fold CV estimates for all ten models. The difference in estimated accuracy between the 10-fold CV and LOLO CV was greater for more flexible models like gradient boosting and random forest. The order of estimated model accuracy depended on the choice of evaluation metric, indicating that 10-fold CV and LOLO CV may select different models or sets of covariates as optimal, which calls into question the reliability of 10-fold CV for model (or variable) selection. These prediction models are designed for interpolating ozone exposure, and are not suited to inferring the effect of wildfires on ozone or extrapolating to predict ozone in other spatial or temporal domains. This is demonstrated by the inability of the best performing models to accurately predict ozone during 2007 southern California wildfires.

Associations between respiratory health and ozone and fine particulate matter during a wildfire event.

Wildfires have been increasing in frequency in the western United States (US) with the 2017 and 2018 fire seasons experiencing some of the worst wildfires in terms of suppression costs and air pollution that the western US has seen. Although growing evidence suggests respiratory exacerbations from elevated fine particulate matter (PM2.5) during wildfires, significantly less is known about the impacts on human health of ozone (O3) that may also be increased due to wildfires. Using machine learning, we created daily surface concentration maps for PM2.5 and O3 during an intense wildfire in California in 2008. We then linked these daily exposures to counts of respiratory hospitalizations and emergency department visits at the ZIP code level. We calculated relative risks of respiratory health outcomes using Poisson generalized estimating equations models for each exposure in separate and mutually-adjusted models, additionally adjusted for pertinent covariates. During the active fire periods, PM2.5 was significantly associated with exacerbations of asthma and chronic obstructive pulmonary disease (COPD) and these effects remained after controlling for O3. Effect estimates of O3 during the fire period were non-significant for respiratory hospitalizations but were significant for ED visits for asthma (RR = 1.05 and 95% CI = (1.022, 1.078) for a 10 ppb increase in O3). In mutually-adjusted models, the significant findings for PM2.5 remained whereas the associations with O3 were confounded. Adjusted for O3, the RR for asthma ED visits associated with a 10 µg/m3 increase in PM2.5 was 1.112 and 95% CI = (1.087, 1.138). The significant findings for PM2.5 but not for O3 in mutually-adjusted models is likely due to the fact that PM2.5 levels during these fires exceeded the 24-hour National Ambient Air Quality Standard (NAAQS) of 35 µg/m3 for 4976 ZIP-code days and reached levels up to 6.073 times the NAAQS, whereas our estimated O3 levels during the fire period only occasionally exceeded the NAAQS of 70 ppb with low exceedance levels. Future studies should continue to investigate the combined role of O3 and PM2.5 during wildfires to get a more comprehensive assessment of the cumulative burden on health from wildfire smoke.

Wildfire smoke exposure under climate change: impact on respiratory health of affected communities.

PURPOSE OF REVIEW: In this review, we describe the current status of the literature regarding respiratory health related to wildfire smoke exposure, anticipated future impacts under a changing climate, and strategies to reduce respiratory health impacts of wildfire smoke. RECENT FINDINGS: Recent findings confirm associations between wildfire smoke exposure and respiratory health outcomes, with the clearest evidence for exacerbations of asthma. Although previous evidence showed a clear association between wildfire smoke and chronic obstructive pulmonary disease, findings from recent studies are more mixed. Current evidence in support of an association between respiratory infections and wildfire smoke exposure is also mixed. Only one study has investigated long-term respiratory health impacts of wildfire smoke, and few studies have estimated future health impacts of wildfires under likely climate change scenarios. SUMMARY: Wildfire activity has been increasing over the past several decades and is likely to continue to do so as climate change progresses, which, combined with a growing population, means that population exposure to and respiratory health impacts of wildfire smoke is likely to grow in the future. More research is needed to understand which population subgroups are most vulnerable to wildfire smoke exposure and the long-term respiratory health impacts of these high pollution events.

It's not easy assessing greenness: A comparison of NDVI datasets and neighborhood types and their associations with self-rated health in New York City.

Growing evidence suggests that exposure to greenness benefits health, but studies assess greenness differently. We hypothesize greenness-health associations vary by exposure assessment method. To test this, we considered four vegetation datasets (three Normalized Difference Vegetation Index datasets with different spatial resolutions and a finely-resolved land cover dataset), and six aggregation units (five radial buffer sizes and self-described neighborhoods) of each dataset. We compared associations of self-rated health and these metrics of greenness among a sample of New York City residents. Associations with self-rated health varied more by aggregation unit than by vegetation dataset; larger buffers and self-described neighborhoods showed more positive associations. Researchers should consider spatial exposure misclassification in future greenness and health research.

Is All Urban Green Space the Same? A Comparison of the Health Benefits of Trees and Grass in New York City.

Living near vegetation, often called "green space" or "greenness", has been associated with numerous health benefits. We hypothesized that the two key components of urban vegetation, trees and grass, may differentially affect health. We estimated the association between near-residence trees, grass, and total vegetation (from the 2010 High Resolution Land Cover dataset for New York City (NYC)) with self-reported health from a survey of NYC adults (n = 1281). We found higher reporting of "very good" or "excellent" health for respondents with the highest, compared to the lowest, quartiles of tree (RR = 1.23, 95% CI = 1.06-1.44) but not grass density (relative risk (RR) = 1.00, 95% CI = 0.86-1.17) within 1000 m buffers, adjusting for pertinent confounders. Significant positive associations between trees and self-reported health remained after adjustment for grass, whereas associations with grass remained non-significant. Adjustment for air pollutants increased beneficial associations between trees and self-reported health; adjustment for parks only partially attenuated these effects. Results were null or negative using a 300 m buffer. Findings imply that higher exposure to vegetation, particularly trees outside of parks, may be associated with better health. If replicated, this may suggest that urban street tree planting may improve population health.

Differential respiratory health effects from the 2008 northern California wildfires: A spatiotemporal approach.

We investigated health effects associated with fine particulate matter during a long-lived, large wildfire complex in northern California in the summer of 2008. We estimated exposure to PM2.5 for each day using an exposure prediction model created through data-adaptive machine learning methods from a large set of spatiotemporal data sets. We then used Poisson generalized estimating equations to calculate the effect of exposure to 24-hour average PM2.5 on cardiovascular and respiratory hospitalizations and ED visits. We further assessed effect modification by sex, age, and area-level socioeconomic status (SES). We observed a linear increase in risk for asthma hospitalizations (RR=1.07, 95% CI=(1.05, 1.10) per 5µg/m(3) increase) and asthma ED visits (RR=1.06, 95% CI=(1.05, 1.07) per 5µg/m(3) increase) with increasing PM2.5 during the wildfires. ED visits for chronic obstructive pulmonary disease (COPD) were associated with PM2.5 during the fires (RR=1.02 (95% CI=(1.01, 1.04) per 5µg/m(3) increase) and this effect was significantly different from that found before the fires but not after. We did not find consistent effects of wildfire smoke on other health outcomes. The effect of PM2.5 during the wildfire period was more pronounced in women compared to men and in adults, ages 20-64, compared to children and adults 65 or older. We also found some effect modification by area-level median income for respiratory ED visits during the wildfires, with the highest effects observed in the ZIP codes with the lowest median income. Using a novel spatiotemporal exposure model, we found some evidence of differential susceptibility to exposure to wildfire smoke.