Perceptions of weather-based pain forecasts and their effect on daily activities.

As studies begin to have more success uncovering the relationships between atmospheric conditions and pain, weather-based pain forecasting becomes more of a reality. In this study, a survey was used to determine if people living with migraines and/or other pain-related conditions are receptive to weather-based pain forecasts. Moreover, we wished to identify whether these forecasts actually impact the decision-making of those who use them. Survey respondents were generally eager to use these novel forecasts. Furthermore, when provided with different scenarios involving weather-based pain forecasts, the respondents' actions were altered. When a hypothetical forecast indicated that the weather was conducive to migraines or other types of pain, many indicated that they would likely take preventative measures (e.g., medication). Additionally, respondents were less likely to continue with a planned activity, regardless of length, as forecast severity increased. The results from this survey highlight the importance of developing and improving weather-based pain forecasting.

Projection of extreme heat- and cold-related mortality in Sweden based on the spatial synoptic classification.

BACKGROUND: Climate change is projected to result in increased heat events and decreased cold events. This will substantially impact human health, particularly when compounded with demographic change. This study employed the Spatial Synoptic Classification (SSC) to categorize daily weather into one of seven types. Here we estimated future mortality due to extremely hot and cold weather types under different climate change scenarios for one southern (Stockholm) and one northern (Jämtland) Swedish region. METHODS: Time-series Poisson regression with distributed lags was used to assess the relationship between extremely hot and cold weather events and daily deaths in the population above 65 years, with cumulative effects (6 days in summer, 28 days in winter), 1991 to 2014. A global climate model (MPI-M-MPI-ESM-LR) and two climate change scenarios (RCP 4.5 and 8.5) were used to project the occurrence of hot and cold days from 2031 to 2070. Place-specific projected mortality was calculated to derive attributable numbers and attributable fractions (AF) of heat- and cold-related deaths. RESULTS: In Stockholm, for the RCP 4.5 scenario, the mean number of annual deaths attributed to heat increased from 48.7 (CI 32.2-64.2; AF = 0.68%) in 2031-2040 to 90.2 (56.7-120.5; AF = 0.97%) in 2061-2070, respectively. For RCP 8.5, heat-related deaths increased more drastically from 52.1 (33.6-69.7; AF = 0.72%) to 126.4 (68.7-175.8; AF = 1.36%) between the first and the last decade. Cold-related deaths slightly increased over the projected period in both scenarios. In Jämtland, projections showed a small decrease in cold-related deaths but no change in heat-related mortality. CONCLUSIONS: In rural northern region of Sweden, a decrease of cold-related deaths represents the dominant trend. In urban southern locations, on the other hand, an increase of heat-related mortality is to be expected. With an increasing elderly population, heat-related mortality will outweigh cold-related mortality at least under the RCP 8.5 scenario, requiring societal adaptation measures.

Mortality risk attributable to high and low ambient temperature in Pune city, India: A time series analysis from 2004 to 2012.

BACKGROUND: Exposure to high and low ambient temperatures is associated with morbidity and mortality across the globe. Most of these studies assessing the effects of non-optimum temperatures on health and have been conducted in the developed world, whereas in India, the limited evidence on ambient temperature and health risks and has focused mostly on the effects of heat waves. Here we quantify short term association between all temperatures and mortality in urban Pune, India. METHODS: We applied a time series regression model to derive temperature-mortality associations based on daily mean temperature and all-cause mortality records of Pune city from year January 2004 to December 2012. We estimated high and low temperature-mortality relationships by using standard time series quasi-Poisson regression in conjunction with a distributed lag non-linear model (DLNM). We calculated temperature attributable mortality fractions for total heat and total cold. FINDINGS: The analysis provides estimates of the total mortality burden attributable to ambient temperature. Overall, 6∙5% [95%CI 1.76-11∙43] of deaths registered in the observational period were attributed to non-optimal temperatures, cold effect was greater 5.72% [95%CI 0∙70-10∙06] than heat 0∙84% [0∙35-1∙34]. The gender stratified analysis revealed that the highest burden among men both for heat and cold. CONCLUSION: Non-optimal temperatures are associated with a substantial mortality burden. Our findings could benefit national, and local communities in developing preparedness and prevention strategies to reduce weather-related impacts immediately due to climate change.

Geographical variability in the relationship between synoptic weather type and emergency department visits for pain across North Carolina.

Bodily pain plagues populations across the globe. Past studies have discovered some links between synoptic weather types and different kinds of pain. These relationships are essential as they can aide in treatment and potentially prevention of pain. In this study, the role of geographical characteristics on the relationships between synoptic weather type and pain were looked at. North Carolina was separated into three geographic sections: Appalachian Mountains, Piedmont Plateau, and Coastal Plain. Over a 7-year period, synoptic weather types and emergency department (ED) visits for various kinds of pain (migraine, fibromyalgia, rheumatoid arthritis, osteoarthritis, and general back pain) were collected. Bootstrapped confidence intervals of the mean number of population-adjusted ED visit rates (per 100,000 persons), for the different synoptic weather types, were compared across the different geographic regions. In the plateau region, Moist Tropical and Moist Moderate weather types were often linked to the highest rates of ED visits, while Polar weather types were frequently associated with the fewest visits. The mountainous portion of the state displayed similar patterns between synoptic weather types and the different forms of pain, with migraine and fibromyalgia being the exceptions. Few statistically significant relationships were noted for the coastal region.

Effect of extreme hot and cold weather on cause-specific hospitalizations in Sweden: A time series analysis.

Considering that several meteorological variables can contribute to weather vulnerability, the estimation of their synergetic effects on health is particularly useful. The spatial synoptic classification (SSC) has been used in biometeorological applications to estimate the effect of the entire suite of weather conditions on human morbidity and mortality. In this study, we assessed the relationships between extremely hot and dry (dry tropical plus, DT+) and hot and moist (moist tropical plus, MT+) weather types in summer and extremely cold and dry (dry polar plus, DP+) and cold and moist (moist polar, MP+) weather types in winter and cardiovascular and respiratory hospitalizations by age and sex. Time-series quasi-Poisson regression with distributed lags was used to assess the relationship between oppressive weather types and daily hospitalizations over 14 subsequent days in the extended summer (May to August) and 28 subsequent days during the extended winter (November to March) over 24 years in 4 Swedish locations from 1991 to 2014. In summer, exposure to hot weather types appeared to reduce cardiovascular hospitalizations while increased the risk of hospitalizations for respiratory diseases, mainly related to MT+. In winter, the effect of cold weather on both cause-specific hospitalizations was small; however, MP+ was related to a delayed increase in cardiovascular hospitalizations, whilst MP+ and DP + increased the risk of hospitalizations due to respiratory diseases. This study provides useful information for the staff of hospitals and elderly care centers who can help to implement protective measures for patients and residents. Also, our results could be helpful for vulnerable people who can adopt protective measures to reduce health risks.

Hot and cold weather based on the spatial synoptic classification and cause-specific mortality in Sweden: a time-stratified case-crossover study.

The spatial synoptic classification (SSC) is a holistic categorical assessment of the daily weather conditions at specific locations; it is a useful tool for assessing weather effects on health. In this study, we assessed (a) the effect of hot weather types and the duration of heat events on cardiovascular and respiratory mortality in summer and (b) the effect of cold weather types and the duration of cold events on cardiovascular and respiratory mortality in winter. A time-stratified case-crossover design combined with a distributed lag nonlinear model was carried out to investigate the association of weather types with cause-specific mortality in two southern (Skåne and Stockholm) and two northern (Jämtland and Västerbotten) locations in Sweden. During summer, in the southern locations, the Moist Tropical (MT) and Dry Tropical (DT) weather types increased cardiovascular and respiratory mortality at shorter lags; both hot weather types substantially increased respiratory mortality mainly in Skåne. The impact of heat events on mortality by cardiovascular and respiratory diseases was more important in the southern than in the northern locations at lag 0. The cumulative effect of MT, DT and heat events lagged over 14 days was particularly high for respiratory mortality in all locations except in Jämtland, though these did not show a clear effect on cardiovascular mortality. During winter, the dry polar and moist polar weather types and cold events showed a negligible effect on cardiovascular and respiratory mortality. This study provides valuable information about the relationship between hot oppressive weather types with cause-specific mortality; however, the cold weather types may not capture sufficiently effects on cause-specific mortality in this sub-Arctic region.

Relationship between synoptic weather type and emergency department visits for different types of pain across the Triangle region of North Carolina.

Many people around the world are impacted by some form of bodily pain. Outside factors, such as weather, are thought to help trigger pain, especially in those who have pain-related conditions. When it comes to human health and comfort, understanding the potential external factors that aide in triggering pain is essential. Identifying such factors makes prevention and treatment of pain more feasible. This study focused on how those who suffer from various pain-related conditions (fibromyalgia, rheumatoid arthritis, osteoarthritis, and general back pain) are impacted by different synoptic weather types (i.e., air masses). Synoptic weather types and emergency department (ED) visits for pain in select central North Carolina counties were collected over a seven-year period to determine a potential relationship. Bootstrapped confidence intervals revealed that moist tropical weather types resulted in the highest number of ED visits for each of the conditions examined, while moist polar weather types often resulted in the fewest. The barometric pressure changes associated with transitional weather types, which are often associated with fronts, did not have any significant relationships with pain.

A new approach to modeling temperature-related mortality: Non-linear autoregressive models with exogenous input.

Temperature-mortality relationships are nonlinear, time-lagged, and can vary depending on the time of year and geographic location, all of which limits the applicability of simple regression models in describing these associations. This research demonstrates the utility of an alternative method for modeling such complex relationships that has gained recent traction in other environmental fields: nonlinear autoregressive models with exogenous input (NARX models). All-cause mortality data and multiple temperature-based data sets were gathered from 41 different US cities, for the period 1975-2010, and subjected to ensemble NARX modeling. Models generally performed better in larger cities and during the winter season. Across the US, median absolute percentage errors were 10% (ranging from 4% to 15% in various cities), the average improvement in the r-squared over that of a simple persistence model was 17% (6-24%), and the hit rate for modeling spike days in mortality (>80th percentile) was 54% (34-71%). Mortality responded acutely to hot summer days, peaking at 0-2 days of lag before dropping precipitously, and there was an extended mortality response to cold winter days, peaking at 2-4 days of lag and dropping slowly and continuing for multiple weeks. Spring and autumn showed both of the aforementioned temperature-mortality relationships, but generally to a lesser magnitude than what was seen in summer or winter. When compared to distributed lag nonlinear models, NARX model output was nearly identical. These results highlight the applicability of NARX models for use in modeling complex and time-dependent relationships for various applications in epidemiology and environmental sciences.

Mortality risks during extreme temperature events (ETEs) using a distributed lag non-linear model.

This study investigates the relationship between all-cause mortality and extreme temperature events (ETEs) from 1975 to 2004. For 50 U.S. locations, these heat and cold events were defined based on location-specific thresholds of daily mean apparent temperature. Heat days were defined by a 3-day mean apparent temperature greater than the 95th percentile while extreme heat days were greater than the 97.5th percentile. Similarly, calculations for cold and extreme cold days relied upon the 5th and 2.5th percentiles. A distributed lag non-linear model assessed the relationship between mortality and ETEs for a cumulative 14-day period following exposure. Subsets for season and duration effect denote the differences between early- and late-season as well as short and long ETEs. While longer-lasting heat days resulted in elevated mortality, early season events also impacted mortality outcomes. Over the course of the summer season, heat-related risk decreased, though prolonged heat days still had a greater influence on mortality. Unlike heat, cold-related risk was greatest in more southerly locations. Risk was highest for early season cold events and decreased over the course of the winter season. Statistically, short episodes of cold showed the highest relative risk, suggesting unsettled weather conditions may have some relationship to cold-related mortality. For both heat and cold, results indicate higher risk to the more extreme thresholds. Risk values provide further insight into the role of adaptation, geographical variability, and acclimatization with respect to ETEs.

Effects of heat waves on daily excess mortality in 14 Korean cities during the past 20 years (1991-2010): an application of the spatial synoptic classification approach.

The aims of this study are to explore the "offensive" summer weather types classified under the spatial synoptic classification (SSC) system and to evaluate their impacts on excess mortality in 14 Korean cities. All-cause deaths per day for the entire population were examined over the summer months (May-September) of 1991-2010. Daily deaths were standardized to account for long-term trends of subcycles (annual, seasonal, and weekly) at the mid-latitudes. In addition, a mortality prediction model was constructed through multiple stepwise regression to develop a heat-health warning system based on synoptic climatology. The result showed that dry tropical (DT) days during early summer caused excess mortality due to non-acclimatization by inhabitants, and moist tropical (MT) plus and double plus resulted in greater spikes of excess mortality due to extremely hot and humid conditions. Among the 14 Korean cities, highly excess mortality for the elderly was observed in Incheon (23.2%, 95%CI 5.6), Seoul (15.8%, 95%CI 2.6), and Jeonju (15.8%, 95%CI 4.6). No time lag effect was observed, and excess mortality gradually increased with time and hot weather simultaneously. The model showed weak performance as its predictions were underestimated for the validation period (2011-2015). Nevertheless, the results clearly revealed the efficiency of relative and multiple-variable approaches better than absolute and single-variable approaches. The results indicate the potential of the SSC as a suitable system for investigating heat vulnerability in South Korea, where hot summers could be a significant risk factor.

Surveying Local Health Departments and County Emergency Management Offices on Cooling Centers as a Heat Adaptation Resource in New York State.

Local agencies in New York State (NYS) set up cooling centers to provide relief from summer-time heat especially for people with limited access to air-conditioning. We aimed to determine cooling center locations in NYS, and explore county agencies' involvement in organizing and promoting utilization of cooling centers. We conducted a survey among county health and emergency preparedness offices in NYS (excluding NYC) and explored official county websites. We identified 377 cooling centers, mostly in metropolitan areas of NYS. Although 47 % of counties listed locations online, only 29 % reported locations via survey. Radio (90 %) and internet (84 %) were popular for information dissemination. Air-conditioning was available at all indoor cooling center facilities. Cooling centers in 13 % of the counties were accessible by either public transportation or shuttles arranged by the facility. About 38 % counties do not consider cooling centers important in their region or promote informal cooling centers. More than a third of New York counties had neither cooling centers nor plans to establish a cooling center as extreme heat was not perceived as a threat in their region.

Sixty years of the International Journal of Biometeorology.

The International Journal of Biometeorology (IJB) has continuously evolved since its first publications in 1957. In this paper, we examine these changes using a database that includes all manuscript titles and author information. A brief history considers the development of the journal and shifts over time. With an interdisciplinary focus, publications draw on a wide array of subdisciplines. Using content analysis, we evaluate the themes found within IJB. Some research themes have maintained prominence throughout the journal's history, while other themes have waxed or waned over time. Similarly, the most influential manuscripts throughout the past 60 years reveal that human biometeorological papers, particularly regarding thermal comfort, have been influential throughout the journal's history, with other themes, including phenology and animal biometeorology, more concentrated in specific periods. Dominated by North America and Europe in the early years, publication authorship has shifted over the last decade to be more globally representative. Recent inclusion of special issues devoted to regional biometeorological issues, as well as to Students and New Professionals, offer insight into the future direction of the IJB.

High-mortality days during the winter season: comparing meteorological conditions across 5 US cities.

While the relationship between weather and human health has been studied from various perspectives, this study examines an alternative method of analysis by examining weather conditions on specific high-mortality days during the winter season. These high-mortality days, by definition, represent days with dramatic increases in mortality and the days with the highest mortality. By focusing solely on high-mortality days, this research examines the relationship between weather variables and mortality through a synoptic climatology, environment-to circulation approach. The atmospheric conditions during high-mortality days were compared to the days prior and the days not classified as high-mortality days. Similar patterns emerged across all five locations despite the spatial and temporal variability. Southern locations had a stronger relationship with temperature changes while northern locations showed a greater relationship to atmospheric pressure. Overall, all high-mortality days were associated with warmer temperatures, decreased pressure, and a greater likelihood of precipitation when compared to the previous subset of days. While the atmospheric conditions were consistent across all locations, the importance of the lag effect should not be overlooked as a contributing factor to mortality during the winter season. Through a variety of diverse, methodological approaches, future studies may build upon these results and explore in more detail the complex relationship between weather situations and the impact of short-term changes in weather and health outcomes.

Modeling complex effects of exposure to particulate matter and extreme heat during pregnancy on congenital heart defects: A U.S. population-based case-control study in the National Birth Defects Prevention Study.

BACKGROUND/OBJECTIVE: Research suggests gestational exposure to particulate matter ≤2.5 µm (PM2.5) and extreme heat may independently increase risk of birth defects. We investigated whether duration of gestational extreme heat exposure modifies associations between PM2.5 exposure and specific congenital heart defects (CHDs). We also explored nonlinear exposure-outcome relationships. METHODS: We identified CHD case children (n = 2824) and non-malformed live-birth control children (n = 4033) from pregnancies ending between 1999 and 2007 in the National Birth Defects Prevention Study, a U.S. population-based multicenter case-control study. We assigned mothers 6-week averages of PM2.5 exposure during the cardiac critical period (postconceptional weeks 3-8) using the closest monitor within 50 km of maternal residence. We assigned a count of extreme heat days (EHDs, days above the 90th percentile of daily maximum temperature for year, season, and weather station) during this period using the closest weather station. Using generalized additive models, we explored logit-nonlinear exposure-outcome relationships, concluding logistic models were reasonable. We estimated joint effects of PM2.5 and EHDs on six CHDs using logistic regression models adjusted for mean dewpoint and maternal age, education, and race/ethnicity. We assessed multiplicative and additive effect modification. RESULTS: Conditional on the highest observed EHD count (15) and at least one critical period day during spring/summer, each 5 µg/m3 increase in average PM2.5 exposure was significantly associated with perimembranous ventricular septal defects (VSDpm; OR: 1.54 [95% CI: 1.01, 2.41]). High EHD counts (8+) in the same population were positively, but non-significantly, associated with both overall septal defects and VSDpm. Null or inverse associations were observed for lower EHD counts. Multiplicative and additive effect modification estimates were consistently positive in all septal models. CONCLUSIONS: Results provide limited evidence that duration of extreme heat exposure modifies the PM2.5-septal defects relationship. Future research with enhanced exposure assessment and modeling techniques could clarify these relationships.

How neighborhood environment modified the effects of power outages on multiple health outcomes in New York state?

Background: Although power outage (PO) is one of the most important consequences of increasing weather extremes and the health impact of POs has been reported previously, studies on the neighborhood environment underlying the population vulnerability in such situations are limited. This study aimed to identify dominant neighborhood environmental predictors which modified the impact of POs on multiple health outcomes in New York State. Methods: We applied a two-stage approach. In the first stage, we used time series analysis to determine the impact of POs (versus non-PO periods) on multiple health outcomes in each power operating division in New York State, 2001-2013. In the second stage, we classified divisions as risk-elevated and non-elevated, then developed predictive models for the elevation status based on 36 neighborhood environmental factors using random forest and gradient boosted trees. Results: Consistent across different outcomes, we found predictors representing greater urbanization, particularly, the proportion of residents having access to public transportation (importance ranging from 4.9-15.6%), population density (3.3-16.1%), per capita income (2.3-10.7%), and the density of public infrastructure (0.8-8.5%), were associated with a higher possibility of risk elevation following power outages. Additionally, the percent of minority (-6.3-27.9%) and those with limited English (2.2-8.1%), the percent of sandy soil (6.5-11.8%), and average soil temperature (3.0-15.7%) were also dominant predictors for multiple outcomes. Spatial hotspots of vulnerability generally were located surrounding New York City and in the northwest, the pattern of which was consistent with socioeconomic status. Conclusion: Population vulnerability during power outages was dominated by neighborhood environmental factors representing greater urbanization.

The individual and synergistic impacts of windstorms and power outages on injury ED visits in New York State.

BACKGROUND: There is little work in assessing the impact of storm events combined with power outage (PO). In this study, we evaluated the individual and synergistic impacts of wind events and PO on overall and subtypes of injuries in New York State (NYS) and by demographics. METHODS: The emergency department (ED) visit data were obtained from the NYS Department of Health from November-April 2005-2013 to identify injury cases, length of stay and care costs. Wind event was defined according to high wind, strong wind or thunderstorm wind defined by NOAA. PO occurrence was defined when PO coverage exceeded the 50th percentile of its distribution. By comparing non-event days, we used distributed lag nonlinear models to evaluate the impacts of wind events, PO, and their combined effect on injuries during the cold season over a 0-3-day lag period, while controlling for time-varying confounders. The differences in critical care indicators between event and non-event days were also evaluated. RESULTS: Overall injuries ED visits (16,628,812) significantly increased during the wind events (highest Risk Ratio (RR): 1.05; 95% CI: 1.02-1.08), and were highest when wind events cooccurred with PO (highest RR: 1.14; 95% CI: 1.10-1.18), but not during PO alone (RR: 1.00; 95%CI: 0.96-1.04). The increase was also observed with all subgroups through Day 2 after the event. Greater risks exist for older adults (≥65 years) and those on Medicaid. After the joint occurrences of wind events and PO, average visits are 0.2 days longer, and cost 13% more, compared to no wind/no PO days. CONCLUSION: There is a significant increase in ED visits, length of stay and cost of injuries during wind events, especially when they coupled with PO and especially among older cases and Medicaid holders. Our findings may be used for planning disaster preparedness and recovery efforts.

Heat-health warning systems: a comparison of the predictive capacity of different approaches to identifying dangerously hot days.

OBJECTIVES: We compared the ability of several heat-health warning systems to predict days of heat-associated mortality using common data sets. METHODS: Heat-health warning systems initiate emergency public health interventions once forecasts have identified weather conditions to breach predetermined trigger levels. We examined 4 commonly used trigger-setting approaches: (1) synoptic classification, (2) epidemiologic assessment of the temperature-mortality relationship, (3) temperature-humidity index, and (4) physiologic classification. We applied each approach in Chicago, Illinois; London, United Kingdom; Madrid, Spain; and Montreal, Canada, to identify days expected to be associated with the highest heat-related mortality. RESULTS: We found little agreement across the approaches in which days were identified as most dangerous. In general, days identified by temperature-mortality assessment were associated with the highest excess mortality. CONCLUSIONS: Triggering of alert days and ultimately the initiation of emergency responses by a heat-health warning system varies significantly across approaches adopted to establish triggers.

How community vulnerability factors jointly affect multiple health outcomes after catastrophic storms.

BACKGROUND: While previous studies uncovered individual vulnerabilities to health risks during catastrophic storms, few evaluated the population vulnerability which is more important for identifying areas in greatest need of intervention. OBJECTIVES: We assessed the association between community factors and multiple health outcomes, and developed a community vulnerability index. METHODS: We retained emergency department visits for several health conditions from the 2005-2014 New York Statewide Planning and Research Cooperative System. We developed distributed lag nonlinear models at each spatial cluster across eight counties in downstate New York to evaluate the health risk associated with Superstorm Sandy (10/28/2012-11/9/2012) compared to the same period in other years, then defined census tracts in clusters with an elevated risk as "risk-elevated communities", and all others as "unelevated". We used machine-learning techniques to regress the risk elevation status against community factors to determine the contribution of each factor on population vulnerability, and developed a community vulnerability index (CVI). RESULTS: Overall, community factors had positive contributions to increased community vulnerabilities to Sandy-related substance abuse (91.35%), injuries (70.51%), cardiovascular diseases (8.01%), and mental disorders (2.71%) but reversely contributed to respiratory diseases (-34.73%). The contribution of low per capita income (max: 22.08%), the percentage of residents living in group quarters (max: 31.39%), the percentage of areas prone to flooding (max: 38.45%), and the percentage of green coverage (max: 29.73%) tended to be larger than other factors. The CVI based on these factors achieved an accuracy of 0.73-0.90 across outcomes. CONCLUSIONS: Our findings suggested that substance abuse was the most sensitive disease susceptible to less optimal community indicators, whereas respiratory diseases were higher in communities with better social environment. The percentage of residents in group quarters and areas prone to flooding were among dominant predictors for community vulnerabilities. The CVI based on these factors has an appropriate predictive performance.

Power Outage: An Ignored Risk Factor for COPD Exacerbations.

BACKGROUND: COPD is the third leading cause of death in the United States, with 16 million Americans currently experiencing difficulty with breathing. Power outages could be life-threatening for those relying on electricity. However, significant gaps remain in understanding the potential impact of power outages on COPD exacerbations. RESEARCH QUESTION: The goal of this study was to determine how power outages affect COPD exacerbations. STUDY DESIGN AND METHODS: Using distributed lag nonlinear models controlling for time-varying confounders, the hospitalization rate during a power outage was compared vs non-outage periods to determine the rate ratio (RR) for COPD and its subtypes at each of 0 to 6 lag days in New York State from 2001 to 2013. Stratified analyses were conducted according to sociodemographic characteristics, season, and clinical severity; changes were investigated in numerous critical medical indicators, including length of stay, hospital cost, the number of comorbidities, and therapeutic procedures between the two periods. RESULTS: The RR of COPD hospitalization following power outages ranged from 1.03 to 1.39 across lag days. The risk was strongest at lag0 and lag1 days and lasted significantly for 7 days. Associations were stronger for the subgroup with acute bronchitis (RR, 1.08-1.69) than for cases of acute exacerbation (RR, 1.03-1.40). Compared with non-outage periods, the outage period was observed to be $4.67 thousand greater in hospital cost and 1.38 greater in the number of comorbidities per case. The average cost (or number of comorbidities) was elevated in all groups stratified according to cost (or number of comorbidities). In contrast, changes in the average length of stay (-0.43 day) and the average number of therapeutic procedures (-0.09) were subtle. INTERPRETATION: Power outages were associated with a significantly elevated rate of COPD hospitalization, as well as greater costs and number of comorbidities. The average cost and number of comorbidities were elevated in all clinical severity groups.