The independent and synergistic impacts of power outages and floods on hospital admissions for multiple diseases

The Impact of Power Outages on Cardiovascular Hospitalizations Among Medicare Fee-for-service Enrollees in New York State, 2017-2018

BACKGROUND

Power outages are common. They can result in exposure to extreme temperatures by shutting off temperature-controlling devices, and thereby also cause stress. Consequently, outages may precipitate cardiovascular disease (CVD)-related hospitalizations. We assessed this relationship among older adults.

METHODS

We leveraged 2017-2018 data from 245,452 New York State Medicare Fee-for-Service beneficiaries (65+ years) with 390,530 CVD hospitalizations. Using NY Department of Public Services data, we calculated total hours without power 1 day, 1-2 days, and 1-3 days before case and control periods, with an outage ZIP Code Tabulation Area (ZCTA)-hour defined based on ≥10% of customers in a ZCTA-hour without power in primary analyses. We used a case-crossover study design and ran conditional logistic regression to assess associations separately within each urbanicity level: New York City (NYC), non-NYC urban, and rural areas. We additionally stratified models by warm versus cool season, individual-level age and sex, and ZCTA-level socioeconomic factors. Secondarily, we considered emergency (n = 298,910) and nonemergency hospitalizations separately.

RESULTS

We generally observed null associations between power outages and all CVD hospitalizations across New York State and within subgroups. For example, in NYC, we observed a rate ratio of 1.05 (95% confidence interval: 0.85, 1.30) for each additional power outage hour 1 day prior.

CONCLUSIONS

The case-crossover design we used eliminated time-fixed confounding, but there were a limited number of exposed cases, limiting statistical power. Future studies should investigate co-occurring severe weather, span additional years, and evaluate other and broader geographic areas.

Hospitalization risks associated with floods in a multi-country study

Floods of unprecedented intensity and frequency have been observed. However, evidence regarding the impacts of floods on hospitalization remains limited. Here we collected daily hospitalization counts during 2000-2019 from 747 communities in Australia, Brazil, Canada, Chile, New Zealand, Taiwan, Thailand and Vietnam. For each community, flooded days were defined as days from the start dates to the end dates of flood events. Lag-response associations between flooded day and daily hospitalization risks were estimated for each community using a quasi-Poisson regression model with a distributed lag nonlinear function. The community-specific estimates were then pooled using a random-effects meta-analysis. Based on the pooled estimates, attributable fractions of hospitalizations due to floods were calculated. We found that hospitalization risks increased and persisted for up to 210 days after flood exposure, with the overall relative risks being 1.26 (95% confidence interval 1.15-1.38) for all causes, 1.35 (1.21-1.50) for cardiovascular diseases, 1.30 (1.13-1.49) for respiratory diseases, 1.26 (1.10-1.44) for infectious diseases, 1.30 (1.17-1.45) for digestive diseases, 1.11 (0.98-1.25) for mental disorders, 1.61 (1.39-1.86) for diabetes, 1.35 (1.21-1.50) for injury, 1.34 (1.21-1.48) for cancer, 1.34 (1.20-1.50) for nervous system disorders and 1.40 (1.22-1.60) for renal diseases. The associations were modified by climate types, flood severity, age, population density and socioeconomic status. Flood exposure contributed to hospitalizations by up to 0.27% from all causes. This study revealed that flood exposure was associated with increased all-cause and ten cause-specific hospitalization risks within up to 210 days after exposure.

Mapping the Vulnerability of Older-Adult Neighborhoods: An Ecological Study of New York State

We examined neighborhood-level demographic, economic, and social characteristics and food and health-services access to gauge the vulnerability of older-adult neighborhoods in New York State (NYS), which is understudied and is significant given the rapid aging of populations worldwide. We conducted descriptive ecological analyses using data from the American Community Survey, historical redlining maps, Social Capital Instruments, U.S. Department of Agriculture food access atlas, ESRI businesses, and Social Determinants of Health. We compared census tracts classified as having high and low levels of older-adult population; among those identified as high-older-adult neighborhoods, we then examined tracts with high and low levels of adult population living alone and in poverty. Our results showed that NYS neighborhoods with large shares of the older adult population are generally faring well in terms of their socioeconomic status, social capital, lack of social isolation, and health services access. However, the older-adult neighborhoods with larger shares of the population living alone and in poverty fare worse, living in areas with poorer socioeconomic status, lower social capital, and considered medically underserved. NYS older adult communities are projected to increase by 2030. Resources should be invested in such areas with vulnerable groups so populations may age in equitable and accessible communities.

Mortality and morbidity risks associated with floods: A systematic review and meta-analysis

BACKGROUND

Flood disasters are projected to increase in severity, duration, and frequency in the context of climate change, and the risks of mortality and morbidity may increase after floods, which will overwhelm health systems.

OBJECTIVE

This study aims to synthesize current epidemiological evidence about the impacts of floods on mortality and morbidity.

METHOD

After performing a systematic literature search from 2000 to 2023, we included studies involving human participants, with exposures of floods, and with outcomes of mortality or morbidity.

RESULTS

In total, 37 studies were included in evidence syntheses. Meta-analyses yielded an overall relative risk of 1.26 (95% confidence interval [CI]: 1.10, 1.46), 1.10 (1.08, 1.13), 1.11 (1.04, 1.20), and 1.38 (1.18, 1.62) for all-cause mortality and morbidities of overall gastrointestinal diseases, diarrhea diseases, and dysentery, respectively. Although meta-analyses were not conducted, evidence from at least three studies consistently supported that exposure to floods was associated with increased risks of malaria and respiratory diseases. The evidence for other outcomes was reported but either limited or uncertain.

CONCLUSION

This study suggests that exposure to floods is associated with increased risks of all-cause mortality and morbidities of overall gastrointestinal diseases, diarrhea diseases, dysentery, malaria, and respiratory diseases, while further research is urgently called.

The Joint Effects of Thunderstorms and Power Outages on Respiratory-Related Emergency Visits and Modifying and Mediating Factors of This Relationship

BACKGROUND

While limited studies have evaluated the health impacts of thunderstorms and power outages (POs) separately, few have assessed their joint effects. We aimed to investigate the individual and joint effects of both thunderstorms and POs on respiratory diseases, to identify disparities by demographics, and to examine the modifications and mediations by meteorological factors and air pollution.

METHODS

Distributed lag nonlinear models were used to examine exposures during three periods (i.e., days with both thunderstorms and POs, thunderstorms only, and POs only) in relation to emergency department visits for respiratory diseases (2005-2018) compared to controls (no thunderstorm/no PO) in New York State (NYS) while controlling for confounders. Interactions between thunderstorms and weather factors or air pollutants on health were assessed. The disparities by demographics and seasons and the mediative effects by particulate matter with aerodynamic diameter ≤ 2.5 μ m (PM 2.5 ) and relative humidity (RH) were also evaluated.

RESULTS

Thunderstorms and POs were independently associated with total and six subtypes of respiratory diseases in NYS [highest risk ratio (RR) = 1.12; 95% confidence interval (CI): 1.08, 1.17], but the impact was stronger when they co-occurred (highest RR = 1.44; 95% CI: 1.22, 1.70), especially during grass weed, ragweed, and tree pollen seasons. The stronger thunderstorm/PO joint effects were observed on chronic obstructive pulmonary diseases, bronchitis, and asthma (lasted 0-10 d) and were higher among residents who lived in rural areas, were uninsured, were of Hispanic ethnicity, were 6-17 or over 65 years old, and during spring and summer. The number of comorbidities was significantly higher by 0.299-0.782/case. Extreme cold/heat, high RH, PM 2.5 , and ozone concentrations significantly modified the thunderstorm-health effect on both multiplicative and additive scales. Over 35% of the thunderstorm effects were mediated by PM 2.5 and RH.

CONCLUSION

Thunderstorms accompanied by POs showed the strongest respiratory effects. There were large disparities in thunderstorm-health associations by demographics. Meteorological factors and air pollution levels modified and mediated the thunderstorm-health effects. https://doi.org/10.1289/EHP13237.

Impact of extreme weather events on healthcare utilization and mortality in the United States

Climate change is intensifying extreme weather events. Yet a systematic analysis of post-disaster healthcare utilization and outcomes for severe weather and climate disasters, as tracked by the US government, is lacking. Following exposure to 42 US billion-dollar weather disasters (severe storm, flood, flood/severe storm, tropical cyclone and winter storm) between 2011 and 2016, we used a difference-in-differences (DID) approach to quantify changes in the rates of emergency department (ED) visits, nonelective hospitalizations and mortality between fee-for-service Medicare beneficiaries in affected compared to matched control counties in post-disaster weeks 1, 1-2 and 3-6. Overall, disasters were associated with higher rates of ED utilization in affected counties in post-disaster week 1 (DID of 1.22% (95% CI, 0.20% to 2.25%; P < 0.020)) through week 2. Nonelective hospitalizations were unchanged. Mortality was higher in affected counties in week 1 (DID of 1.40% (95% CI, 0.08% to 2.74%; P = 0.037)) and persisted for 6 weeks. Counties with the greatest loss and damage experienced greater increases in ED and mortality rates compared to all affected counties. Thus, billion-dollar weather disasters are associated with excess ED visits and mortality in Medicare beneficiaries. Tracking these outcomes is important for adaptation that protects patients and communities, health system resilience and policy.