A framework for evaluating clinical artificial intelligence systems without ground-truth annotations.

A clinical artificial intelligence (AI) system is often validated on data withheld during its development. This provides an estimate of its performance upon future deployment on data in the wild; those currently unseen but are expected to be encountered in a clinical setting. However, estimating performance on data in the wild is complicated by distribution shift between data in the wild and withheld data and the absence of ground-truth annotations. Here, we introduce SUDO, a framework for evaluating AI systems on data in the wild. Through experiments on AI systems developed for dermatology images, histopathology patches, and clinical notes, we show that SUDO can identify unreliable predictions, inform the selection of models, and allow for the previously out-of-reach assessment of algorithmic bias for data in the wild without ground-truth annotations. These capabilities can contribute to the deployment of trustworthy and ethical AI systems in medicine.

A systematic approach towards missing lab data in electronic health records: A case study in non-small cell lung cancer and multiple myeloma.

Real-world data derived from electronic health records often exhibit high levels of missingness in variables, such as laboratory results, presenting a challenge for statistical analyses. We developed a systematic workflow for gathering evidence of different missingness mechanisms and performing subsequent statistical analyses. We quantify evidence for missing completely at random (MCAR) or missing at random (MAR), mechanisms using Hotelling's multivariate t-test, and random forest classifiers, respectively. We further illustrate how to apply sensitivity analyses using the not at random fully conditional specification procedure to examine changes in parameter estimates under missing not at random (MNAR) mechanisms. In simulation studies, we validated these diagnostics and compared analytic bias under different mechanisms. To demonstrate the application of this workflow, we applied it to two exemplary case studies with an advanced non-small cell lung cancer and a multiple myeloma cohort derived from a real-world oncology database. Here, we found strong evidence against MCAR, and some evidence of MAR, implying that imputation approaches that attempt to predict missing values by fitting a model to observed data may be suitable for use. Sensitivity analyses did not suggest meaningful departures of our analytic results under potential MNAR mechanisms; these results were also in line with results reported in clinical trials.

Considerations for the Use of Machine Learning Extracted Real-World Data to Support Evidence Generation: A Research-Centric Evaluation Framework.

A vast amount of real-world data, such as pathology reports and clinical notes, are captured as unstructured text in electronic health records (EHRs). However, this information is both difficult and costly to extract through human abstraction, especially when scaling to large datasets is needed. Fortunately, Natural Language Processing (NLP) and Machine Learning (ML) techniques provide promising solutions for a variety of information extraction tasks such as identifying a group of patients who have a specific diagnosis, share common characteristics, or show progression of a disease. However, using these ML-extracted data for research still introduces unique challenges in assessing validity and generalizability to different cohorts of interest. In order to enable effective and accurate use of ML-extracted real-world data (RWD) to support research and real-world evidence generation, we propose a research-centric evaluation framework for model developers, ML-extracted data users and other RWD stakeholders. This framework covers the fundamentals of evaluating RWD produced using ML methods to maximize the use of EHR data for research purposes.

Health Benefits of Strategies for Carbon Mitigation in US Transportation, 2017‒2050.

Objectives. To quantify health benefits and carbon emissions of 2 transportation scenarios that contrast optimum levels of physical activity from active travel and minimal air pollution from electric cars. Methods. We used data on burden of disease, travel, and vehicle emissions in the US population and a health impact model to assess health benefits and harms of physical activity from transportation-related walking and cycling, fine particulate pollution from car emissions, and road traffic injuries. We compared baseline travel with walking and cycling a median of 150 weekly minutes for physical activity, and with electric cars that minimized carbon pollution and fine particulates. Results. In 2050, the target year for carbon neutrality, the active travel scenario avoided 167 000 deaths and gained 2.5 million disability-adjusted life years, monetized at $1.6 trillion using the value of a statistical life. Carbon emissions were reduced by 24% from baseline. Electric cars avoided 1400 deaths and gained 16 400 disability-adjusted life years, monetized at $13 billion. Conclusions. To achieve carbon neutrality in transportation and maximize health benefits, active travel should have a prominent role along with electric vehicles in national blueprints. (Am J Public Health. 2022; 112(3):426-433. https://doi.org/10.2105/AJPH.2021.306600).

Environmental Injustice and Industrial Chicken Farming in Maryland.

Maryland's growing chicken industry, including concentrated animal feeding operations (CAFOs) and meat processing plants, raises a number of concerns regarding public health and environmental justice. Using hot spot analysis, we analyzed the totality of Maryland's CAFOs and meat processing plants and those restricted to the Eastern Shore to assess whether communities of color and/or low socioeconomic status communities disproportionately hosted these types of facilities at the census tract level. We used zero-inflated regression modeling to determine the strength of the associations between environmental justice variables and the location of CAFOs and meatpacking facilities at the State level and on the Eastern Shore. Hot spot analyses demonstrated that CAFO hot spots on the Eastern Shore were located in counties with some of the lowest wealth in the State, including the lowest ranking county-Somerset. Zero-inflated regression models demonstrated that increases in median household income across the state were associated with a 0.04-unit reduction in CAFOs. For every unit increase in the percentage of people of color (POC), there was a 0.02-unit increase in meat processing facilities across the state. The distribution of CAFOs and meat processing plants across Maryland may contribute to poor health outcomes in areas affected by such production, and contribute to health disparities and health inequity.

Climate change, extreme events, and increased risk of salmonellosis: foodborne diseases active surveillance network (FoodNet), 2004-2014.

BACKGROUND: Infections with nontyphoidal Salmonella cause an estimated 19,336 hospitalizations each year in the United States. Sources of infection can vary by state and include animal and plant-based foods, as well as environmental reservoirs. Several studies have recognized the importance of increased ambient temperature and precipitation in the spread and persistence of Salmonella in soil and food. However, the impact of extreme weather events on Salmonella infection rates among the most prevalent serovars, has not been fully evaluated across distinct U.S. regions. METHODS: To address this knowledge gap, we obtained Salmonella case data for S. Enteriditis, S. Typhimurium, S. Newport, and S. Javiana (2004-2014; n = 32,951) from the Foodborne Diseases Active Surveillance Network (FoodNet), and weather data from the National Climatic Data Center (1960-2014). Extreme heat and precipitation events for the study period (2004-2014) were identified using location and calendar day specific 95th percentile thresholds derived using a 30-year baseline (1960-1989). Negative binomial generalized estimating equations were used to evaluate the association between exposure to extreme events and salmonellosis rates. RESULTS: We observed that extreme heat exposure was associated with increased rates of infection with S. Newport in Maryland (Incidence Rate Ratio (IRR): 1.07, 95% Confidence Interval (CI): 1.01, 1.14), and Tennessee (IRR: 1.06, 95% CI: 1.04, 1.09), both FoodNet sites with high densities of animal feeding operations (e.g., broiler chickens and cattle). Extreme precipitation events were also associated with increased rates of S. Javiana infections, by 22% in Connecticut (IRR: 1.22, 95% CI: 1.10, 1.35) and by 5% in Georgia (IRR: 1.05, 95% CI: 1.01, 1.08), respectively. In addition, there was an 11% (IRR: 1.11, 95% CI: 1.04-1.18) increased rate of S. Newport infections in Maryland associated with extreme precipitation events. CONCLUSIONS: Overall, our study suggests a stronger association between extreme precipitation events, compared to extreme heat, and salmonellosis across multiple U.S. regions. In addition, the rates of infection with Salmonella serovars that persist in environmental or plant-based reservoirs, such as S. Javiana and S. Newport, appear to be of particular significance regarding increased heat and rainfall events.

Food insecurity and compound environmental shocks in Nepal: Implications for a changing climate.

Food insecurity is a key global health challenge that is likely to be exacerbated by climate change. Though climate change is associated with an increased frequency of extreme weather events, little is known about how multiple environmental shocks in close succession interact to impact household health and well-being. In this paper, we assess how earthquake exposure followed by monsoon rainfall anomalies affect food insecurity in Nepal. We link food security data from the 2016 Nepal Demographic and Health Survey to data on shaking intensity during the 2015 Gorkha earthquake and rainfall anomalies during the 2015 monsoon season. We then exploit spatial variation in exposure to the earthquake and monsoon rainfall anomalies to isolate their independent and compound effects. We find that earthquake exposure alone was not associated with an increased likelihood of food insecurity, likely due in part to effective food aid distribution. However, the effects of rainfall anomalies differed by severity of earthquake exposure. Among households minimally impacted by the earthquake, low rainfall was associated with increased food insecurity, likely due to lower agricultural productivity in drought conditions. Among households that experienced at least moderate shaking, greater rainfall was positively associated with food insecurity, particularly in steep, mountainous areas. In these locations, rainfall events disproportionately increased landslides, which damaged roads, disrupted distribution of food aid, and destroyed agricultural land and assets. Additional research on the social impacts of compound environmental shocks is needed to inform adaptation strategies that work to improve well-being in the face of climate change.

Enteric Viruses and Pepper Mild Mottle Virus Show Significant Correlation in Select Mid-Atlantic Agricultural Waters.

Enteric viruses (EVs) are the largest contributors to foodborne illnesses and outbreaks globally. Their ability to persist in the environment, coupled with the challenges experienced in environmental monitoring, creates a critical aperture through which agricultural crops may become contaminated. This study involved a 17-month investigation of select human EVs and viral indicators in nontraditional irrigation water sources (surface and reclaimed waters) in the Mid-Atlantic region of the United States. Real-time quantitative PCR was used for detection of Aichi virus, hepatitis A virus, and norovirus genotypes I and II (GI and GII, respectively). Pepper mild mottle virus (PMMoV), a common viral indicator of human fecal contamination, was also evaluated, along with atmospheric (air and water temperature, cloud cover, and precipitation 24 h, 7 days, and 14 days prior to sample collection) and physicochemical (dissolved oxygen, pH, salinity, and turbidity) data, to determine whether there were any associations between EVs and measured parameters. EVs were detected more frequently in reclaimed waters (32% [n = 22]) than in surface waters (4% [n = 49]), similar to PMMoV detection frequency in surface (33% [n = 42]) and reclaimed (67% [n = 21]) waters. Our data show a significant correlation between EV and PMMoV (R2 = 0.628, P < 0.05) detection levels in reclaimed water samples but not in surface water samples (R2 = 0.476, P = 0.78). Water salinity significantly affected the detection of both EVs and PMMoV (P < 0.05), as demonstrated by logistic regression analyses. These results provide relevant insights into the extent and degree of association between human (pathogenic) EVs and water quality data in Mid-Atlantic surface and reclaimed waters, as potential sources for agricultural irrigation. IMPORTANCE Microbiological analysis of agricultural waters is fundamental to ensure microbial food safety. The highly variable nature of nontraditional sources of irrigation water makes them particularly difficult to test for the presence of viruses. Multiple characteristics influence viral persistence in a water source, as well as affecting the recovery and detection methods that are employed. Testing for a suite of viruses in water samples is often too costly and labor-intensive, making identification of suitable indicators for viral pathogen contamination necessary. The results from this study address two critical data gaps, namely, EV prevalence in surface and reclaimed waters of the Mid-Atlantic region of the United States and subsequent evaluation of physicochemical and atmospheric parameters used to inform the potential for the use of indicators of viral contamination.

Global Population Exposed to Extreme Events in the 150 Most Populated Cities of the World: Implications for Public Health.

Climate change driven increases in the frequency of extreme heat events (EHE) and extreme precipitation events (EPE) are contributing to both infectious and non-infectious disease burden, particularly in urban city centers. While the share of urban populations continues to grow, a comprehensive assessment of populations impacted by these threats is lacking. Using data from weather stations, climate models, and urban population growth during 1980-2017, here, we show that the concurrent rise in the frequency of EHE, EPE, and urban populations has resulted in over 500% increases in individuals exposed to EHE and EPE in the 150 most populated cities of the world. Since most of the population increases over the next several decades are projected to take place in city centers within low- and middle-income countries, skillful early warnings and community specific response strategies are urgently needed to minimize public health impacts and associated costs to the global economy.

Impact of high precipitation and temperature events on the distribution of emerging contaminants in surface water in the Mid-Atlantic, United States.

Extreme weather events induced by climate change have potential to impact water quality and have received increasing attention from surface water source management perspectives. However, it remains unclear how such phenomenon may influence concentration of emerging contaminants (ECs) in surface water that are vital source of irrigation. In the present study, we investigated the impact of high precipitation and ambient temperature on the distribution of ECs in surface water samples (N = 250) from Mid-Atlantic region, collected between 2016 and 2018. We analyzed the water samples using a liquid chromatography tandem mass spectrometry (LC-MS/MS) based method. We then investigated how the detection frequencies and concentrations of ten emerging contaminants were influenced by high precipitation and temperature events in the previous day or 7 days prior to the sampling events using a generalized additive model (GAM). We observed that heavy rainfalls occurring within 24 h before sampling increased the concentration/likelihood of detection of the ECs in surface waters, likely due to surface runoffs, remobilization from soil/sediment and sewage overflows. The impact of high precipitation during previous seven days varied across chemicals. Likewise, the detection frequency and concentration of most analytes increased with increasing temperature, in previous day of sampling event, likely due to enhanced solubility in water. Long-term high temperature events appeared to decrease the detection of the most tested ECs probably due to enhanced degradation. However, the potential risk of unknown degradation products cannot be ignored. Our results indicate potential decline of water quality after extreme weather events which may have implications for water source management under changing climate.

Role of extreme weather events and El Niño Southern Oscillation on incidence of Enteric Fever in Ahmedabad and Surat, Gujarat, India.

INTRODUCTION: Enteric Fever (EF) affects over 14.5 million people every year globally, with India accounting for the largest share of this burden. The water-borne nature of the disease makes it prone to be influenced as much by unsanitary living conditions as by climatic factors. The detection and quantification of the climatic effect can lead to improved public health measures which would in turn reduce this burden. METHODOLOGY: We obtained a list of monthly Widal positive EF cases from 1995 to 2017 from Ahmedabad and Surat Municipalities. We obtained population data, daily weather data, and Oceanic Niño Index values from appropriate sources. We quantified the association between extreme weather events, phases of El Niño Southern Oscillations (ENSO) and incidence of EF. RESULTS: Both cities showed a seasonal pattern of EF, with cases peaking in early monsoon. Risk of EF was affected equally in both cities by the monsoon season -- Ahmedabad (35%) and Surat (34%). Extreme precipitation was associated with 5% increase in EF in Ahmedabad but not in Surat. Similarly, phases of ENSO had opposite effects on EF across the two cities. In Ahmedabad, strong El Niño months were associated with 64% increase in EF risk while strong La Niña months with a 41% reduction in risk. In Surat, strong El Niño was associated with 25% reduction in risk while moderate La Niña with 21% increase in risk. CONCLUSIONS: Our results show that the risk of EF incidence in Gujarat is highly variable, even between the two cities only 260 kms apart. In addition to improvements in water supply and sewage systems, preventive public health measures should incorporate variability in risk across season and phases of ENSO. Further studies are needed to characterize nationwide heterogeneity in climate-mediated risk, and to identify most vulnerable populations that can benefit through early warning systems.

Longitudinal Assessment of the Dynamics of Escherichia coli, Total Coliforms, Enterococcus spp., and Aeromonas spp. in Alternative Irrigation Water Sources: a CONSERVE Study.

As climate change continues to stress freshwater resources, we have a pressing need to identify alternative (nontraditional) sources of microbially safe water for irrigation of fresh produce. This study is part of the center CONSERVE, which aims to facilitate the adoption of adequate agricultural water sources. A 26-month longitudinal study was conducted at 11 sites to assess the prevalence of bacteria indicating water quality, fecal contamination, and crop contamination risk (Escherichia coli, total coliforms [TC], Enterococcus, and Aeromonas). Sites included nontidal freshwater rivers/creeks (NF), a tidal brackish river (TB), irrigation ponds (PW), and reclaimed water sites (RW). Water samples were filtered for bacterial quantification. E. coli, TC, enterococci (∼86%, 98%, and 90% positive, respectively; n = 333), and Aeromonas (∼98% positive; n = 133) were widespread in water samples tested. Highest E. coli counts were in rivers, TC counts in TB, and enterococci in rivers and ponds (P < 0.001 in all cases) compared to other water types. Aeromonas counts were consistent across sites. Seasonal dynamics were detected in NF and PW samples only. E. coli counts were higher in the vegetable crop-growing (May-October) than nongrowing (November-April) season in all water types (P < 0.05). Only one RW and both PW sites met the U.S. Food Safety Modernization Act water standards. However, implementation of recommended mitigation measures of allowing time for microbial die-off between irrigation and harvest would bring all other sites into compliance within 2 days. This study provides comprehensive microbial data on alternative irrigation water and serves as an important resource for food safety planning and policy setting.IMPORTANCE Increasing demands for fresh fruit and vegetables, a variable climate affecting agricultural water availability, and microbial food safety goals are pressing the need to identify new, safe, alternative sources of irrigation water. Our study generated microbial data collected over a 2-year period from potential sources of irrigation (rivers, ponds, and reclaimed water sites). Pond water was found to comply with Food Safety Modernization Act (FSMA) microbial standards for irrigation of fruit and vegetables. Bacterial counts in reclaimed water, a resource that is not universally allowed on fresh produce in the United States, generally met microbial standards or needed minimal mitigation. We detected the most seasonality and the highest microbial loads in river water, which emerged as the water type that would require the most mitigation to be compliant with established FSMA standards. This data set represents one of the most comprehensive, longitudinal analyses of alternative irrigation water sources in the United States.

Association Between Changes in Timing of Spring Onset and Asthma Hospitalization in Maryland.

Importance: Ongoing climate change is affecting the health of communities across the globe. While direct consequences, including morbidity and mortality tied to increases in the frequency of extreme weather events, have received significant attention, indirect health effects, particularly those associated with climate change-driven disruptions in ecosystems, are less understood. Objective: To investigate how ongoing changes in the timing of spring onset related to climate change are associated with rates of asthma hospitalization in Maryland. Design, Setting, and Participants: This cross-sectional study of 29 257 patients with asthma used general additive (quasi Poisson) and mixed-effect (negative binomial) models to investigate the association between changes in the timing of spring onset, detected using satellite observations, and the risk of asthma hospitalization in Maryland from 2001 to 2012. Data analysis was conducted from January 2016 to March 2019. Exposures: Phenology data, derived from the National Aeronautics and Space Administration's Moderate Resolution Imaging Spectroradiometer, were used to calculate location-specific median dates for start of season from 2001 to 2012. How the start of season for a given year and location deviated from the long-term average was calculated and categorized as very early, early, normal, or late. Main Outcomes and Measures: Daily asthma hospitalization in Maryland during the spring season (ie, March to May). Results: There were 108 358 total asthma hospitalizations during the study period, of which 29 257 (27.0%; 14 379 [49.1%] non-Hispanic black patients; 17 877 [61.1%] women) took place during springtime. In the unadjusted model, very early (incident rate ratio [IRR], 1.17; 95% CI, 1.07-1.28) and late (IRR, 1.07; 95% CI, 1.00-1.15) onset of spring were associated with increased risk of asthma hospitalization. When the analysis was adjusted for extreme heat events and concentrations of particulate matter with an aerodynamic diameter less than 2.5 µm, the risk remained significant for very early spring onset (IRR, 1.10; 95% CI, 1.02-1.20) but not for late spring onset (IRR, 1.03; 95% CI, 0.97-1.11). Conclusions and Relevance: These results suggest that ongoing changes in the timing of spring onset, which are related to climate variability and change, are associated with asthma hospitalization. Given the high burden of allergic diseases and the number of individuals sensitized to tree pollen, these findings serve as a wake-up call to public health and medical communities regarding the need to anticipate and adapt to the ongoing changes in the timing and severity of the spring allergy season.

Prevalence of Salmonella and Listeria monocytogenes in non-traditional irrigation waters in the Mid-Atlantic United States is affected by water type, season, and recovery method.

Irrigation water contaminated with Salmonella enterica and Listeria monocytogenes may provide a route of contamination of raw or minimally processed fruits and vegetables. While previous work has surveyed specific and singular types of agricultural irrigation water for bacterial pathogens, few studies have simultaneously surveyed different water sources repeatedly over an extended period of time. This study quantified S. enterica and L. monocytogenes levels (MPN/L) at 6 sites, including river waters: tidal freshwater river (MA04, n = 34), non-tidal freshwater river, (MA05, n = 32), one reclaimed water holding pond (MA06, n = 25), two pond water sites (MA10, n = 35; MA11, n = 34), and one produce wash water site (MA12, n = 10) from September 2016-October 2018. Overall, 50% (84/168) and 31% (53/170) of sampling events recovered S. enterica and L. monocytogenes, respectively. Results showed that river waters supported significantly (p < 0.05) greater levels of S. enterica than pond or reclaimed waters. The non-tidal river water sites (MA05) with the lowest water temperature supported significantly greater level of L. monocytogenes compared to all other sites; L. monocytogenes levels were also lower in winter and spring compared to summer seasons. Filtering 10 L of water through a modified Moore swab (MMS) was 43.5 (Odds ratio, p < 0.001) and 25.5 (p < 0.001) times more likely to recover S. enterica than filtering 1 L and 0.1 L, respectively; filtering 10 L was 4.8 (p < 0.05) and 3.9 (p < 0.05) times more likely to recover L. monocytogenes than 1L and 0.1 L, respectively. Work presented here shows that S. enterica and L. monocytogenes levels are higher in river waters compared to pond or reclaimed waters in the Mid-Atlantic region of the U.S., and quantitatively shows that analyzing 10 L water is more likely recover pathogens than smaller samples of environmental waters.

Association of Extreme Heat Events With Hospital Admission or Mortality Among Patients With End-Stage Renal Disease.

Importance: Extreme heat events (EHEs) are increasing in frequency, duration, and intensity, and this trend is projected to continue as part of ongoing climate change. There is a paucity of data regarding how EHEs may affect highly vulnerable populations, such as patients with end-stage renal disease (ESRD). Such data are needed to inform ESRD patient management guidelines in a changing climate. Objectives: To investigate the association between EHEs and the risk of hospital admission or mortality among patients with ESRD and further characterize how this risk may vary among races/ethnicities or patients with preexisting comorbidities. Design, Setting, and Participants: This study used hospital admission and mortality records of patients with ESRD who underwent hemodialysis treatment at Fresenius Kidney Care clinics in Boston, Massachusetts; Philadelphia, Pennsylvania; or New York, New York, from January 1, 2001, to December 31, 2012. Data were analyzed using a time-stratified case-crossover design with conditional Poisson regression to investigate associations between EHEs and risk of hospital admission or mortality among patients with ESRD. Data were analyzed from July 1, 2017, to March 31, 2019. Exposures: Calendar day- and location-specific 95th-percentile maximum temperature thresholds were calculated using daily meteorological data from 1960 to 1989. These thresholds were used to identify EHEs in each of the 3 cities during the study. Main Outcomes and Measures: Daily all-cause hospital admission and all-cause mortality among patients with ESRD. Results: The study included 7445 patients with ESRD (mean [SD] age, 61.1 [14.1] years; 4283 [57.5%] men), among whom 2953 deaths (39.7%) and 44 941 hospital admissions (mean [SD], 6.0 [7.5] per patient) were recorded. Extreme heat events were associated with increased risk of same-day hospital admission (rate ratio [RR], 1.27; 95% CI, 1.13-1.43) and same-day mortality (RR, 1.31; 95% CI, 1.01-1.70) among patients with ESRD. There was some heterogeneity in risk, with patients in Boston showing statistically significant increased risk for hospital admission (RR, 1.15; 95% CI, 1.00-1.31) and mortality (RR, 1.45; 95% CI, 1.04-2.02) associated with cumulative exposure to EHEs, while such risk was absent among patients with ESRD in Philadelphia. While increases in risks were similar among non-Hispanic black and non-Hispanic white patients, findings among Hispanic and Asian patients were less clear. After stratifying by preexisting comorbidities, cumulative lag exposure to EHEs was associated with increased risk of mortality among patients with ESRD living with congestive heart failure (RR, 1.55; 95% CI, 1.27-1.89), chronic obstructive pulmonary disease (RR, 1.60; 95% CI, 1.24-2.06), or diabetes (RR, 1.83; 95% CI, 1.51-2.21). Conclusions and Relevance: In this study, extreme heat events were associated with increased risk of hospital admission or mortality among patients with ESRD, and the association was potentially affected by geographic region and race/ethnicity. Future studies with larger populations and broader geographic coverage are needed to better characterize this variability in risk and inform ESRD management guidelines and differential risk variables, given the projected increases in the frequency, duration, and intensity of EHEs.

Using Google Street View to examine associations between built environment characteristics and U.S. health outcomes.

Neighborhood attributes have been shown to influence health, but advances in neighborhood research has been constrained by the lack of neighborhood data for many geographical areas and few neighborhood studies examine features of nonmetropolitan locations. We leveraged a massive source of Google Street View (GSV) images and computer vision to automatically characterize national neighborhood built environments. Using road network data and Google Street View API, from December 15, 2017-May 14, 2018 we retrieved over 16 million GSV images of street intersections across the United States. Computer vision was applied to label each image. We implemented regression models to estimate associations between built environments and county health outcomes, controlling for county-level demographics, economics, and population density. At the county level, greater presence of highways was related to lower chronic diseases and premature mortality. Areas characterized by street view images as 'rural' (having limited infrastructure) had higher obesity, diabetes, fair/poor self-rated health, premature mortality, physical distress, physical inactivity and teen birth rates but lower rates of excessive drinking. Analyses at the census tract level for 500 cities revealed similar adverse associations as was seen at the county level for neighborhood indicators of less urban development. Possible mechanisms include the greater abundance of services and facilities found in more developed areas with roads, enabling access to places and resources for promoting health. GSV images represents an underutilized resource for building national data on neighborhoods and examining the influence of built environments on community health outcomes across the United States.

Effects of heat on first-ever strokes and the effect modification of atmospheric pressure: A time-series study in Shenzhen, China.

BACKGROUND: Stroke is a leading cause of death globally. Extreme temperatures may induce stroke, but evidence on the effects of heat on first-ever strokes is not clear. Low air pressure can lead to depression and an increase in blood pressure, and it may exacerbate the health impact of heat. In this study, we aimed to evaluate the effects of heat on first-ever strokes, the possible sensitive populations, and the effect of modification of atmospheric pressure. METHODS: We collected data on 142,569 first-ever strokes during 2005-2016 in Shenzhen, a coastal city in southern China, with subtropical oceanic monsoon climate. We fitted a time-series Poisson model in our study, estimating the association between daily mean temperature and first-ever strokes in hot months, with a distributed lag non-linear model with 7 days of lag. We calculated strokes attributable to heat in various gender, age groups, household register types, stroke subtypes, and atmospheric pressure levels. RESULTS: Heat had a significant cumulative association with first-ever strokes, and the risk of strokes increased with the rise in temperature after it was higher than 30 °C (the 85th percentile). In total, 1.95% (95% empirical CI 0.63-3.20%) of first-ever strokes were attributable to high temperature. The attributable fraction and attributable number of heat were statistically significant in male, female, middle-aged and old patients, immigrant patients, and CBI patients. The fraction attributable to heat was 3.33% in the low atmospheric pressure group, and the number of estimated daily attributable strokes at low atmospheric pressure levels was higher than that of medium and high atmospheric pressure levels (p < 0.01). CONCLUSIONS: High temperatures in hot months may trigger first-ever strokes, and low atmospheric pressure may exacerbate the effect. We mainly found associations between heat and first-ever strokes for intracerebral hemorrhage, middle-aged and old patients, as well as immigrant patients.

Associations between alteration in plant phenology and hay fever prevalence among US adults: Implication for changing climate.

Plant phenology (e.g. timing of spring green-up, flowering) is among the most sensitive indicator of ecological response to ongoing climate variability and change. While previous studies have documented changes in the timing of spring green-up and flowering across different parts of the world, empirical evidence regarding how such ongoing ecological changes impact allergic disease burden at population level is lacking. Because earlier spring green-up may increase season length for tree pollen, we hypothesized that early onset of spring (negative anomaly in start of season (SOS)) will be associated with increased hay fever burden. To test this, we first calculated a median cardinal date for SOS for each county within the contiguous US for the years 2001-2013 using phenology data from the National Aeronautics and Space Administration's Moderate Resolution Imaging Spectroradiometer (MODIS). We categorized yearly deviations in SOS for each county from their respective long-term averages as: very early (>3 wks early), early (1-3 wks early), average (within 1 wk), late (1-3 wks late) and very late (> 3 wks late). We linked these data to 2002-2013 National Health Interview Survey data, and investigated the association between changes in SOS and hay fever prevalence using logistic regression. We observed that adults living in counties with a very early onset of SOS had a 14% higher odds of hay fever compared to the reference group, i.e. those living in counties where onset of spring was within the normal range (Odds Ratios (OR): 1.14. 95% Confidence Interval (CI): 1.03-1.27). Likewise, adults living in counties with very late onset of SOS had a 18% higher odds hay fever compared to the reference group (OR: 1.18, CI: 1.05-1.32). Our data provides the first-ever national scale assessment of the impact of changing plant phenology-linked to ongoing climate variability and change-on hay fever prevalence. Our findings are likely tied to changes in pollen dynamics, i.e early onset of spring increases the duration of exposure to tree pollen, while very late onset of spring increases the propensity of exposure because of simultaneous blooming.

Applying the concept of "number needed to treat" to the formulation of daily ambient air quality standards.

The World Health Organization sets up the Ambient Air Quality Guidelines mainly based on short-term and long-term health effects of air pollution. Previous studies, however, have generally revealed a non-threshold concentration-response relationship between air pollution and health, making it difficult to determine a concentration, below which no obvious health effects can be observed. Here we proposed a novel approach based on the concept of "number needed to treat", specifically, we calculated the reduction in air pollution concentrations needed to avoid one death corresponding to different hypothetical concentration standards; the one with the smallest value would be the most practical concentration standard. As an example, we applied this approach to the daily standard of ambient PM2.5 (particulate matter with aerodynamic diameter ≤2.5 µm) in four Chinese cities. The calculation was based on the association between daily mortality and ambient PM2.5, which was examined by a generalized additive model with adjustment of important covariates. Significant associations were observed between PM2.5 and mortality. Our analyses suggested that it is appropriate to have 50 µg/m3 as the daily standard of ambient PM2.5 for the study area, compared to the current standard of which were directly adopted from the national standard of 75 µg/m3. This novel approach should be considered when planning and/or revising the ambient air quality guidelines/standards.

Groundwater level changes with a focus on agricultural areas in the Mid-Atlantic region of the United States, 2002-2016.

Climate change impacts all water sources, including high quality groundwater that supplies agricultural irrigation in many regions of the United States. This study assessed groundwater level changes in the U.S. Mid-Atlantic region with a focus on cultivated areas. Trends of groundwater level were estimated using linear regression, and examined for shallow, medium, and deep depths across physiographic regions of Mid-Atlantic. A hotspot analysis was conducted to identify spatial clusters of wells with rising or declining groundwater levels. In addition, differences in the percentage of cultivated area with declining groundwater between cultivated land categories was examined at the county level. From 2002-2016, the Mid-Atlantic region had an overall decline in groundwater level (0.06 m/yr, 95% CI: 0.03, 0.09) although groundwater changes varied by physiographic regions. The Coastal Plain physiographic region was dominated by declining groundwater wells (48%) and had the most significant groundwater level declines (0.23 m/yr, 95% CI: 0.19, 0.26). Significant groundwater level rises were detected in Southern Virginia adjacent to the Chesapeake Bay (0.92 m/yr on average), which could be due to the cessation of groundwater withdrawal from one of the region's largest groundwater users. In the Mid-Atlantic region, shallow groundwater was found to have slight rising trends (0.08 m, p < 0.05) while deeper groundwater showed distinctive declining trends (1.36 m, p < 0.05) between 2002 and 2016. There were significantly more cultivated areas with declining groundwater levels (88% vs. 35%, p < 0.05) in counties with high percentages of cropland (> 50%) compared to areas covered by less cropland. As climate and human pressures increase, it will be critical to identify and evaluate alternative water sources, such as reclaimed water, to sustain agricultural production and protect groundwater resources.