Distribution of the environmental and socioeconomic risk factors on COVID-19 death rate across continental USA: a spatial nonlinear analysis

Geographically weighted machine learning for modeling spatial heterogeneity in traffic crash frequency and determinants in US

Spatial analyses of traffic crashes have drawn much interest due to the nature of the spatial dependence and spatial heterogeneity in the crash data. This study makes the best of Geographically Weighted Random Forest (GW-RF) model to explore the local associations between crash frequency and various influencing factors in the US, including road network attributes, socio-economic characteristics, and land use factors collected from multiple data sources. Special emphasis is put on modeling the spatial heterogeneity in the effects of a factor on crash frequency in different geographical areas in a data-driven way. The GW-RF model outperforms global models (e.g. Random Forest) and conventional geographically weighted regression, demonstrating superior predictive accuracy and elucidating spatial variations. The GW-RF model reveals spatial distinctions in the effects of certain factors on crash frequency. For example, the importance of intersection density varies significantly across regions, with high significance in the southern and northeastern areas. Low-grade road density emerges as influential in specific cities. The findings highlight the significance of different factors in influencing crash frequency across zones. Road network factors, particularly intersection density, exhibit high importance universally, while socioeconomic variables demonstrate moderate effects. Interestingly, land use variables show relatively lower importance. The outcomes could help to allocate resources and implement tailored interventions to reduce the likelihood of crashes.

Exploring spatiotemporal pattern in the association between short-term exposure to fine particulate matter and COVID-19 incidence in the continental United States: a Leroux-conditional-autoregression-based strategy

Background

Numerous studies have demonstrated that fine particulate matter (PM2.5) is adversely associated with COVID-19 incidence. However, few studies have explored the spatiotemporal heterogeneity in this association, which is critical for developing cost-effective pollution-related policies for a specific location and epidemic stage, as well as, understanding the temporal change of association between PM2.5 and an emerging infectious disease like COVID-19.

Methods

The outcome was state-level daily COVID-19 cases in 49 native United States between April 1, 2020 and December 31, 2021. The exposure variable was the moving average of PM2.5 with a lag range of 0-14 days. A latest proposed strategy was used to investigate the spatial distribution of PM2.5-COVID-19 association in state level. First, generalized additive models were independently constructed for each state to obtain the rough association estimations, which then were smoothed using a Leroux-prior-based conditional autoregression. Finally, a modified time-varying approach was used to analyze the temporal change of association and explore the potential causes spatiotemporal heterogeneity.

Results

In all states, a positive association between PM2.5 and COVID-19 incidence was observed. Nearly one-third of these states, mainly located in the northeastern and middle-northern United States, exhibited statistically significant. On average, a 1 μg/m3 increase in PM2.5 concentration led to an increase in COVID-19 incidence by 0.92% (95%CI: 0.63-1.23%). A U-shaped temporal change of association was examined, with the strongest association occurring in the end of 2021 and the weakest association occurring in September 1, 2020 and July 1, 2021. Vaccination rate was identified as a significant cause for the association heterogeneity, with a stronger association occurring at a higher vaccination rate.

Conclusion

Short-term exposure to PM2.5 and COVID-19 incidence presented positive association in the United States, which exhibited a significant spatiotemporal heterogeneity with strong association in the eastern and middle regions and with a U-shaped temporal change.

What influence farmers' relative poverty in China: A global analysis based on statistical and interpretable machine learning methods

Poverty eradication has always been a major challenge to global development and governance, which received widespread attention from each country. With the completion poverty alleviation task in 2020, relative poverty governance becomes an important issue to be solved in China urgently. Because of a large population, poor infrastructures, insufficient resources, and long-term uneven development raising the living standard of farmers in rural areas is critical to China's success in realizing moderate prosperity. Therefore, identifying the poor farmers, exploring the influence factors to relative poverty, and clarifying its effect mechanism in rural areas are significant for the subsequent poverty governance. Most of the previous studies adopted the method of apriori assuming the factor system and verifying the hypothesis. We innovatively constructed a relative poverty index system consistent with China's actual conditions, selecting all the possible variables that could affect relative poverty based on the existing literature, including individual characteristics, psychological endowment, and geographical environment, and rebuilt an experimental database. Then, through data processing and data analysis, the main factors influencing the relative poverty of farmers were systematically sorted out based on the machine learning method. Finally, 25 chosen influencing factors were discussed in detail. Research findings show that: 1) Machine learning algorithm is proved it could be well applied in relative poverty fields, especially XGBoost, which achieves 81.9% accuracy and the score of ROC_AUC reaches 0.819. 2) This study sheds light on many new research directions in applying machine learning for relative poverty research, besides, the paper offers an integral framework and beneficial reference for target identification using machine learning algorithms. 3) In addition, by utilizing the interpretable tools, the "black-box" of ML become transparent through PDP and SHAP explanation, it also reveals that machine learning models can readily handle the non-linear association relationship.

Spatial machine learning for predicting physical inactivity prevalence from socioecological determinants in Chicago, Illinois, USA

The increase in physical inactivity prevalence in the USA has been associated with neighborhood characteristics. While several studies have found an association between neighborhood and health, the relative importance of each component related to physical inactivity or how this value varies geographically (i.e., across different neighborhoods) remains unexplored. This study ranks the contribution of seven socioecological neighborhood factors to physical inactivity prevalence in Chicago, Illinois, using machine learning models at the census tract level, and evaluates their predictive capabilities. First, we use geographical random forest (GRF), a recently proposed nonlinear machine learning regression method that assesses each predictive factor's spatial variation and contribution to physical inactivity prevalence. Then, we compare the predictive performance of GRF to geographically weighted artificial neural networks, another recently proposed spatial machine learning algorithm. Our results suggest that poverty is the most important determinant in the Chicago tracts, while on the other hand, green space is the least important determinant in the rise of physical inactivity prevalence. As a result, interventions can be designed and implemented based on specific local circumstances rather than broad concepts that apply to Chicago and other large cities.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10109-023-00415-y.

Explaining the higher COVID-19 mortality rates among disproportionately Black counties: A decomposition analysis

Background

Why is COVID-19 mortality higher in counties with a disproportionately large (>13.4%) share of Black residents (hereafter "Black counties") relative to others ("non-Black counties")? Existing literature points to six categories of determinants: (1) social distancing, (2) COVID-19 testing, (3) socioeconomic characteristics, (4) environmental characteristics, (5) prevalence of (pre-existing) chronic health conditions, and (6) demographic characteristics. The relative importance of these determinants has not yet been thoroughly examined.

Methods

We built a dataset consisting of 21 sub-indicators across the six categories of determinants for 3108 US counties and their COVID-19 mortality over the period of January 22, 2020-December 31, 2020. Applying the Gelbach's decomposition, we quantified which determinants were most (or least) associated with the COVID-19 mortality disparity between Black and non-Black counties.

Results

We find that COVID-19 death rates were 26 percent higher in Black counties compared to non-Black counties. This disparity was almost completely explained by the six categories of determinants included in our model. Decomposition analyses indicate that county-level demographic and population health characteristics explained most of this disparity. Among all sub-indicators considered, the greater proportion of females and smaller proportion of rural residents in Black counties were the two largest contributors to the COVID-19 mortality gap between Black and non-Black counties. Proportions of diabetic residents, uninsured residents, and the degree of income inequality also significantly contributed to the gap in COVID-19 mortality.

Conclusion

The COVID-19 mortality gap between Black and non-Black counties was largely explained by pre-pandemic differences in demographic and population health characteristics. Policies aiming to reduce the prevalence of chronic conditions and uninsured residents in Black counties would have helped narrow the COVID-19 mortality gap between Black and non-Black counties in 2020.

The association of demographic and socioeconomic factors with COVID-19 during pre- and post-vaccination periods: A cross-sectional study of Virginia

Sociodemographic factors have been found to be associated with the transmission of coronavirus disease 2019 (COVID-19), yet most studies focused on the period before the proliferation of vaccination and obtained inconclusive results. In this cross-sectional study, the infections, deaths, incidence rates, case fatalities, and mortalities of Virginia's 133 jurisdictions during the pre-vaccination and post-vaccination periods were compared, and their associations with demographic and socioeconomic factors were studied. The cumulative infections and deaths and medians of incidence rates, case fatalities, and mortalities of COVID-19 in 133 Virginia jurisdictions were significantly higher during the post-vaccination period than during the pre-vaccination period. A variety of demographic and socioeconomic risk factors were significantly associated with COVID-19 prevalence in Virginia. Multiple linear regression analysis suggested that demographic and socioeconomic factors contributed up to 80% of the variation in the infections, deaths, and incidence rates and up to 53% of the variation in the case fatalities and mortalities of COVID-19 in Virginia. The demographic and socioeconomic determinants differed during the pre- and post-vaccination periods. The developed multiple linear regression models could be used to effectively characterize the impact of demographic and socioeconomic factors on the infections, deaths, and incidence rates of COVID-19 in Virginia.

Modifiable contributing factors to COVID-19: A comprehensive review

The devastating complications of coronavirus disease 2019 (COVID-19) result from an individual's dysfunctional immune response following the initial severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Multiple toxic stressors and behaviors contribute to underlying immune system dysfunction. SARS-CoV-2 exploits the dysfunctional immune system to trigger a chain of events ultimately leading to COVID-19. The current study identifies eighty immune system dysfunction-enabling toxic stressors and behaviors (hereafter called modifiable contributing factors (CFs)) that also link directly to COVID-19. Each CF is assigned to one of the five categories in the CF taxonomy shown in Section 3.3.: Lifestyle (e.g., diet, substance abuse); Iatrogenic (e.g., drugs, surgery); Biotoxins (e.g., micro-organisms, mycotoxins); Occupational/Environmental (e.g., heavy metals, pesticides); Psychosocial/Socioeconomic (e.g., chronic stress, lower education). The current study shows how each modifiable factor contributes to decreased immune system capability, increased inflammation and coagulation, and increased neural damage and neurodegeneration. It is unclear how real progress can be made in combatting COVID-19 and other similar diseases caused by viral variants without addressing and eliminating these modifiable CFs.

COVID-19 mortality and deprivation: pandemic, syndemic, and endemic health inequalities

COVID-19 has exacerbated endemic health inequalities resulting in a syndemic pandemic of higher mortality and morbidity rates among the most socially disadvantaged. We did a scoping review to identify and synthesise published evidence on geographical inequalities in COVID-19 mortality rates globally. We included peer-reviewed studies, from any country, written in English that showed any area-level (eg, neighbourhood, town, city, municipality, or region) inequalities in mortality by socioeconomic deprivation (ie, measured via indices of multiple deprivation: the percentage of people living in poverty or proxy factors including the Gini coefficient, employment rates, or housing tenure). 95 papers from five WHO global regions were included in the final synthesis. A large majority of the studies (n=86) found that COVID-19 mortality rates were higher in areas of socioeconomic disadvantage than in affluent areas. The subsequent discussion reflects on how the unequal nature of the pandemic has resulted from a syndemic of COVID-19 and endemic inequalities in chronic disease burden.

Real-Time Analysis of Predictors of COVID-19 Infection Spread in Countries in the European Union Through a New Tool

Objectives: Real-time data analysis during a pandemic is crucial. This paper aims to introduce a novel interactive tool called Covid-Predictor-Tracker using several sources of COVID-19 data, which allows examining developments over time and across countries. Exemplified here by investigating relative effects of vaccination to non-pharmaceutical interventions on COVID-19 spread.

Methods: We combine >100 indicators from the Global COVID-19 Trends and Impact Survey, Johns Hopkins University, Our World in Data, European Centre for Disease Prevention and Control, National Centers for Environmental Information, and Eurostat using random forests, hierarchical clustering, and rank correlation to predict COVID-19 cases.

Results: Between 2/2020 and 1/2022, we found among the non-pharmaceutical interventions “mask usage” to have strong effects after the percentage of people vaccinated at least once, followed by country-specific measures such as lock-downs. Countries with similar characteristics share ranks of infection predictors. Gender and age distribution, healthcare expenditures and cultural participation interact with restriction measures.

Conclusion: Including time-aware machine learning models in COVID-19 infection dashboards allows to disentangle and rank predictors of COVID-19 cases per country to support policy evaluation. Our open-source tool can be updated daily with continuous data streams, and expanded as the pandemic evolves.

Nonlinear associations between environmental factors and lipid levels in middle-aged and elderly population in China: A national cross-sectional study

BACKGROUND

Blood lipid is an important factor affecting cardiovascular disease in middle-aged and elderly people. At present, the associations between environmental factors and blood lipid level in elderly people has been controversial, and the nonlinear effect of their relationship is lack of research.

METHODS

This study used data from a national cross-sectional survey of blood lipid levels in 13,354 subjects and data from environmental monitoring sites. Logistic regression was used to measure the relationship between the basic characteristics of the study population and blood lipid levels. After controlling the confounding factors, the nonlinear associations between environmental factors and blood lipid levels of middle-aged and elderly people in different geographical regions were studied by random forest model.

RESULTS

The risk of dyslipidemia is significantly higher in middle-aged women, obese people, elderly people, and urban people. Smoking and alcohol consumption increase the risk. The associations between environmental factors and lipid levels of middle-aged and elderly people are nonlinear, the correlation effect between air pollutants and blood lipid level is mainly shown in northern China, and the correlation between meteorological factors and blood lipid level is more obvious in southern China.

CONCLUSIONS

This study shows that the associations between environmental factors and lipid levels in middle-aged and elderly population are nonlinear and have regional differences. Therefore it should be considered in optimizing the allocation of public health resources and preventing and controlling environmental exposure of middle-aged and elderly population.

A machine learning-driven spatio-temporal vulnerability appraisal based on socio-economic data for COVID-19 impact prevention in the U.S. counties

A mature and hybrid machine-learning model is verified by mature empirical analysis to measure county-level COVID-19 vulnerability and track the impact of the imposition of pandemic control policies in the U.S. A total of 30 county-level social, economic, and medical variables and a timeline of the imposed policies constitutes a COVID-19 database. A hybrid feature-selection model composed of four machine-learning algorithms is developed to emphasize the regional impact of community features on the case fatality rate (CFR). A COVID-19 vulnerability index (COVULin) is proposed to measure the county's vulnerability, the effects of model's parameters on mortality, and the efficiency of control policies. The results showed that the dense counties in which minority groups represent more than 45% of the population and those with poverty rates greater than 24% were the most vulnerable counties during the first and the last pandemic peaks, respectively. Highly-correlated CFR and COVULin scores indicated a close agreement between the model outcomes and COVID-19 impacts. Counties with higher poverty and uninsured rates were the most resistant to government intervention. It is anticipated that the proposed model can play an essential role in identifying vulnerable communities and help reduce damages during long-term alike disasters.

Geographically weighted regression model for physical, social, and economic factors affecting the COVID-19 pandemic spreading

This study aims to analyze the spatial distribution of the epidemic spread and the role of the physical, social, and economic characteristics in this spreading. A geographically weighted regression (GWR) model was built within a GIS environment using infection data monitored by the Iraqi Ministry of Health records for 10 months from March to December 2020. The factors adopted in this model are the size of urban interaction areas and human gatherings, movement level and accessibility, and the volume of public services and facilities that attract people. The results show that it would be possible to deal with each administrative unit in proportion to its circumstances in light of the factors that appear in it. So, there will not be a single treatment for all areas with different urban characteristics, which sometimes helps not to stop social and economic life due to the imposition of a comprehensive ban on movement and activities. Therefore, there will be other supportive policies other than the ban, depending on the urban indicators for each region, such as reducing external movement from it or relying on preventing public activities only.

Air pollution and COVID-19 mortality and hospitalization: An ecological study in Iran

Exposure to air pollution can exacerbate the severe COVID-19 conditions, subsequently causing an increase in the death rate. In this study, we investigated the association between long-term exposure to air pollution and risks of COVID-19 hospitalization and mortality in Arak, Iran. Air pollution data was obtained from air quality monitoring stations located in Arak, including particulate matter (PM), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), ozone (O₃) and carbon monoxide (CO). Daily numbers of Covid-19 cases including hospital admissions (hospitalization) and deaths (mortality) were obtained from a national data registry recorded by Arak University of Medical Sciences. A Poisson regression model with natural spline functions was applied to set the effects of air pollution on COVID-19 hospitalization and mortality. The percent change of COVID-19 hospitalization per 10 μg/m³ increase in PM_2.5 and PM₁₀ were 8.5% (95% CI 7.6 to 11.5) and 4.8% (95% CI 3 to 6.5), respectively. An increase of 10 μg/m³ in PM_2.5 resulting in 5.6% (95% CI: 3.1-8.3%) increase in COVID-19 mortality. The percent change of hospitalization (7.7%, 95% CI 2.2 to 13.3) and mortality (4.5%, 95% CI 0.3 to 9.5) were positively significant per one ppb increment in SO₂, while NO₂, O₃ and CO were inversely associated with hospitalization and mortality. Our findings strongly suggesting that a small increase in long-term exposure to PM_2.5, PM₁₀ and SO2 elevating risks of hospitalization and mortality related to COVID-19.

The impact of geo-environmental factors on global COVID-19 transmission: A review of evidence and methodology

Studies on Coronavirus Disease 2019 (COVID-19) transmission indicate that geo-environmental factors have played a significant role in the global pandemic. However, there has not been a systematic review on the impact of geo-environmental factors on global COVID-19 transmission in the context of geography. As such, we reviewed 49 well-chosen studies to reveal the impact of geo-environmental factors (including the natural environment and human activity) on global COVID-19 transmission, and to inform critical intervention strategies that could mitigate the worldwide effects of the pandemic. Existing studies frequently mention the impact of climate factors (e.g., temperature and humidity); in contrast, a more decisive influence can be achieved by human activity, including human mobility, health factors, and non-pharmaceutical interventions (NPIs). The above results exhibit distinct spatiotemporal heterogeneity. The related analytical methodology consists of sensitivity analysis, mathematical modeling, and risk analysis. For future studies, we recommend highlighting geo-environmental interactions, developing geographically statistical models for multiple waves of the pandemic, and investigating NPIs and care patterns. We also propose four implications for practice to combat global COVID-19 transmission.

A Geographically Weighted Random Forest Approach to Predict Corn Yield in the US Corn Belt

Crop yield prediction before the harvest is crucial for food security, grain trade, and policy making. Previously, several machine learning methods have been applied to predict crop yield using different types of variables. In this study, we propose using the Geographically Weighted Random Forest Regression (GWRFR) approach to improve crop yield prediction at the county level in the US Corn Belt. We trained the GWRFR and five other popular machine learning algorithms (Multiple Linear Regression (MLR), Partial Least Square Regression (PLSR), Support Vector Regression (SVR), Decision Tree Regression (DTR), and Random Forest Regression (RFR)) with the following different sets of features: (1) full length features; (2) vegetation indices; (3) gross primary production (GPP); (4) climate data; and (5) soil data. We compared the results of the GWRFR with those of the other five models. The results show that the GWRFR with full length features (R2 = 0.90 and RMSE = 0.764 MT/ha) outperforms other machine learning algorithms. For individual categories of features such as GPP, vegetation indices, climate, and soil features, the GWRFR also outperforms other models. The Moran’s I value of the residuals generated by GWRFR is smaller than that of other models, which shows that GWRFR can better address the spatial non-stationarity issue. The proposed method in this article can also be potentially used to improve yield prediction for other types of crops in other regions.

Socioeconomic and environmental factors of poverty in China using geographically weighted random forest regression model

Correlations between socioeconomic factors and poverty in regression models do not reflect actual relationships, especially when data exhibit patterns of spatial heterogeneity. Spatial regression models can estimate the relationships between socioeconomic factors and poverty in defined geographical areas, explaining the imbalanced distribution of poverty, but the relationships between these factors and poverty are not always linear however, and conventional simple linear local regression models do not accurately capture these nonlinear relationships. To fill this gap, we used a local regression method, geographically weighted random forest regression (GW-RFR), that integrates a spatial weight matrix (SWM) and random forest (RF). The GW-RFR evaluates the spatial variations in the nonlinear relationships between variables. A county-level poverty data set of China was employed to estimate the performance of the GW-RFR against the random forest (RF). In this poverty application, the value of [Formula: see text] was 0.128 higher than that of the RF, the NRMSE value was 1.6% lower than the RF, and the MAE value was 0.295 lower than the RF. These results showed that the relationship between poverty factors and poverty varies with space at the county level in China, and the GW-RFR was suitable for dealing with nonlinear relationships in local regression analysis.

Socioeconomic Inequalities in COVID-19 Incidence During Different Epidemic Phases in South Korea

Objective

Area-level socioeconomic status (SES) is associated with coronavirus disease 2019 (COVID-19) incidence. However, the underlying mechanism of the association is context-specific, and the choice of measure is still important. We aimed to evaluate the socioeconomic gradient regarding COVID-19 incidence in Korea based on several area-level SES measures.

Methods

COVID-19 incidence and area-level SES measures across 229 Korean municipalities were derived from various administrative regional data collected between 2015 and 2020. The Bayesian negative binomial model with a spatial autocorrelation term was used to estimate the incidence rate ratio (IRR) and relative index of inequality (RII) of each SES factor, with adjustment for covariates. The magnitude of association was compared between two epidemic phases: a low phase (<100 daily cases, from May 6 to August 14, 2020) and a rebound phase (>100 daily cases, from August 15 to December 31, 2020).

Results

Area-level socioeconomic inequalities in COVID-19 incidence between the most disadvantaged region and the least disadvantaged region were observed for nonemployment rates [RII = 1.40, 95% credible interval (Crl) = 1.01–1.95] and basic livelihood security recipients (RII = 2.66, 95% Crl = 1.12–5.97), but were not observed for other measures in the low phase. However, the magnitude of the inequalities of these SES variables diminished in the rebound phase. A higher area-level mobility showed a higher risk of COVID-19 incidence in both the low (IRR = 1.67, 95% Crl = 1.26–2.17) and rebound phases (IRR = 1.28, 95% Crl = 1.14–1.44). When SES and mobility measures were simultaneously adjusted, the association of SES with COVID-19 incidence remained significant but only in the low phase, indicating they were mutually independent in the low phase.

Conclusion

The level of basic livelihood benefit recipients and nonemployment rate showed social stratification of COVID-19 incidence in Korea. Explanation of area-level inequalities in COVID-19 incidence may not be derived only from mobility differences in Korea but, instead, from the country's own context.

The Geographical Distribution and Influencing Factors of COVID-19 in China

The study of the spatial differentiation of COVID-19 in cities and its driving mechanism is helpful to reveal the spatial distribution pattern, transmission mechanism and diffusion model, and evolution mechanism of the epidemic and can lay the foundation for constructing the spatial dynamics model of the epidemic and provide theoretical basis for the policy design, spatial planning and implementation of epidemic prevention and control and social governance. Geodetector (Origin version, Beijing, China) is a great tool for analysis of spatial differentiation and its influencing factors, and it provides decision support for differentiated policy design and its implementation in executing the city-specific policies. Using factor detection and interaction analysis of Geodetector, 15 indicators of economic, social, ecological, and environmental dimensions were integrated, and 143 cities were selected for the empirical research in China. The research shows that, first of all, risks of both infection and death show positive spatial autocorrelation, but the geographical distribution of local spatial autocorrelation differs significantly between the two. Secondly, the inequalities in urban economic, social, and residential environments interact with COVID-19 spatial heterogeneity, with stronger explanatory power especially when multidimensional inequalities are superimposed. Thirdly, the spatial distribution and spread of COVID-19 are highly spatially heterogeneous and correlated due to the complex influence of multiple factors, with factors such as Area of Urban Construction Land, GDP, Industrial Smoke and Dust Emission, and Expenditure having the strongest influence, the factors such as Area of Green, Number of Hospital Beds and Parks, and Industrial NOx Emissions having unignorable influence, while the factors such as Number of Free Parks and Industrial Enterprises, Per-GDP, and Population Density play an indirect role mainly by means of interaction. Fourthly, the factor interaction effect from the infected person’s perspective mainly shows a nonlinear enhancement effect, that is, the joint influence of the two factors is greater than the sum of their direct influences; but from the perspective of the dead, it mainly shows a two-factor enhancement effect, that is, the joint influence of the two factors is greater than the maximum of their direct influences but less than their sum. Fifthly, some suggestions are put forward from the perspectives of building a healthy, resilient, safe, and smart city, providing valuable reference and decision basis for city governments to carry out differentiated policy design.

Ranking the importance of demographic, socioeconomic, and underlying health factors on US COVID-19 deaths: A geographical random forest approach

A growing number of studies show that the uneven spatial distribution of COVID-19 deaths is related to demographic and socioeconomic disparities across space. However, most studies fail to assess the relative importance of each factor to COVID-19 death rate and, more importantly, how this importance varies spatially. Here, we assess the variables that are more important locally using Geographical Random Forest (GRF), a local non-linear regression method. Through GRF, we estimated the non-linear relationships between the COVID-19 death rate and 29 socioeconomic and health-related factors during the first year of the pandemic in the USA (county level). GRF outputs are compared to global (Random Forest and OLS) and local (Geographically Weighted Regression) models. Results show that GRF outperforms all models and that the importance of variables highly varies by location. For example, lack of health insurance is the most important factor in one-third (34.86%) of the US counties. Most of these counties are (concentrated mainly in the Midwest region and South region). On the other hand, no leisure-time physical activity is the most important primary factor for 19.86% of the US counties. These counties are found in California, Oregon, Washington, and parts of the South region. Understanding the location-based characteristics and spatial patterns of socioeconomic and health factors linked to COVID-19 deaths is paramount for policy designing and decision making. In this way, interventions can be designed and implemented based on the most important factors locally, avoiding thus general guidelines addressed for the entire nation.

Phenotype Discovery and Geographic Disparities of Late-Stage Breast Cancer Diagnosis across U.S. Counties: A Machine Learning Approach

BACKGROUND

Disparities in the stage at diagnosis for breast cancer have been independently associated with various contextual characteristics. Understanding which combinations of these characteristics indicate highest risk, and where they are located, is critical to targeting interventions and improving outcomes for patients with breast cancer.

METHODS

The study included women diagnosed with invasive breast cancer between 2009 and 2018 from 680 U.S. counties participating in the Surveillance, Epidemiology, and End Results program. We used a machine learning approach called Classification and Regression Tree (CART) to identify county "phenotypes," combinations of characteristics that predict the percentage of patients with breast cancer presenting with late-stage disease. We then mapped the phenotypes and compared their geographic distributions. These findings were further validated using an alternate machine learning approach called random forest.

RESULTS

We discovered seven phenotypes of late-stage breast cancer. Common to most phenotypes associated with high risk of late-stage diagnosis were high uninsured rate, low mammography use, high area deprivation, rurality, and high poverty. Geographically, these phenotypes were most prevalent in southern and western states, while phenotypes associated with lower percentages of late-stage diagnosis were most prevalent in the northeastern states and select metropolitan areas.

CONCLUSIONS

The use of machine learning methods of CART and random forest together with geographic methods offers a promising avenue for future disparities research.

IMPACT

Local interventions to reduce late-stage breast cancer diagnosis, such as community education and outreach programs, can use machine learning and geographic modeling approaches to tailor strategies for early detection and resource allocation.

Contributing factors common to COVID‑19 and gastrointestinal cancer

The devastating complications of coronavirus disease 2019 (COVID-19) result from the dysfunctional immune response of an individual following the initial severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Multiple toxic stressors and behaviors contribute to underlying immune system dysfunction. SARS-CoV-2 exploits the dysfunctional immune system to trigger a chain of events, ultimately leading to COVID-19. The authors have previously identified a number of contributing factors (CFs) common to myriad chronic diseases. Based on these observations, it was hypothesized that there may be a significant overlap between CFs associated with COVID-19 and gastrointestinal cancer (GIC). Thus, in the present study, a streamlined dot-product approach was used initially to identify potential CFs that affect COVID-19 and GIC directly (i.e., the simultaneous occurrence of CFs and disease in the same article). The nascent character of the COVID-19 core literature (~1-year-old) did not allow sufficient time for the direct effects of numerous CFs on COVID-19 to emerge from laboratory experiments and epidemiological studies. Therefore, a literature-related discovery approach was used to augment the COVID-19 core literature-based ‘direct impact’ CFs with discovery-based ‘indirect impact’ CFs [CFs were identified in the non-COVID-19 biomedical literature that had the same biomarker impact pattern (e.g., hyperinflammation, hypercoagulation, hypoxia, etc.) as was shown in the COVID-19 literature]. Approximately 2,250 candidate direct impact CFs in common between GIC and COVID-19 were identified, albeit some being variants of the same concept. As commonality proof of concept, 75 potential CFs that appeared promising were selected, and 63 overlapping COVID-19/GIC potential/candidate CFs were validated with biological plausibility. In total, 42 of the 63 were overlapping direct impact COVID-19/GIC CFs, and the remaining 21 were candidate GIC CFs that overlapped with indirect impact COVID-19 CFs. On the whole, the present study demonstrates that COVID-19 and GIC share a number of common risk/CFs, including behaviors and toxic exposures, that impair immune function. A key component of immune system health is the removal of those factors that contribute to immune system dysfunction in the first place. This requires a paradigm shift from traditional Western medicine, which often focuses on treatment, rather than prevention.

Universal risk phenotype of US counties for flu-like transmission to improve county-specific COVID-19 incidence forecasts

The spread of a communicable disease is a complex spatio-temporal process shaped by the specific transmission mechanism, and diverse factors including the behavior, socio-economic and demographic properties of the host population. While the key factors shaping transmission of influenza and COVID-19 are beginning to be broadly understood, making precise forecasts on case count and mortality is still difficult. In this study we introduce the concept of a universal geospatial risk phenotype of individual US counties facilitating flu-like transmission mechanisms. We call this the Universal Influenza-like Transmission (UnIT) score, which is computed as an information-theoretic divergence of the local incidence time series from an high-risk process of epidemic initiation, inferred from almost a decade of flu season incidence data gleaned from the diagnostic history of nearly a third of the US population. Despite being computed from the past seasonal flu incidence records, the UnIT score emerges as the dominant factor explaining incidence trends for the COVID-19 pandemic over putative demographic and socio-economic factors. The predictive ability of the UnIT score is further demonstrated via county-specific weekly case count forecasts which consistently outperform the state of the art models throughout the time-line of the COVID-19 pandemic. This study demonstrates that knowledge of past epidemics may be used to chart the course of future ones, if transmission mechanisms are broadly similar, despite distinct disease processes and causative pathogens.

Accurate case count forecasts in an epidemic is non-trivial, with the spread of infectious diseases being modulated by diverse hard-to-model factors. This study introduces the concept of a universal risk phenotype for US counties that predictably increases the risk of person-to-person transmission of influenza-like illnesses; universal in the sense that it is pathogen-agnostic provided the transmission mechanism is similar to that of seasonal influenza. We call this the Universal Influenza-like Transmission (UnIT) score, which accounts for unmodeled effects by automatically leveraging subtle geospatial patterns underlying the flu epidemics of the past. It is a phenotype of the counties themselves, as it characterizes how the transmission process is differentially impacted in different geospatial contexts. Grounded in information-theory and machine learning, the UnIT score reduces the need to manually identify every factor that impacts the case counts. Applying to the COVID-19 pandemic, we show that incidence patterns from a past epidemic caused by an appropriately-chosen distinct pathogen can substantially inform future projections. Our forecasts consistently outperform the state of the art models throughout the time-line of the COVID-19 pandemic, and thus is an important step to inform policy decisions in current and future pandemics.

PM2.5 as a major predictor of COVID-19 basic reproduction number in the USA

Many studies have proposed a relationship between COVID-19 transmissibility and ambient pollution levels. However, a major limitation in establishing such associations is to adequately account for complex disease dynamics, influenced by e.g. significant differences in control measures and testing policies. Another difficulty is appropriately controlling the effects of other potentially important factors, due to both their mutual correlations and a limited dataset. To overcome these difficulties, we will here use the basic reproduction number (R₀) that we estimate for USA states using non-linear dynamics methods. To account for a large number of predictors (many of which are mutually strongly correlated), combined with a limited dataset, we employ machine-learning methods. Specifically, to reduce dimensionality without complicating the variable interpretation, we employ Principal Component Analysis on subsets of mutually related (and correlated) predictors. Methods that allow feature (predictor) selection, and ranking their importance, are then used, including both linear regressions with regularization and feature selection (Lasso and Elastic Net) and non-parametric methods based on ensembles of weak-learners (Random Forest and Gradient Boost). Through these substantially different approaches, we robustly obtain that PM_2.5 is a major predictor of R₀ in USA states, with corrections from factors such as other pollutants, prosperity measures, population density, chronic disease levels, and possibly racial composition. As a rough magnitude estimate, we obtain that a relative change in R₀, with variations in pollution levels observed in the USA, is typically ~30%, which further underscores the importance of pollution in COVID-19 transmissibility.

Investigating the roles of meteorological factors in COVID-19 transmission in Northern Italy

The novel COVID-19 is a highly invasive, pathogenic, and transmittable disease that has stressed the health care sector and hampered global development. Information of other viral respiratory diseases indicates that COVID-19 transmission could be affected by varying weather conditions; however, the impact of meteorological factors on the COVID-19 death counts remains unexplored. By investigating the impact of meteorological factors (absolute humidity, relative humidity, and temperature), this study will contribute both theoretically and practically to the concerned domain of pandemic management to be better prepared to control the spread of the disease. For this study, data is collected from 23 February to 31 March 2020 for Milan, Northern Italy, one of the badly hit regions by COVID-19. The generalized additive model (GAM) is applied, and a nonlinear relationship is examined with penalized spline methods. A sensitivity analysis is conducted for the verification of model results. The results reveal that temperature, relative humidity, and absolute humidity have a significant but negative relationship with the COVID-19 mortality rate. Therefore, it is possible to postulate that cool and dry environmental conditions promote virus transmission, leading to an increase in COVID-19 death counts. The results may facilitate health care policymakers in developing and implementing effective control measures in a timely and efficient way.

Spatio-temporal characteristics and control strategies in the early period of COVID-19 spread: a case study of the mainland China

COVID-19 has caused huge impacts on human health and the economic operation of the world. Analyzing and summarizing the early propagation law can help reduce the losses caused by public health emergencies in the future. Early data on the spread of COVID-19 in 30 provinces (autonomous regions and municipalities) of mainland China except for Hubei, Hong Kong, Macao, and Taiwan were selected in this study. Spatio-temporal analysis, inflection point analysis, and correlation analysis are used to explore the spatio-temporal characteristics in the early COVID-19 spread. The results suggested that (1) the total confirmed cases have risen in an "S"-shaped curve over time, and the daily new cases have first increased and finally decreased; (2) the spatial distributions of both total and daily new cases show a trend of more in the east and less in the west, with a "multi-center agglomeration distribution" around Hubei Province and some major cities; (3) the spatial agglomeration of total confirmed cases has been increasing over time, while that of the daily new cases shows much more obvious in the mid-stage; and (4) timely release of the first-level public health emergency response can accelerate the emergence of the epidemic inflection point. The above analysis results have a specific reference value for the government's policy-making and measures to face public health emergencies.

Common contributing factors to COVID-19 and inflammatory bowel disease

The devastating complications of coronavirus disease 2019 (COVID-19) result from an individual's dysfunctional immune response following the initial severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Multiple toxic stressors and behaviors contribute to underlying immune system dysfunction. SARS-CoV-2 exploits the dysfunctional immune system to trigger a chain of events ultimately leading to COVID-19. We have previously identified many contributing factors (CFs) (representing toxic exposure, lifestyle factors and psychosocial stressors) common to myriad chronic diseases. We hypothesized significant overlap between CFs associated with COVID-19 and inflammatory bowel disease (IBD), because of the strong role immune dysfunction plays in each disease. A streamlined dot-product approach was used to identify potential CFs to COVID-19 and IBD. Of the fifty CFs to COVID-19 that were validated for demonstration purposes, approximately half had direct impact on COVID-19 (the CF and COVID-19 were mentioned in the same record; i.e., CF---→COVID-19), and the other half had indirect impact. The nascent character of the COVID-19 core literature (∼ one year old) did not allow sufficient time for the direct impacts of many CFs on COVID-19 to be identified. Therefore, an immune system dysfunction (ID) literature directly related to the COVID-19 core literature was used to augment the COVID-19 core literature and provide the remaining CFs that impacted COVID-19 indirectly (i.e., CF---→immune system dysfunction---→COVID-19). Approximately 13000 potential CFs for myriad diseases (obtained from government and university toxic substance lists) served as the starting point for the dot-product identification process. These phrases were intersected (dot-product) with phrases extracted from a PubMed-derived IBD core literature, a nascent COVID-19 core literature, and the COVID-19-related immune system dysfunction (ID) core literature to identify common ID/COVID-19 and IBD CFs. Approximately 3000 potential CFs common to both ID and IBD, almost 2300 potential CFs common to ID and COVID-19, and over 1900 potential CFs common to IBD and COVID-19 were identified. As proof of concept, we validated fifty of these ∼3000 overlapping ID/IBD candidate CFs with biologic plausibility. We further validated 24 of the fifty as common CFs in the IBD and nascent COVID-19 core literatures. This significant finding demonstrated that the CFs indirectly related to COVID-19 -- identified with use of the immune system dysfunction literature -- are strong candidates to emerge eventually as CFs directly related to COVID-19. As discussed in the main text, many more CFs common to all these core literatures could be identified and validated. ID and IBD share many common risk/contributing factors, including behaviors and toxic exposures that impair immune function. A key component to immune system health is removal of those factors that contribute to immune system dysfunction in the first place. This requires a paradigm shift from traditional Western medicine, which often focuses on treatment, rather than prevention.

Exploration of spatial-temporal varying impacts on COVID-19 cumulative case in Texas using geographically weighted regression (GWR)

Since COVID-19 is extremely threatening to human health, it is significant to determine its impact factors to curb the virus spread. To tackle the complexity of COVID-19 expansion on a spatial-temporal scale, this research appropriately analyzed the spatial-temporal heterogeneity at the county-level in Texas. First, the impact factors of COVID-19 are captured on social, economic, and environmental multiple facets, and the communality is extracted through principal component analysis (PCA). Second, this research uses COVID-19 cumulative case as the dependent variable and the common factors as the independent variables. According to the virus prevalence hierarchy, the spatial-temporal disparity is categorized into four quarters in the GWR analysis model. The findings exhibited that GWR models provide higher fitness and more geodata-oriented information than OLS models. In El Paso, Odessa, Midland, Randall, and Potter County areas in Texas, population, hospitalization, and age structures are presented as static, positive influences on COVID-19 cumulative cases, indicating that they should adopt stringent strategies in curbing COVID-19. Winter is the most sensitive season for the virus spread, implying that the last quarter should be paid more attention to preventing the virus and taking precautions. This research is expected to provide references for the prevention and control of COVID-19 and related infectious diseases and evidence for disease surveillance and response systems to facilitate the appropriate uptake and reuse of geographical data.

Space-Time Cluster's Detection and Geographical Weighted Regression Analysis of COVID-19 Mortality on Texas Counties

As COVID-19 run rampant in high-density housing sites, it is important to use real-time data in tracking the virus mobility. Emerging cluster detection analysis is a precise way of blunting the spread of COVID-19 as quickly as possible and save lives. To track compliable mobility of COVID-19 on a spatial-temporal scale, this research appropriately analyzed the disparities between spatial-temporal clusters, expectation maximization clustering (EM), and hierarchical clustering (HC) analysis on Texas county-level. Then, based on the outcome of clustering analysis, the sensitive counties are Cottle, Stonewall, Bexar, Tarrant, Dallas, Harris, Jim hogg, and Real, corresponding to Southeast Texas analysis in Geographically Weighted Regression (GWR) modeling. The sensitive period took place in the last two quarters in 2020 and the first quarter in 2021. We explored PostSQL application to portray tracking Covid-19 trajectory. We captured 14 social, economic, and environmental impact's indices to perform principal component analysis (PCA) to reduce dimensionality and minimize multicollinearity. By using the PCA, we extracted five factors related to mortality of COVID-19, involved population and hospitalization, adult population, natural supply, economic condition, air quality or medical care. We established the GWR model to seek the sensitive factors. The result shows that adult population, economic condition, air quality, and medical care are the sensitive factors. Those factors also triggered high increase of COVID-19 mortality. This research provides geographical understanding and solution of controlling COVID-19, reference of implementing geographically targeted ways to track virus mobility, and satisfy for the need of emergency operations plan (EOP).

Machine Learning on the COVID-19 Pandemic, Human Mobility and Air Quality: A Review

The ongoing COVID-19 global pandemic is touching every facet of human lives (e.g., public health, education, economy, transportation, and the environment). This novel pandemic and non-pharmaceutical interventions of lockdown and confinement implemented citywide, regionally or nationally are affecting virus transmission, people's travel patterns, and air quality. Many studies have been conducted to predict the diffusion of the COVID-19 disease, assess the impacts of the pandemic on human mobility and on air quality, and assess the impacts of lockdown measures on viral spread with a range of Machine Learning (ML) techniques. This literature review aims to analyze the results from past research to understand the interactions among the COVID-19 pandemic, lockdown measures, human mobility, and air quality. The critical review of prior studies indicates that urban form, people's socioeconomic and physical conditions, social cohesion, and social distancing measures significantly affect human mobility and COVID-19 viral transmission. During the COVID-19 pandemic, many people are inclined to use private transportation for necessary travel to mitigate coronavirus-related health problems. This review study also noticed that COVID-19 related lockdown measures significantly improve air quality by reducing the concentration of air pollutants, which in turn improves the COVID-19 situation by reducing respiratory-related sickness and deaths. It is argued that ML is a powerful, effective, and robust analytic paradigm to handle complex and wicked problems such as a global pandemic. This study also explores the spatio-temporal aspects of lockdown and confinement measures on coronavirus diffusion, human mobility, and air quality. Additionally, we discuss policy implications, which will be helpful for policy makers to take prompt actions to moderate the severity of the pandemic and improve urban environments by adopting data-driven analytic methods.

Exploring spatiotemporal effects of the driving factors on COVID-19 incidences in the contiguous United States

Since December 2019, the world has witnessed the stringent effect of an unprecedented global pandemic, coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). As of January 29,2021, there have been 100,819,363 confirmed cases and 2,176,159 deaths reported. Among the countries affected severely by COVID-19, the United States tops the list. Research has been conducted to discuss the causal associations between explanatory factors and COVID-19 transmission in the contiguous United States. However, most of these studies focus more on spatial associations of the estimated parameters, yet exploring the time-varying dimension in spatial econometric modeling appears to be utmost essential. This research adopts various relevant approaches to explore the potential effects of driving factors on COVID-19 counts in the contiguous United States. A total of three global spatial regression models and two local spatial regression models, the latter including geographically weighted regression (GWR) and multiscale GWR (MGWR), are performed at the county scale to take into account the scale effects. For COVID-19 cases, ethnicity, crime, and income factors are found to be the strongest covariates and explain most of the variance of the modeling estimation. For COVID-19 deaths, migration (domestic and international) and income factors play a critical role in explaining spatial differences of COVID-19 deaths across counties. Such associations also exhibit temporal variations from March to July, as supported by better performance of MGWR than GWR. Both global and local associations among the parameters vary highly over space and change across time. Therefore, time dimension should be paid more attention to in the spatial epidemiological analysis. Among the two local spatial regression models, MGWR performs more accurately, as it has slightly higher Adj. R² values (for cases, R² = 0.961; for deaths, R² = 0.962), compared to GWR's Adj. R² values (for cases, R² = 0.954; for deaths, R² = 0.954). To inform policy-makers at the nation and state levels, understanding the place-based characteristics of the explanatory forces and related spatial patterns of the driving factors is of paramount importance. Since it is not the first time humans are facing public health emergency, the findings of the present research on COVID-19 therefore can be used as a reference for policy designing and effective decision making.

Geographically weighted machine learning model for untangling spatial heterogeneity of type 2 diabetes mellitus (T2D) prevalence in the USA

Type 2 diabetes mellitus (T2D) prevalence in the United States varies substantially across spatial and temporal scales, attributable to variations of socioeconomic and lifestyle risk factors. Understanding these variations in risk factors contributions to T2D would be of great benefit to intervention and treatment approaches to reduce or prevent T2D. Geographically-weighted random forest (GW-RF), a tree-based non-parametric machine learning model, may help explore and visualize the relationships between T2D and risk factors at the county-level. GW-RF outputs are compared to global (RF and OLS) and local (GW-OLS) models between the years of 2013-2017 using low education, poverty, obesity, physical inactivity, access to exercise, and food environment as inputs. Our results indicate that a non-parametric GW-RF model shows a high potential for explaining spatial heterogeneity of, and predicting, T2D prevalence over traditional local and global models when inputting six major risk factors. Some of these predictions, however, are marginal. These findings of spatial heterogeneity using GW-RF demonstrate the need to consider local factors in prevention approaches. Spatial analysis of T2D and associated risk factor prevalence offers useful information for targeting the geographic area for prevention and disease interventions.

Methodological limitations in studies assessing the effects of environmental and socioeconomic variables on the spread of COVID-19: a systematic review

BACKGROUND

While numerous studies have assessed the effects of environmental (meteorological variables and air pollutants) and socioeconomic variables on the spread of the COVID-19 pandemic, many of them, however, have significant methodological limitations and errors that could call their results into question. Our main objective in this paper is to assess the methodological limitations in studies that evaluated the effects of environmental and socioeconomic variables on the spread of COVID-19.

MAIN BODY

We carried out a systematic review by conducting searches in the online databases PubMed, Web of Science and Scopus up to December 31, 2020. We first excluded those studies that did not deal with SAR-CoV-2 or COVID-19, preprints, comments, opinion or purely narrative papers, reviews and systematic literature reviews. Among the eligible full-text articles, we then excluded articles that were purely descriptive and those that did not include any type of regression model. We evaluated the risk of bias in six domains: confounding bias, control for population, control of spatial and/or temporal dependence, control of non-linearities, measurement errors and statistical model. Of the 5631 abstracts initially identified, we were left with 132 studies on which to carry out the qualitative synthesis. Of the 132 eligible studies, we evaluated 63.64% of the studies as high risk of bias, 19.70% as moderate risk of bias and 16.67% as low risk of bias.

CONCLUSIONS

All the studies we have reviewed, to a greater or lesser extent, have methodological limitations. These limitations prevent conclusions being drawn concerning the effects environmental (meteorological and air pollutants) and socioeconomic variables have had on COVID-19 outcomes. However, we dare to argue that the effects of these variables, if they exist, would be indirect, based on their relationship with social contact.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1186/s12302-021-00550-7.