Spatiotemporal Regression Model Research Articles

The impact of measurement error in modeled ambient particles exposures on health effect estimates in multilevel analysis

We sampled 1,000 small administrative areas in London, United Kingdom, and simulated the "true" underlying daily exposure surfaces for PM10 and PM2.5 for 2009-2013 incorporating temporal variation and spatial covariance informed by the extensive London monitoring network. We added measurement error assessed by comparing measurements at fixed sites and predictions from spatiotemporal land-use regression (LUR) models; dispersion models; models using satellite data and applying machine learning algorithms; and combinations of these methods through generalized additive models. Two health outcomes were simulated to assess whether the bias varies with the effect size. We applied multilevel Poisson regression to simultaneously model the effect of long- and short-term pollutant exposure. For each scenario, we ran 1,000 simulations to assess measurement error impact on health effect estimation. For long-term exposure to particles, we observed bias toward the null, except for traffic PM2.5 for which only LUR underestimated the effect. For short-term exposure, results were variable between exposure models and bias ranged from -11% (underestimate) to 20% (overestimate) for PM10 and of -20% to 17% for PM2.5. Integration of models performed best in almost all cases. No single exposure model performed optimally across scenarios. In most cases, measurement error resulted in attenuation of the effect estimate.

RETRACTED ARTICLE: Spatiotemporal evolution characteristics of China’s cold chain logistics resources and agricultural product using remote sensing perspective

ABSTRACT The cold chain logistics is a special distribution process with respect to agricultural products that are sensitive to temperature. In this approach, remote sensing data and spatial analysis are the main components of china’s cold chain logistics capabilities, particularly on agricultural products. Therefore, The Cold Chain Logistics analysis on agricultural products in various provinces is evaluated using a Spatial Exploratory Data Analysis and Geo-detector approach. Hence, The characteristically Remote Sensing Spatio-Temporal Logistical Regression (RS-STLR) model has been developed to characterize both the cold chain's logistic resource and provincial differences. This paper provides a perspective on the growth and variations in Chinese logistics services for cold chain logistics. The research has been conducted at a lab-scale which established an alternate interpretation between different regional variations in logistic capital of the ecological features of the cold chain logistics. The aim of this analysis is to validate the condition of the cold chain market based on the quantitative system of risk assessment on agricultural produces.

Prediction of inter-fractional radiotherapy dose plans with domain translation in spatiotemporal embeddings.

External beam radiation therapy fractions have become extremely complex and tedious procedures to plan, due to stringent requirements of delivering the highest radiation dose to the tumor while maximally avoiding organs at risk. However, due to anatomic and/or biological changes between fractions, dose re-optimization may be needed. Re-optimization is a time-consuming task which is typically triggered based on subjective visual assessment by an experienced physician. To address limitations in this process, we introduce a predictive framework which learns the evolution of tumor anatomy as well as inter-fractional dose delivery variations for head and neck cancers. First, joint low-dimensional discriminant embeddings maximizing the separation between responsive and non-responsive groups to external beam radiotherapy plans are constructed from deep neural networks in order to capture patient-specific dose modulations with respect to anatomical variations. Then, latent representations are fed to a domain-level adversarial network to translate observed anatomical changes into dosimetric variations, which aims to enforce local semantic consistency in the overall translation. Dose distribution trajectories are represented in a group-average piecewise-geodesic setting to handle anatomical variations during therapy, using a quadratic optimization to perform curve regression. At test time, an annotated baseline CT is projected onto the latent space and translated to dose domain, from which a spatiotemporal regression model is constructed using parallel transport trajectories defined from closest samples. This allows to predict dosimetry changes during the course of treatment. The model was trained on 337 cases and tested on 50 separate patients using sequential CT and associated dosimetry data, with the probabilistic framework yielding a Dice score of 92% and an overall dose difference of 1.2Gy in organs at risk and tumor volume over the course of multi-day treatment course, with a 5% reduction in delivered fraction segments.

Detecting regime shifts in communities using estimated rates of change

AbstractRegime shifts (periods of rapid change punctuating longer periods of lower variability) are observed in a wide range of ecosystems, and effective fisheries management requires the ability to detect these shifts. Detecting shifts is straightforward in single-species time series when transitions are detectable as periods of rapid change. However, shifts in complex and spatially structured communities may be harder to detect. We propose an approach to characterize community regime shifts, using nonparametric spatiotemporal regression models to estimate three indicators of community change (the among-species mean rate of change, mean per-capita rate of change, and standard deviation of per-capita rate of change). These indicators can detect shifts that result in either changes in abundance or composition. We applied our approach to a 37-year community biomass time series that spans the Newfoundland Shelf groundfish collapse. Our method detected a community shift earlier than alternative single-indicator breakpoint approaches and gave additional insight into the spatiotemporal pattern of the change, including detecting three separate periods of change and characterizing the first locations to show signs of recovery. The indicators applied in this study provide novel insights into Newfoundland groundfish dynamics and should be useful in the characterization of other abrupt ecological changes.

Local and regional temporal trends (2013\u20132019) of canine Ehrlichia spp. seroprevalence in the USA

BackgroundIn the USA, there are several Ehrlichia spp. of concern including Ehrlichia canis, Ehrlichia ewingii, Ehrlichia chaffeensis, Ehrlichia muris eauclarensis, and “Panola Mountain Ehrlichia”. Of these, E. canis is considered the most clinically relevant for domestic dogs, with infection capable of causing acute, subclinical, and chronic stages of disease. Changes in climate, land use, habitats, and wildlife reservoir populations, and increasing contact between both human and dog populations with natural areas have resulted in the increased risk of vector-borne disease throughout the world.MethodsA Bayesian spatio-temporal binomial regression model was applied to serological test results collected from veterinarians throughout the contiguous USA between January 2013 and November 2019. The model was used to quantify both regional and local temporal trends of canine Ehrlichia spp. seroprevalence and identify areas that experienced significant increases in seroprevalence.ResultsRegionally, increasing seroprevalence occurred within several states throughout the central and southeastern states, including Missouri, Arkansas, Mississippi, Alabama, Virginia, North Carolina, Georgia and Texas. The underlying local trends revealed increasing seroprevalence at a finer scale. Clusters of locally increasing seroprevalence were seen from the western Appalachian region into the southern Midwest, along the Atlantic coast in New England, parts of Florida, Illinois, Wisconsin and Minnesota, and in a couple areas of the Mountain region. Clusters of locally decreasing seroprevalence were seen throughout the USA including New York and the mid-Atlantic states, Texas, the Midwest, and California.ConclusionsCanine Ehrlichia spp. seroprevalence is increasing in both endemic and non-endemic areas of the USA. The findings from this study indicate that dogs across a wide area of the USA are at risk of exposure and these results should provide veterinarians and pet owners with the information they need to make informed decisions about prevention of tick exposure.

Mapping malaria seasonality in Madagascar using health facility data

BackgroundMany malaria-endemic areas experience seasonal fluctuations in case incidence as Anopheles mosquito and Plasmodium parasite life cycles respond to changing environmental conditions. Identifying location-specific seasonality characteristics is useful for planning interventions. While most existing maps of malaria seasonality use fixed thresholds of rainfall, temperature, and/or vegetation indices to identify suitable transmission months, we construct a statistical modelling framework for characterising the seasonal patterns derived directly from monthly health facility data.MethodsWith data from 2669 of the 3247 health facilities in Madagascar, a spatiotemporal regression model was used to estimate seasonal patterns across the island. In the absence of catchment population estimates or the ability to aggregate to the district level, this focused on the monthly proportions of total annual cases by health facility level. The model was informed by dynamic environmental covariates known to directly influence seasonal malaria trends. To identify operationally relevant characteristics such as the transmission start months and associated uncertainty measures, an algorithm was developed and applied to model realisations. A seasonality index was used to incorporate burden information from household prevalence surveys and summarise ‘how seasonal’ locations are relative to their surroundings.ResultsPositive associations were detected between monthly case proportions and temporally lagged covariates of rainfall and temperature suitability. Consistent with the existing literature, model estimates indicate that while most parts of Madagascar experience peaks in malaria transmission near March–April, the eastern coast experiences an earlier peak around February. Transmission was estimated to start in southeast districts before southwest districts, suggesting that indoor residual spraying should be completed in the same order. In regions where the data suggested conflicting seasonal signals or two transmission seasons, estimates of seasonal features had larger deviations and therefore less certainty.ConclusionsMonthly health facility data can be used to establish seasonal patterns in malaria burden and augment the information provided by household prevalence surveys. The proposed modelling framework allows for evidence-based and cohesive inferences on location-specific seasonal characteristics. As health surveillance systems continue to improve, it is hoped that more of such data will be available to improve our understanding and planning of intervention strategies.

Bayesian Space-Time Analysis of Brain Cancer Incidence in Southern Ontario, Canada: 2010-2013.

Canada has one of the highest incidence rates of brain cancer in the world. This study investigates the space–time variation of brain cancer risk across Southern Ontario, Canada. A Bayesian spatio-temporal regression model is used to estimate the relative risk of brain cancer in the 12 spatial health units of Southern Ontario over a four-year period (2010–2013). This work also explores the association between brain cancer and two potential risk factors: traumatic head injury (THI) and excess body fat (EBF). Across all areal units from 2010–2013, results show that the relative risk of brain cancer ranged from 0.83 (95% credible interval (CI) 0.74–0.91) to 1.26 (95% CI 1.13–1.41). Over the years, the eastern and western health units had persistently higher risk levels compared to those in the central areas. Results suggest that areas with elevated THI rates and EBF levels were also potentially associated with higher brain cancer relative risk. Findings revealed that the mean temporal trend for cancer risk progression in the region smoothly decreased over time. Overall, 50% of the health units displayed area-specific trends which were higher than the region’s average, thus indicating a slower decrease in cancer rates for these areas in comparison to the mean trend.

Application of an advanced spatiotemporal model for PM2.5 prediction in Jiangsu Province, China

Either long- or short-term of fine particle (PM2.5) exposure is associated with adverse health effects especially for children. Primary school students spend much time in schools whereas PM2.5 prediction for such fine-scale places remains a demanding task, letalone a combined prediction with high temporal resolution. The study aimed to estimate PM2.5 levels of different time scales for primary schools in Jiangsu Province, China. Hourly PM2.5 measurements within the academic year (Sept. 2016-June 2017) were collected from 72 routine monitoring sites. Together with PM2.5 emission inventory and dozens of geographic variables, an advanced spatiotemporal land use regression (LUR) model was employed to estimate PM2.5 concentrations of biweekly, seasonal and academic year levels in Jiangsu Province at 2457 primary school locations. 10-fold cross-validation verified high prediction ability with squared correlations RCV2 of 0.72 for temporal and 0.71 for spatial changes. PM2.5 levels in primary schools in Nanjing and Nantong were >10% higher than that of the corresponding cities while pollution levels in primary schools in Xuzhou were >20% lower. For 10 out of the 13 cities in Jiangsu, PM2.5 levels for primary schools surpassed 70μg/m3 in winter. Schools in Lianyungang, Zhenjiang and Huai'an suffered the most. This study demonstrated the fine-scale prediction ability of the novel spatiotemporal LUR model, as well as the potential and necessity to apply it in epidemiological studies. It also verified the emergency of pollution control for primary schools from cities such as Lianyungang and Zhenjiang.

Concurrent spatiotemporal daily land use regression modeling and missing data imputation of fine particulate matter using distributed space-time expectation maximization

In this study, a spatiotemporal land use regression (LUR) model using distributed space-time expectation maximization (D-STEM) software was developed. We trained the model using daily mean ambient particulate matter ≤2.5 μm (PM2.5) data measured hourly in 2015 at 30 regulatory monitoring network stations within the megacity of Tehran, Iran. Since a substantial amount of measured data were missing (48% of the total number of daily PM2.5 observations), we used the D-STEM to impute missing data and compared the missing imputation performance between different fitted models and the mean substitution method. We used h-block cross-validation (h-block CV) method in order to account for spatial autocorrelation in the model building and validation. In the imputation of missing data, the D-STEM LUR model had a mean absolute percentage error (MAPE) of 25.3%, outperforming the mean substitution method, which resulted in MAPE of 28.3%. The spatiotemporal R-squared was 0.73 and the average CV R-squared of 2-block and 5-block cross-validations was 0.60. These values were 0.68 and 0.47 when the spatial aspect of the LUR model was assessed, and 0.995 and 0.992 when the temporal aspect of the LUR model was assessed. This study demonstrated the competence of D-STEM software in spatiotemporal modeling, missing data imputation, and mapping of daily ambient PM2.5 at a very high spatial resolution (20 m × 20 m). These estimations are available for future research, especially for epidemiological studies on short- and/or long-term health effects of ambient PM2.5. Generally, we found D-STEM as a promising tool for spatiotemporal LUR modeling of ambient air pollution, especially for those models that rely on regulatory network monitoring stations with a considerable amount of missing data.

Multivariate spatiotemporal modeling of drug- and alcohol-poisoning deaths in New York City, 2009–2014

Drug- and alcohol-poisoning deaths remain current public health problems. Studies to date have typically focused on individual-level predictors of drug overdose deaths, and there remains a limited understanding of the spatiotemporal patterns and predictors of the joint outcomes. We use a hierarchical Bayesian spatiotemporal multivariate Poisson regression model on data from (N = 167) ZIP-codes between 2009 and 2014 in New York City to examine the spatiotemporal patterns of the joint occurrence of drug (opioids) and alcohol-poisoning deaths, and the covariates associated with each outcome. Results indicate that rates of both outcomes were highly positively correlated across ZIP-codes (cross-correlation: 0.57, 95% credible interval (CrI): 0.29, 0.77). ZIP-codes with a higher prevalence of heavy drinking had higher alcohol-poisoning deaths (relative risk (RR):1.63, 95% CrI: 1.26, 2.05) and drug-poisoning deaths (RR: 1.29, 95% CrI: 1.03, 1.59). These spatial patterns may guide public health planners to target specific areas to address these co-occurring epidemics.

Area deprivation and notifiable infectious diseases in Germany: A longitudinal small-area analysis

Abstract Background Socioeconomic inequalities may affect the infectious disease incidence. We studied the association between area deprivation and incidence of notifiable infectious diseases in Germany to understand spatio-temporal patterns and the effects of societal factors on disease epidemiology. Methods Using national surveillance data of 401 districts from 2001 to 2017, we examined the incidence of infectious diseases using spatiotemporal Bayesian regression models. We analyzed eight disease classes: blood-borne viral hepatitis, gastrointestinal, vaccine preventable, vector-borne, zoonotic, other bacterial, other infectious, and overall burden of infectious diseases. As explanatory factors we considered area deprivation (measured by the German Index of Multiple Deprivation), fraction of non-nationals, sex, age, and spatiotemporal effects. Results A risk gradient across deprivation quintiles was observed for the overall burden of infectious diseases. The relative risk (RR) for gastrointestinal diseases in areas with medium and high deprivation relative to low deprivation was 1.65 (95%-credible interval [CrI] 1.01-2.54) and 2.64 (1.22-4.98), respectively. The RR for vector-borne diseases was 1.89 (1.27-2.73) in districts with high deprivation compared to areas with low deprivation. Lower risks in highly deprived areas relative to low deprived areas were identified in vaccine-preventable diseases (RR = 0.39; 0.14-0.88) and zoonoses (RR = 0.69; 0.48-0.96). For blood-borne viral hepatitis, other bacterial, and other infectious diseases no association with area deprivation was observed. Spatial risks of infections were predominantly concentrated in eastern parts of Germany and changed marginally over time. Conclusions The risks of infections tend to be higher in more deprived areas and in eastern parts of Germany, but they varied by class of disease. Our results can guide measures of infectious disease control and prevention by considering spatial risks and deprivation. Key messages Area deprivation has both positive and inverse associations with the incidences of infectious diseases in Germany. Regions with increased risks may benefit from targeted public health measures. Spatial risks of infections tended to be higher in eastern regions of Germany. Disparities in the incidence of infectious diseases may be still present between western and eastern Germany.

Bayesian modeling of spatiotemporal patterns of TB-HIV co-infection risk in Kenya

BackgroundTuberculosis (TB) and Human Immunodeficiency Virus (HIV) diseases are globally acknowledged as a public health challenge that exhibits adverse bidirectional relations due to the co-epidemic overlap. To understand the co-infection burden we used the case notification data to generate spatiotemporal maps that described the distribution and exposure hypotheses for further epidemiologic investigations in areas with unusual case notification levels.MethodsWe analyzed the TB and TB-HIV case notification data from the Kenya national TB control program aggregated for forty-seven counties over a seven-year period (2012–2018). Using spatiotemporal poisson regression models within the Integrated Nested Laplace Approach (INLA) paradygm, we modeled the risk of TB-HIV co-infection. Six competing models with varying space-time formulations were compared to determine the best fit model. We then assessed the geographic patterns and temporal trends of coinfection risk by mapping the posterior marginal from the best fit model.ResultsOf the total 608,312 TB case notifications, 194,129 were HIV co-infected. The proportion of TB-HIV co-infection was higher in females (39.7%) than in males (27.0%). A significant share of the co-infection was among adults aged 35 to 44 years (46.7%) and 45 to 54 years (42.1%). Based on the Bayesian Defiance Information (DIC) and the effective number of parameters (pD) comparisons, the spatiotemporal model allowing space-time interaction was the best in explaining the geographical variations in TB-HIV coinfection. The model results suggested that the risk of TB-HIV coinfection was influenced by infrastructure index (Relative risk (RR) = 5.75, Credible Interval (Cr.I) = (1.65, 19.89)) and gender ratio (RR = 5.81e−04, Cr. I = (1.06e−04, 3.18e−03). The lowest and highest temporal relative risks were in the years 2016 at 0.9 and 2012 at 1.07 respectively. The spatial pattern presented an increased co-infection risk in a number of counties. For the spatiotemporal interaction, only a few counties had a relative risk greater than 1 that varied in different years.ConclusionsWe identified elevated risk areas for TB/HIV co-infection and fluctuating temporal trends which could be because of improved TB case detection or surveillance bias caused by spatial heterogeneity in the co-infection dynamics. Focused interventions and continuous TB-HIV surveillance will ensure adequate resource allocation and significant reduction of HIV burden amongst TB patients.

Effects of prenatal exposure to air particulate matter on the risk of preterm birth and roles of maternal and cord blood LINE-1 methylation: A birth cohort study in Guangzhou, China

BackgroundEpidemiological studies have found that increased risk of preterm birth (PTB) is associated with higher prenatal exposure to PM10 and PM2.5, but few studies have been conducted to assess the impacts of extremely fine particulate matter (PM1) which may have more toxic effects than other types of ambient particulate air pollution (PM). Several studies have separately investigated the associations between DNA methylation and PTB risk and PM. Maternal LINE-1 methylation level negatively correlated with prenatal exposure to PM and risk of PTB. A comprehensive picture is lacking regarding the associations between prenatal exposure to PM, LINE-1 methylation, and risk of PTB. ObjectivesThis study aimed to estimate the effects of exposure to ambient PM (PM10, PM2.5, and PM1) of different sizes during pregnancy on risk of PTB, identify susceptible exposure windows, and illustrate the roles of LINE-1 methylation in the associations between PM and PTB risk. MethodsThe Birth Cohort Study on Prenatal Environments and Offspring Health (PEOH) has been ongoing since 2016 in Guangzhou, China. A total of 4928 pregnant women were recruited during early pregnancy, and 4278 (86.8%) were successfully followed-up. Each individual weekly exposure to PM10 and PM2.5 from 3 months before pregnancy to childbirth was assessed using a spatiotemporal land use regression model, and the weekly PM1 exposure was estimated by employing a generalized additive model. Maternal and cord blood LINE-1 methylation levels (%5mC) were tested using bisulfite-PCR pyrosequencing. A distributed lag nonlinear model incorporated with a Cox proportional hazard model was applied to assess the effect of weekly-specific maternal PM exposure on PTB risk, and a multiple-linear regression model was employed to investigate the associations between PM exposure and LINE-1 methylation levels of maternal and cord bloods. We also assessed the associations between LINE-1 methylation levels and PTB risk by using a logistic regression model. ResultsThe risk of PTB was positively associated with PM2.5 and PM1 concentrations during the 12th to 20th gestational weeks, and the strongest association was in the fourth quartile (Q4) versus the first quartile (Q1) and observed during the 16th gestational week (PM2.5: harzard ratio [HR] = 1.18, 95%CI: 1.04–1.35, IQR = 11.94 μg/m3. PM1: HR = 1.20, 95%CI: 1.03–1.39, IQR = 11.36 μg/m3). We observed significantly negative associations of PM10(β = −0.51%5mC per 10 μg/m3, P = 0.014), PM2.5 (β = −0.66%5mC per 10 μg/m3, P = 0.032) and PM1 (β = −0.67%5mC per 10 μg/m3, P = 0.032) concentrations with cord blood LINE-1 methylation levels, and a negative association between PM1 concentration and maternal LINE-1 methylation level (β = −0.86%5mC per 10 μg/m3, P = 0.034). ConclusionHigher prenatal exposure to PM1 and PM2.5 during the 12th to 20th gestational weeks was associated with increased risk of PTB. Maternal and fetal LINE-1 methylation alternation might be an underlying mechanism of PM that increasing the risk of PTB.

Semiparametric spatiotemporal variable coefficient regression model

In order to study the semi-parametric spatiotemporal variable coefficient regression model, firstly, the parametric and non-parametric models are introduced. Based on the spatiotemporal variable coefficient model and estimation method, a semi-parametric spatiotemporal variable coefficient model is established. According to the two-step estimation simulation experiment of semi-parametric space-time variable coefficient model, three groups of numerical experiments of semi-parametric spatiotemporal variable coefficient model are carried out. The experimental results show that the semi-parametric regression model has both the advantages of parametric model and non-parametric model, and the estimated value is very close to the real value. After combining the semi-parametric and time-varying coefficient regression model, the fitting effect of the model is better, and the estimation accuracy is obviously improved.

Regional and local temporal trends in the prevalence of canine heartworm infection in the contiguous United States: 2012\u20132018

BackgroundCanine heartworm disease is a potentially fatal disease for which treatment is financially burdensome for many pet owners. Prevention is strongly advocated by the veterinary community along with routine testing for infection during annual wellness examinations. Despite the availability of efficacious chemoprophylaxis, recent reports have suggested that the incidence of heartworm disease in domestic dogs is increasing.ResultsUsing data from tests for heartworm infection in the USA from January 2012 through September 2018, a Bayesian spatio-temporal binomial regression model was used to estimate the regional and local temporal trends of heartworm infection prevalence. The area with the largest increase in regional prevalence was found in the Lower Mississippi River Valley. Regional prevalence increased throughout the southeastern states and northward into Illinois and Indiana. Local (county-level) prevalence varied across the USA, with increasing prevalence occurring along most of the Atlantic coast, central United States, and western states. Clusters of decreasing prevalence were present along the Mississippi Alluvial Plain (a historically endemic area), Oklahoma and Kansas, and Florida.ConclusionsCanine heartworm infection prevalence is increasing in much of the USA, both regionally and locally, despite veterinarian recommendations on prevention and testing. Additional steps should be taken to protect dogs, cats and ferrets. Further work is needed to identify the driving factors of the locally decreasing prevalence present along the Mississippi Alluvial plain, Florida, and other areas.

Prediction outcomes for anterior vertebral body growth modulation surgery from discriminant spatiotemporal manifolds.

Anterior vertebral body growth modulation (AVBGM) is a minimally invasive surgical technique that gradually corrects spine deformities while preserving lumbar motion. However, identifying suitable patients for surgery is based on clinical judgment and surgical experience. This process would be facilitated by the identification of patients responding to AVBGM prior to surgery using data-driven models trained on previous instrumented cases. We introduce a statistical framework for predicting the surgical outcomes following AVBGM in adolescents with idiopathic scoliosis. A discriminant manifold is first constructed to maximize the separation between responsive and non-responsive groups of patients treated with AVBGM for scoliosis. The model then uses subject-specific correction trajectories based on articulated transformations in order to map spine correction profiles to a group-average piecewise-geodesic path. Spine correction trajectories are described in a piecewise-geodesic fashion to account for varying times at follow-up examinations, regressing the curve via a quadratic optimization process. To predict the evolution of correction, a baseline reconstruction is projected onto the manifold, from which a spatiotemporal regression model is built from parallel transport curves inferred from neighboring exemplars. The model was trained on 438 reconstructions and tested on 56 subjects using 3D spine reconstructions from follow-up examinations, with the probabilistic framework yielding accurate results with differences of [Formula: see text] in main curve angulation and a classification rate of 83.2%, and generating models similar to biomechanical simulations. The proposed method achieved a higher prediction accuracy and improved the modeling of spatiotemporal morphological changes in surgical patients treated with AVBGM.

Gut microbiota partially mediates the effects of fine particulate matter on type 2 diabetes: Evidence from a population-based epidemiological study

BackgroundExperimental studies have indicated that alterations in the gut microbiota might play a role in the pathway of diabetes induction resulting from particulate matter pollution with aerodynamic diameters < 2.5 μm (PM2.5). However, few human studies have examined such experimental findings. Here, we examine the mediating effects of gut microbial dysbiosis on the associations between PM2.5 and particulate matter pollution with aerodynamic diameters < 1 μm (PM1) on diabetes using the Guangdong Gut Microbiome Project (GGMP) dataset. MethodsA multistage cluster sampling method was employed to recruit adult participants from communities in Guangdong. Each participant was interviewed using a questionnaire, fasting blood and stool samples were collected, and the exposure to air pollutants was assessed using a spatiotemporal land-use regression model. The mediation analysis was conducted to estimate the associations among air pollutants, gut microbiota diversity and diabetes. ResultsBoth PM2.5 and PM1 were positively associated with the risks of impaired fasting glucose (IFG) or type 2 diabetes and negatively associated with alpha diversity indices of the gut microbiota. The mediation analyses indicated that the associations of PM2.5 and PM1 with the risk of type 2 diabetes were partially mediated by the decrease in gut microbiota diversity. Moreover, we found that 79 (PM2.5 on IFG), 84 (PM2.5 on type 2 diabetes), 83 (PM1 on IFG) and 89 (PM1 on type 2 diabetes) bacterial taxa could partially mediate the associations of PM2.5 and PM1 with IFG and type 2 diabetes, respectively. The relative abundance of most Firmicutes, Proteobacteria and Verrucomicrobia bacteria were negatively associated with particulate matter (PM) concentrations and the risks of diabetes. ConclusionsLong-term exposure to PM may increase the risk of diabetes, and alterations in the gut microbiota partially explained these associations.

Spatio-temporal additive regression model selection for urban water demand

Understanding the factors influencing urban water use is critical for meeting demand and conserving resources. To analyze the relationships between urban household-level water demand and potential drivers, we develop a method for Bayesian variable selection in partially linear additive regression models, particularly suited for high-dimensional spatio-temporally dependent data. Our approach combines a spike-and-slab prior distribution with a modified version of the Bayesian group lasso to simultaneously perform selection of null, linear, and nonlinear models and to penalize regression splines to prevent overfitting. We investigate the effectiveness of the proposed method through a simulation study and provide comparisons with existing methods. We illustrate the methodology on a case study to estimate and quantify uncertainty of the associations between several environmental and demographic predictors and spatio-temporally varying household-level urban water demand in Tampa, FL.

A new centered spatio-temporal autologistic regression model with an application to local spread of plant diseases

We propose a new spatio-temporal autologistic centered model for binary data on a lattice. Centering allows the self-regression coefficients to be interpreted by separating the large-scale structure from the small-scale structure. One of the coefficients determines the overall level (or average) of the process, the second determines the spatial autocorrelation. We discuss the existence of the joint distribution of the process and carry out numerical studies to highlight the interest of this type of centering. We suggest using the estimator that maximises the Pseudo Likelihood (denoted Maximum Pseudo Likelihood Estimator (MPLE) in the following) and we give a method for choosing the neighbourhood structure. We run simulations studies that show that the estimation method and model selection method work well. The method is applied to model and fit epidemiological data on Esca disease in a vineyard in the Bordeaux region.

Quantifying the relationship between human Lyme disease and Borrelia burgdorferi exposure in domestic dogs.

Lyme disease (LD) is the most common vector-borne disease in the United States. Early confirmatory diagnosis remains a challenge, while the disease can be debilitating if left untreated. Further, the decision to test is complicated by under-reporting, low positive predictive values of testing in non-endemic areas and travel, which together exacerbate the difficulty in identification of newly endemic areas or areas of emerging concern. Spatio-temporal analyses at the national scale are critical to establishing a baseline human LD risk assessment tool that would allow for the detection of changes in these areas. A well-established surrogate for human LD incidence is canine LD seroprevalence, making it a strong candidate covariate for use in such analyses. In this paper, Bayesian statistical methods were used to fit a spatio-temporal spline regression model to estimate the relationship between human LD incidence and canine seroprevalence, treating the latter as an explanatory covariate. A strong non-linear monotonically increasing association was found. That is, this analysis suggests that mean incidence in humans increases with canine seroprevalence until the seroprevalence in dogs reaches approximately 30%. This finding reinforces the use of canines as sentinels for human LD risk, especially with respect to identifying geographic areas of concern for potential human exposure.