Spatio-temporal Disease Mapping Research Articles

Spatiotemporal Trends of Birth Defects in North Carolina, 2003-2015.

Birth defects are a leading cause of infant mortality in the United States, but little is known about causes of many types of birth defects. Spatiotemporal disease mapping to identify high-prevalence areas, is a potential strategy to narrow the search for potential environmental and other causes that aggregate over space and time. We described the spatial and temporal trends of the prevalence of birth defects in North Carolina during 2003-2015, using data on live births obtained from the North Carolina Birth Defects Monitoring Program. By employing a Bayesian space-time Poisson model, we estimated spatial and temporal trends of non-chromosomal and chromosomal birth defects. During 2003-2015, 52,524 (3.3%) of 1,598,807 live births had at least one recorded birth defect. The prevalence of non-chromosomal birth defects decreased from 3.8% in 2003 to 2.9% in 2015. Spatial modeling suggested a large geographic variation in non-chromosomal birth defects at census-tract level, with the highest prevalence in southeastern North Carolina. The strong spatial heterogeneity revealed in this work allowed to identify geographic areas with higher prevalence of non-chromosomal birth defects in North Carolina. This variation will help inform future research focused on epidemiologic studies of birth defects to identify etiologic factors.

Big problems in spatio-temporal disease mapping: Methods and software

Background and objective: Fitting spatio-temporal models for areal data is crucial in many fields such as cancer epidemiology. However, when data sets are very large, many issues arise. The main objective of this paper is to propose a general procedure to analyze high-dimensional spatio-temporal areal data, with special emphasis on mortality/incidence relative risk estimation.Methods: We present a pragmatic and simple idea that permits hierarchical spatio-temporal models to be fitted when the number of small areas is very large. Model fitting is carried out using integrated nested Laplace approximations over a partition of the spatial domain. We also use parallel and distributed strategies to speed up computations in a setting where Bayesian model fitting is generally prohibitively time-consuming or even unfeasible.Results:Using simulated and real data, we show that our method outperforms classical global models. We implement the methods and algorithms that we develop in the open-source R package bigDM where specific vignettes have been included to facilitate the use of the methodology for non-expert users.Conclusions: Our scalable methodology proposal provides reliable risk estimates when fitting Bayesian hierarchical spatio-temporal models for high-dimensional data.

Variance partitioning in spatio-temporal disease mapping models.

Bayesian disease mapping, yet if undeniably useful to describe variation in risk over time and space, comes with the hurdle of prior elicitation on hard-to-interpret random effect precision parameters. We introduce a reparametrized version of the popular spatio-temporal interaction models, based on Kronecker product intrinsic Gaussian Markov random fields, that we name the variance partitioning model. The variance partitioning model includes a mixing parameter that balances the contribution of the main and interaction effects to the total (generalized) variance and enhances interpretability. The use of a penalized complexity prior on the mixing parameter aids in coding prior information in an intuitive way. We illustrate the advantages of the variance partitioning model using two case studies.

Spatio-temporal disease risk estimation using clustering-based adjacency modelling.

Conditional autoregressive models are typically used to capture the spatial autocorrelation present in areal unit disease count data when estimating the spatial pattern in disease risk. This correlation is represented by a binary neighbourhood matrix based on a border sharing specification, which enforces spatial correlation between geographically neighbouring areas. However, enforcing such correlation will mask any discontinuities in the disease risk surface, thus impeding the detection of clusters of areas that exhibit higher or lower risks compared to their neighbours. Here we propose novel methodology to account for these clusters and discontinuities in disease risk via a two-stage modelling approach, which either forces the clusters/discontinuities to be the same for all time periods or allows them to evolve dynamically over time. Stage one constructs a set of candidate neighbourhood matrices to represent a range of possible cluster/discontinuity structures in the data, and stage two estimates an appropriate structure(s) by treating the neighbourhood matrix as an additional parameter to estimate within a Bayesian spatio-temporal disease mapping model. The effectiveness of our novel methodology is evidenced by simulation, before being applied to a new study of respiratory disease risk in Greater Glasgow, Scotland from 2011 to 2017.

Space-time interactions in Bayesian disease mapping with recent tools: Making things easier for practitioners.

Spatio-temporal disease mapping studies the distribution of mortality or incidence risks in space and its evolution in time, and it usually relies on fitting hierarchical Poisson mixed models. These models are complex for practitioners as they generally require adding constraints to correctly identify and interpret the different model terms. However, including constraints may not be straightforward in some recent software packages. This paper focuses on NIMBLE, a library of algorithms that contains among others a configurable system for Markov chain Monte Carlo (MCMC) algorithms. In particular, we show how to fit different spatio-temporal disease mapping models with NIMBLE making emphasis on how to include sum-to-zero constraints to solve identifiability issues when including spatio-temporal interactions. Breast cancer mortality data in Spain during the period 1990-2010 is used for illustration purposes. A simulation study is also conducted to compare NIMBLE with R-INLA in terms of parameter estimates and relative risk estimation. The results are very similar but differences are observed in terms of computing time.

Big Problems in Spatio-Temporal Disease Mapping: Methods and Software

Big Problems in Spatio-Temporal Disease Mapping: Methods and Software

Spatiotemporal trends in dementia‐associated deaths, what do the data reveal?

AbstractBackgroundData on dementia mortality are routinely collected by governments and used in health research. However, historical underrecording of dementia on death certificates has limited the utility of this widely available data source. Considerable effort has been invested in raising awareness of dementia, increasing diagnosis rates, and improving dementia recording on death certificates. Updated versions of the ICD‐10 rules for coding the underlying cause of death, implemented in Scotland in 2011 and 2017, have increasingly assigned deaths that mention both dementia and respiratory or urinary tract infections towards an underlying cause of dementia, if no other underlying medical condition is mentioned. Since previous cohort studies have shown geographical variation in dementia risk, this study utilises spatiotemporal modelling to investigate changes in dementia‐associated deaths in Scotland.MethodData on dementia and Alzheimer’s deaths and mid‐year population estimates available from the National Records of Scotland for the time period 2000‐2019 were used to build a Bayesian spatiotemporal disease mapping model. The data were grouped by the 32 Council areas and year, whilst stratification by age and gender were built into the model. The analyses were run in R using Integrated Nested Laplace Approximation methodology.ResultThe data show strong temporal and spatial trends in the relative risk of dementia‐associated death. The relative risk increases progressively over the time period studied, with a 2.3‐fold increase (95% Confidence Interval: 2.1‐2.5) between 2000 and 2019, reaching levels just below predicted dementia prevalence in recent years. The relative risk varies geographically and is highest in Scotland’s industrialised central belt region and lowest in the rural areas and islands, with a relative risk of 1.9 (95% CI: 1.6‐2.3) between the lowest and highest risk Council areas.ConclusionThere are factors, other than the step changes in underlying cause of death coding, driving the progressively increasing temporal trend in relative risk of dementia‐associated death. These factors could be a combination of increased dementia awareness, a larger proportion of people being diagnosed, greater recording on death certificates and, possibly, increasing prevalence. The strong spatial variation in relative risk in Scotland indicates that there are geographically‐varying factors affecting this risk.

Spatio-temporal small area surveillance of the COVID-19 pandemic

The emergence of COVID-19 requires new effective tools for epidemiological surveillance. Spatio-temporal disease mapping models, which allow dealing with small units of analysis, are a priority in this context. These models provide geographically detailed and temporally updated overviews of the current state of the pandemic, making public health interventions more effective. These models also allow estimating epidemiological indicators highly demanded for COVID-19 surveillance, such as the instantaneous reproduction number Rt, even for small areas. In this paper, we propose a new spatio-temporal spline model particularly suited for COVID-19 surveillance, which allows estimating and monitoring Rt for small areas. We illustrate our proposal on the study of the disease pandemic in two Spanish regions. As a result, we show how tourism flows have shaped the spatial distribution of the disease in these regions. In these case studies, we also develop new epidemiological tools to be used by regional public health services for small area surveillance.

Bayesian spatio-temporal joint disease mapping of Covid-19 cases and deaths in local authorities of England

The overwhelming spatio-temporal nature of the spread of the ongoing Covid-19 pandemic demands urgent attention of data analysts and model developers. Modelling results obtained from analytical tool development are essential to understand the ongoing pandemic dynamics with a view to helping the public and policy makers. The pandemic has generated data on a huge number of interesting statistics such as the number of new cases, hospitalisations and deaths in many spatio-temporal resolutions for the analysts to investigate. The multivariate nature of these data sets, along with the inherent spatio-temporal dependencies, poses new challenges for modellers. This article proposes a two-stage hierarchical Bayesian model as a joint bivariate model for the number of cases and deaths observed weekly for the different local authority administrative regions in England. An adaptive model is proposed for the weekly Covid-19 death rates as part of the joint bivariate model. The adaptive model is able to detect possible step changes in death rates in neighbouring areas. The joint model is also used to evaluate the effects of several socio-economic and environmental covariates on the rates of cases and deaths. Inclusion of these covariates points to the presence of a north–south divide in both the case and death rates. Nitrogen dioxide, the only air pollution measure used in the model, is seen to be significantly positively associated with the number cases, even in the presence of the spatio-temporal random effects taking care of spatio-temporal dependencies present in the data. The proposed models provide excellent fits to the observed data and are seen to perform well for predicting the location specific number of deaths a week in advance. The structure of the models is very general and the same framework can be used for modelling other areally aggregated temporal statistics of the pandemics, e.g. the rate of hospitalisation.

Chronic high risk of intimate partner violence against women in disadvantaged neighborhoods: An eight-year space-time analysis

We conducted a small-area ecological longitudinal study to analyze neighborhood contextual influences on the spatio-temporal variations in intimate partner violence against women (IPVAW) risk in a southern European city over an eight-year period. We used geocoded data of IPVAW cases with associated protection orders (n = 5867) in the city of Valencia, Spain (2011–2018). The city's 552 census block groups were used as the neighborhood units. Neighborhood-level covariates were: income, education, immigrant concentration, residential instability, alcohol outlet density, and criminality. We used a Bayesian autoregressive approach to spatio-temporal disease mapping. Neighborhoods with low levels of income and education and high levels of residential mobility and criminality had higher relative risk of IPVAW. Spatial patterns of high risk of IPVAW persisted over time during the eight-year period analyzed. Areas of stable low risk and with increasing or decreasing risk were also identified. Our findings link neighborhood disadvantage to the existence and persistence over time of spatial inequalities in IPVAW risk, showing that high risk of IPVAW can become chronic in disadvantaged neighborhoods. Our analytic approach provides specific risk estimates at the small-area level that are informative for intervention purposes, and can be useful to assess the effectiveness of prevention efforts in reducing IPVAW.

An Autoregressive Disease Mapping Model for Spatio-Temporal Forecasting

One of the more evident uses of spatio-temporal disease mapping is forecasting the spatial distribution of diseases for the next few years following the end of the period of study. Spatio-temporal models rely on very different modeling tools (polynomial fit, splines, time series, etc.), which could show very different forecasting properties. In this paper, we introduce an enhancement of a previous autoregressive spatio-temporal model with particularly interesting forecasting properties, given its reliance on time series modeling. We include a common spatial component in that model and show how that component improves the previous model in several ways, its predictive capabilities being one of them. In this paper, we introduce and explore the theoretical properties of this model and compare them with those of the original autoregressive model. Moreover, we illustrate the benefits of this new model with the aid of a comprehensive study on 46 different mortality data sets in the Valencian Region (Spain) where the benefits of the new proposed model become evident.

Spatial-Temporal Modelling of Disease Risk Accounting for PM2.5 Exposure in the Province of Pavia: An Area of the Po Valley

Spatio-temporal Bayesian disease mapping is the branch of spatial epidemiology interested in providing valuable risk estimates in certain geographical regions using administrative areas as statistical units. The aim of the present paper is to describe spatio-temporal distribution of cardiovascular mortality in the Province of Pavia in 2010 through 2015 and assess its association with environmental pollution exposure. To produce reliable risk estimates, eight different models (hierarchical log-linear model) have been assessed: temporal parametric trend components were included together with some random effects that allowed the accounting of spatial structure of the region. The Bayesian approach allowed the borrowing information effect, including simpler model results in the more complex setting. To compare these models, Watanabe–Akaike Information Criteria (WAIC) and Leave One Out Information Criteria (LOOIC) were applied. In the modelling phase, the relationship between the disease risk and pollutants exposure (PM2.5) accounting for the urbanisation level of each geographical unit showed a strong significant effect of the pollutant exposure (OR = 1.075 and posterior probability, or PP, >0.999, equivalent to p < 0.001). A high-risk cluster of Cardiovascular mortality in the Lomellina subareas in the studied window was identified.

Bayesian estimation of multivariate Gaussian Markov random fields with constraint.

This article concerns with conditionally formulated multivariate Gaussian Markov random fields (MGMRF) for modeling multivariate local dependencies with unknown dependence parameters subject to positivity constraint. In the context of Bayesian hierarchical modeling of lattice data in general and Bayesian disease mapping in particular, analytic and simulation studies provide new insights into various approaches to posterior estimation of dependence parameters under "hard" or "soft" positivity constraint, including the well-known strictly diagonal dominance criterion and options of hierarchical priors. Hierarchical centering is examined as a means to gain computational efficiency in Bayesian estimation of multivariate generalized linear mixed effects models in the presence of spatial confounding and weakly identified model parameters. Simulated data on irregular or regular lattice, and three datasets from the multivariate and spatiotemporal disease mapping literature, are used for illustration. The present investigation also sheds light on the use of deviance information criterion for model comparison, choice, and interpretation in the context of posterior risk predictions judged by borrowing-information and bias-precision tradeoff. The article concludes with a summary discussion and directions of future work. Potential applications of MGMRF in spatial information fusion and image analysis are briefly mentioned.

On estimation the relative risk of small area and visualization spatio-temporal map

ObjectiveThe purpose is to propose a serial of approach for estimation for disease risk for ILI in "small area" and present the risk values by spatio-temporal disease mapping or an interactive visualization with HTML format.IntroductionDisease mapping is a method used to descript the geographical variation in risk (heterogeneity of risk) and to provide the potential reason (factors or confounders) to explain the distribution. Possibly the most famous uses of disease mapping in epidemiology were the studies by John Snow of the cholera epidemics in London. Accurate estimation relative risk of small areas such as mortality and morbidity, by different age, ethnic group, interval and regions, is important for government agencies to identify hazards and mitigate disease burden. Recently, as the innovative algorithms and the available software, more and more disease risk index has been pouring out. This abstract will provide several estimation risk index, from raw incidence to model-based relative risks, and use visual approach to display them.MethodsAll the data are from a syndromic surveillance and real-time early warning system in the Yunnan province in the China. For brief introduction aim, we are using the ILI (Influenza-like illness) data in December 2017 in one county. The relative risks of disease in small area are including: raw incidence, a standardized morbidity ratio (SMR), Empirical Bayes smoothing estimation relative risk (EB-RR) and the Besag-York-Mollio model (BYM). The incidence in each small area is common used for descriptive the risk but fail to comparable directly since the different population at risk in each area. SMR is a good way to deal with this incomparability. But SMR can give rise to imprecisely estimate in areas with small populations. Empirical Bayes estimation approach has been used for smoothing purpose and can be seen as a compromise between relative risks and P-values. However, all above approaches are inept to have spatial or spatio-temporal structure in mind. BYM based the Bayesian inference can handle both the area-specific spatial structured component (such as intrinsic conditional autoregressive component) and the exchangeable random effect (unstructured component). All the analyses are implemented in the R software with INLA package (http://www.r-inla.org). The outcome of relative risk estimation with visual way and interactive maps showing are using ggplot2 and leaflet packages.Results1, the spatio-temporal raw cases of ILI from 2017/12/01 to 2017/12/31 is Fig.12. the SMR and EB-RR estimation RR of ILI are in Fig.2 and Fig.33. the most excited is the interactive visualization with HTML format for all the risk indexes is visited http://rpubs.com/ynsxx/424814 in detail. And the screenshot is Fig.4ConclusionsSmall area disease risk estimation is important for disease prevention and control. The faster function of computer with power R software can lead to advance in disease mapping, allowing for complex spatio-temporal models and communicate the results with visualization way.

Showcasing Cancer Incidence and Mortality in Rural ZCTAs Using Risk Probabilities via Spatio-Temporal Bayesian Disease Mapping

Health departments are seeking new ways to determine when and where limited resources should be allocated to achieve maximum benefit for the population. In this work, we demonstrate how one state health department worked to create relative risk measures of cancer incidence, late-stage cancer incidence and mortality incidence displayed in an easy to read map using spatio-temporal statistical tools. The data included age, sex, cancer type and stage, and ZIP Code Tabulation Area (ZCTA) for every incidence and death from 2004 to 2015. Eight types of cancer were selected for analysis: breast, cervical, colorectal, liver, lung, non-Hodgkin lymphoma (NHL), prostate, and melanoma. The risk maps were designed to illustrate areas of the state where risk for developing or dying from certain cancers was higher than the state average, and to show how trends are evolving over time. A hierarchical Bayesian log-normal Poisson regression model, with effects for ZCTA, time period, and a space-time interaction was implemented. The spatial effects accounted for spatial correlation using an intrinsic conditional auto-regressive model, and the time effects used an autoregressive model. Through the model, we were able to achieve reliable estimates of relative risk per ZCTA and time period, even for small population ZCTAs with few, if any, cases during the time period. Furthermore, we calculated a measure of risk probability for each ZCTA, relative to the state average. Results from two cancers are discussed in this manuscript, but all 24 results are available on the project website.

Some recent work on multivariate Gaussian Markov random fields

Some recent work on conditional formulation of multivariate Gaussian Markov random fields is presented. The focus is on model constructions by compatible conditionals and coregionalization. Special attention is given to multivariate generalizations of univariate models. Beginning with univariate model constructions, a survey of key approaches to formulating multivariate extensions is presented. Two challenges in the formulation and implementation of multivariate models are highlighted: (1) entanglement of spatial and non-spatial components, and (2) enforcement for positivity condition. Managing the two challenges by decomposition, separation, and constrained parameterization is discussed. Also highlighted is the challenge of flexible modeling of (conditional) cross-spatial dependencies and, in particular, asymmetric cross-spatial dependencies. Interpretation of asymmetric cross-spatial dependencies is also discussed. A coregionalization framework which connects and unifies the various lines of model development is presented. The framework enables a systematic development of a broad range of models via linear and spatially varying coregionalization, respectively, with extensions to locally adaptive models. Formulation of multivariate models over variable-specific lattices is discussed. Selected models are illustrated with examples of Bayesian multivariate and spatiotemporal disease mapping. Potential applications of coregionalization models in imaging analysis, covariance modeling, dimension reduction, and latent variable analysis are briefly mentioned.

Bivariate spatiotemporal disease mapping of cancer of the breast and cervix uteri among Iranian women.

Cervical cancer in women is one of the most common cancers and breast cancer has grown dramatically in recent years. The purpose of this study was to map the incidence of breast and cervix uteri cancer among Iranian women over a 6-year period (2004-2009) searching for trend changes and risk factors. Cancer incidence data were extracted from the annual reports of the National Cancer Registry in Iran. Hierarchical Bayesian models, including random spatial and temporal effects was utilized together with bivariate, spatio-temporal shared component modelling. The provinces Tehran, Isfahan, Mazandaran and Gilan were found to have the highest relative risk (RR) of breast cancer, while the highest RR of cervix uteri cancer was observed in Tehran, Golestan, Khuzestan and Khorasan Razavi. Shared risk factors (smoking component) between the two cancers were seen to have the highest influence in Tehran, Khorasan Razavi, Yazd, Isfahan, Golestan, Khuzestan, Fars and Mazandaran, while the least were observed in Kohgiluyeh Boyerahmad. Apparent differences and distinctions between high-risk and low-risk provinces reveal a pattern of obvious dispersion for these cancers in Iran that should be considered when allocating healthcare resources and services in different areas.

Smartphone-Based Mobile Detection Platform for Molecular Diagnostics and Spatiotemporal Disease Mapping.

Rapid and quantitative molecular diagnostics in the field, at home, and at remote clinics is essential for evidence-based disease management, control, and prevention. Conventional molecular diagnostics requires extensive sample preparation, relatively sophisticated instruments, and trained personnel, restricting its use to centralized laboratories. To overcome these limitations, we designed a simple, inexpensive, hand-held, smartphone-based mobile detection platform, dubbed "smart-connected cup" (SCC), for rapid, connected, and quantitative molecular diagnostics. Our platform combines bioluminescent assay in real-time and loop-mediated isothermal amplification (BART-LAMP) technology with smartphone-based detection, eliminating the need for an excitation source and optical filters that are essential in fluorescent-based detection. The incubation heating for the isothermal amplification is provided, electricity-free, with an exothermic chemical reaction, and incubation temperature is regulated with a phase change material. A custom Android App was developed for bioluminescent signal monitoring and analysis, target quantification, data sharing, and spatiotemporal mapping of disease. SCC's utility is demonstrated by quantitative detection of Zika virus (ZIKV) in urine and saliva and HIV in blood within 45 min. We demonstrate SCC's connectivity for disease spatiotemporal mapping with a custom-designed website. Such a smart- and connected-diagnostic system does not require any lab facilities and is suitable for use at home, in the field, in the clinic, and particularly in resource-limited settings in the context of Internet of Medical Things (IoMT).

Estimating relative risk for dengue disease in Peninsular Malaysia using INLA

Study in spatio-temporal disease mapping models give a great worth in epidemiology, in describing the pattern of disease incidence across geographical space and time. This paper studies generalized linear mixed models (GLMM) for the analysis of spatial and temporal variability of dengue disease rates. For spatio-temporal study, the models accommodate spatially correlated random effects as well as temporal effects together with the space time interaction. The space time interaction is used to capture any additional effects that are not explained by the main factors of space and time. However, as study including time dimension is quite complex for disease mapping, the temporal effects that only relate to structured and unstructured time pattern are considered in these models as initial screening in studying disease pattern and time trend. The models are fitted within a hierarchical Bayesian framework using Integrated Nested Laplace Approximation (INLA) methodology. For this study, there are three main objectives. First, to choose the best model that represent the disease phenomenon. Second, to estimate the relative risk of disease based on the model selected and lastly, to visualize the risk spatial pattern and temporal trend using graphical representation. The models are applied to monthly dengue fever data in Peninsular Malaysia reported to Ministry of Health Malaysia for year 2015 by district level.

Mapping child maltreatment risk: a 12-year spatio-temporal analysis of neighborhood influences

Background‘Place’ matters in understanding prevalence variations and inequalities in child maltreatment risk. However, most studies examining ecological variations in child maltreatment risk fail to take into account the implications of the spatial and temporal dimensions of neighborhoods. In this study, we conduct a high-resolution small-area study to analyze the influence of neighborhood characteristics on the spatio-temporal epidemiology of child maltreatment risk.MethodsWe conducted a 12-year (2004–2015) small-area Bayesian spatio-temporal epidemiological study with all families with child maltreatment protection measures in the city of Valencia, Spain. As neighborhood units, we used 552 census block groups. Cases were geocoded using the family address. Neighborhood-level characteristics analyzed included three indicators of neighborhood disadvantage—neighborhood economic status, neighborhood education level, and levels of policing activity—, immigrant concentration, and residential instability. Bayesian spatio-temporal modelling and disease mapping methods were used to provide area-specific risk estimations.ResultsResults from a spatio-temporal autoregressive model showed that neighborhoods with low levels of economic and educational status, with high levels of policing activity, and high immigrant concentration had higher levels of substantiated child maltreatment risk. Disease mapping methods were used to analyze areas of excess risk. Results showed chronic spatial patterns of high child maltreatment risk during the years analyzed, as well as stability over time in areas of low risk. Areas with increased or decreased child maltreatment risk over the years were also observed.ConclusionsA spatio-temporal epidemiological approach to study the geographical patterns, trends over time, and the contextual determinants of child maltreatment risk can provide a useful method to inform policy and action. This method can offer a more accurate description of the problem, and help to inform more localized prevention and intervention strategies. This new approach can also contribute to an improved epidemiological surveillance system to detect ecological variations in risk, and to assess the effectiveness of the initiatives to reduce this risk.