Computational Techniques for Spatial Logistic Regression with Large Datasets

In epidemiological work, outcomes are frequently non-normal, sample sizes may be large, and effects are often small. To relate health outcomes to geographic risk factors, fast and powerful methods for fitting spatial models, particularly for non-normal data, are required. We focus on binary outcomes, with the risk surface a smooth function of space. We compare penalized likelihood models, including the penalized quasi-likelihood (PQL) approach, and Bayesian models based on fit, speed, and ease of implementation. A Bayesian model using a spectral basis representation of the spatial surface provides the best tradeoff of sensitivity and specificity in simulations, detecting real spatial features while limiting overfitting and being more efficient computationally than other Bayesian approaches. One of the contributions of this work is further development of this underused representation. The spectral basis model outperforms the penalized likelihood methods, which are prone to overfitting, but is slower to fit and not as easily implemented. Conclusions based on a real dataset of cancer cases in Taiwan are similar albeit less conclusive with respect to comparing the approaches. The success of the spectral basis with binary data and similar results with count data suggest that it may be generally useful in spatial models and more complicated hierarchical models.

A spatial-temporal approach to surveillance of prostate cancer disparities in population subgroups

BACKGROUND: Prostate cancer mortality disparities exist among racial/ethnic groups in the United States, yet few studies have explored the spatiotemporal trend of the disease burden. To better understand mortality disparities by geographic regions over time, the present study analyzed the geographic variations of prostate cancer mortality by three Texas racial/ethnic groups over a 22-year period. METHODS: The Spatial Scan Statistic developed by Kulldorff et al was used. Excess mortality was detected using scan windows of 50% and 90% of the study period and a spatial cluster size of 50% of the population at risk. Time trend was analyzed to examine the potential temporal effects of clustering. Spatial queries were used to identify regions with multiple racial/ethnic groups having excess mortality. RESULTS: The most likely area of excess mortality for blacks occurred in Dallas-Metroplex and upper east Texas areas between 1990 and 1999; for Hispanics, in central Texas between 1992 and 1996: and for non-Hispanic whites, in the upper south and west to central Texas areas between 1990 and 1996. Excess mortality persisted among all racial/ethnic groups in the identified counties. The second scan revealed that three counties in west Texas presented an excess mortality for Hispanics from 1980-2001. Many counties bore an excess mortality burden for multiple groups. There is no time trend decline in prostate cancer mortality for blacks and non-Hispanic whites in Texas. CONCLUSION: Disparities in prostate cancer mortality among racial/ethnic groups existed in Texas. Central Texas counties with excess mortality in multiple subgroups warrant further investigation.

Spatial-temporal clustering of companion animal enteric syndrome: detection and investigation through the use of electronic medical records from participating private practices

SUMMARY There is interest in the potential of companion animal surveillance to provide data to improve pet health and to provide early warning of environmental hazards to people. We implemented a companion animal surveillance system in Calgary, Alberta and the surrounding communities. Informatics technologies automatically extracted electronic medical records from participating veterinary practices and identified cases of enteric syndrome in the warehoused records. The data were analysed using time-series analyses and a retrospective space-time permutation scan statistic. We identified a seasonal pattern of reports of occurrences of enteric syndromes in companion animals and four statistically significant clusters of enteric syndrome cases. The cases within each cluster were examined and information about the animals involved (species, age, sex), their vaccination history, possible exposure or risk behaviour history, information about disease severity, and the aetiological diagnosis was collected. We then assessed whether the cases within the cluster were unusual and if they represented an animal or public health threat. There was often insufficient information recorded in the medical record to characterize the clusters by aetiology or exposures. Space-time analysis of companion animal enteric syndrome cases found evidence of clustering. Collection of more epidemiologically relevant data would enhance the utility of practice-based companion animal surveillance.

Taxonomy, phylogeny and molecular epidemiology of Echinococcus multilocularis: From fundamental knowledge to health ecology.

Alveolar echinococcosis, caused by the tapeworm Echinococcus multilocularis, is one of the most severe parasitic diseases in humans and represents one of the 17 neglected diseases prioritised by the World Health Organisation (WHO) in 2012. Considering the major medical and veterinary importance of this parasite, the phylogeny of the genus Echinococcus is of considerable importance; yet, despite numerous efforts with both mitochondrial and nuclear data, it has remained unresolved. The genus is clearly complex, and this is one of the reasons for the incomplete understanding of its taxonomy. Although taxonomic studies have recognised E. multilocularis as a separate entity from the Echinococcus granulosus complex and other members of the genus, it would be premature to draw firm conclusions about the taxonomy of the genus before the phylogeny of the whole genus is fully resolved. The recent sequencing of E. multilocularis and E. granulosus genomes opens new possibilities for performing in-depth phylogenetic analyses. In addition, whole genome data provide the possibility of inferring phylogenies based on a large number of functional genes, i.e. genes that trace the evolutionary history of adaptation in E. multilocularis and other members of the genus. Moreover, genomic data open new avenues for studying the molecular epidemiology of E. multilocularis: genotyping studies with larger panels of genetic markers allow the genetic diversity and spatial dynamics of parasites to be evaluated with greater precision. There is an urgent need for international coordination of genotyping of E. multilocularis isolates from animals and human patients. This could be fundamental for a better understanding of the transmission of alveolar echinococcosis and for designing efficient healthcare strategies.

Spatial analysis of water body distance and West Nile Virus human cases in Houston, TX

West Nile Virus (WNV) is an arboviral disease that has affected hundreds of residents in Harris County, Texas since its introduction in 2002. Persistent infection, lingering sequelae and other long-term symptoms of patients reaffirm the need for prevention of this important vector-borne disease. This study aimed to determine if living within 400m of a water body increases one’s odds of infection with WNV. Additionally, we wanted to determine if one’s proximity to a particular water type or water body source increased one’s odds of infection with WNV.^ 145 cases’ addresses were abstracted from the initial interview and consent records from a cohort of patients (Epidemiology of Arboviral Encephalitis in Houston study, HSC-SPH-03-039). After applying inclusion criteria, 140 cases were identified for analysis. 140 controls were selected for analysis using a population proportionate to size model and US Census Bureau data. MapMarker USA v14 was used to geocode the cases’ addresses. Both cases’ and controls’ coordinates were uploaded onto a Harris County water shapefile in MapInfo Professional v9.5.1. Distance in meters to the closest water source, closest water source type, and closest water source name were recorded.^ Analysis of Variance (p=0.329, R2 = 0.0034) indicated no association between water body distance and risk of WNV disease. Living near a creek (x2 = 11.79, p < 0.001), or the combined group of creek and gully (x 2 = 14.02, p < 0.001) were found to be strongly associated with infection of WNV. Living near Cypress Creek and its feeders (x2 = 15.2, p < 0.001) was found to be strongly associated with WNV infection. We found that creek and gully habitats, particularly Cypress Creek, were preferential for the local disease transmitting Culex quinquefasciatus and reservoir avian population.^

A Bayesian analysis for spatial processes with application to mapping

In geographical epidemiology, maps of disease rates and disease risk provide a spatial perspective for researching disease etiology. For rare diseases or when the population base is small, the rate and risk estimates may be unstable. Empirical Bayesian (EB) methods have been used to spatially smooth the estimates by permitting an area estimate to "borrow strength" from its neighbors. Such EB methods include the use of a Gamma model, of a James-Stein estimator, and of a conditional autoregressive (CAR) process. A fully Bayesian analysis of the CAR process is proposed. One advantage of this fully Bayesian analysis is that it can be implemented simply by using repeated sampling from the posterior densities. Use of a Markov chain Monte Carlo technique such as Gibbs sampler was not necessary. Direct resampling from the posterior densities provides exact small sample inferences instead of the approximate asymptotic analyses of maximum likelihood methods (Clayton & Kaldor, 1987). Further, the proposed CAR model provides for covariates to be included in the model. A simulation demonstrates the effect of sample size on the fully Bayesian analysis of the CAR process. The methods are applied to lip cancer data from Scotland, and the results are compared. ^

Spatial analysis of pertussis cases in Harris County from 2005-2010

Objective: This study examined the recent trends and characteristics of reported pertussis in Harris County from 2005-2010. ^ Methods: The study population included surveillance data from all reported pertussis cases from January 1, 2005 to December 31, 2010 to Harris County Public Health and Environmental Services (HCPHES). We calculated incidence and attack rates for varying age groups, race/ethnicity, and gender. Spatial analyses were conducted of hot spot and cluster of incident cases in Harris County census tracts. Maps were constructed using geographic information system. ^ Results: Age-specific incidence rates of reported cases of pertussis were highest among infants under a year of age and lowest among adults age 20 and older. Hispanics represented the most cases reported compared to any other race or ethnic group (42% of 483 cases). Age-adjusted rates were highest in 2009 at 9.81 cases per 100,000 population. Only 31.2% of people received at least four of the recommended five doses of vaccine. Spatial analyses revealed statistically significant clusters within the northeast region of Harris County. ^ Conclusions: Hispanic infants are the most at risk group for pertussis. Although 70% of cases had a history of immunization, 41.8% of infants were appropriately vaccinated for their age. Increased vaccination coverage may decrease the incidence of pertussis.^

Spatial analysis of cardiovascular disease mortality and air pollution in Harris County, Texas

The association between fine particulate matter air pollution (PM2.5) and cardiovascular disease (CVD) mortality was spatially analyzed for Harris County, Texas, at the census tract level. The objective was to assess how increased PM2.5 exposure related to CVD mortality in this area while controlling for race, income, education, and age. An estimated exposure raster was created for Harris County using Kriging to estimate the PM2.5 exposure at the census tract level. The PM2.5 exposure and the CVD mortality rates were analyzed in an Ordinary Least Squares (OLS) regression model and the residuals were subsequently assessed for spatial autocorrelation. Race, median household income, and age were all found to be significant (p<0.05) predictors in the model. This study found that for every one μg/m3 increase in PM2.5 exposure, holding age and education variables constant, an increase of 16.57 CVD deaths per 100,000 would be predicted for increased minimum exposure values and an increase of 14.47 CVD deaths per 100,000 would be predicted for increased maximum exposure values. This finding supports previous studies associating PM2.5 exposure with CVD mortality. This study further identified the areas of greatest PM2.5 exposure in Harris County as being the geographical locations of populations with the highest risk of CVD (i.e., predominantly older, low-income populations with a predominance of African Americans). The magnitude of the effect of PM2.5 exposure on CVD mortality rates in the study region indicates a need for further community-level studies in Harris County, and suggests that reducing excess PM2.5 exposure would reduce CVD mortality.^

Using spatial linear models with SAR and CAR structure to examine Texas lung cancer incidence rates

Scholars have found that socioeconomic status was one of the key factors that influenced early-stage lung cancer incidence rates in a variety of regions. This thesis examined the association between median household income and lung cancer incidence rates in Texas counties. A total of 254 individual counties in Texas with corresponding lung cancer incidence rates from 2004 to 2008 and median household incomes in 2006 were collected from the National Cancer Institute Surveillance System. A simple linear model and spatial linear models with two structures, Simultaneous Autoregressive Structure (SAR) and Conditional Autoregressive Structure (CAR), were used to link median household income and lung cancer incidence rates in Texas. The residuals of the spatial linear models were analyzed with Moran's I and Geary's C statistics, and the statistical results were used to detect similar lung cancer incidence rate clusters and disease patterns in Texas.^

Higher-order co-occurrences for exploratory point pattern analysis and decision tree clustering on spatial data

Analyzing geographical patterns by collocating events, objects or their attributes has a long history in surveillance and monitoring, and is particularly applied in environmental contexts, such as ecology or epidemiology. The identification of patterns or structures at some scales can be addressed using spatial statistics, particularly marked point processes methodologies. Classification and regression trees are also related to this goal of finding "patterns" by deducing the hierarchy of influence of variables on a dependent outcome. Such variable selection methods have been applied to spatial data, but, often without explicitly acknowledging the spatial dependence. Many methods routinely used in exploratory point pattern analysis are2nd-order statistics, used in a univariate context, though there is also a wide literature on modelling methods for multivariate point pattern processes. This paper proposes an exploratory approach for multivariate spatial data using higher-order statistics built from co-occurrences of events or marks given by the point processes. A spatial entropy measure, derived from these multinomial distributions of co-occurrences at a given order, constitutes the basis of the proposed exploratory methods. © 2010 Elsevier Ltd.

Higher-order co-occurrences for exploratory point pattern analysis and decision tree clustering on spatial data

Analyzing geographical patterns by collocating events, objects or their attributes has a long history in surveillance and monitoring, and is particularly applied in environmental contexts, such as ecology or epidemiology. The identification of patterns or structures at some scales can be addressed using spatial statistics, particularly marked point processes methodologies. Classification and regression trees are also related to this goal of finding "patterns" by deducing the hierarchy of influence of variables on a dependent outcome. Such variable selection methods have been applied to spatial data, but, often without explicitly acknowledging the spatial dependence. Many methods routinely used in exploratory point pattern analysis are2nd-order statistics, used in a univariate context, though there is also a wide literature on modelling methods for multivariate point pattern processes. This paper proposes an exploratory approach for multivariate spatial data using higher-order statistics built from co-occurrences of events or marks given by the point processes. A spatial entropy measure, derived from these multinomial distributions of co-occurrences at a given order, constitutes the basis of the proposed exploratory methods. © 2010 Elsevier Ltd.

Improving the health of Australians by applying evidence from behavioural epidemiology to urban design projects

Rapid urban population growth in Australia requires an expansion of supporting hard and soft infrastructure. In the State of Victoria, directing this growth are a number of urban design and planning mechanisms that provide a ‘blueprint for development and investment’. Although topics revolving around physical health are present in these and other planning related documents, largely absent from this literature are ‘tools’ to assist decision makers in determining whether or not an urban setting supports physical health and provides opportunities for physical activity. Insufficient physical activity is a risk factor contributing to Australia’s growing and significant burden of chronic disease including cardiovascular disease, Type 2 diabetes and overweight/obesity. The potential of the built environment to influence population-level physical activity is well recognised. A key element in Victoria’s planning framework that can help address these health concerns is the provision and redevelopment of open space(s) in urban areas that provide opportunities for people of all ages and abilities to engage in physical activity. However, in the realisation of these settings, evidence informing the design of urban open space(s) that promote opportunities for physical activity is needed to produce evidence based decision making. Using the three geo-spatial visioning layers embedded in Victoria’s planning framework (i.e. Growth Area Framework Plans, Precinct Structure Plans and Planning Permits) as positioning instruments, this paper merges the fields of behavioural epidemiology and urban design to: i) provide a brief overview of current research relating to design of open space to optimise usage and physical activity, ii) consider what type of evidence relating to features of open space is needed to help inform decision makers, iii) consider the methods and procedures practitioners may use to incorporate evidence in to their planning, and iv) discuss the geo-spatial development level that the respective data can best assist decision making to achieve positive gains in physical health.