Adaptive kernel estimation of spatial relative risk

Davies, Tilman M.; Hazelton, Martin L.

doi:10.1002/sim.3995

Cited by 88 publications

(126 citation statements)

References 23 publications

Supporting

Mentioning

112

Contrasting

Unclassified

Order By: Relevance

“…Let x 1 , x 2 , …, x n be the geographical coordinates of n cases in the study area. The kernel density estimate of cases is written as (Davies and Hazelton, 2010),…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Spatial variation in mortality risk for hematological malignancies near a petrochemical refinery: A population-based case-control study

Salvo

Meneghini

Vieira

et al. 2015

Environmental Research

View full text Add to dashboard Cite

Introduction The study investigated the geographic variation of mortality risk for hematological malignancies (HMs) in order to identify potential high-risk areas near an Italian petrochemical refinery. Material and methods A population-based case-control study was conducted and residential histories for 171 cases and 338 sex- and age-matched controls were collected. Confounding factors were obtained from interviews with consenting relatives for 109 HM deaths and 267 controls. To produce risk mortality maps, two different approaches were applied. We mapped (1) adptive kernel density relative risk estimation (KDE) for case-control studies which estimates a spatial relative risk function using the ratio between cases and controls’ densities, and (2) estimated odds ratios for case-control study data using generalized additive models (GAMs) to smooth the effect of location, a proxy for exposure, while adjusting for confounding variables. Results No high-risk areas for HM mortality were identified among all subjects (men and women combined), by applying both approaches. Using the adaptive KDE approach, we found a significant increase in death risk only among women in a large area 2–6 km southeast of the refinery and the application of GAMs also identified a similarly-located significant high-risk area among women only (global p-value<0.025). Potential confounding risk factors we considered in the GAM did not alter the results. Conclusion Both approaches identified a high-risk area close to the refinery among women only. Those spatial methods are useful tools for public policy management to determine priority areas for intervention. Our findings suggest several directions for further research in order to identify other potential environmental exposures that may be assessed in forthcoming studies based on detailed exposure modeling.

show abstract

“…Let x 1 , x 2 , …, x n be the geographical coordinates of n cases in the study area. The kernel density estimate of cases is written as (Davies and Hazelton, 2010),…”

Section: Methodsmentioning

confidence: 99%

“…To reach the aim, we applied and compared two methods: (1) smoothing of location using Generalized Additive Models (GAMs) to produce risk maps adjusted for potential confounders (Vieira et al, 2002; Webster et al, 2006) and (2) adaptive kernel density relative risk estimation (KDE) (Davies and Hazelton, 2010). …”

Section: Introductionmentioning

confidence: 99%

Spatial variation in mortality risk for hematological malignancies near a petrochemical refinery: A population-based case-control study

Salvo

Meneghini

Vieira

et al. 2015

Environmental Research

View full text Add to dashboard Cite

show abstract

“…To solve the problem of variable population sizes in fixed radius filters, several papers (Davies and Hazelton 2010;Talbot et al 2000;Tiwari and Rushton 2004) proposed using adaptive spatial filters that adjusts the radius of the filter based on population density and maintains a constant population size within each filter to stabilize estimation errors in the disease rates. These works, however, did not provide any criteria for optimal population sizes.…”

Section: Spatially Adaptive Filtersmentioning

confidence: 99%

Validation tests of an improved kernel density estimation method for identifying disease clusters

2011

View full text Add to dashboard Cite

The spatial filter method, which belongs to the class of kernel density estimation methods, has been used to make morbidity and mortality maps in several recent studies. We propose improvements in the method to include spatially adaptive filters to achieve constant standard error of the relative risk estimates; a staircase weight method for weighting observations to reduce estimation bias; and a parameter selection tool to enhance disease cluster detection performance, measured by sensitivity, specificity, and false discovery rate. We test the performance of the method using Monte Carlo simulations of hypothetical disease clusters over a test area of four counties in Iowa. The simulations include different types of spatial disease patterns and high-resolution population distribution data. Results confirm that the new features of the spatial filter method do substantially improve its performance in realistic situations comparable to those where the method is likely to be used.

show abstract

“…[8] proposed a method to better estimate the relative risk of diseases using an adaptive bandwidth method for density estimation that is specially catered for estimating relative risks (i.e. the ratio of risks).…”

Section: Related Workmentioning

confidence: 99%

“…The adaptive bandwidth method of [8] is available in an R package sparr. We applied it to our data sets but the method did not seem to work well, with extreme values around the borders of our study region.…”

Section: Related Workmentioning

confidence: 99%

Comparing spatial counts characterizing human mobility using ratio maps

Loh

2013

Statistics and Its Interface

View full text Add to dashboard Cite

Given two spatial data sets collected over the same geographical region, a common question of interest is the differences between these two data sets. The data may have been collected at two time points and the differences represent how the data has changed over the intervening time period. Or they may be data on two sub-populations and the differences then describe where and how these two subpopulations vary relative to each other. We consider the problem of objectively highlighting the differences between two spatial data sets, and showing these differences on a map. Specifically, we assume that the data sets are collected over a lattice, and model the values of one set as a function of those of the other data set, taking into account the inherent spatial structure. The model is a hierarchical model with spatial errors whose specification controls the amount of smoothing applied to the data. Additional covariates may be added to the model to explore more complex relationships between the two data sets. We apply the method to study the number of commuters to a town as obtained from cellular Call Detail Records, comparing them to Census data, and also model the commuter numbers with underlying population and distance away from the town. We also study differences in the number of visitors to the town during a St. Patrick's Day parade with visitor numbers on other days.

show abstract

Adaptive kernel estimation of spatial relative risk

Cited by 88 publications

References 23 publications

Spatial variation in mortality risk for hematological malignancies near a petrochemical refinery: A population-based case-control study

Spatial variation in mortality risk for hematological malignancies near a petrochemical refinery: A population-based case-control study

Validation tests of an improved kernel density estimation method for identifying disease clusters

Comparing spatial counts characterizing human mobility using ratio maps

Contact Info

Product

Resources

About