Border analysis for spatial clusters

Abstract

BackgroundThe spatial scan statistic is widely used by public health professionals in the detection of spatial clusters in inhomogeneous point process. The most popular version of the spatial scan statistic uses a circular-shaped scanning window. Several other variants, using other parametric or non-parametric shapes, are also available. However, none of them offer information about the uncertainty on the borders of the detected clusters.MethodWe propose a new method to evaluate uncertainty on the boundaries of spatial clusters identified through the spatial scan statistic for Poisson data. For each spatial data location i, a function F(i) is calculated. While not a probability, this function takes values in the [0, 1] interval, with a higher value indicating more evidence that the location belongs to the true cluster.ResultsThrough a set of simulation studies, we show that the F function provides a way to define, measure and visualize the certainty or uncertainty of each specific location belonging to the true cluster. The method can be applied whether there are one or multiple detected clusters on the map. We illustrate the new method on a data set concerning Chagas disease in Minas Gerais, Brazil.ConclusionsThe higher the intensity given to an area, the higher the plausibility of that particular area to belong to the true cluster in case it exists. This way, the F function provides information from which the public health practitioner can perform a border analysis of the detected spatial scan statistic clusters. We have implemented and illustrated the border analysis F function in the context of the circular spatial scan statistic for spatially aggregated Poisson data. The definition is clearly independent of both the shape of the scanning window and the probability model under which the data is generated. To make the new method widely available to users, it has been implemented in the freely available SaTScan software www.satscan.org.

Spatial clustering and contextual factors associated with hospitalisation and deaths due to COVID-19 in Sweden: a geospatial nationwide ecological study

“…Cluster detection is essential in epidemiological surveillance since it indicates areas with excess disease incidence, prevalence or mortality. 12 The frequent reports of the disease in Sweden indicate that the hospitalisations and deaths do not show a homogeneous geospatial distribution in the country. 13 However, to the best of our knowledge, no nationwide study has characterised the spatial disparity of COVID-19 hospitalisations and mortality and its association with underlying factors.…”

Geo-Epidemiology of Malaria at the Health Area Level, Dire Health District, Mali, 2013–2017

“…An elliptical window was centered on the different health areas with a radius ranging from 1 to 50% of the population [33]. The Oliveira test was performed to account for the edge effects of the detected hotspots and coldspots [34]. Malaria incidence was mapped based on this spatio-temporal information, and the location of high and low risk health areas by time period was added to the generated maps.…”

Contrasting Methods of Measurement in Spatial Analyses Examining the Alcohol Environment and Child Maltreatment

“…Statistical significance is assessed by obtaining p-values for each cluster through Monte Carlo simulation, which compares the rank of the maximum likelihood test from the dataset with the maximum likelihoods from randomly generated simulated datasets (Kulldorff, 2015). SaTScan also generates a relative risk ratio for each cluster, defined as the estimated risk in the cluster divided by the estimated risk outside the cluster (Kulldorff, 2015) Because spatial scan statistics result in clusters of general location and size, but with uncertain borders (Oliveira et al, 2018), we ran our models utilizing the Oliveira's F function, which SaTScan calculates by generating M random datasets (we chose M = 1000) with the expected cases in location I set to equal the number of observed cases. The spatial scan statistic is applied to each random dataset using the original population counts, and F(i) is defined as the proportion (a value between [0, 1]) of the most likely clusters in M runs that contain the location I (Oliveira et al, 2018).…”

“…SaTScan also generates a relative risk ratio for each cluster, defined as the estimated risk in the cluster divided by the estimated risk outside the cluster (Kulldorff, 2015) Because spatial scan statistics result in clusters of general location and size, but with uncertain borders (Oliveira et al, 2018), we ran our models utilizing the Oliveira's F function, which SaTScan calculates by generating M random datasets (we chose M = 1000) with the expected cases in location I set to equal the number of observed cases. The spatial scan statistic is applied to each random dataset using the original population counts, and F(i) is defined as the proportion (a value between [0, 1]) of the most likely clusters in M runs that contain the location I (Oliveira et al, 2018). We included the block in the border of the cluster if I > 0.75.…”