An up-to-date review of scan statistics

Abolhassani, Ali; Prates, Marcos O.

doi:10.1214/21-ss132

Cited by 16 publications

(4 citation statements)

References 130 publications

(238 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a univariate framework, for example, Gaussian (Kulldorff, Huang, and Konty 2009), ordinal (Jung, Kulldorff, and Klassen 2007), zero-inflated (Cançado, da Silva, and (Cucala et al 2017;Frévent et al 2021;Frévent et al 2023;Kulldorff et al 2007;Neill and Cooper 2010;Smida et al 2022). The reader is referred to Abolhassani and Prates (2021) for a comprehensive review of spatial scan statistics. These methods have been widely applied in many fields, such as epidemiology (Genin et al 2020;Green et al 2006;Khan et al 2021;Marciano et al 2018), environmental science (Shi, Liu, and Zhong 2022;Wan et al 2020), oncology (Leiser et al 2020), criminology (Minamisava et al 2009), and astronomy (De La Fuente Marcos and De La Fuente Marcos 2008).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Shared‐Frailty Spatial Scan Statistic Model for Time‐to‐Event Data

Frévent,

Ahmed,

Dabo‐Niang

et al. 2024

Biometrical J

View full text Add to dashboard Cite

Spatial scan statistics are well‐known methods widely used to detect spatial clusters of events. Furthermore, several spatial scan statistics models have been applied to the spatial analysis of time‐to‐event data. However, these models do not take account of potential correlations between the observations of individuals within the same spatial unit or potential spatial dependence between spatial units. To overcome this problem, we have developed a scan statistic based on a Cox model with shared frailty and that takes account of the spatial dependence between spatial units. In simulation studies, we found that (i) conventional models of spatial scan statistics for time‐to‐event data fail to maintain the type I error in the presence of a correlation between the observations of individuals within the same spatial unit and (ii) our model performed well in the presence of such correlation and spatial dependence. We have applied our method to epidemiological data and the detection of spatial clusters of mortality in patients with end‐stage renal disease in northern France.

show abstract

Section: Introductionmentioning

confidence: 99%

“…2022). The reader is referred to Abolhassani and Prates (2021) for a comprehensive review of spatial scan statistics. These methods have been widely applied in many fields, such as epidemiology (Genin et al.…”

Section: Introductionmentioning

confidence: 99%

A Shared‐Frailty Spatial Scan Statistic Model for Time‐to‐Event Data

Frévent,

Ahmed,

Dabo‐Niang

et al. 2024

Biometrical J

View full text Add to dashboard Cite

show abstract

“…• Astronomy [36], [5], [18], [28] • Image analysis [21], [56], [3] • Criminology [37], [49], [23], [17], [19] • Ecology [51], [39] • Geography [41], [6], [38], [47] • Pattern recognition [22] • Biology [20], [42] • Epidemiology [46], [29], [16], [55] • Public transportation [52], [2], [24] To have an overview of the scan statistics, it is suggested to refer to [1] in which a comprehensive study is done. It includes:…”

Section: Introductionmentioning

confidence: 99%

Lambert-Uniform Scan Statistic: Hot Spot Detection of Covid-19 Using a New One-Parameter Scan Statistic

Abolhassani

2023

Preprint

View full text Add to dashboard Cite

The spatial scan statistic is widely applied for detection of spatial clusters. These types of clustersare particularly important in epidemiology. Timely warning about the spread of a disease in thesociety can prevent a disaste.Spatial scan statistic can be applied on different types of spatial data to detect the location ofclusters. However, only the two-parameter Beta distribution is conducted on ratios, percentagesand bounded spatial data in the case of hotspot detection which is based on the regression modeland needs explanatory variables.In this study we focus on the recently proposed one-parameter Lambert-Uniform distributionto construct a new scan statistic for spatial bounded data. We present algorithms for detectingirregular and circular spatial clusters based on the new Lambert-Uniform scan statistic. Theadvantages of Lambert-Uniform scan over the Beta scan are at least in two aspects. First, theLambert-Uniform distribution has only one parameter, hence, in the process of clustering, oneneeds to estimate less parameters comparing Beta scan. Second, unlike Beta scan, the LambertUniform scan does not require regression explanatory variables.We consider different scenarios for spatial cluster detection and check the efficiency of LambertUniform scan statistic with the help of common criteria in cluster detection, i.e., precision, recall,bias value, and F 1. We also discuss the importance of the number of data in each cell of themap. We propose to apply MCMC method in sparsely populated maps before performing clusterdetection procedure. The simulation results show that the Lambert-Uniform spatial scan statisticis a suitable method for clustering the bounded spatial data. As a practical example, we considera map of Covid-19 disease and perform clustering based on circular and irregular shape scan.

show abstract

“…It has been extended to work with various types of datasets (Abolhassani et al. 2021), such as Poisson‐distributed (Kulldorff 1997), binomial‐distributed (Kulldorff 1997), zero‐inflated (Cançado et al. 2014, 2017), survival (Huang et al.…”

Section: Introductionmentioning

confidence: 99%

Detecting the Spatial Clustering of Exposure–Response Relationships with Estimation Error: A Novel Spatial Scan Statistic

Wang

Zhang

et al. 2023

Biometrics

View full text Add to dashboard Cite

Detecting the spatial clustering of the exposure–response relationship (ERR) between environmental risk factors and health‐related outcomes plays important roles in disease control and prevention, such as identifying highly sensitive regions, exploring the causes of heterogeneous ERRs, and designing region‐specific health intervention measures. However, few studies have focused on this issue. A possible reason is that the commonly used cluster‐detecting tool, spatial scan statistics, cannot be used for multivariate spatial datasets with estimation error, such as the ERR, which is often defined by a vector with its covariance estimated by a regression model. Such spatial datasets have been produced in abundance in the last decade, which suggests the importance of developing a novel cluster‐detecting tool applicable for multivariate datasets with estimation error. In this work, by extending the classic scan statistic, we developed a novel spatial scan statistic called the estimation‐error‐based scan statistic (EESS), which is applicable for both univariate and multivariate datasets with estimation error. Then, a two‐stage analytic process was proposed to detect the spatial clustering of ERRs in practical studies. A published motivating example and a simulation study were used to validate the performance of EESS. The results show that the clusters detected by EESS can efficiently reflect the clustering heterogeneity and yield more accurate ERR estimates by adjusting for such heterogeneity.

show abstract

An up-to-date review of scan statistics

Cited by 16 publications

References 130 publications

A Shared‐Frailty Spatial Scan Statistic Model for Time‐to‐Event Data

A Shared‐Frailty Spatial Scan Statistic Model for Time‐to‐Event Data

Lambert-Uniform Scan Statistic: Hot Spot Detection of Covid-19 Using a New One-Parameter Scan Statistic

Detecting the Spatial Clustering of Exposure–Response Relationships with Estimation Error: A Novel Spatial Scan Statistic

Contact Info

Product

Resources

About