Skip to main content

Table 3 Applied cluster detection methods according to spatial resolution and global/local estimation (n = 27 studies)

From: An overview of geospatial methods used in unintentional injury epidemiology

Method

Spatial resolution

Global/

local

Total studiesa

Injury category (number of studies)

References

Kernel density estimation

point

local

10

Road traffic (n = 7)

Falls (n = 2)

Drowning(n = 1)

(Cinnamon et al. 2011; Dai et al. 2013; Lai et al. 2009b; Lai et al. 2011; Lawrence et al. 2015; Morency and Cloutier 2006; Nagata et al. 2011; Schuurman et al. 2009; Slaughter et al. 2014; Weiner and Tepas 2009)

Nearest neighbour

hierarchical

point

global

4

Falls (n = 3)

Drowning (n = 1)

(Lai et al. 2009a; Lai et al. 2009b; Lai et al. 2011; Shenoi et al. 2015)

Nearest neighbour

index

point

global

1

Road traffic (n = 1)

(Nunn and Newby 2015)

Spatial scan statistics

point or areal

local

4

Falls (n = 2)

Occupational (n = 1) Poisoning(n = 1)

(Dey et al. 2010; Forst et al. 2015; Nkhoma et al. 2004; Yiannakoulias et al. 2003)

Moran’s I

areal

global

13

Road traffic (n = 8)

Falls (n = 1)

Burns (n = 2)

Drowning (n = 1)

Occupational (n = 1)

(de Pina et al. 2008; Erdogan 2009; Forst et al. 2015; Goltsman et al. 2014; Heng et al. 2015; Jones et al. 2008; Lassarre and Thomas 2005; Lawrence et al. 2015; Nunes and Nascimento 2012; Nunn and Newby 2015; Poulos et al. 2012; Shenoi et al. 2015; Silva et al. 2011)

Geary’s c

areal

global

2

Road traffic (n = 2)

(Erdogan 2009; Lassarre and Thomas 2005)

Local indicators of spatial association

areal

local

5

Road traffic (n = 3)

Drowning (n = 1)

Falls (n = 1)

(Dai et al. 2013; de Pina et al. 2008; Erdogan 2009; Nunes and Nascimento 2012; Nunn and Newby 2015)

Getis Ord statistics

areal

local

4

Road traffic (n = 3)

Burn (n = 1)

(Erdogan 2009; Goltsman et al. 2014; Slaughter et al. 2014; Statter et al. 2011)

  1. atotal number of studies by injury category is not equal to (n = 27) because some studies applied more than one method in a single study