Skip to main content

Table 3 Comparisons of spatial clusters from multiple cluster identification methods

From: Methods used in the spatial analysis of tuberculosis epidemiology: a systematic review

Author, year

Methods

Outcome

Conclusion

Alene, K, 2017 [49]

Local Moran’s I

Getis and Ord

Clustered

Clustered

50% similarity (two non-significant clusters identified by LISA)

Álvarez-Hernández, G., et al. 2010 [145]

Local Moran’s I

Besag and Newell

No significant Clustered

Widely conflicting

Dangisso M, et al. 2015 [26]

Getis and Ord

Spatial scan statistic

Clustered

Clustered

Similar overall pattern, but marked differences by years

Feske, M., et al. 2011 [93, 178]

Getis and Ord

GWR residuals

Clustered

Heterogeneous

Similar overall pattern, but some local differences

Ge E, et al. 2016 [139]

Getis and Ord

Spatial scan statistic

Clustered

Clustered

Similar overall pattern, but differences in some locations and across time

Haase I, et al. 2007 [2]

Hotspot analysis

SaTScan

Clustered

Clustered

Similar overall pattern, but some local differences

Hassarangsee S, et al. 2015 [138

LISA

Spatial scan statistic

Clustered

Clustered

Very similar, but not identical

Li L, et al. 2016 [135]

LISA

Spatial scan statistic

No significant cluster, Clustered

Widely conflicting

Maceiel ELN, et al. 2010 [131]

LISA, Getis and Ord

Model prediction

Clustered

Heterogeneous

Widely conflicting

Wubuli A, et al. 2015 [16]

LISA

Getis and Ord

Clustered

Clustered

Similar overall pattern, but some local differences

Wang T, et al 2016 [102]

Spatial scan statistic

Getis and Ord

Clustered

Clustered

Similar overall pattern, but some local differences

  1. GWR geographically weighted regression; LISA local indicators of spatial association