Besag & Newell's Test
|
|
|
|
Indications/Recommendations for use: Scan for clusters of minimum
size (k). This test is designed for data aggregated into regions
with small population sizes.
|
|
Description: The population size containing the closest
k cases surrounding each region is compared to that expected
under a common disease frequency.
|
|
Test statistic: The number of regions (L)
needed to include the nearest k cases surrounding each region.
Significance is evaluated with a poisson distribution for each potential
cluster.
|
Stat:
|
|
Null Hypothesis: The number of cases in a region follows a Poisson
distribution with a common rate.
|
Ho:
|
|
Alternative Hypothesis: Not the Null Hypothesis.
|
Ha:
|
|
GeoMed Inputs: Region-level data with a centroid coordinate, case
counts, and population at risk per region; cluster size k to detect.
|
|
GeoMed Outputs: Tabled information includes test statistic and
p-value for each significant cluster; mapped information provides cluster
locations.
|
|
Example Analysis
|
Reference: Besag, J. and Newell, J. 1991. The detection of clusters
in rare diseases, Journal of the Royal Statistical Society, Series A,
154:143-155.
|