smacpod (version 2.6)

spscan.test: Spatial Scan Test

Description

spscan.test performs the spatial scan test of Kulldorf (1997) for case/control point data.

Usage

spscan.test(
  x,
  case = 2,
  nsim = 499,
  alpha = 0.1,
  maxd = NULL,
  cl = NULL,
  longlat = FALSE
)

Value

Returns a list of length two of class

scan. The first element (clusters) is a list containing the significant, non-overlapping clusters, and has the the following components:

coords

The centroid of the significant clusters.

r

The radius of the window of the clusters.

pop

The total population in the cluser window.

cases

The observed number of cases in the cluster window.

expected

The expected number of cases in the cluster window.

smr

Standarized mortaility ratio (observed/expected) in the cluster window.

rr

Relative risk in the cluster window.

propcases

Proportion of cases in the cluster window.

loglikrat

The loglikelihood ratio for the cluster window (i.e., the log of the test statistic).

pvalue

The pvalue of the test statistic associated with the cluster window.

Various additional pieces of information are included for plotting, printing

Arguments

x

A ppp object with marks for the case and control groups.

case

The name of the desired "case" group in levels(x$marks). Alternatively, the position of the name of the "case" group in levels(x$marks). Since we don't know the group names, the default is 2, the second position of levels(x$marks). x$marks is assumed to be a factor. Automatic conversion is attempted if it is not.

nsim

The number of simulations from which to compute the p-value. A non-negative integer. Default is 499.

alpha

The significance level to determine whether a cluster is signficant. Default is 0.1.

maxd

The radius of the largest possible cluster to consider. Default is NULL, i.e., half the maximum intercentroid distance.

cl

A cluster object created by makeCluster, or an integer to indicate number of child-processes (integer values are ignored on Windows) for parallel evaluations (see Details on performance). It can also be "future" to use a future backend (see Details), NULL (default) refers to sequential evaluation.

longlat

A logical value indicating whether Euclidean distance (FALSE) or Great Circle (WGS84 ellipsoid, FALSE) should be used. Default is FALSE, i.e., Euclidean distance.

Author

Joshua French

Details

The test is performed using the random labeling hypothesis. The windows are circular and extend from the observed data locations. The clusters returned are non-overlapping, ordered from most significant to least significant. The first cluster is the most likely to be a cluster. If no significant clusters are found, then the most likely cluster is returned (along with a warning).

Setting cl to a positive integer MAY speed up computations on non-Windows computers. However, parallelization does have overhead cost, and there are cases where parallelization results in slower computations.

References

Kulldorff M., Nagarwalla N. (1995) Spatial disease clusters: Detection and Inference. Statistics in Medicine 14, 799-810.

Kulldorff, M. (1997) A spatial scan statistic. Communications in Statistics -- Theory and Methods 26, 1481-1496.

Waller, L.A. and Gotway, C.A. (2005). Applied Spatial Statistics for Public Health Data. Hoboken, NJ: Wiley.

Examples

data(grave)
# apply scan method
out = spscan.test(grave, case = "affected", nsim = 99)
# print scan object
out
print(out, extra = TRUE)
# summarize results
summary(out)
# plot results
plot(out, chars = c(1, 20), main = "most likely cluster")
# extract clusters from out 
# each element of the list gives the location index of the events in each cluster
clusters(out)
# get warning if no significant cluster
out2 = spscan.test(grave, case = 2, alpha = 0.001, nsim = 99)