BEST PRACTICES - SPACE-TIME DISEASE SURVEILLANCE TOOLS

CrimeStat (version 3.0)

Features

Spatial statistics
Local and global spatial autocorrelation

Knox test
Mantel test
Spatial-temporal moving average

Risk adjusted nearest neighbour hierarchical clustering
Text and graph output

Comments

CrimeStat offers many spatial analysis functions, but few methods for space-time analysis, nor can purely spatial methods be implemented prospectively. Designed for the spatial analysis of crime, CrimeStat offers excellent facility for exploratory spatial data analysis, including many tests for spatial clustering. Many input file formats are supported in CrimeStat, including DBF and shapefiles, and input data can be specified as primary (cases and counts), secondary (controls and population) and reference grids.

Selected case studies

We were unable to find examples of CrimeStat as part of a surveillance system or space-time analysis.

Surveillance Objective(s)

Data exploration

User Expertise

CrimeStat is easy to use. The methods require varying degrees of statistical sophistication for proper interpretation. Extensive documentation on the methods and parameters is provided for download, with sample data. CrimeStat is a good program for those wishing to learn more about spatial analysis in general. For disease surveillance, CrimeStat is of limited value.

Key Considerations

Not suitable for cluster detection in a prospective context
Identify existence and scale of space-time interaction via the Knox and Mantel tests
Limited visualization capabilities (requires integration with a GIS)

More Information

Website: http://www.icpsr.umich.edu/CRIMESTAT/

Key Resource: CrimeStat III User Workbook: http://www.icpsr.umich.edu/CRIMESTAT/workbook.html.