BEST PRACTICES - SPACE-TIME DISEASE SURVEILLANCE TOOLS

R-Surveillance

Features

Time series surveillance models
Bayesian predictive posterior probabilities

Functions for simulating surveillance data to evaluate detection algorithms
Time varying Poisson CUSUM

Documentation is not helpful unless familiar with R and the methods implemented

Comments

The surveillance package implements a variety of algorithms for outbreak detection of disease surveillance data with a focus on temporal surveillance. Implementation within R adds the ability to easily integrate the methods with other statistical functions and graphing options in R. However, this flexibility comes at the cost of user expertise. R is a command line driven program that requires substantial investment in mastering. However, this allows the methods implemented in the surveillance package to be integrated into an automated, ongoing surveillance system.

Selected case studies

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

Surveillance Objective(s)

Outbreak detection, Algorithm testing

User Expertise

Users need to be familiar with statistical surveillance models and the use of R statistical programming software.

Key Considerations

Advanced knowledge of statistical modelling required
Full suite of statistical functions available in R
Flexible data input and output functions
Command line driven (no GUI)
Easily automated
Very limited spatial surveillance capability

More Information

Website: http://surveillance.r-forge.r-project.org/

Key Resource: Höhle, M. 2007. surveillance: An R package for the monitoring of infectious diseases. Computational Statistics 22, no. 4: 571-582.