Effect of relative risk and cluster configuration on the power of the one-dimensional scan statistic

Statistical Power for Postlicensure Medical Product Safety Data Mining

OBJECTIVE

To perform sample size calculations when using tree-based scan statistics in longitudinal observational databases.

METHODS

Tree-based scan statistics enable data mining on epidemiologic datasets where thousands of disease outcomes are organized into hierarchical tree structures with automatic adjustment for multiple testing. We show how to evaluate the statistical power of the unconditional and conditional Poisson versions. The null hypothesis is that there is no increase in the risk for any of the outcomes. The alternative is that one or more outcomes have an excess risk. We varied the excess risk, total sample size, frequency of the underlying event rate, and the level of across-the-board health care utilization. We also quantified the reduction in statistical power resulting from specifying a risk window that was too long or too short.

RESULTS

For 500,000 exposed people, we had at least 98 percent power to detect an excess risk of 1 event per 10,000 exposed for all outcomes. In the presence of potential temporal confounding due to across-the-board elevations of health care utilization in the risk window, the conditional tree-based scan statistic controlled type I error well, while the unconditional version did not.

DISCUSSION

Data mining analyses using tree-based scan statistics expand the pharmacovigilance toolbox, ensuring adequate monitoring of thousands of outcomes of interest while controlling for multiple hypothesis testing. These power evaluations enable investigators to design and optimize implementation of retrospective data mining analyses.

[Basic principles of statistical methods applied to public health surveillance data analysis: a review]

Statistical methods in public health surveillance data analysis are important for detecting temporal and/or spatial disease clusters, which can indicate an outbreak or an epidemic. In this paper we present a review and literature survey introducing the public health surveillance process and focusing on some statistical methods which are indicated to the data analysis in routine procedure. Comparative studies of these methods are also analyzed.

The power of focused tests to detect disease clustering

Statistical tests have been proposed for determining whether incident cases of adverse health effects are 'clustered' together. Several procedures, termed 'focused', specifically analyse disease surveillance data around pre-specified putative sources of environmental hazard. Little has been done to compare the performance of various proposed methods on actual models of clustering. Analytic power functions are derived for three tests of focused clustering. These functions are based on the probabilistic structure of the clustering tests and do not require simulation. The three tests are compared with respect to statistical power on hypothetical data where monotone multiplicative increases in disease risk near a putative hazard define disease clusters of varying intensity.