Studies of disease patterns frequently result from allegations made by concerned citizens who identify patterns within small structured environments such as households, families, sibships, wards, cellblocks, classrooms, job types, and locations within a building. The detection of disease clusters is a serious problem in community medicine. Foremost among the difficulties faced by health professionals is that the data sets often are small and inferences are based on a relative handful of observations. It is crucial for the health professional to know what tests are best in these small-environment problems, and to have these methods available in a user-friendly computer package. The objective of this project is to implement new, rapid, and exact combinatorial expressions for analysis of patterns of disease. The results of Phase I investigations into the power of both these and other previously published methods in structured small environments will be used to recommend different tests for different kinds of problems and different amounts Of data, and to provide a user-friendly, interactive program EPIC (Exact Probabilities for Incidence Clusters) that includes an intuitive interface and a thorough set of tutorials and guidelines to help the professional choose the best test for a particular problem.
Public health professionals often analyze alleged disease clusters in small environments where samples are rare. Based on Phase I findings, we will in Phase II develop EPIC, a microcomputer program for optimal and often exact statistical analysis of clusters in small, structured environments. Many of the techniques implemented have not before been available in the research and regulatory communities.