Early detection of aberrations in incoming data streams can be of critical importance in a variety of medical, public health, and bio-security applications: adverse reactions to prescription drugs, outbreaks of infectious diseases, clusters of environmentally triggered health problems, etc. Failing to discover any of these effects can lead to otherwise avoidable deaths. Many mathematical and statistical techniques have been developed to try and identify problems as early as possible after they've begun. This program will provide a basic overview of these methods and under which sorts of circumstances each is best employed. We will also cover a few of the more interesting potential new methods currently being investigated for use in the field.

As part of the program, we'll also discuss how these methods can be "fooled" inadvertently by practical constraints of data collection, surveillance, and human behavior.

This program will be appropriate for teaching faculty, anyone involved in educating or leading human, agricultural animal, or veterinary health workers involved in surveillance data collection and/or analysis. Time will be spent on many aspects of the problem, making this program appropriate for those already having expertise in either math and statistics, or biology and medicine, but wishing to gain an understanding of the other and how these fields inter-relate within the scope of biosurveillance. This program will provide both theoretical understanding and practical skills.

Prerequisites:

Some background in statistics and some previous exposure to any field in which empirically collected data is quantitatively analyzed will be helpful, but not strictly necessary.

References: