|
Activity Number:
|
308
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 6, 2003 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #301928 |
|
Title:
|
Rapid Detection of Increased Rates of Disease in Small Regions
|
|
Author(s):
|
Peter A. Rogerson*+
|
|
Companies:
|
University of Buffalo
|
|
Address:
|
Wilkeson Hall, Buffalo, NY, 14261,
|
|
Keywords:
|
surveillance ; cumulative sums ; multiple testing ; local statistics
|
|
Abstract:
|
When there is interest in simultaneous monitoring of disease rates in many regions, one approach is to use independent monitoring schemes for each region, with a Bonferroni adjustment for the multiple testing. Another alternative would be to employ a multivariate cumulative sum system. Yet another possibility is to monitor the local statistics based upon a regional value and the values in surrounding regions. I focus upon this latter approach to optimize the detection of change when change occurs across more than one region. I employ both cumulative sum methods and Shewhart charts to detect changes in disease rates. A complicating factor is that the local statistics are spatially autocorrelated. A key question is how to adjust appropriately for the multiple testing. Some adjustment of the monitoring threshold is necessary, or too many false alarms will be indicated. A Bonferroni adjustment is too conservative. The alternative approaches are illustrated by using both simulated data and data on breast cancer mortality in the northeastern United States.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2003 program |
