 | The 131st Annual Meeting (November 15-19, 2003) of APHA |
|
| The 131st Annual Meeting (November 15-19, 2003) of APHA |
Lori C. Hutwagner, PhD, DHS/ESB, Agency for Toxic Substance and Disease Registry, 1600 Clifton Road NE, MS- E31, Atlanta, GA 30333, 404-498-0575, lch1@cdc.gov and Howard S. Burkom, PhD, National Security Technology Department, Johns Hopkins Applied Physics Laboratory, 11100 Johns Hopkins Road, Laurel, MD 20723.
The international war against terrorism has heightened concerns over the possibility of covert attacks using weaponized pathogens. Public health agencies are using various types of consumer data to increase their capability of early warning against outbreaks of infectious disease. This presentation discusses aspects of using automatic alerting algorithms at the local or county level to monitor these data on a daily basis. It addresses the questions: “what should we be counting, what alerting methods are appropriate for our data, and what are the obstacles?” The selection of outcome variables depends on the nature of the data. For emergency and outpatient visit data, syndromic classification systems have been adopted so that a small number of syndrome counts can be tracked. For less specific, nontraditional data sources such as school absenteeism and over-the-counter remedy sales, data analyses can guide the selection of counts to be monitored. Specific examples will show how data features such as sparse counts and weekly and seasonal variations affect the performance of available statistical methods, including the widely used CDC EARS package. The known behavior of the monitored data counts can thus help in the choice of alerting algorithms with desired sensitivity and specificity. In some cases simple statistical tests may be automated to make the algorithm choice adaptive. We will also present algorithm performance analysis methods that can help to determine alerting thresholds given constraints on the expected rate of false alerts.
Learning Objectives:
Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.