DIMACS Working Group on Adverse Event/Disease Reporting, Surveillance, and Analysis I
DIMACS Subgroup on Adverse Event/Disease Reporting, Surveillance, and Analysis
Thursday, February 19, 2004
8:15 - 9:00 Breakfast and Registration
9:00 - 9:15 Welcome and Introductory Comments
Fred Roberts, DIMACS Director
9:15 - 10:00 Overview of Syndromic Surveillance
Henry Rolka, CDC
(This session will be primarily to provide orientation and context to the problem for
those in the audience who may not have been directly involved in applications in this
area. In parallel during this session the Multivariate Surveillance Panelists will
convene and prepare their session. Panel membership is open. The panel will moderated
by David Siegrist and present perspectives on multivariate surveillance from practitioners.)
10:00 - 10:45 Panel Presentation I
moderated by David Siegrist
Multiple Data Sources in the Harvard/Pilgrim
Biosurveillance System
Ken Kleiman, Harvard University
Monitoring of Multivariate Data in ESSENCE
Biosurveillance Systems
Howard Burkom, APL
Overview of WSARE, Fast Spatial Scan and PANDA
Andrew Moore, CMU
Site Surveillance Using Differential Detection
Murray Campbell, IBM
Simulation of outbreak signals for multiple data
source detectors
David Buckeridge, Stanford University
Power Evaluation of the Spatial Scan Statistic for Multiple Data Stream
Luiz Duczmal, Harvard Medical School and Universidade Federal de Minas Gerais, Brazil
10:45 - 11:15 Break
11:15 - 12:00 Panel Presentation II
moderated by David Siegrist
12:00 - 2:00 Lunch and Breakout Sessions
Breakout Sessions
Break into several (~5) groups. Each group has a moderator. Each group is to discuss a
particular question/task:
- What are the advantages and disadvantages of combining univariate analyses as against
a single multivariate analysis?
- How can one deal with varying quality issues in multivariate surveillance data streams?
- How best can national and local surveillance data systems be used for multiple
purposes - both broad-based regional surveillance and local case management?
- Describe a set of ordered management procedures for responding to a detected data
aberration in a non-traditional early indication data source (egs. OTC, absenteeism data).
Include decision point threads for validation up to alerting multiple agency officials.
Please forward suggestions for other breakout topics to either Henry Rolka (hrr2@crc.gov) or
David Madigan (madigan@stat.rutgers.edu).
Session II: Alternative Perspectives
Speakers, please provide titles as soon as you can.
2:00 - 2:40 Borrowing methodology from industry:
Methods of Statistical Process Control
for Public Health Surveillance
Karen Kafadar, University of Colorado
2:40 - 3:20 Multiplicity and sequential testing
Chris Jennison, University of Bath, England
3:20 - 3:40 Break
3:40 - 4:20 Eamonn Keogh
4:20 - 5:00 Visualization Frameworks to Detection the Presence of Unusual Structure
in High-Dimensional Data
Jeff Solka, George Mason University
5:00 Wine and Cheese
Friday, February 20, 2004
Session III: Whither?
8:45 - 9:00 Perspectives on surveillance (TBA)
9:00 - 10:00 Breakout groups reconvene and prepare summaries
10:00 - 11:00 Breakout group reports and general discussion
11:00 - 11:30 Break
11:30 - 12:30 Summary discussion and recommendations
Previous: Participation
Workshop Index
DIMACS Homepage