Syndromic Surveillance in Montreal: An Overview of Practice and Research

1
Syndromic Surveillance in Montreal An Overview
of Practice and Research

• David Buckeridge, MD PhD
• Epidemiology and Biostatistics, McGill University
• Surveillance Team, Montreal Public Health
• QPHI Surveillance Meeting
• KFLA Public Health, Kingston, ON
• June 13th, 2008

2
Syndromic Surveillance in Montreal (ou, Vigie
Multirisque)
Counts, Native coding schemes, ISDS consensus
syndromes
Routine SaTScan, alerts for shared addresses
Daily review of analysis results, not clear
protocol
Telehealth 911 Calls Hospital Reportable
3
Vigie Multirisque Data Sources

• Emergency Departments
• Currently All 22 ED in Montreal via web form,
  total counts, no diagnosis or chief complaint
• Future Automated feeds under development, triage
  code and level, chief complaint, postal code
• EMS Dispatch and Billing
• Long-Term Care
• Tele Health
• Reportable Diseases

4
Vigie Multirisque Dashboard
5
Vigie Multirisque Analysis
6
Vigie Multirisque Analysis
7
Vigie Multirisque Descriptive
8
Surveillance Research
9
Syndromic Surveillance Research
Subsets of admin data for ILI surveillance
10
Looking for the Leading ILI Indicator in Billing
Data
11
Syndromic Surveillance Research
Accuracy of ICD codes and syndromes in ambulatory
practice
Subsets of admin data for ILI surveillance
12
Syndromic Surveillance Research
1. Selecting the best algorithm 2. 3.
Accuracy of ICD codes and syndromes in ambulatory
practice
Subsets of admin data for ILI surveillance
13
Building the Knowledge-Base for Algorithm
Selection
14
Syndromic Surveillance Research
1. Selecting the best algorithm 2. Looking for
connected cases 3.
Accuracy of ICD codes and syndromes in ambulatory
practice
Subsets of admin data for ILI surveillance
15
System Architecture
16
Organizing Data by Person, Place and Time
Spatial Database
PostGreSQL / PostGIS DB
Episode Onset Date Disease Type
Contact
Person MADO Name Birthdate
Situation Role (Home, Work, School, ) Active
Date
Place Address X, Y Place Type (Residence,
Workplace)
17
Address Validation and Correction in a Public
Health System
18
Dracones Query Form
19
(No Transcript)
20
Dracones SaTScan Results
21
Syndromic Surveillance Research
1. Selecting the best algorithm 2. Looking for
connected cases 3. Spatial TB clusters
Accuracy of ICD codes and syndromes in ambulatory
practice
Optimal decision making after an alarm
Subsets of admin data for ILI surveillance
22
Using Surveillance Information to Manage
Outbreaks Effectively

• Much research on the statistical accuracy of
  aberrancy detection algorithms
• Little attention to what happens next
• Some attempts to describe response protocols
  (e.g., flow chart, wait a day)
• No quantitative modeling of response
• Rational response is important
• Small window to obtain benefit
• Surveillance information uncertain

23
The Traditional Surveillance Alert Response Model
Environmental Data
Knowledge
Detection Method
No Intervention
Intervention
No Alert
Alert
No Outbreak
No
Wait
Yes
No
Review Records
Yes
Investigate
No
Yes
Confirm
24
Identifying an Optimal Policy

• The goal is to identify a policy, or a mapping
  from a belief state (probability distribution
  over states) to actions
• The belief state, provides the same information
  as maintaining the complete history
• Value iteration is used to solve POMDP

25
Applying a POMDP to Surveillance

• S - True outbreak state No Outbreak, D1, .
• O - Output from detection algorithm 0,1
• A - Possible public health actions
• T(s,a,s) - Impact of actions given the state
• R(s,a) - Costs of actions and outbreak states

Action
Transition
Do nothing Review records Investigate
cases Declare outbreak
(Izadi M Buckeridge DL, 2007)
26
POMDP Policy Dominates Ad Hoc Policy
27
Syndromic Surveillance Research
1. Selecting the best algorithm 2. Looking for
connected cases 3. Spatial TB clusters
Accuracy of ICD codes and syndromes in ambulatory
practice
Optimal decision making after an alarm
Subsets of admin data for ILI surveillance
28
Automated and Traditional Surveillance for
Waterborne Outbreaks
Syndromic Surveillance
S
Historical Tele- health and ED Data
Tele-health
Infectious (Asymptomatic)
S
Analysis by Public Health
O
S,R
S
O
O
Latent Infected
Infectious (Symptomatic)
ED
Outbreak Detection
R
Stool Test
Analysis by Public Health
R
R
R
R
Out-patient
Historical Case Reports
Exposure
Disease
Health Care Utilization
Reportable Disease Surveillance
29
Modeling Dispersion of Microorganisms
30
Modeling Infection Mobility
31
Mobility-Weighted Infection Probability by Home
Address
32
Modeling Disease, Visits, Testing, Reporting to
Public Health
33
Evaluating the Effect of Surveillance Enhancements
34
For more information