Transcript Real-time Outbreak and Disease Surveillance (RODS)

Staffing RODS in Ohio
February 23rd, 2006
Biosurveillance Information Exchange Working Group
Rutgers University
Piscataway, NJ
Loren Shaffer, MPH
[email protected]
Early Event Surveillance
Supervisor
Our Current Systems
• Real-time Outbreak and Disease Surveillance
(RODS) System
• BioSense
• Influenza-like Illness Sentinel Providers
• School Absenteeism
• Wildlife Disease Surveillance
• Ohio Disease Reporting System
• Early Aberration Reporting System (EARS)
RODS
• Developed at the University of Pittsburgh
• Utilizes a hospital’s existing HL-7 socket to receive realtime data
• Interprets free-text chief complaint from emergency
department registration database and groups into one of
seven prodromes
• Collects information from retailers about sales of over-thecounter medications
• Provides immediate analysis for trends and anomalies
suggestive of disease outbreaks
• Secure web access by users
HIPAA
• Export of data for public
health surveillance is
permitted (45 CFR
164.512)
• Prior written authorization
of individual is not
required (45 CFR
164.512)
End Users of RODS in Ohio
• Total = 174
– State Health Department = 6
– Local Health Department = 87
• 88 of 135 LHDs represented
– Hospital = 77
– Other = 4
RODS Alerts
• February 2004 to February 2006
– 310 unique alerts received
•
•
•
•
Seasonal Elevation – 40%
Outbreaks – 22%
Data Errors – 9%
No Cause Identified – 28%
– 53% handled at State level
– 47% forwarded to local epidemiologist
RODS Alerts and Reported Outbreaks in Ohio
2/1/2004 - 2/1/2006
Date of RODS
Alert
12/24/2004
Syndrome
County
Outbreak Cause
Number Ill
Respiratory
Statewide
CO Poisoning
1/11/2005
Gastrointestinal
Franklin
Norovirus
7
1/24/2005
1/11/2005
Gastrointestinal
Franklin
Norovirus
3
2/16/2005
1/30/2005
Gastrointestinal
Mahoning
Norovirus
5
2/9/2005
3/3/2005
Gastrointestinal
Mahoning
C perfringens
4
3/14/2005
10/26/2005
Gastrointestinal
Preble
Norovirus
~300
11/4/2005
11/23/2005
Rash
Jefferson
Scabies
5
not reported
12/27/2005
Gastrointestinal
Montgomery
Campylobacter
3
1/30/2006
8/3/2005
8/4/2005
8/5/2005
Gastrointestinal
Constitutional
Gastrointestinal
Hamilton
Cryptosporidiosis
~750
8/22/2005
1/9/2006
Constitutional
Scioto
Scromboid Poisoning
3
1/30/2006
Respiratory
Erie
Influenza-like Illness
24
not reported
1/18/2006
*Information about the number ill and date reported was obtained from the Ohio Disease Reporting System (ODRS)
~50
Date Reported to
ODH
not reported
What are we doing with RODS?
• Monitoring for patterns suggestive of an
infectious disease outbreak
• Influenza surveillance study & efforts
• Retrospective outbreak study (improved
specificity)
• Adaptation to animal-based sources for
emerging infectious disease surveillance
• Real-time clinical data analysis
• Environmental surveillance
• Specialized investigations
Influenza
Influenza
Summary of Correlation Analysis
2003/2004 and 2004/2005 Influenza Seasons
Data
Thermometer
Cough & Cold
Pediatric Cough Syrup
Adult Cough Syrup
Pediatric Cold Relief
Adult Cold Relief
Pediatric Antifever
Adult Antifever
VA Clinic Visits
DoD Clinic Visits
Class B
Respiratory ED Visits
Constitutional ED Visits
ILI Sentinel Providers
Franklin
Montgomery
Summit
Average
Lead
Correlation Lead
Correlation Lead
Correlation Lead
Correlation
(weeks) Coefficient (weeks) Coefficient (weeks) Coefficient (weeks) Coefficient
4
0.5908
4
0.6254
3
0.5226
3.67
0.5796
4
0.5592
5
0.5728
3
0.4522
4.00
0.5281
4
0.5079
4
0.5487
3
0.4591
3.67
0.5052
4
0.2133
4
0.2938
3
0.3042
3.67
0.2704
4
0.5730
4
0.5236
4
0.4541
4.00
0.5169
4
0.5985
4
0.5875
4
0.4783
4.00
0.5548
5
0.5524
5
0.4626
3
0.3788
4.33
0.4646
5
0.3224
1
0.3844
5
0.1901
3.67
0.2990
7
0.3081
7
0.3159
-1
0.4470
4
0.5107
0
0.2594
1.00
0.4057
4
0.5544
5*
0.5227
4.00
0.5386
4
0.5133
4*
0.4603
4.00
0.4868
4.33
0.4536
5
0.2311
5
0.5591
3
0.5705
(4 w /o
(0.5648 w /o
School Absences
Laboratory Analysis
* = Data From 1 Influenza Season
10
0.2852
4*
5
0.4540
0.5943
5
3
0.3965
0.5966
Franklin)
Franklin)
7.50
4.00
0.3786
0.5955
Known Outbreak
ED Registrations
P(↑Consitutional|Outbreak N)= xc
P(↑Botulinic|Outbreak N)= xb
P(↑Respiratory|Outbreak N)= xr
P(↑GI|Outbreak N) = xg
P(↑Hemorrhagic|Outbreak N)= xh
P(↑Neurologic|Outbreak N)= xn
P(↑Rash|Outbreak N)= xra
OTC Sales
P(↑Thermometer|Outbreak N)= xt
P(↑Cold Relief|Outbreak N)= xcr
P(↑Lozenges|Outbreak N)= xl
P(↑Chest Rubs|Outbreak N) = xchr
P(↑Bronchial Remedies|Outbreak N)= xbr
P(↑Antifever|Outbreak N)= xa
P(↑Other Categories|Outbreak N)= x
Zoonotic Disease Surveillance
• Adapt RODS to …
– Veterinary Diagnostics Laboratory
• Lab orders batched & transmitted via ftp
• Parsed into syndrome categories based
on previous mapping of tests
• Spatial/temporal analysis at the ZIP
code-level
– Veterinary Clinic
• Reason for Visit field included
• Near real-time export via ftp
• Co-co classifier parses free text into
syndrome category
• Spatial/temporal analysis at the ZIP
code-level
Areas for Attention & Assistance
Continuous
Continuous
Emergency
Department
Chief Complaint and
Patient Demographics
In HL7 messages
Categorize
Into
Syndromes
Daily
Daily
Totals
Totals
Higher Than
Expected?
Review
Trends
Historical
Models
Alert
Updated monthly
Valid
Alert?
Initiate
Response
Identified Needs
•
•
•
•
User Training → 1:1 & Conference Workshops
User Protocols → Developed Protocol Document
User Support → 24/7 Coverage Pilot
24/7 Coverage → System Enhancement to Permit
Alerts to Jurisdictions
• Advisory Group(s) → Expansion of ODRS Advisory
Committee
• Integration with Other Surveillance Systems →
Ongoing Evaluations and Study Projects
• Local Health Department Resources → ?
Questions?
Loren Shaffer
[email protected]
614-728-3343