Syndromic surveillance hrsa

Syndromic Surveillance for Early Event Detection with Real Time Online Disease

Surveillance (RODS)

Houston Department of Health and Human Services

Wes McNeely MS, MPH, Epidemiology Supervisor (PHP)A.J. Rosario MD, MPH Systems Analyst (PHP)

Definition

Syndromic surveillance for early outbreak detection is an investigational approach where health department staff…monitor disease indicators continually (real-time) or at least daily (near real-time) to detect outbreaks of diseases earlier and more completely than might otherwise be possible with traditional public health methods. The distinguishing characteristic of syndromic surveillance is the use of indicator data types.

--Excerpt from MMWR May 7, 2004 / 53(RR05)1-11

Hardware - PHINPHIN - Public Health Information Network

Server Cluster uses Microsoft Windows Advanced Server 2003 platform

Web Server, Application Server, Database

Database Server hosts Oracle 9i

Resides at HDHHS

Software - RODS

Hospital Sends Data Hospital admits patient into Emergency

Room (ER) and generates a computer record of visit

Hospital formats ER admissions record into HL-7 or other “messaging” format

Hospital “message router” uploads record to HDHHS PHIN through a secure Virtual Private Network connection (VPN)

“Message”

MSH1|^~\&||UPMC(hosp)||RODS|200003171458||ADT^A04|20000317145841029270(message control ID)2|P|2.3|<CR>3PID|||||||^017(age)|M|||^^^^12345^^^^|<CR>PV1||E||D||||||||||||||||||||||||||||||||||||||||200109130045(admittime)||<CR>PV2|||ABDOMINAL PAIN(chief complaint)|<CR>DG1||I9|789.09|ABDOMINAL PAIN-SITE NEC(diagnosis desc.)|<CR>NK1||||^^^^12345^^|||EMP|<CR>IN1||||||||||||||||||||||||||||||||||||||||||||^^^^12345<CR><ETX>4

RODS Receives Data RODS “listens” for new data uploaded into

folder and “consumes” the message

RODS chops up the text containing the complaint information (“parses the string”)

RODS classifies the chopped up complaint text using “Natural Language Processing” (NLP) into one (1) of seven (7) possible syndromes related to bioterrorism

Complaint text classification (CoCo)

Syndrome Classification Accuracy

“The sensitivity of a naive Bayes classifier using triage diagnoses was 0.63, the specificity was 0.94, and the area under the ROC curve was 0.82”.....

“We conclude that a naive Bayes classifier of free-text triage diagnosis data provides more sensitive and earlier detection of cases [than other methods].”

“The sensitivity achieved should be sufficient for syndromic surveillance system designed to detect moderate to large epidemics.”

Ivanov, O., M. M. Wagner, et al. (2002). "Accuracy of three classifiers of acute gastrointestinal syndrome for syndromic surveillance." Proc AMIA Symp: 345-9.

SyndromesRODS assigns syndromic categories to the complaint

1. Gastrointestinal2. Constitutional3. Respiratory4. Rash5. Hemorrhagic6. Botulinic7. Neurological8. Other

DatabaseRODS imports the data with syndrome field

added into Oracle 9i database

MSH1|^~\&||UPMC(hosp)||RODS|200003171458||ADT^A04|20000317145841029270(message control ID)2|P|2.3|<CR>3PID|||||||^017(age)|M|||^^^^12345^^^^|<CR>PV1||E||D||||||||||||||||||||||||||||||||||||||||200109130045(admittime)||<CR>PV2|||ABDOMINAL PAIN(chief complaint)|<CR>DG1||I9|789.09|ABDOMINAL PAIN-SITE NEC(diagnosis desc.)|<CR>NK1||||^^^^12345^^|||EMP|<CR>IN1||||||||||||||||||||||||||||||||||||||||||||^^^^12345<CR><ETX>4

Pretty data

Pretty data

Pretty data

Graphs

RODS Analyzes the Data

RODS analyzes the data to detect unusual patterns every 4 hours Time Series

Recursive Least Squares (RLS) CuSum Linear Models

Spatial analysis Scan Statistics

Multivariate analysis WSARE (What’s Strange About Recent Events?)

Alerts

Output

Output

Syndromic Trends

Graph3 Gastrointestinal

Graph2 Neurological

Graph4 Botulinic

Graph1 Respiratory

Graph1 RespiratoryBURN-IN PHASE

Interface Upgrades

Version 4.0

Upgrades1

Upgrades2

Upgrades3

Upgrades4

Upgrades5

Contact Information

Olushola.adeleye@cityofhouston.net713-794-9181Epidemiologist Specialist

Shola Adeleye, MD, MPH

Wesley.mcneely@cityofhouston.net713-794-9181Epidemiology Supervisor

Wes McNeely, MS, MPH

annie.chen@cityofhouston.net713-798-0884DBAAnnie Chen, OCA

arnulfo.rosario@cityofhouston.net713-558-2493Systems Analyst IVAJ Rosario, MD, MPH

RODS Laboratory

jue@generalbiodefense.com 917-254-7278User Interface, Developer

Jeremy Espino, MD

dhenry@cbmi.pitt.edu412-648-6738Project ManagerDaphne Henry

HDHHS

Extra info - RODS Syndromes Botulinic – ocular abnormalities, difficulty speaking,

and difficulty swallowing. Constitutional – non-localized, systemic problems

including fever, chills, body aches, flu symptoms. Gastrointestinal – pain or cramps anywhere in the

abdomen, nausea, vomiting, and diarrhea. Hemorrhagic – bleeding from any site, vomiting

blood, nose bleed, or gastrointestinal bleeding. Neurological – non-psychiatric complaints that

relate to brain function. Rash – any rash such as macular, papular,

vesicular, petechial, purpuric, or hives. Respiratory – problems of the nose, throat, and

lungs. The presence of both clod and flu symptoms is counted in this category, not constitutional.

Extra info - Sensitivity and Specificity

Test Result+ -

Has

Conditio

n

N

o Y

es

Extra info - ROC Curve“In signal detection theory, a receiver operating characteristic (ROC) is a graphical plot of the sensitivity (fraction of true positives - TP) vs. 1-specificity (the fraction of false positives - FP) for a binary classifier system as its discrimination threshold is varied.” http://en.wikipedia.org/wiki/ROC_Curve