1	"Robert W" Robert W. Haley, M.D. Epidemiology Division Department of Internal Medicine UT Southwestern Medical Center
2	What Is Syndromic Surveillance? Definition Monitor acute syndromes, not diagnoses Purposes Detect bioterrorism attacks early Improve routine reportable disease surveillance Advantages over Laboratory Surveillance Quicker response and much less expensive New diseases like SARS have no lab test.
3	Bioterrorism Pathogens by CDC Categories Category A Anthrax (Bacillus anthracis) Botulism (Clostridium botulinum toxin) Plague (Yersinia pestis) Smallpox (variola major) and smallpox vaccination (vaccinia) Tularemia (Francisella tularensis) Viral hemorrhagic fevers (filoviruses [e.g., Ebola, Marburg] and arenaviruses [e.g., Lassa, Machupo]) Category B Brucellosis (Brucella species) Epsilon toxin of Clostridium perfringens Food safety threats (e.g., Salmonella species, Escherichia coli O157:H7, Shigella) Glanders (Burkholderia mallei) Melioidosis (Burkholderia pseudomallei) Psittacosis (Chlamydia psittaci) Q fever (Coxiella burnetii) Ricin toxin from Ricinus communis (castor beans) Staphylococcal enterotoxin B Typhus fever (Rickettsia prowazekii) Viral encephalitis (alphaviruses [e.g., Venezuelan equine encephalitis, eastern equine encephalitis, western equine encephalitis]) Water safety threats (e.g., Vibrio cholerae, Cryptosporidium parvum) Category C Emerging infectious diseases such as Nipah virus and hantavirus
4	Examples of Presenting Syndromes of Bioterrorism Infections Influenza-like Anthrax, smallpox, plague, tularemia, brucellosis, viral hemorrhagic fevers Pulmonary Plague, psittacosis, Q fever, influenza, mellioidosis, ricin Gastrointestinal Anthrax, salmonella, shigella, cholera, staph enterotoxin Systemic illness Anthrax, smallpox, plague, tularemia, brucellosis, viral hemorrhagic fevers, typhus Sepsis/DIC Anthrax, plague, viral hemorrhagic fevers, typhus Encephalitis/Meningitis Viral encephalitides (VEE, WEE, EEE), anthrax Rash-illness Smallpox, viral hemorrhagic fevers, typhus, anthrax, tularemia, mellioidosis
5	What is the problem that Syndromic Surveillance addresses? The earlier a bioterrorism attack is identified the more casualties can be prevented.
6	The Window of Opportunity Tens of thousands could be infected. Prophylaxis of millions is not practical. A 2-5 day window of opportunity to identify the problem and prophylax the exposed to avoid catastrophic situation. Failing this, we will have a mass casualty situation to deal with. The government is putting billions into preparedness. Must use it wisely.
7	What Would a Massive Anthrax Attack Be Like with 50,000 Exposed?
8	What Would a Massive Anthrax Attack Be Like in Dallas? Days 1 and 2 - Nothing happens, terrorists escape. Days 3 and 4 - 100 and 500 present with flu-like symptoms, anthrax not diagnosed. Day 5 - 1,000 new cases come in, the first 50 deaths, blood cultures from first cases prove anthrax. Press announces epidemic, panic ensues. Day 6 - 1,800 new cases, 700 deaths, ICU/ hospitals full, dying sent home, run on antibiotics. Day 7 - 2,800 new cases, 1,000 deaths, no Abs left in N. Texas, body disposal, commerce stops.
9	What Would a Massive Anthrax Attack Be Like in Dallas? Second week - epidemic peaks on 11th day with 4,500 new cases, 3,600 deaths, most dying without care. Total death toll to date: 25,000. Third week, new cases drops from 3,800 on Mon. to 500 on Fri. Total deaths to date: 35,000. Weeks 4-7 - new cases continue at 200-500 per day. By now, doctors prescribing antibiotics for least cough and fever and case-fatality drops from 100% to 50%. >7th week - cases trail off, environment cleanup.
10	What Would a Massive Anthrax Attack Be Like with 50,000 Exposed?
11	Public Health Response Plan Early recognition - maximize window of opportunity. Syndromic surveillance an important component. Microbiologic diagnosis - early confirmation / Ab sens. Epidemiologic investigation – ID the exposed groups Delivery of antibiotics to high risk groups If prophylaxis fails, mass casualty management Intensive care for critically ill patients
12	Approaches to Syndromic Surveillance Original systems depended on ED physicians to identify and computer-enter syndrome diagnoses into a secure Website where trends were analyzed. Next idea was to download ICD-9 codes from hospital and clinic computer to centralized national systems. Newest idea is to use daily automated download of routine ED records to a LHD computer that analyzes for syndromic trends at the local level. Use of “natural language processing” to extract components of syndromes from text in ED notes. Constructs “statistical syndromes” for trending. Encrypted identifier links maintained for look-back investigations and local mapping to street level.
13	Recent Progress in Techniques for Syndromic Surveillance Recognition that special data-entry by ED physicians is not suitable for routine SS. “Natural language processing” computer methods. Improved mathematical algorithms for thresholding from control chart techniques are recognizing a higher percentage of true outbreaks and raising fewer false alerts. Dual use of SS for bioterrorist attacks and for natural outbreaks, notifiable disease reporting, and quality improvement.
14	Validation of Automated Syndromic Surveillance Study in progress to determine usefulness of syndromic surveillance in detecting natural outbreaks of infectious diseases. 5 private hospitals in a medium sized city daily download ED computer records to LHD computer which analyzes trends and issues alerts. “Gold standard” compiled from trends of all viral and bacterial isolates and LHD disease surveillance.
15	Syndromes Studied Influenza-like Gastrointestinal Systemic illness Sepsis/DIC Encephalitis Rash-illness Hepatitis
16	Alert Levels Green – low risk Blue – attention Yellow – caution Orange – high risk Red – extreme risk
17	Flu-like Syndrome
18	GI Syndrome
19	Encephalitis Syndrome
20	Sepsis Syndrome
21	Systemic Illness Syndrome
22	Rash Illness Syndrome
23	Summary of Results To Date Over 12 months, 29 flags for of the 8 syndromes. 17 flags for Influenza-Like Syndrome All but 1 coincided with influenza or RSV outbreaks. Detected exponential phase of local influenza epidemic earlier than existing influenza surveillance systems. 3 flags for Gastroenteritis Syndrome Two flags coincided with rotavirus outbreaks; one unsolved. One of these occurred at the peak of the influenza outbreak. 1 flag for Encephalitis Syndrome Coincided with cluster of febrile seizures and sepsis/confusion. Probably influenza complications.
24	Benefits to Hospital of Participating in Syndromic Surveillance Demonstrates compliance with JCAHO Standard EC.1.4.f Requires hospital to participate in “planning . . . to facilitate sharing of information.” Negative data is vitally important to reassure the public when no epidemic is occurring during a crisis elsewhere. Opportunities for ED data to contribute to quality improvement.
25	Benefits to Hospital of Analyzing Its Own Syndromic Surveillance Data Rapid look-back to patients in unusual syndrome clusters makes it easy to investigate outbreaks. Daily data export reduces workload of reporting notifiable diseases. Faster recognition of patients needing evaluation for isolation. Avoids surprises when the health department calls.
26	Role of Hospital BT Coordinator Should be a member of the local/regional BT preparedness tack force. Should be involved in selecting the region’s syndromic surveillance system. Should be involved in analyzing the hospital’s own syndromic surveillance data. Should function as the liaison to the local health department in investigating suspicious clusters. Only local staff can expeditiously investigate the findings and take action rapidly in a crisis.
27	Alternative Models of Syndromic Surveillance
28	Advantages of the Hierarchical Model over the Centralized Model Hospital – LHD partnership allows sharing of encrypted-identifiable ED data, not allowed by HIPAA to centralized systems. LHD can ask hospitals to trace-back suspicious clusters to the involved patients. Can map involved patients down to the address level. Will more rapidly detect the most likely local outbreaks. Allows dual use for purposes besides syndromic surveillance. Can automatically transmits data to local, state and federal public health agencies with the appropriate privacy restrictions.
29	"Robert W" Robert W. Haley, M.D. Epidemiology Division Department of Internal Medicine UT Southwestern Medical Center