| OCR Text |
Show Detection of an Influenza Epidemic Using an Electronic Medical Record-based Bioterrorism Surveillance System during the 2002 Winter Olympics Kurt T. Hegmann, M.D., M.P.H.; Kristine B. Hegmann, M.S.P.H., C.I.C.; Aaron M. Wendelboe, M.S.P.H.; Stephen C. Alder, Ph.D.; Alisa R. Haws, B.A.; George L. White Jr., Ph.D. Epidemic surveillance has traditionally, though inefficiently, relied upon paper records and phone calls from physicians. This has resulted hi recognition of infectious disease epidemics at or near the end of epidemic curves. The 2002 Winter Olympics provided an opportunity to implement a bioterrorism surveillance system using an electronic medical record (EMR) developed by EPIC of Madison, WI. This EMR allowed for daily updates of diagnoses and, as such, near real-time detection of an epidemic. Data were available for both the 2000-01 and 2001-02 influenza seasons, allowing for comparisons. Statistical process control was used to monitor daily ICD-9 coded diagnoses across a clinic system (n ~ 100,000; 5 clinics) for syndromic patterns of disease consistent with bioterrorism or natural epidemics. The EMR additionally allowed for immediate investigation of potential cases for other attributes, such as vital signs, physical examination findings, laboratory results and treatment prescribed. Nineteen infectious diseases were under surveillance (including anthrax, botulism, brucellosis, cholera, cryptosporidium, E. coli, febrile illness with rash, hepatitis A, influenza, meningitis/encephalitis/NEC delirium, pertussis, plague, respiratory infection with fever, salmonella, sepsis/NEC death with fever history, shigella, small pox, tuberculosis, and tularemia) using a tiered system of ICD-9 diagnoses that were progressively less specific to the disease. This tiered system was used to see if broader classification of diseases could accurately predict epidemics as well as more narrowly defined disease-specific diagnoses. During the course of this surveillance, three cases of Anthrax were retrospectively identified as having been diagnosed in late 2001. No clinical manifestations occurred hi these cases. These individuals were postal workers with alleged powder exposures. An influenza epidemic was identified hi January 2002. The peak of the 2002 whiter influenza epidemic was substantially larger than that seen during the whiter of 2001. This was most likely due to a Type A influenza season during 2002 as opposed to the predominantly Type B season seen in 2001. hi the course of surveillance, elevated GI symptoms were observed, investigated and unexpectedly found attributable to influenza. A significant proportion of cases (rapid test positive) were affected by nausea (16.6%), vomiting (17.9%) and diarrhea (4.9%) that may have resulted hi additional cases with significant gastrointestinal symptoms remaining undiagnosed. Subsequent analyses of the prior influenza season (2000-01) found still higher prevalences of nausea (31.1%) and vomiting (24.4%) hi those testing positive. Additionally, antiviral prescriptions were significantly greater hi 2002 than 2001 (52.4% vs. 17.8%, p<0.05), likely augmented hi part by real-time feedback to clinicians. We suspect that the increased vigilance surrounding the Olympics also contributed to increased anti-viral prescription rates. However, rapid testing of high probability cases decreased substantially hi 2002 (Odds ratio =0.35, 95% CI = 0.25, 0.50). especially after information regarding the epidemic was disseminated to providers, suggesting greater reliance on clinical diagnosis. It appears that accuracy and clarity of future messages to providers hi similar situations will be needed. Additionally the probability of detecting a second epidemic event may be reduced due to potentially decreased testing of cases after any communication of an initial event, hi 2000-01. there were 228 cases diagnosed with influenza, of which 71 (31.1%) had a rapid influenza test administered and 45 (63.4%) tested positive. During 2001-02, there were 582 influenza cases, 272 (46.7%) tested with a rapid influenza test and 223 (82.0%) tested positive (pO.OOl comparing all 2000-01 with 2001-02 cases). Those testing negative were excluded from the subsequent symptoms analyses. Improved outbreak surveillance using such systems appears promising for future major events and routine communicable diseases. ABOUT THE AUTHORS Kurt T. Hegmann, MD, MPH,Rocky Mountain Center for Occupational and Environmental Health at the University of Utah. Kristine B. Hegmann, MSPH, CIC, St. Mark Hospital, Salt Lake City, UT. Aaron M. Wendelboe, MSPH, Public Health Program, University of Utah. Stephen C. Alder, PhD, Faculty member in Department of Family and Preventative Medicine, University of Utah. Alisa R. Haws, BA, Public Health Program, University of Utah. George L. White, PhD, Faculty member in Department of Family and Preventative Medicine, University of Utah. 131 Utah's Health: An Annual Review Volume DC |
