OBJECTIVE
A “whole-system facsimile” recreates a complex automated biosurveillance system running prospectively on real historical datasets. We systematized this approach to compare the performance of otherwise identical surveillance systems that used alternative statistical outbreak detection approaches, those used by CDC’s BioSense syndromic system or a popular scan statistics.
Respiratory viruses cause substantial morbidity and costly resource utilization among young children, especially during the winter months. Accurate estimates of the impact of these viruses are important in guiding prevention efforts and measuring the impact of public health interventions. Previous studies have focused on the rate of hospitalizations resulting from viral infections, particularly those attributable to influenza virus for which a vaccine is available, but have not included healthcare use in the emergency department (ED) nor considered the impact of other viruses such as respiratory syncytial virus (RSV), for which limited preventative methods are available. We used ED surveillance data for acute respiratory infection to measure the population-based impact of specific viruses.
Objective
To use surveillance data to estimate resource utilization and parental lost productivity associated with influenza and RSV infections among young children.
The BioSense system receives patient level clinical data from > 370 hospitals and 1100 ambulatory care Departments of Defense and Veterans Affairs medical facilities. Visits are assigned as appropriate to 78 sub-syndromes, including respiratory syncytial virus (RSV). Among infants and children < 1 year of age, RSV is the most common cause of bronchiolitis and pneumonia; 0.5% to 2% require hospitalization. Increasingly, RSV is also recognized as a major cause of pneumonia in elderly adults.
Objective
To analyze final diagnosis data available to BioSense and determine its potential utility for surveillance of RSV illness.
We have previously shown that timeliness of detection is influenced both by the data source (e.g., ambulatory vs. emergency department) and demographic characteristics of patient populations (e.g., age). Because epidemic waves are thought to move outward from large cities, patient distance from an urban center also may affect disease susceptibility and hence timing of visits. Here, we describe spatial models of local respiratory illness spread across two major metropolitan areas and identify recurring early hotspots of risk. These models are based on methods that explicitly track illness as a traveling wave across local geography.
Objective
To characterize yearly spatial epidemic waves of respiratory illness to identify early hotspots of infection.
In order to detect influenza outbreaks, the New York State Department of Health emergency department (ED) syndromic surveillance system uses patients’ chief complaint (CC) to assign visits to respiratory and fever syndromes. Recently, the CDC developed a more specific set of “sub-syndromes” including one that included only patients with a CC of flu or having a final ICD9 diagnosis of flu. Our own experience was that although flu may be a common presentation in the ED during the flu season, it is not commonly diagnosed as such. Emergency physicians usually use a symptomatic diagnosis in preference, probably because rapid testing is generally unavailable or may not change treatment. The flu subsyndrome is based on a specific ICD9 code for influenza. It is unknown whether patient visits that meet these restrictive criteria are sufficiently common to be of use, or whether patients who identify themselves as having the flu are correct.
Objective
Our objective was to examine the CC and ICD9 classifiers for the influenza sub-syndrome to assess the frequency of visits and the agreement between the CC, ICD9 code and chart review for these patient visits.
Syndromic surveillance systems can detect increases in respiratory and gastrointestinal illness, but diagnosis of etiologic agents can be delayed due to difficult, time-consuming identification and low rates of testing for viral pathogens. Rapid diagnostic (RD) assays may aid in early identification and characterization of large outbreaks by allowing decision makers to “rule in” or “rule out” potential etiologic agents.
Objective
This paper describes preliminary results and implementation lessons learned from a RD testing pilot project. The project’s purpose is to prospectively collect diagnostic data on common causes of community-wide illness in order to supplement syndromic surveillance in New York City.
The Centers for Disease Control and Prevention BioSense has developed chief complaint (CC) and ICD9 sub syndrome classifiers for the major syndromes for early event detection and situational awareness. The prevalence of these sub-syndromes in the emergency department population and the performance of these CC classifiers have been little studied. Chart reviews have been used in the past to study this type of question but because of the large number of cases to review, the labor involved would be prohibitive. Therefore, we used an ICD9 code classifier for a syndrome as a surrogate by chart reviews to estimate the performance of a CC classifier.
Objective
To determine the prevalence of the sub-syndromes based on the ICD9 classifiers, and to determine the sensitivity, specificity, positive predictive value and negative predictive value of CC classifiers for the sub-syndromes associated with the respiratory and gastrointestinal syndromes using the ICD9 classifier as the criterion standard.
Previously we used an “N-Gram” classifier for syndromic surveillance of emergency department (ED) chief complaints (CC) in English for bioterrorism. The classifier is trained on a set of ED visits for which both the ICD diagnosis code and CC are available by measuring the associations of text fragments within the CC (e.g. 3 characters for a “3-gram”) with a syndromic group of ICD codes. Because the ICD system is language independent, the technique has the potential advantage of rapid automated deployment in multiple languages. Our objective was to apply the N-Gram method to a training set of Turkish ED data to create a Turkish CC classifier for the respiratory syndrome (RESP) and determine its performance in a test set.
Objective
To determine how closely the performance of an ngram CC classifier for the RESP syndrome matched the performance of the ICD9 classifier.
Syndromic surveillance had been implemented in Dongcheng District with a view to probing into the feasibility of establishing a syndromic surveillance system in major Chinese cities, sieving syndromic surveillance indicators applicable to the eruption of infectious respiratory tract and digestive tract diseases, and attempting the operating method of data collection in different locations such as hospital and drug stores in Dongcheng of Beijing China.
Objective
The project has fund donated by World Bank under joint management of WHO and Ministry of Health of P.R.China , The target was try to build up a syndromic surveillance system in Beijing.
To evaluate whether syndromic surveillance can detect localized outbreaks of lower-respiratory infections (LRI's) with limited numbers of alarms in time.