Fortunately the need for this type of surveillance has occurred at the same time computerized records have become more prevalent in hospitals, pharmacies, Emergency Departments, and EMS systems across the country. This allows for automated collection, transfer, and processing of information. The challenge is to develop a system that recognizes an abnormal event in a timely manner in order to reduce morbidity and mortality. The developers of these systems must deal with issues of privacy, archiving, issuing alerts, sensitivity, validity, and cost.
The Emergency Department has been the primary focus of syndromic surveillance implementation. Other data sources that have been reported include clinical impressions of ambulance log sheets, over-the-counter drug sales, consumer health hotline telephone calls, ambulatory visit care record, and school or work absenteeism.
ED Logs: Chief Complaints
In 1999 Wayne State University developed and tested the feasibility of a web-based surveillance program based on patients' chief complaints in the Emergency Department. Triage data was sorted into five chief complaint groups. Separate chief complaint groups were identified for all Category A threats. For example, the anthrax group consists of cough, dyspnea, fever, lethargy, pleuritic chest pain, headache, upper respiratory infection, weakness or fatigue, vomiting, and generalized abdominal pain. Historical data was used as a control. Using the anthrax criteria, they were able to successfully identify the influenza outbreak within the first week. They emphasize, as do other public agencies, that syndromic surveillance can only detect potential threat exposures. Once an aberrant signal is detected it is essential that public heath officials follow-up with field investigations.
Traditionally, the nurses at triage record chief complaints in free-text format. This results in a large degree of variability. Patients express their symptoms in different terminology, which is then recorded by the triage nurse. Aronsky demonstrated that all chief complaints to the ED could be categorized into 57 categories. Day utilized computer algorithms to recognize a combination of words, word fragments, and word pat terns to link free-text-compliant fields to 20 reason-for-visit categories. They suggest that reason-for-visit taxonomy is well suited for syndromic surveillance, as patients who look similar on presentation will be grouped together for analysis.
ICD-9 codes are universally used in the United States and are considered diagnos-tically more accurate than chief complaints. There is a great deal of latitude that the coder can exercise when picking a particular code. As a result, an increase in incidence of a disease may not be recognized as cases are distributed across a number of related codes. For example, an asthmatic with a chief complaint of difficulty breathing might be labeled with a number of ICD-9 codes: wheezing (786.07), acute asthma (493), acute bronchiolitis (466.1), and viral infection (480). Additional disadvantages of ICD-9 codes are the delay from the time care is provided to the time a code is assigned as well as the biases of physicians and coders as they seek to enhance revenues.
The Department of Defense's ESSENCE program has developed broad syndrome groups using ICD-9 codes that approximate natural infectious disease outbreaks or bioterrorism. Syndrome groups were created using all possible codes. These syndrome groups are now being used for routine surveillance at all military medical treatment facilities. These syndrome definitions and associated ICD-9 coded syndrome groups can be used in syndromic surveillance systems to allow for comparability and evaluation among programs.
Beitel and colleagues at Children's Hospital in Boston compared the sensitivity and specificity of ED chief complaints and ICD-9 codes. Both chief complaints and ICD-9 codes demonstrated excellent specificity and moderate specificity for all respiratory infections. They were also able to demonstrate that adding the chief complaint codes of fever and earache could raise the sensitivity but lower the specificity. Although ICD-9 codes are superior to chief complaints, the timeliness of chief complaints and the possibility of improving the accuracy and/or sensitivity of chief complaint codes make them an attractive alternative.
A challenge for all systems is to establish the correct balance between sensitivity and specificity. Systems that are too sensitive will result in frequent false alarms. This will tax the resources of public health officials as well as EDs and other clinicians. Frequent expenditures of resources investigating false alarms could potentially erode the confidence of the medical professionals utilizing the system. From a practical point of view, no public health system will be able to sustain a system that identifies more alarms than it can muster the resources to investigate. However, a system that is too specific and lacks sensitivity risks missing a significant event. The degree of sensitivity needs to be flexible and allow for a decrease in threshold levels during periods of high concern. To date no system has provided an early warning of a bioterrorist event. However, the sensitivity of these systems can be tested using naturally occurring outbreaks such as annual influenza epidemics and through simulation models.
Was this article helpful?
Do You Suffer From ASTHMA Chronic asthma is a paralyzing, suffocating and socially isolating condition that can cause anxiety that can trigger even more attacks. Before you know it you are caught in a vicious cycle Put an end to the dependence on inhalers, buying expensive prescription drugs and avoidance of allergenic situations and animals. Get control of your life again and Deal With Asthma Naturally