Watson, WA; Litovitz, TL; Belson, M; Funk, AB; Patel, M; Schier, JG; Kilbourne, E; Rubin, C
Toxicosurveillance enhances the utility of data collected by poison centers during poisoned patient case management. A fundamental approach of toxicosurveillance is detecting aberrances in the gathered data. The application of various statistical methodologies allows for the early detection of aberrances and patterns of toxicity that may have broad public health significance. It is then up to the investigator to determine what additional information is necessary to determine whether a clinically significant public health incident, hazardous or contaminated product, emerging pattern of substance abuse, or an intentional chemical release is present. Since a toxic event may not be confined within political boundaries or poison center coverage areas, a major objective of toxicosurveillance is to include large, naturally-defined geographical areas in order to detect widespread events. This fact argues for coordinated toxicosurveillance using data from multiple poison centers. At the same time, surveillance at the local level is also necessary and a critical step in identification of outbreaks that would be lost against the background noise when only national data is utilized. Using examples from aberrant values detected by TESS, this presentation will focus on the key components of toxicosurveillance involving multiple poison centers. The key components of surveillance include timely collection and transfer of data to a central site, continuous and systematic assessment of the data, and dissemination of the results with public-health significance in order to contain events and reduce morbidity and mortality. Effective toxicosurveillance requires consistency in poison center definitions, training, and operations, resulting in uniform data collection. Additionally, the assessment of data for signals must include adequate resources to perform the desired assessment in a timely manner, at regular intervals, with clear definitions of the processes and outcomes desired. An evaluation process will help determine the optimal surveillance methodology for detection of toxic events. TESS was developed so there would be a uniform data set and standard definitions, replacing multiple pre-existing systems. The AAPCC member poison centers and committees have processes in place to standardize any modifications to data fields. Development of an automated upload of any new or updated case, whether open or closed, with upload occurring by ftp about every 4 to 10 minutes, was the first step in the initiation of continuous toxicosurveillance. With an impending war in Iraq and threatened increase in chem/ bioterrorism, real-time toxicosurveillance was initiated by AAPCC in March 2003 at the urgent request of the CDC. Current programs include those monitoring the volume of cases reported to each poison center and the frequency of clinical effects reported in exposed patients. Case counts (for volume and clinical effects) for the immediately preceding time intervals (hours or days) are compared at 2 to 24 hour intervals to those expected based on their specific baseline during similar time periods over the last 3 years. Cases that generate outlier signals are reviewed by a clinical toxicologist to determine if unusual, worrisome patterns exist. The pattern review focuses on the substance, outcome, and geographical distribution of events. Surveillance case definitions are used to identify specific types of exposures, such as cyanide, arsenic, botulism, nerve agents, and paraquat. Finally, exposure frequency for specific groups of products (e.g. contaminated water, food poisoning/food product cases, and carbon monoxide cases) is continuously monitored. All of these toxicosurveillance approaches are continuously evaluated and updated to determine time from case reporting to detection, false positive and false negative rates, and usefulness of findings. Various statistical approaches, including ARIMA modeling, Bayesian methods, and concurrent space-time detection of clusters are being assessed to determine what approaches are valid, effective, and efficient. The current toxicosurveillance programs have identified a number of local chemical events, including covert exposure to arsenic of a small group of people, chem/ bioterrorism preparedness exercises sponsored by local and federal agencies, and exposure to environment toxins. The electrical power blackout of the northeastern United States and adjacent areas of Canada in August 2003 resulted in significant changes in the call types reported to poison centers serving the impacted population; these changes were easily detected by toxicosurveillance. Although most AAPCC toxicosurveillance programs implemented to date are based on national data for all US poison centers, AAPCC is also providing surveillance of local data for individual poison centers. One approach includes a pilot program which enables poison centers to receive and evaluate surveillance findings for their region. Toxicosurveillance with poison center data supplements other public health or security surveillance processes. The availability of real-time surveillance has increased the role and visibility of US poison centers. This may lead to increased recognition and support for poison centers in addition to improving public health.