Preliminary Design of an Internet-Based Graphical Surveillance Information System

As part of an effort to develop tools to help monitor and manage the Air Force's world-wide health delivery systems, the USAF Surgeon General's Office has been developing a web-based statistical surveillance information system in order to provide near real-time management and clinical information. The intent is to incorporate statistical and related scientific methods that can help identify unusual events and patterns of concern in large, highly distributed, and near continual data flows. This system combines graphical analysis, query ability, and embedded semi-automated intelligence for examining process performance at any level of aggregation or stratification. The objectives, current status, latest version, successes to-date, future directions, and barriers are discussed and demonstrated. SYSTEM DESCRIPTION & FUNCTIONALITY Current focus is on development of an Epidemiology Tracking System (ETS) to monitor disease and other health event rates via statistical process control (SPC) and other pattern detection methods. Ultimately, the complete ETS is envisioned to consist of the following three primary functions or modules: 1. Automatic alerts of irregular events via statistical changes in disease or other health event rates; 2. Ad hoc analysis capability to query and graphically analyze data in several possible manners; 3. Forecating of illness and injury trends in order to better predict and plan for future needs. A key objective of the alerting module is to build into ETS various types of "automated intelligence" that guide users towards potentially important information within the data. For example, automated alerts currently are generated based on SPC concepts in order to identify any processes or sub-processes that contain statistical signals of irregular data patterns. The ultimate goal is to provide an appealing graphical information environment that allows users to conduct ad hoc exploration while also guiding them to discover important information via embedded intelligence. CURRENT STATUS & EVALUATION Upon accessing the ETS website, a list displays any recent conditions that violate SPC criteria for unnatural events or patterns. By clicking on various links, users then can investigate these alerts further at worldwide, regional, theater, or facility levels and stratified by overall rates, medical diagnostic category (MDC), or by specific ICD-9. Smart features such as color-coded data points on bar charts and SPC charts also lead users on "guided surfing" data explorations towards anomalies at higher or lower aggregation levels. Other recent SPC applications include adverse events, emergency services utilization, incomplete medical records, and missing vaccinations. These preliminary uses indicate that significant potential exists to identify problematic situations, rate changes, and differences between facilities, locations, etc. Feedback also has been very encouraging, with several suggestions for expansion currently undergoing review and revision as appropriate. Further information on the status and functionality of this tool can be found in [1].