Using Dynamic Graph Matching and Gravity Models for Early Detection of Bioterrorist Attacks

Timely detection of a bioterrorist attack is of profound significance for efficient emergency public health management. Various systems currently exist which are capable of detecting the biologic agents prior to (e.g. biosensors) and after exposure (syndromic surveillance) but suffer from limitations like high cost and false positives (Stoto et al., Williams). In this paper, we use novel dynamic graph matching and gravity models to formulate a more precise and efficient methodology for detection. The problem is complicated by the similarity of anthrax and small pox symptoms to common diseases like influenza, chickenpox, airborne characteristics of these agents (that increases the risk of infection spreading to proximal regions), and non uniform distribution of terrorism risk among areas belonging to the same region. Our methodology will analyze patient symptom data available at hospitals using dynamic graph matching algorithms. We propose a heuristic that dynamically updates the template graphs based on patient data before applying matching algorithms, a unique feature of this study. Successful matches will be used to update counters that generate alerts once the counters surpass the threshold values. We develop a heuristic that uses a gravity model to group hospitals in a region into clusters based on the population they serve. Hospitals grouped together as a cluster affect counters that are local to the population they serve and generate alarms to the Public Health Department when they surpass the set threshold values. In addition, we use the fact that some symptoms are unique to these agents to make our algorithms more robust. These models could be used to develop practical applications for agencies such as DHS due to its ability to increase not just the likelihood of detection of a bioterrorism attack but also to identify with greater precision the location(s) of the attack. With minor modification they could also be used to plan for other disasters/epidemics such as SARS, and bird flu.