Diseases such as avian influenza, severe acute respiratory syndrome (SARS) and Creutzfeldt-Jacob syndrome represent a new era of biological threats. Nowadays, these hazards breed, mutate and evolve at tremendous speed. Furthermore, they may spread out at the same speed as which we travel. This reveals an urgent need for an agent capable of dealing with such threats. Data warehouses are databases which provide decision support by on-line analytical processing (OLAP) techniques. We present the architecture for an effective information system infrastructure enabling the prediction and near real-time detection of disease outbreaks, using knowledge extraction algorithms to explore a symptoms/diseases data warehouse in a continuous and active form. To collect such data, we take advantage of the Internet and features existing in today?s common communication devices such as personal computers, portable digital assistants and cellular phones. We present a case-simulation based on a small country, showing the system can detect an outbreak within hours or even minutes after its physical occurrence, alerting health decision makers and providing quick interaction and feedback between all users. The architecture is also functionally independent from its geographical dimension.