Situational Awareness and Preparedness for Public Health Incidences using Reasoning Engines (SAPPHIRE) is a public health informatics research and development program at CBPHIR. The program has two main objectives:

• To introduce an innovative approach for conceptualization the public health information network (PHIN) and the technology infrastructures supporting it.
• To design and implement a robust situation awareness and surveillance system for public health preparedness and biosecurity using knowledge representation and reasoning engines (with a particular focus on influenza).

CBPHIR has implemented the core SAPPHIRE engine using data inputs from eight Memorial Hermann Healthcare System hospitals. These hospitals account for more than 30% of all emergency care visits in the metropolitan Houston area. Triage data and all nurse notes taken during patient stay at these hospitals are transmitted to SAPPHIRE system in near real-time (< 10 minutes). The SAPPHIRE engine uses OnLine Analytical Processing (OLAP) techniques for multidimensional exploration and navigation in datasets.

Ongoing development efforts include implementing exponentialy weighted moving averages and more advanced time-series analyses for signal detection. A semantic/syntactic natural language processing agent is being developed to process free-text entries (such as chief complaints) and semantically tag the information using a combination of UMLS and SAPPHIRE ontologies. Bayesian classification engines are employed for more specific classification of observations due to eight syndromes with highest relevance to bioterrorism and to classify respiratory patients into probable flu, avian flu, or pneumonia classes.

Once the SAPPHIRE ontologies are complete, we will implement a direct-interaction user interface to use domain knowledge representations for intelligent user interactions and for automated query formulation (SAPPHIRE Monitor). This will enable dynamic disease modeling and interactive querying of the knowledgebase by public health experts.