Transcript EON project
Using Ontologies in Clinical Decision Support Applications Samson W. Tu Stanford Medical Informatics Stanford University 1 Main points 2 Information technology has the potential to advance patient care by improving clinician adherence to clinical practice guidelines Principled architecture that separates ontologies, knowledge bases, and problemsolving components allows development and deployment of maintainable complex software systems EON and ATHENA projects demonstrate use of ontologies in clinical decision support applications EON project NLM-funded project at Stanford (PI: Dr. Musen) Develop methodology, ontologies, and software components for creating decision-support system for guideline-based care Use Protégé knowledge-acquisition methodology and tool for construction of Develop software components that assist clinicians in specific tasks 3 Domain concept ontologies Patient information model Guideline knowledge bases Therapy-advisory and eligibility-determination servers Database mediator for time-oriented queries Explanation and visualization facilities EON architecture Patient Database Servers Temporal Mediator Protégé Knowledge Base EON Guideline Ontology Medical Domain Ontology Patient Data Model Guidelines Protégé 4 Clients Clients Protocol Eligibility Checker Yenta Eligibility Yenta Client Therapy Advisory Server Advisory Yenta Yenta Client ATHENA project Funded by VA Research Service HSR&D (PIs: Drs. Hoffman and Goldstein, VA clinicians and Stanford faculties) Hypothesized that guideline-based interventions in management of hypertension can Deployed and evaluated at primary care VA clinics in 9 geographically diverse cities over a 15-month clinical trial Results 5 Change physicians’ prescribing behavior Change patient outcome Expert clinicians maintain hypertension knowledge base using Protégé Clinicians interacted with the ATHENA Hypertension Advisory at 54% of all patient visits Impact on prescribing behavior and change patient outcome being analyzed Building ATHENA system from EON components VA CPRS EON Servers VA DHCP Patient Database Data Converter nightly data extraction Guideline Knowledge Base 6 Temporal Mediator Guideline Interpreter Protégé ATHENA Clients ATHENA Clients Event Event Monitor Monitor Advisory Advisory Client Client ATHENA GUI What the Clinician Sees… 7 ATHENA HTN Advisory BP targets Primary recommendation Drug recommendation 8 ATHENA HTN Advisory: More Info 9 ATHENA HTN Advisory: Link to evidence base 10 EON ontologies Generic data types (generalize to HL7 data types) Medical concept ontology (generalizes to standard terminologies) Patient information model (generalizes to HL7 RIM) Guideline ontology 11 Physician-maintained hypertension knowledge base 12 Benefits of ontology-based clinical information systems 13 Separation of declarative domain knowledge and procedural problem-solving knowledge allow Content experts to maintain knowledge bases Standardization of ontologies that leads to sharing and interoperability Semantically rich ontologies allow sophisticated reasoning and decision support e.g., automatic concept classification based on description logic e.g., detailed drug recommendations based on computable model of clinical practice guidelines