Using the architecture of ATHENA-CDS, a Knowledge-based System (KBS) that provides guideline-based treatment recommendations, we developed a framework with nine categories to identify the potential sources of errors in KBS. Using 206 test cases, we found 69 discrepancies: 12 Knowledge-base, 4 Mapping Tables, 7 Testing Database, 6 Testing Environment, 20 Software Specification, and 20 Domain Expert. We hope this framework will aid in the categorization of errors for other Knowledge-based Clinical Decision Support systems.
Describe the new knowledge and additional skills the participant will gain after attending your presentation.: Understand the potential sources of errors uncovered during software testing of Knowledge-based Clinical Decision Support systems and be able to classify the errors into categories and groups.
Omar Usman (Presenter)
VA Palo Alto Health Care System
Michael Ashcraft, VA Palo Alto Health Care System
Connie Oshiro, VA Palo Alto Health Care System
Justin Chambers, VA Palo Alto Health Care System
Susana Martins, VA Palo Alto Health Care System
Samson Tu, Stanford University
Amy Robinson, VA Palo Alto Health Care System
Mary Goldstein, VA Palo Alto Health Care System