The use of expert systems for improving test use and enhancing the accuracy of diagnosis

• HIT Description: Computerized decision support. More info...
  
• Purpose of Study: Investigate the effect of computerized decision support systems on laboratory testing.
  
• Years of study: 2002
  
• Study Design: Comparison of physician test ordering with and without computerized decision support system for 14 clinical vignettes.
  
• Outcomes: Test ordering.
  

• Settings: Six physicians with 10 to 20 years experience working in private general practice.
• Intervention: A computerized decisions support system called the Laboratory Advisory System (LAS).
• Evaluation Method: Test ordering based on clinical vignettes.
• Description: The HIT system was the Laboratory Advisory System that provides electronic knowledge support for the ordering and interpretation of laboratory tests.
• Interoperability: Laboratory Advisory System receives the results of tests completed in the lab as an HL7 string and are stored in the LAS relational database. Results are then posted on an electronic reporting system that is similar to many commercial systems (such as abaton.com and doctorchart).
• Healthcare Utilization: In comparison to usual care across the 14 clinical vignettes the average number of tests ordered was about half .With the LAS system the physicians ordered 17.8 test compared with 32.7 test without, with a mean number of samples (venopunctures) of 5.8 with the LAS compared with 7.5 without.
• Quality of Care and Patient Safety Outcome: Diagnostic accuracy of physicians using the LAS was 100% compared to 66% without.
• Changes in healthcare costs: Cost of tests using the LAS system was $40 per patient less than the costs of tests ordered without use of the system.