Validation of the medical expert system PNEUMON-IA

Controlling for Chance Agreement in the Validation of Medical Expert Systems with No Gold Standard: PNEUMON-IA and RENOIR Revisited

“…Both PNEUMON-IA and RENOIR expert systems and the application of distance measures to their respective validation studies have been extensively described elsewhere (26,27). We will briefly summarize our previous work, before focusing on the kappa coefficient approach proposed in the present study.…”

Expert Systems in Clinical Microbiology

“…In one study, clinicians using an expert system (compared with conventional practice) ordered fewer laboratory tests during the diagnostic process, completed the diagnostic workup with fewer sample collections, generated lower laboratory costs, shortened the time required to reach a diagnosis, showed closer adherence to established clinical practice guidelines, and exhibited a more uniform and diagnostically successful investigation (231). Expert systems have been used to diagnose a variety of clinical conditions, including community-acquired pneumonias (9,279), septicemia (203), female genital disease and abdominal pain (263), urinary tract infection (51,281), viral (121) and infantile (90) meningitis, febrile tropical diseases (27), chronic prostatitis (31), infective endocarditis (76), and infectious diseases (266). Some systems have used fuzzy-logic methods (14); others have used the World Wide Web (76).…”

Case-based learning of plans and goal states in medical diagnosis

“…The selection of cases has been done by the teacher, and follows a stratied random procedure based on the diagnostic groups of the WHO (Verdaguer et al, 1992). We used 460 attributes to represent the domain of pneumonia.…”

“…The threshold has been experimentally found from a current set of cases and domain solutions available, and set up to 0.96 (see (Verdaguer, 1989;Verdaguer et al, 1992)). the goal state of , then = is the domain solution of and .…”