Purpose: Many physicians find the management of supraventricular tachyarrhythmia (SVT) in the elderly complex and challenging. With the use of artificial intelligence theory, we developed an interactive computer expert system, TACHY, to recommend therapies and warn physicians of potential therapeutic side effects.
Methods: We developed a knowledge base that stores guidelines for the management of SVT in the elderly. After the diagnosis of current SVT was input into the computer, TACHY generated a list of therapeutic options as hypotheses. TACHY then prompted the user to provide current patient-specific clinical information, and the optimal therapeutic option was then selected. Potential therapeutic side effects were also displayed. TACHY was tested in a retrospective and a prospective study. The retrospective study, comprising 96 patients with 126 episodes of SVT, was performed to determine the concordance of therapy between TACHY and cardiologists. A prospective study in 18 patients with 26 episodes of SVT was also performed to validate TACHY's recommendations in restoring sinus rhythm.
Results: In the retrospective study the concordance between TACHY and the cardiologist's first therapeutic option was 95.9% (121 of 126 episodes). In the prospective study sinus rhythm was restored in 18 (69%) of 26 episodes of SVT by the first therapeutic option recommended by TACHY. In the remaining eight episodes use of the second or third suggestion of TACHY was successful in controlling the ventricular response to SVT. No adverse reaction to the therapeutic options suggested by TACHY occurred.
Conclusion: TACHY, a knowledge-based computer expert system that simulates human decision making, produced promising results in the management of SVT in the elderly.