Related Experiment Videos
The decision support system for a therapeutic department physician as an example of Dr. Watson Type System
S Goldberg1, D Saveliev, E Tsibulkin
1Urgent Pediatrics Ltd., 32, S. Deriabina Str. Yekaterinburg 620014 Russia.
Abstract:
The System Analysis Guide (Sage) is decision support system intended for highly-skilled physicians. The SAGe estimates the effectiveness and adequacy of treatment delivered by physicians. The traditional technique was not suitable for the solution of such complex tasks, and, as a basic model, the technology of the Dr. Watson Type System (Dr.WTS) (S. Goldberg, L. Meshalkin, 1990) was selected. Dr.WTS creates an informational medium that promotes correct decision making by a specialist. With that end in view, it is necessary to: a) get all the necessary information for decision making; b) stimulate the user by production of a conflict situation for the given informational space; c) direct the user's reflections toward a correct decision; d) provide the user with an compelling motive for continuing to work with the Dr.WTS program. Similar methods are actively used for the organization of creative process conditions by psychologists, system analysis specialists, and for invention machines. A distinctive feature of the Dr.WTS is the presence of an inference engine. Though the decisions themselves are not offered, Dr.WTS creates the enumerated conditions on the basis of decision making algorithms together with the result of the specialists work. Another characteristic is the system self-tuning on the problem which is solved, not in invention machines, but by the concrete specialist. The SAGe is written in C++ on an Intel 386SX-based computer and consists of five subsystems: "Diagnostics," "Treatment Effectiveness," "Treatment Adequacy," "Integral assessment of patients in the department," and "Information import." The "Diagnostics" subsystem gives a physician the opportunity to assess all information about a patient from the point of view of compensating the organism's basic systems. The SAGe organizes the presentation of the information and notes the signs which are not used in the physician's diagnostic schema (as interpreted by the SAGe). The "Treatment Effectiveness" subsystem makes it possible to compare the state severity dynamics with the dynamics of treatment character and intensity. The physician has the possibility of including the dynamics of any quantitative or qualitative indices in the analysis. The SAGe proposes the dynamics of signs which, from its "point" of view, are the most important for the choice of treatment and assessment of treatment efficacy. The "Treatment Adequacy" subsystem gives the physician an opportunity to create an archive of his own a priori and a posteriori reflections about the treatment course, causes of a state aggravation, treatment ineffectiveness, non-coincidence of prognoses, etc. The SAGe reflects the basic dynamic information about a patient and the dynamics of a physician's reflections in the form of a flower growth picture. The SAGe, stressing various details of the picture according to its notions, gives a physician an opportunity to assess divergences between his reflections at different moments in time and between the real dynamics of a disease. Thus the physician receives a notion of harmony patient treatment which is close to the adequacy of notions. The SAGe's most important role, both for stimulating a physician's intellectual activity and for the work of decision-making algorithms, is played by demanding that the physician make independent short-term and long-term prognostic assessments of a disease's development. These prognostications are then compared with reality. At present, the SAGe has been passed for approbation to some of the physicians of Yekaterinburg and Peterburg. The next task in developing the SAGe is a full-volume realization of mechanisms for algorithmic self-tuning with a concrete user.