Correspondences between biomathematical and causal models for clinical decision making.

Abstract

Due to incompleteness and other uncertainties, biomathematical models are unsuitable for direct use in clinical decision making. In this research work, we develop a procedure to derive clinical decision-making causal models from mathematical representation. The process involves obtaining the determination ordering for an incompletely specified system of equations. The concept of determination ordering is extended to dynamic systems of equations, in order to derive clinically usable models. The procedure to transform biomathematical models into causal representation has been machine-implemented for fluid flow models of the eye. A case-structured natural language system (CHRONOS) has been developed to accept, process, and store causal as well as biomathematical models. The system obtains the determination ordering for the biomathematical models and stores their causal representation. The system has the capability to compare the causal models. The deductive capabilities of the system can be used by a clinician to consult the diagnostic reasonings of the biomathematical and causal models.