Speculative Computation with constraint processing for the generation of clinical scenarios

Abstract

Clinical decision making often involves making decisions in situations of uncertainty. Clinical Decision Support Systems are tools devised to help in such moments, but the information may not be available during the decision process. Be it because of communication failure or errors in data input, the truth is that it would be beneficial to present the most likely clinical scenarios to a physician, given the incompleteness of the information. Speculative Computation offers a way to structure such a scenario generation process. This work presents a framework for clinical decision support with disjunctive constraint processing that acts as an interface with computer-interpretable versions of Clinical Practice Guidelines. Being a reasoning process based on defaults, it has to rely on a default generation process. For that we propose Bayesian Networks. The interaction between the different components of the system resulted in a process capable of generating clinical scenarios.