Implementation of a ‘Useful’ Information Measure for Healthcare Decision Making

Abstract

Cryptography, molecular biology, natural language processing, and statistical inference are a few fields where information theory is used. It’s also utilized in medical science. In this article, we show how principles from information theory have been applied to improve medical decision-making in various ways. We begin with an overview of information theory and the notions of available data and entropy. In this study, we show how ‘useful’ relative entropy may be utilized to determine which diagnostic seems to be the utmost useful at a given stage of analysis. Shannon information may be utilized to determine the range of standards of medical reports across which the test offers relevant information about the patient’s condition when the result is binary. Of course, this isn’t the only approach available, but it may create a visually appealing representation. Next, the article introduces the more complex ideas of ‘useful’ conditional entropy and ‘useful’ mutual information, demonstrating how they may be used to prioritize clinical testing and uncover redundancies. Finally, we evaluate our findings thus far and suggest that providing a well-informed framework for the extensive use of theory of information to clinical managerial problems is worthwhile.