Page - 24 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

theory is to provide a sound basis for optimising the amount of information that can be extracted from a specific situation. Many of the outcomes from the study of information theory have been reduced to engineering practice in a wide range of disciplines, from Artificial Intelligence and Machine Learning, to cybernetics and complexity science. So, the question naturally arises: could it be used to inform the practice of medicine. This is the topic of the current chapter. Central to Information Theory is the study of situations where one agent (the transmitter) conveys some message over a channel to another agent (the receiver). This is typically performed by having the transmitter send a series of partial messages. In the case of the Internet, for example, the Transmission Control Protocol (TCP) defines how a message may be broken down into packets before sending, enabling the resulting packets to be reassembled in the correct order by the receiver. Each of these partial messages can be thought of as resolving some measure of uncertainty in the receiver as to the content of the original message. The measure of uncertainty resolved by a partial message is its information content. Let us start with a schematic of a general communication system, redrawn after Shannon’s original paper [9]. Figure 1. Schematic of a general information system. We start with an information source, which generates a message or sequence of messages which are intended to be communicated to a destination. The destination is assumed to be remote from the information source. Hence, the message needs to be converted by a transmitter into a signal that is in a suitable form to be transmitted through some channel, after which the received signal is converted back into a suitable format by a receiver to enable it to be easily interpreted at the destination. The challenge of communication theory is to understand how information that is transmitted from the source can be completely and correctly received and interpreted by the destination. This is a challenge because in general any communication channel will have an associated noise source that may corrupt, to a greater or lesser extent, the transmitted signal before it is received (by altering or even losing components of the signal). Furthermore, we cannot be certain that the transmitter and the receiver are perfect converters of message to signal, and signal to message, respectively. In this chapter, we will show how viewing diagnosis as embedded within a communication system can lead to an information theoretic perspective on medical diagnosis. Each test or intervention can be seen as a partial message leading towards the desired complete message that provides sufficient information to confirm a diagnosis. At any stage in an investigation, one would then select the next test as the one that would Transmitter Channel Information Source Receiver Destination Noise Source message message signal received signal P.Krause / InformationTheoryandMedicalDecisionMaking24