Quantitative Evaluation for Uncertainty of Information About Patients' Injury Severity in a Hospital Disaster: A Simulation Study Using Shannon's Information Theory.

Abstract

Reducing uncertainty about information on injury severity or number of patients is an important concern in managing equipment and rescue personnel in a disaster setting. A simplified disaster model was designed using Shannon's Information Theory to study the uncertainty of information in a triage scenario. Hypothesis A disaster triage scene with a specific number of injured patients represents a source of information regarding the extent of patients' disability. It is possible to quantify uncertainty of information regarding patients' incapacity as entropy if the information source and information arising from the source in Information Theory can be adapted to the disaster situation and the information on patients' incapacity that arises. Five different scenarios of a fire disaster in a hospital were modeled. Information on patients' extent of impairment was converted to numerical values in relation to available equipment and the number of rescue personnel. Victims were 10 hospitalized patients with conditions of unknown severity. Triage criteria were created arbitrarily and consisted of four categories from Level 1 (able to walk) to Level 4 (cardiac arrest). The five situations were as follows: (1) Case 1: no triage officer; (2) Case 2: one triage officer; (3) Case 3: one triage officer and a message that six patients could walk; (4) Case 4: one triage officer and a message that all patients could obey commands; and (5) Case 5: one triage officer and a message that all patients could walk. Entropy in all cases and the amount of information newly given in Cases 2 through 5 were calculated. Entropies in Cases 1 through 5 were 5.49, 2.00, 1.60, 1.00, and 0.00 bits/symbol, respectively. These values depict the uncertainty of probability of the triage categories arising in each situation. The amount of information for the triage was calculated as 3.49 bits (ie, 5.49 minus 2.00). In the same manner, the amount of information for the messages in Cases 3 through 5 was calculated as 0.4, 1.0, and 2.0 bits, respectively. These amounts of information indicate a reduction in uncertainty regarding the probability of the triage levels arising. It was possible to quantify uncertainty of information about the extent of disability in patients at a triage location and to evaluate reduction of the uncertainty by using entropy based on Shannon's Information Theory.