Application of fuzzy mathematical model of decision-making for the selection of optimal surgical tactics in patients with non-tumor obstructive jaundice

Abstract

Aim. Creation of a model of fuzzy logic for predicting the risk of postoperative complications and the choice of individual optimal surgical tactics in obstructive jaundice caused by choledocholithiasis. Methods. At the first stage, we determined the most prognostically significant factors affecting the risk of postoperative complications. In accordance with these factors, linguistic variables were introduced: X1 - patient’s age; X2 - duration of jaundice; X3 - temperature; X4 - comorbidities; X5 - the level of liver dysfunction; X6 - CD4+ in the blood; X7 - interleukin-2 in the serum; Y - level of risk. The intervals of their changes were determined. Fuzzi Logic Toolbox Matlab soft was used to achieve the determined aim. The values of input variables were introduced into the model, transformed in the «Phaser» block and then the rule base of the fuzzy inference system was formed by the expert method. As a result, the level of risk is determined and the choice of surgical tactics is made: (1) risk is absent or low (A); (2) doubtful risk (B) - if the risk assessment in the dynamics after preoperative therapy decreases, then tactics A, if the score does not decrease or increases, then tactics C; (3) high and very high risk (C) - an unequivocal choice of stage tactics. Results. According to the defined level of risk, in 92 patients a one-stage procedure was used, while 58 underwent a two-stage intervention. Due to the developed fuzzy mathematical model, forecasting of the optimal choice of surgical tactics is achieved, which significantly improves the results of treatment. Conclusion. The developed fuzzy mathematical model makes it possible to differentiate the choice of surgical tactics for a particular patient and thereby reduce the incidence of postoperative complications from 29.0 to 4.7% and mortality from 11.0 to 1.3%.