Forecasting Surgical Outcomes Using a Fuzzy-Based Decision System

Document Type: Research Paper

Author

PhD Candidate; Department of Information Technology Management, Science and Research Branch, Islamic Azad University, Tehran, IR Iran

Abstract

Background and objectives: The kidneys of chronic kidney disease (CKD) patients do not have enough function and hemodialysis (HD) is a common procedure for their treatment. HD requires vascular access surgery (VAS) and arteriovenous fistula (AVF) is a low-complication method in VAS. However, different rates of AVF failure have been reported worldwide which can cause repeating surgeries and patient hospitalization. The goal of this study was to provide a system with the ability to predict VAS outcomes to reduce failures of surgeries.
 
Methods: The data of created AVF for 195 CKD patients – consisting 131 males (67.18%) and 64 females (32.82%), and aged from 15 to 87 years - were studied. Our provided system is based on “Fuzzy Inference System” (FIS) and learns rules by extracted results of decision tree algorithm.
Results: The number of diabetic patients was 73 and 117 persons had hypertension. Their hemoglobin range was 4.9 to 16. Their systolic blood pressure (BP) and diastolic BP were in the ranges [95-230] and [60-120], respectively. Using provided fuzzy control system, these results were investigated: (i) When the systolic BP increases, the AVF maturation improves (ii) In the young patients, the rate of AVF failure is higher than older patients; (iii) Growing patient from “Young” to “Middle-aged” causes switching from “AVF failure” status to “late Maturation”; (iv) In aged patients, high systolic BP with low diastolic BP, shifts from “late” AVF maturation to better statuses namely “good” and “excellent”.
Conclusion: Using FIS can forecast surgery outcomes and thus reduce risk factors of patients. In the present developed fuzzy system, surgeons can configure the risk ranges of patient’s parameters before vascular surgery and configure changeable factors based on estimating postoperative outcomes.

Keywords