Deep Learning for Intradialytic Hypotension Prediction in Hemodialysis Patients

Abstract

Intradialytic hypotension is a common problem during hemodialysis treatment. Despite several clinical variables have been authenticated for associations during dialysis session, the interaction effects between variables has not yet been presented. Our study aimed to investigate clinical factors associated with intradialytic hypotension by deep learning. A total of 279 participants with 780 hemodialysis sessions on an outpatient in a hospital-facilitated hemodialysis center were enrolled in March 2018. Associations between clinical factors and intradialytic hypotension were determined using linear regression method and deep neural network. A full-adjusted model indicated that intradialytic hypotension is positively associated with body mass index (Beta = 0.17, p = 0.028), hypertension comorbidity (Beta = 0.17, p = 0.008), and ultrafiltration amount (Beta = 0.31, p < 0.001), and is inversely associated with the ultrafiltration rate in a hemodialysis session (Beta = −0.30, p = 0.001). The 4-factor locus obtained by the deep neural network reached the maximum performance metrics evaluation (accuracy = 64.97± 0.94; true positive rate = 87.97 ± 2.73; positive predictive value = 66.74 ± 0.98; Matthews correlation coefficient = 0.19 ± 0.03). The prediction model obtained by the deep learning scheme could be a potential tool for the management of intradialytic hypotension. INDEX TERMS Hemodialysis, deep learning, intradialytic hypotension.

Machine-Learning Model to Predict the Intradialytic Hypotension Based on Clinical-Analytical Data

“…First, the XGBoost model achieved more accurate results than the rest of the ML models. This model also achieved better results for all metrics compared to previous work [40], where only the clinical variables of the patient HD were used, and presented better results than other studies [33], [35]- [37]. A possible explanation for these good results may lie in the novel incorporation of the clinical and analytical variables that were most influential, according to the applied ensemble-trees and the domain knowledge of nephrology experts in the creation of the ML predictor model.…”

“…In [35], a deep neural network (DNN) model was proposed with the potential to determine the clinical factors that are related to the occurrence of IDH during an HD session. The researchers collected demographic data, HD clinical variables and laboratory data to identify factors associated with IDH.…”

System of integrating biosignals during hemodialysis: the CONTINUAL (Continuous mOnitoriNg viTal sIgN dUring hemodiALysis) registry

“…The large quantity of biosignals necessitates advanced or novel analytics that range from collection to interpretation [ 24 ]. Machine learning, including deep learning, is a rapidly developing branch of artificial intelligence that has shown promise for use in clinics [ 13 , 25 ]. A major limitation in utilizing biosignals for artificial intelligence-based clinical purposes is the lack of data storage [ 26 ].…”

Dialysis adequacy predictions using a machine learning method

“…Various studies have also used DL to investigate hemodialysis. Akl et al 14 suggested decades ago that the neural network can achieve artificial-intelligent dialysis control, and studies on intradialytic hypotension predictions [15][16][17][18] , the optimal dry weight setting 19 , and anemia control 20 for hemodialysis have been presented. DL in research has also expanded to other kidney diseases to predict acute kidney injury outcomes 21,22 and hyperkalemia 23 .…”