Construction of an Early Alert System for Intradialytic Hypotension before Initiating Hemodialysis Based on Machine Learning

Abstract

Introduction: Intradialytic hypotension (IDH) is prevalent and associated with high hospitalization and mortality rates. The purpose of this study was to explore the risk factors for IDH and use artificial intelligence to establish an early alert system before hemodialysis sessions to identify patients at high risk of IDH. Materials and methods: We obtained data on 314534 hemodialysis sessions conducted at Sichuan Provincial People's Hospital from the Renal Disease Treatment Information System. IDH was defined as a systolic blood pressure drop≥20 mmHg, a mean arterial pressure drop≥10 mmHg during dialysis, or the occurrence of clinical hypotensive events requiring nursing intervention. After pre-processing, the data were randomly divided into training (80%) and testing (20%) sets. Four interpolation methods, three feature selection methods, and 18 machine learning algorithms were used to construct predictive models. The area under the receiver operating characteristic curve (AUC) was the main indicator for evaluating the performance of the models, while Shapley Additive ExPlanation (SHAP) was used to explain the contribution of each variable to the best predictive model. Results: A total of 3906 patients and 314534 dialysis sessions were included, of which 142237 cases showed IDH (incidence rate, 45.2%). Nineteen parameters were identified through artificial intelligence feature screening. They included age, pre-dialysis weight, dry weight, pre-dialysis blood pressure, heart rate, prescribed ultrafiltration, blood cell counts (neutrophil, lymphocyte, monocyte, eosinophil, lymphocyte, and platelet counts), hematocrit, serum calcium, creatinine, urea, glucose, and uric acid. Random forest, gradient boosting, and logistic regression were the three best models, and the AUCs were 0.812 (95% confidence interval[CI], 0.811-0.813), 0.748 (95% CI, 0.747-0.749), and 0.743 (95% CI, 0.742-0.744), respectively. Conclusion: Our dialysis software-based artificial intelligence alert system can be used to predict IDH occurrence, enabling the initiation of relevant interventions.

External validation of the prediction model of intradialytic hypotension: a multicenter prospective cohort study

“…The outcomes revealed the superior performance of the original column-line graph model in the external validation, boasting an AUC value of 0.746 (95% CI: 0.718-0.775), aligning closely with the observations of Hong et al. who reported an AUC of 0.743 [ 32 ]. Nevertheless, it was discerned that the discriminative and calibration capabilities of all three initial prediction models were notably overestimated.…”

Construction and Application of Machine Learning Models for Predicting Intradialytic Hypotension

“…Particularly well-suited for tasks involving categorical and heterogeneous data [29], CatBoost emerged as the top-performing machine learning algorithm in our study across all IDH de nitions. In prior studies on machine learning-based IDH prediction, Daqing Hong et al [15] utilized 18 machine learning algorithms, with the top three being random forest (AUC = 0.812, 95% CI = 0.811-0.813), gradient boosting (AUC = 0.748, 95% CI = 0.747-0.749), and logistic regression (AUC = 0.743, 95% CI = 0.742-0.744). Notably, our study incorporated the CatBoost algorithm, which was absent in their analysis.…”

“…Secondly, the assessment of IDH in our study lacked the inclusion of clinical symptoms due to the unavailability of valid clinical symptom data. Notably, a limited number of studies incorporated clinical symptoms [15,34],…”

“…Firstly, the absence of consensus on IDH de nitions has led to variations in the de nitions employed across current research [11,12]. Secondly, the spectrum of clinical features gathered in these studies is diverse, encompassing blood test indices, medication pro les, hemodialysis-related parameters, co-morbidities, and cardiothoracic ratios, but none have incorporated echocardiography-related parameters [13][14][15][16]. However, this study highlights the potential of echocardiographic assessments as a predictive tool for IDH [17].…”

A Klotho-Based Machine Learning Model for Prediction of both Kidney and Cardiovascular Outcomes in Chronic Kidney Disease

“…Machine learning (ML) is now widely used in the construction of clinical prediction models for specific tasks, ultimately yielding more accurate predictive models [21]. In nephrology, ML has found utility in identifying high-risk patients [22], managing transplant recipients [23], making prescribing decisions for hemodialysis patients [24], and more. Harnessing ML's potential in kidney disease enhances clinicians' ability to predict ESKD and CKD [25].…”