Predicting Laboratory Test Outcomes for Patients on Maintenance Hemodialysis Using Cellular Bioelectrical Measurements

Background: Patients with end-stage kidney disease (ESKD) frequently experience complications such as anemia, malnutrition, and cardiovascular issues. Serological tests, which are invasive and not routinely conducted, play a crucial role in medical assessments. A non-invasive, convenient method for predicting these test results could significantly enhance patient monitoring. Methods: This study develops machine learning models to predict key serological test results using non-invasive bioelectrical impedance measurements, a routine clinical procedure for ESKD patients.The study employed two machine learning models, Support Vector Machine (SVM) and Random Forest (RF), to predict key serological tests from cellular bioelectrical indicators. Data from 688 patients, comprising 3,872 paired biochemical–bioelectrical records, were used for model validation. Results: Both SVM and RF models demonstrated effective classification of key serological results (albumin, phosphorus, parathyroid hormone) into low, normal, and high. RF generally exhibited superior performance compared to SVM, except in predicting calcium levels in women. Conclusion: The machine learning models effectively estimated serological test results for maintenance hemodialysis patients based on bioelectrical impedance parameters.Confusion matrices for albumin, calcium, parathyroid hormone, phosphorus, and hemoglobin levels classified by the model in both genders: (a) Support Vector Machine, (b) Random Forest.