22Oct 2024
Machine Learning Predicts GIB Risk in Hemodialysis Patients

Machine Learning Predicts GIB Risk in Hemodialysis Patients

Gastrointestinal bleeding (GIB) poses significant challenges in patients with kidney failure. The INSPIRE group evaluated whether machine learning could predict the 180-day GIB hospitalization risk in hemodialysis (HD) patients. Using an HD dataset from the U.S. (2017-2020), the eXtreme Gradient Boosting (XGBoost) and logistic regression models analyzed 451,579 patients. XGBoost achieved an area under the receiver operating curve (AUROC) of 0.74, outperforming logistic regression's 0.68. Both models identified factors like age and anemia disturbances as risk predictors. These findings highlight the potential of machine learning for early GIB risk detection, necessitating further validation

  • #nephrology

Like

Save

Share