Machine learning for the prediction of all-cause mortality in patients with sepsis-associated acute kidney injury during hospitalization

Abstract

BackgroundSepsis-associated acute kidney injury (S-AKI) is considered to be associated with high morbidity and mortality, a commonly accepted model to predict mortality is urged consequently. This study used a machine learning model to identify vital variables associated with mortality in S-AKI patients in the hospital and predict the risk of death in the hospital. We hope that this model can help identify high-risk patients early and reasonably allocate medical resources in the intensive care unit (ICU).MethodsA total of 16,154 S-AKI patients from the Medical Information Mart for Intensive Care IV database were examined as the training set (80%) and the validation set (20%). Variables (129 in total) were collected, including basic patient information, diagnosis, clinical data, and medication records. We developed and validated machine learning models using 11 different algorithms and selected the one that performed the best. Afterward, recursive feature elimination was used to select key variables. Different indicators were used to compare the prediction performance of each model. The SHapley Additive exPlanations package was applied to interpret the best machine learning model in a web tool for clinicians to use. Finally, we collected clinical data of S-AKI patients from two hospitals for external validation.ResultsIn this study, 15 critical variables were finally selected, namely, urine output, maximum blood urea nitrogen, rate of injection of norepinephrine, maximum anion gap, maximum creatinine, maximum red blood cell volume distribution width, minimum international normalized ratio, maximum heart rate, maximum temperature, maximum respiratory rate, minimum fraction of inspired O2, minimum creatinine, minimum Glasgow Coma Scale, and diagnosis of diabetes and stroke. The categorical boosting algorithm model presented significantly better predictive performance [receiver operating characteristic (ROC): 0.83] than other models [accuracy (ACC): 75%, Youden index: 50%, sensitivity: 75%, specificity: 75%, F1 score: 0.56, positive predictive value (PPV): 44%, and negative predictive value (NPV): 92%]. External validation data from two hospitals in China were also well validated (ROC: 0.75).ConclusionsAfter selecting 15 crucial variables, a machine learning-based model for predicting the mortality of S-AKI patients was successfully established and the CatBoost model demonstrated best predictive performance.

Development and validation of a machine-learning model for predicting the risk of death in sepsis patients with acute kidney injury

“…In addition, the random forest algorithm was proved to perform better than logistic regression models using SAPS II, APACHE IV, and SOFA. Zhou et al [ 13 ] also developed multiple machine learning models for mortality prediction of S-AKI patients using the MIMIC-IV dataset, and the CatBoost model was reported to perform the best. However, only model discriminability was assessed during performance evaluation, and no sample size calculation was performed to justify the sufficiency of the training sample size.…”

“…Some models were developed using the AKI dataset for predicting the risk of death in SAKI patients, which is subject to bias in cohort selection. Furthermore, many studies did not calculate the sample size before constructing the final models [ [10] , [11] , [12] , [13] ]. This issue easily leads to ineffectiveness of the model.…”

Improving predictions: Enhancing in-hospital mortality forecast for ICU patients with sepsis-induced coagulopathy using a stacking ensemble model

“…Measuring these 2 parameters in the early stages of SIC patients can aid in assessing renal function, identifying sepsis-related acute kidney injury, and predicting disease progression and prognosis. [ 31 ] As an essential physiological parameter in the human body, SpO 2 serves to assess the circulatory system functionality and stands as a crucial monitoring indicator in early resuscitation protocols. [ 32 ] Lara Hessels et al’s study revealed a close association between potassium disturbances and in-hospital mortality, which persisted even after adjusting for disease severity and acute kidney injury (AKI).…”

Artificial intelligence and machine learning’s role in sepsis-associated acute kidney injury

“…Artificial intelligence (AI) and machine learning (ML) are rapidly emerging as transformative tools for diagnosing and managing AKI patients [12][13][14][15][16][17][18][19][20][21]. Compared to traditional methods, ML algorithms can reveal patterns beyond human discernment and enhance SA-AKI prediction accuracy by analyzing vast datasets [22][23][24][25][26][27]. Furthermore, ML enables earlier SA-AKI detection than traditional approaches, allowing timely, appropriate intervention and improved outcomes [12][13][14][15][16][17][18][19][20]24,[28][29][30][31][32].…”

“…Contemporary research increasingly explores AI/ML's capabilities to advance precision medicine and tailored SA-AKI care. The integration of these technologies promises to usher in a new era of early detection and optimized therapeutic interventions for SA-AKI [22][23][24][25][26][27]. Several stateof-the-art studies and initiatives are currently underway, highlighting the adoption of these technologies in various clinical settings, each aiming to address the profound challenges posed by SA-AKI with a degree of sophistication previously unattainable [12][13][14][15][16][17][18][19][20].…”