Development of a prediction score for in-hospital mortality in COVID-19 patients with acute kidney injury: a machine learning approach

Abstract

Acute kidney injury (AKI) is frequently associated with COVID-19 and it is considered an indicator of disease severity. This study aimed to develop a prognostic score for predicting in-hospital mortality in COVID-19 patients with AKI (AKI-COV score). This was a cross-sectional multicentre prospective cohort study in the Latin America AKI COVID-19 Registry. A total of 870 COVID-19 patients with AKI defined according to the KDIGO were included between 1 May 2020 and 31 December 2020. We evaluated four categories of predictor variables that were available at the time of the diagnosis of AKI: (1) demographic data; (2) comorbidities and conditions at admission; (3) laboratory exams within 24 h; and (4) characteristics and causes of AKI. We used a machine learning approach to fit models in the training set using tenfold cross-validation and validated the accuracy using the area under the receiver operating characteristic curve (AUC-ROC). The coefficients of the best model (Elastic Net) were used to build the predictive AKI-COV score. The AKI-COV score had an AUC-ROC of 0.823 (95% CI 0.761–0.885) in the validation cohort. The use of the AKI-COV score may assist healthcare workers in identifying hospitalized COVID-19 patients with AKI that may require more intensive monitoring and can be used for resource allocation.

Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation

“…In this study, we combined ML-based models with electronic medical record data to retrieve information on various clinical characteristics that affect in-hospital mortality, and we determined that the RF, XGB, and GBM models exhibited the most favorable discriminative ability with AUROCs of 0.816, 0.806, and 0.823, respectively. The accuracy of the ML models used in this study in predicting in-hospital mortality in the patients receiving CRRT is comparable to that reported in previous studies [ 7 , 10 , 22 , 23 , 24 , 25 ]. In particular, our model exhibited higher discriminative power than did the prediction model developed by Kang et al [ 7 ].…”

“…Pattharanitima et al used ML and deep learning to build a model for predicting renal replacement therapy-free survival and demonstrated that the long short-term memory model with multilayer perceptron architecture exhibited high discrimination performance with an AUC of 0.70 [ 13 ]. Studies have investigated the accuracy of other ML-based mortality prediction models in intensive care settings, including in patients with lactic acidosis [ 23 ], mechanically ventilated patients [ 24 ], and those with COVID-19 and AKI [ 25 ]. These findings emphasize the importance of ML in predicting outcomes in critical care settings, especially for patients receiving CRRT.…”

Development of a Risk Score for AKI onset in COVID-19 Patients: COV-AKI Score

“…In this way, our score is applied at the time of patient admission to the ICU to predict AKI development during ICU stay, thus selecting a more complex population of patients with worse prognosis and greater chance of developing organ dysfunctions, including kidney failure. Also, other authors did not determine a specific point in time where the score should be applied and combined different populations of patients in the same study, such as patients admitted to the emergency room, ward and the intensive care unit [17,[35][36][37]. Furthermore, Lu et al [36], describes that the overall prediction performance by Area under Receiver Operating Charactheristic Curve (AUC) was good at day 0, and moderate at day -1 and -2, a finding with doubtful clinical significance, since at that moment (day 0) there is no sufficient time to adopt preventive measures that minimize the risk of developing AKI during hospitalisation.…”

Application of interpretable machine learning for early prediction of prognosis in acute kidney injury

“…Accurate and timely prediction of mortality for AKI is required to identify patients at high risk of clinical deterioration so that preventive measures can be taken in a timely manner, which may reduce mortality. Several studies have attempted to establish prognostic models among AKI patients with ML methods and showed a modest prognostic yield [5] , [7] , [8] , [28] , [29] , [30] , [31] , [32] , [33] , [34] . For example, a study from the US constructed a prognostic model for predicting 60-day mortality in critically ill patients with AKI.…”