Proposed clinical models outperform standard scoring tools for mortality, kidney recovery
Click Here to Manage Email Alerts
According to this study, researchers developed clinical models to predict mortality and kidney recovery among critically ill patients with incident AKI in the ICU that outperformed standard scoring tools.
Additional validation is required to support the implementation of the proposed models, the researchers wrote.
“Recent studies clearly showed that clinical risk-prediction models are helpful in predicting AKI onset in different critically ill populations. Even though these tools are of importance for the prevention, early diagnosis, and timely management of AKI, external validation of model performance and implementation of these tools for guiding interventions to prevent AKI or its progression are lacking,” Javier A. Neyra, MD, MSCS, from the University of Kentucky Medical Center, and colleagues wrote. They added, “The aim of this study was to develop and validate useful clinical models for the risk-stratification of critically ill adult patients that suffered from AKI, for their individual risk of hospital mortality and major adverse kidney events.”
In a multicenter, retrospective cohort study, Neyra and colleagues evaluated 7,354 adults admitted to the ICU at the University of Kentucky Hospital between March 2009 and February 2017 and 2,233 adults admitted to the ICU at the University of Texas Southwestern Medical Center between May 2009 and February 2017. All patients developed AKI within the first 3 days of ICU admission.
Using data from the first 3 days of patients’ ICU stays, researchers created clinical models to predict hospital mortality and major adverse kidney events occurring up to 120 days after hospital discharge. The initial model included 71 validated clinical variables that were chosen from data extracted from electronic health records.
Using 10-fold cross validation and external validation, researchers measured the predictive performance of the models. Additionally, investigators utilized Shapley Additive Explanations framework to interpret feature importance in the models.
The proposed clinical models were compared with corresponding baseline models.
Analyses revealed one proposed model that included 15 features outperformed the Sequential Organ Failure Assessment score for the prediction of hospital mortality. Similarly, another model that included 14 features outperformed Kidney Disease: Improving Global Outcomes (KDIGO) AKI severity staging for the prediction of major adverse kidney events.
“The proposed clinical models exhibited good performance for outcome prediction and could enable risk-stratification for timely interventions that promote or enhance kidney recovery, both during ICU stay (KDIGO AKI bundle, resource allocation) and after discharge from the hospital (specialized follow-up care),” Neyra and colleagues wrote. “We provide an online platform available to the public for applicability of the clinical models at the bedside. Additional external validation is needed to warrant wide implementation of these clinical models.”
Collapse
Click Here to Manage Email Alerts