JMIR Publications

ABSTRACT

Background:

Community-acquired acute kidney injury (CA-AKI)-associated hospitalizations impose significant health care needs and in-hospital mortality. However, most risk predictions have focused on AKI in a specific group of patients during hospitalization; there is limited knowledge on the baseline risk in the general population for preventing CA-AKI-associated hospitalization.

Objective:

To gain further insight into risk exploration, this study aimed to develop, validate, and set a scoring system to facilitate health professionals in enabling early recognition and intervention to prevent permanent kidney damage using different machine learning techniques.

Methods:

A nested case-control study design was employed using electronic health records derived from a group of Chang Gung Memorial Hospitals in Taiwan from 2010-2017 to identify 234,867 adults with at least two measures of serum creatinine at hospital admissions. Patients were classified into derivation cohort (2010-2016) and temporal validation cohort (2017). Patients with first episode of CA-AKI at hospital admission were classified into the case group and those without CA-AKI in the control group. A total of 47 potential candidate variables, including age, gender, prior use of nephrotoxic medications, Charlson comorbid conditions, commonly measured laboratory results, and recent use of health services were tested to develop a CA-AKI hospitalization risk model. Permutation-based selection with extreme gradient boost (XGBoost) and least absolute shrinkage and selection operator (LASSO) algorithms both determined the same top 10 important features.

Results:

The discriminative ability of the risk model was assessed by the area under the receiver operating characteristic curve (AUC), and the predictive CA-AKI risk model derived by logistic regression algorithm achieved an AUC of 0.767 (95% confidence interval [CI] 0.764–0.770) on derivation and of 0.761 on validation for any stage of AKI with 19.2% and 96.1% of the positive and negative predictive values. The risk model for the prediction of CA-AKI stages 2 and 3 had 0.818 of AUC on validation cohort with 13.3% and 98.4% of the positive and negative predictive values. The metrics above were evaluated at the cut-off value of 7.993, which was a threshold to discriminate the risk of AKI.

Conclusions:

Machine learning-generated risk score model can identify patients at risk of developing CA-AKI related hospitalization through a routine care data driven approach. The validated multivariate risk assessment tool could help clinicians to stratify patients in primary care, and to provide monitoring and early intervention to prevent AKI and improve the quality of AKI care in the general population. Clinical Trial: NA