Characterizing the temporal changes in association between modifiable risk factors and acute kidney injury with multi-view analysis

Abstract

Background: Acute kidney injury (AKI) is a common life-threatening clinical syndrome in hospitalized patients. Advances in machine learning has demonstrated success in AKI risk prediction using electronic health records (EHRs). However, to prevent AKI, it is critical to identify clinically modifiable factors and understand their impact at different prevention windows.

Method: We extracted 4129 clinical variables including demographics, social history, past diagnoses, procedures, labs, medications, vitals from EHRs for a cohort of 144,084 eligible inpatient encounters. We developed a multi-view learning framework for XGBoost (MV-XGB) to enhance algorithm attention on modifiable factors. To study effects of modifiable factors at different time points, we built AKI prediction models at 24-hours, 48-hours, 72-hours before AKI onset. To characterize the temporal changes in effect of modifiable factors on AKI, we derived two indicators, inter-class score-difference and exposed-score-difference, based on SHAP values to compare effects of modifiable factors in different windows.

Result: MV-XGB effectively increased attention on modifiable factors (explained 92.4%-94.1% inter-class score-difference, i.e., predictive difference between AKI and non-AKI samples) while maintaining good predictive performance (AUROCs were 0.854, 0.798, 0.765 in models for 24-48-72 h AKI prediction respectively). We observed that 62% of predicted odds-ratio difference between AKI and non-AKI patients in 24 h can be explained by factors occurring between 24 and 72 h. Among the important modifiable factors, electrolyte balance explained 38.3% of the inter-class score difference increase between 24 h and 72 h, followed by high-risk medications (13.7%), care strategy (12.1%), blood pressure (10%), infection (7.8%), and anemia (5.4%). Effects of cardiac surgery or condition, respiratory ventilation, and anemia remained important longer than 72 h.

Conclusion: Better understanding of the clinically modifiable factors is important to AKI prevention. The proposed multi-view learning approach improved the identification of modifiable factors of AKI and allowed characterization of the temporal dynamics of their potential benefit in intervention.

Keywords: Acute kidney injury; Electronic health records; Machine learning; Multi-view analysis; Prevention.

Fig. 1.

Proportional change of important predictors between original and multi-view XGB.

Fig. 2.

Proportional change of source of inter-class prediction score difference between original and multi-view XGB.

Fig. 3.

Calibration comparison between original and multi-view XGB.

Fig. 4.

Comparison of Top-20 modifiable features between MV-XGB for 24 h and 72 h.

Fig. 5.

Factors related to inter-class score difference increase in different time windows. Results in this figure were based on union of top-50 modifiable features in MV-XGB for 24–72 h. Inter-class score difference of factors and its share (%) in all important predictors is presented. The percentages were calculated by dividing change of inter-class score different in a specific class of important factors by change of inter-class score different in all important predictors. In (A)-(C), only effect change of predictors increased inter-class score difference from 72 h to 24 h were considered.