Use of disease embedding technique to predict the risk of progression to end-stage renal disease

Fang Zhou, Avrum Gillespie, Djordje Gligorijevic, Jelena Gligorijevic, Zoran Obradovic

Abstract: Highlights•A simple but accurate model that automatically generates variables from EHRs.•Interpretable variables well summarize information of comorbidities.•The model helps clinicians more easily interpret the prediction results.•The model helps to tailor discussions with patients about their CKD progression.