Key result
LightGBM model predicts physical frailty from EHRs in hospitalized CVD patients with ~0.80 AUC.
Why the study?
The study aimed to develop a deep learning model to predict phenotypic physical frailty from electronic medical record data in patients with cardiovascular disease.
Does an AI-driven deep learning model using electronic health records accurately predict phenotypic physical frailty and prognosis in hospitalized patients with cardiovascular disease?
Cohort (n=2,705)
No
Does an AI-driven deep learning model using electronic health records accurately predict phenotypic physical frailty and prognosis in hospitalized patients with cardiovascular disease?
Deep learning models using routine electronic health record data can accurately predict phenotypic physical frailty and are strongly associated with long-term prognosis in hospitalized patients with cardiovascular disease.
EHR-based frailty prediction may aid risk stratification in hospitalized CVD patients; leaves open prospective validation before clinical adoption.
Background: This study aimed to create a deep learning model for predicting phenotypic physical frailty from electronic medical record information in patients with cardiovascular disease. Methods and Results: This single-center retrospective study enrolled patients who could be assessed for physical frailty according to cardiovascular health study criteria (25.5% [691/2,705] of the patients were frail). Patients were randomly separated for training (Train set: 80%) and validation (Test set: 20%) of the deep learning model. Multiple models were created using LightGBM, random forest, and logistic regression for deep learning, and their predictive abilities were compared. The LightGBM model had the highest accuracy (in a Test set: F1 score 0.561; accuracy 0.726; area under the curve of the receiver operating characteristics [AUC] 0.804). These results using only commonly used blood biochemistry test indices (in a Test set: F1 score 0.551; accuracy 0.721; AUC 0.793) were similar. The created models were consistently and strongly associated with physical functions at hospital discharge, all-cause death, and heart failure-related readmission. Conclusions: Deep learning models derived from large sample sizes of phenotypic physical frailty have shown good accuracy and consistent associations with prognosis and physical functions.
No takes yet. Share an insight, caveat, or question.
Yamashita et al. (2024) conducted a cohort in Cardiovascular disease (n=2,705). LightGBM deep learning frailty prediction model was evaluated on Prediction of phenotypic physical frailty (AUC in Test set) (AUC 0.804). A LightGBM deep learning model predicted phenotypic physical frailty from electronic medical records in hospitalized patients with cardiovascular disease with an AUC of 0.804.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:
More from Circulation Reports