Label attention meets deep learning: a novel model for ICD-11 automatic classification of Chinese EMRs
Peer-Reviewed Publication
Updates every hour. Last Updated: 20-Jun-2026 14:16 ET (20-Jun-2026 18:16 GMT/UTC)
The study proposes LA-TextCNN-BiLSTM, an ICD-11 automatic coding model using MC-BERT and label attention. Experiments on clinical records show 83.86% accuracy, 75.82% macro-F1, and 82.83% micro-F1. The model effectively handles long-distance dependency and Chinese semantic diversity in electronic medical records.
Veteran male athletes who have spent years training at high intensity may be at greater risk of serious heart problems while exercising, new University of Leeds research shows. The study shows that male endurance athletes aged over 50 may be more likely to experience abnormal heart rhythms during training if they already have scarring in their heart. The team says their research shows that exercise itself is not dangerous - but the findings suggest that the presence of scar tissue in the heart increases the risk of potentially dangerous heart problems during physical activity.
Both the new weight loss drugs and bariatric (weight loss) surgery improve body composition in patients with obesity by inducing a moderate loss of fat-free mass (including lean muscle) along with a substantial reduction in fat, researchers at Vanderbilt Health have found.