A Shared Decision-Making System for Diabetes Medication Choice Utilizing Electronic Health Record Data

摘要: The use of a shared decision-making (SDM) process in antihyperglycemic medication strategy decisions is necessary due to the complexity conditions diabetes patients. Knowledge guidelines used as decision aids clinical situations, and during this process, no patient health are considered. In paper, we propose an SDM system framework for type-2 mellitus (T2DM) patients that not only contains knowledge abstracted from but also employs multilabel classification model uses class-imbalanced electronic record (EHR) data aims provide recommended list available medications help physicians have conversation. EHR serve decision-support component helps reach more intuitive understanding current allows tailoring each patient, leading effective SDM. Real-world 2542 T2DM inpatient EHRs were substituted by 77 features eight output labels, i.e., medications, these utilized build validate recommendation model. exhibited stable performance every single-label showed ability predict minority positive cases which average recall value classes was 0.9898. As whole classifier, demonstrated outstanding performance, with scores 0.0941 Hamming Loss , 0.7611 ${\rm{Accuracy}}_{{\rm{exam}}}$ 0.9664 ${\rm{Recall}}_{{\rm{exam}}}$ 0.8269 $F_{{\rm{exam}}}$ .