대한신장학회

Skip Navigation: Skip to contents

KOR
ENG
전체메뉴보기

간행물 검색

현재 페이지 경로

HOME
간행물
간행물 검색

논문분류	춘계학술대회 초록집
제목	Development and Internal Validation of a Clinical Decision Support System for Maintenance Hemodialysis Prescription Using Longitudinal Real-World Data
저자	Jin Eop Kim
출판정보	2026; 2026(1):
키워드	hemodialysis, drug prescription, supervised machine learning, classification
초록	Objectives: Nephrologists face substantial workload related to the monthly review of laboratory findings and medication adjustments for patients undergoing maintenance hemodialysis. Although prescribing practices are guided by established recommendations such as KDIGO and K-DOQI guidelines, real-world implementation varies among physicians and is often influenced by accumulated clinical experience. To enhance consistency and support clinical workflow, we developed and internally validated a clinical decision support system using longitudinal real-world prescription data. Methods: We analyzed approximately 180,000 paired laboratory and prescription records collected between 2006 and 2023 at a teaching hospital. Structured prescription protocols were first formalized and incorporated into a supervised machine learning framework. The model was trained to generate medication recommendations based on current laboratory results and prior prescription history. Internal validation was performed using patient-level data separation to prevent information leakage. Model performance was assessed using accuracy, precision, recall, and F1-score. Results: The final model demonstrated high concordance with historical prescriptions, achieving performance metrics exceeding 0.99 across all evaluation indices. In additional validation analyses, predicted prescriptions were largely consistent with established protocol logic. The system was integrated into a user interface that allows physicians to review and finalize recommended prescriptions. Conclusion: This clinical decision support system may assist in standardizing maintenance hemodialysis prescription processes and reducing repetitive cognitive workload. Prospective evaluation is warranted to determine its impact on clinical outcomes and physician workflow efficiency.
원문(PDF)	PDF 원문보기

목록

위로가기