A Framework of Hybrid Recommender System for Personalized Clinical Prescription

General practitioners are faced with a great challenge of clinical prescription owing to the increase of new drugs and their complex functions to different diseases. A personalized recommender system can help practitioners discover mass of medical knowledge hidden in history medical records to deal with information overload problem in prescription. To support practitioner's decision making in prescription, this paper proposes a framework of a hybrid recommender system which integrates artificial neural network and case-based reasoning. Three issues are considered in this system framework: (1) to define a patient's need by giving his/her symptom, (2) to mine features from free text in medical records and (3) to analyze temporal efficiency of drugs. The proposed recommender system is expected to help general practitioners to improve their efficiency and reduce risks of making errors in daily clinical consultation with patients.