Modeling a semantic recommender system for medical prescriptions and drug interaction detection

The administration of appropriate drugs to patients is one of the most important processes of treatment and requires careful decision-making based-on the current conditions of the patient and its history and symptoms. In many cases, patients may require more than one drug, or in addition to having a previous illness and receiving the drug, they need new drugs for the new illness, which may increase medical errors in the administration of the drug and the adverse drug events(ADE) such as drug interactions for the patient.

In this article, the stages of designing and describing the requirements and the modeling of the ontology-based semantic recommender system of the prescribing physician and the discovery of the ADEs were presented. First, the requirements of the system were extracted and described in detail and then, based on the extracted requirements, the modeling of the system using the Unified Modeling Language of UML2.0 was discussed. Then, according to the extracted requirements for the discovery of ADEs, a proper ontology was designed for the system and implemented by Protégé software. In order to evaluate the functions of recommendation and discovering ADEs (interactions), a prototype was developed using Java language, and a collection of rules for reasoning and discovering interactions and ADEs were gathered.

The results of the system performance evaluation for the functions of detecting ADEs and medication recommendation suggests improvement of the proposed approach to 9.25% and 11.3% in the precision criterion, 29% and 60.6% in the recall, and 26% (respectively, approaches to the detection of ADEs and drug recommendations).

The use of this system as a computerized physician ordering entry can, in addition to helping physicians to prescribe a more accurate prescription, reduce the risks to the health of patients resulting from medical errors in the prescribing phase.