Development of a Fuzzy Logic-Based Intelligent System for Drug Interaction Detection in Polypharmacy

polypharmacy, the simultaneous use of multiple medications, increases the risk of adverse drug interactions, especially in elderly or chronically ill patients. Traditional methods often struggle to address the uncertainty and variability in patient-specific factors like dosage, age, organ function, and comorbidities. This work presents a fuzzy logic-based intelligent system designed to detect and evaluate potential drug interactions in polypharmacy scenarios. By modelling clinical knowledge with fuzzy rules, the system interprets imprecise inputs, such as “high dose” or “poor organ function” and provides interaction severity scores along with clinical recommendations. The architecture combines a fuzzy inference engine, a drug interaction database (e.g., DrugBank or RxNorm), and patient data inputs including medications, dosages, age, renal/liver function, comorbidities, and a user-friendly web interface. The system supports healthcare professionals by offering interpretable risk assessments and decision support to enhance medication safety. The system architecture is established on three major levels which include an Input Layer that identifies patient demographics, laboratory results and detailed drug data; a Fuzzy Logic Core that deploys the input data on a predetermined set of fuzzy membership features and expert IF-THEN rules; and Output Layer that reports the risk assessment and clinical suggestions. To test the system five test cases involving patient age, renal functioning and medication load were chosen as representatives. The outcomes showed that the performance is high, and the average precision was 91.4%, recall was 89.7%, and F1-score was 90.5%. The system was able to assist in the prediction of DDI risk level such as considering complex clinical situations, as well as being able to predict the levels of DDI risk such as the level of risk being High and Very High, proving itself to be potentially helpful in protecting patient safety. Through the efficient provision of dynamic, personalized evaluations instead of use of the static lists of interactions, such fuzzy logic system is a scaly and understandable solution to the contemporary clinical decision support. Implementation was done in python using scikitfuzzy, and a FastAPI-based interface was developed for clinical usability. Fuzzy logic was applied to model imprecise variables (e.g., “high dose”, “elderly”). A rule-based system assessed drug interaction severity using inference and centroid defuzzification. Pairwise drug combinations were evaluated, and overall interaction risk was computed for each patient.The system successfully detected high-risk interactions with interpretable outputs, showing improved usability and clinical relevance. Fuzzy logic provides a flexible, interpretable, and effective framework for drug interaction detection in polypharmacy, supporting safer prescribing practices in complex scenarios