Computational Drug Recommendation Approaches toward Safe Polypharmacy

Adverse drug reactions (ADRs) induced from high-order drug-drug interactions (DDIs) due to polypharmacy - simultaneous use of multiple drugs - represent a significant public health problem. Unfortunately, computational efforts to facilitate decision making for safe polypharmacy, particularly to assist safe multi-drug prescribing, are lacking. We formally formulate the to-avoid and safe drug recommendation problems for multi-drug prescriptions. We investigate preliminary computational approaches to tackling these problems, utilizing a minimum set of available prescription data from a large population, and demonstrate their potentials in assisting safe-polypharmacy decision making once richer data (e.g., electronic medical records, omics data and pathology data) are available. We develop a joint model with a recommendation component and an ADR label prediction component to conduct to-avoid and safe drug recommendation. We also develop real drug-drug interaction datasets and corresponding evaluation protocols to facilitate future computational research on safe polypharmacy.