Personalized Drug Information and Recommendation System Using Gradient Boosting Algorithm (GBM)

The growing complexity in pharmaceutical treatments and the call for personalized patients care have been a challenge to healthcare professionals. A drug recommendation system with rich drug information, as well as patient characteristics and medical history, was designed and tested in this study. All data utilized for the investigation were extracted from EHRs and pharmacy databases. Using 80% of the data, a GBM model was built, and it was then tested with the remaining 20% of the data. The system dramatically improved outcomes of treatment: it greatly reduced ADRs by 25%, and improved adherence to medication by 30%. The mean dosage error decreased by 40%, and in 90% of cases, there were potential drug-drug interactions. This study answers the possibilities that GBM-based personalized medicine can enhance patient safety, improved decision making, and optimized treatment strategies. Given that an estimated 75% of healthcare costs consist of chronic disease management, this paradigm is well-positioned to do better than previous strategies in reducing healthcare spending.