The opioid crisis, a pressing global public health issue, has led to a significant rise in overdose deaths, particularly among individuals under 50, with profound social and economic impacts. This study proposes a comprehensive forecasting system to predict drug use and overdose trends by integrating diverse data sources, including police reports, social network data, medical records, and sewage-based drug epidemiology. Utilizing Recurrent Neural Networks (RNNs), the system aims to identify individuals at risk of opioid abuse by analysing demographic information, medical histories, and prescription records, while distinguishing between therapeutic and harmful usage. Emphasizing privacy protection, ethical data handling, and model interpretability, this approach seeks to enhance the accuracy and timeliness of overdose risk predictions. The findings have the potential to inform clinical decision-making, shape public health policies, and drive targeted interventions to mitigate the opioid epidemic