Detecting adverse Drug-Drug Interactions (DDIs) is critical for patient safety. Graph-based models have advanced DDI prediction by integrating heterogeneous biomedical knowledge. However, most fail to account for patient heterogeneity. Recent work emphasizes that population factors (e.g. age, gender, weight) strongly influence drug outcomes. We built a Neo4j knowledge graph by integrating Canada Vigilance ADR reports, DrugBank, and other biomedical sources. Nodes represent drugs, targets (genes/proteins), diseases, ADR terms, and patient attributes. Patient demographics (age group, sex, weight category) were one-hot or binary encoded and linked to relevant nodes. Embedding patient demographics into the DDI knowledge graph enables more personalized predictions and supports precision medicine goals. This approach can stratify DDI risk by subpopulation and guide safer, individualized prescribing. Key challenges include integrating real-world patient data at scale, handling missing or noisy attributes, and preserving privacy. Nonetheless, demographic-aware knowledge graphs offer a promising path for personalized drug safety assessment.