Pharmacovigilance is a critical measure to ensure drug safety, particularly challenging in the context of antibody drugs-complex macromolecular biologics, which due to their intricate adverse reaction monitoring and management needs. As a causal analysis tool grounded in probabilistic inference, Bayesian networks offer significant advantages in handling complex data relationships and uncertainties. Focusing on the characteristics of antibody drugs, including high target specificity, patterns of adverse reactions, and immunogenicity, this paper explores how to integrate multi-source data using Bayesian networks. Furthermore, it provides a preliminary investigation into the feasibility of causal inference and risk prediction for adverse events under conditions of variable missingness and complex interactions. The study aims to offer both a reference and methodological foundation for pharmacovigilance research on antibody drugs.