[Bayesian methods in real-world studies]

Real-world studies have been widely used in clinical and public health research to evaluate effectiveness, assess risks, and inform health policy decision-making. However, real-world data typically lack randomization and controlled interventions, are often heterogeneous, and exhibit substantial population heterogeneity, posing major challenges for uncertainty assessment and multi-source evidence integration in statistical inference. Bayesian methods, by combining prior information with observed data, provide a flexible and coherent framework for statistical inference in complex data environments. The paper systematically reviews the major applications of Bayesian methods in real-world studies, with a particular focus on the theoretical foundations, modeling strategies, and applicability of state-of-the-art information borrowing approaches. In addition, we summarize recent advances in Bayesian causal inference, hierarchical modeling, and related methods in real-world research. Finally, this paper discusses the methodological issues that need to be addressed when applying Bayesian methods to real-world research, providing a reference for related research.