The growing complexity and uncertainty in health care enterprise management calls for an innovative system-oriented modeling paradigm. When applying traditional probabilistic approaches such as Markov process models, Bayesian network models, and various simulation models to manage and analyze such a complex integrated service system, challenges arise due to the lack of a well-controlled framework to connect these models and process evidence. This article focuses on MEDIA, a Model and Evidence Driven Integrated Analysis, that interfaces system boundaries to reuse existing models, incorporate new models, and integrate evidence efficiently and effectively within an enterprise framework. MEDIA relies on ontology-dependency modeling, probabilistic evidence integration, and adaptive quality assurance, with the support of recent advancements in computer science and systems engineering. Viewed through the lens of an integrated business solution, MEDIA is aimed at emerging operational systems engineering (OSE) applications in service enterprises. Fed with operational evidence, a variety of analysis and management tasks can then be carried out for the practitioners. Aside from an intensive survey of the extensive literature, a case study to illustrate MEDIA application for risk management of health service systems, with interconnected models developed for a generic health care system, will be highlighted in the presentation.