Clinical research is increasingly relying on information gathered and managed in different database systems and institutions. Distributed data collection and management processes in such settings can be extremely complex and lead to a range of issues involving the integrity and accuracy of the distributed data. To address this challenge, we propose a middleware framework for assessing the data integrity and correctness in federated environments. The framework has two main elements: (1) a test model describing the dependencies between and constraints on data sources and datasets, and (2) a family of testing techniques that create and execute test cases based on the model.