Secondary use of data for research purposes is becoming increasingly important to improve medical research and thus patient care. In this context, linking different data sources provides a unique opportunity to gain a comprehensive overview of a patient's medical history. The interoperability of different data sets can be ensured by using the standardized data model OMOP. However, the record linkage of separate OMOP databases is an open problem. In this paper, we present an approach to link disparate datasets in OMOP using German clinical data and claims data as an example. We highlight three challenges and propose solutions for them to ensure accurate and efficient data integration.