PURPOSE:  While patient-level healthcare databases continue to proliferate globally, the challenge to any researcher is how to find the optimal database(s) to solve the research question at hand. Electronic health records (EHR), disease registries, claims, and patient-reported surveys each have their strengths and weaknesses. Oftentimes, the ideal solution is a combination of fields from one or more databases. The integration of these databases can present significant challenges around sample size, patient privacy, and mixed data types; however, if combined carefully and accurately, it can provide huge benefits in understanding more about a disease or patient characteristics. This workshop will address different examples of successful integration of patient-level health care databases in a variety of applications. DESCRIPTION:  The first topic will be on the current state of real-world evidence (RWE) databases and the typical ways in which these datasets are used in health outcomes research, including the strengths and limitations of each dataset. Different approaches for integrating datasets will also be discussed. The second topic will present a novel approach to combining disparate data sources. Topic 3 will demonstrate examples of how different data sources are integrated for use in health outcomes research. For each case study, the presenter will discuss the reason why database integration was necessary, some of the challenges faced performing the integration, how they overcame these challenges, and the value that the final integrated data brought to the research question at hand. The workshop will also encourage audience participation around successes or failures in health care data integration they have experienced and the methodologies they have utilized. The discussion leaders will provide their thoughts on these examples from the audience. This workshop will be valuable to healthcare researchers and biostatisticians working with different RWE databases and are seeking novel approaches to creating a more complete dataset.