Creating a Globally Distributed Multinational Dialysis Database - The ApolloDialDb Initiative.

Introduction: Large amounts of data are captured during dialysis, yet multinational datasets are scarce because of challenges in harmonizing and integrating clinical data, as well as complying with data protection regulations across the world. A global kidney care provider, Fresenius Medical Care, approached this challenge and finalized the creation of an anonymized dialysis database, coined ApolloDialDb (Apollo). We report on the approach used for database creation and detail dialysis patient characteristics globally. Methods: To create this globally distributed multinational database, data from different electronic clinical systems were extracted, covering routinely collected medical information from dialysis clinics worldwide. This data were harmonized, and then anonymized following a reidentification risk assessment conducted by the external company Privacy Analytics, Ontario, Canada. The data was consolidated and is stored in a central cloud environment and will be updated periodically. Results: Apollo captures data from January 2018 to March 2021 from 40 countries and 543,169 patients worldwide (4.6% in Asia-Pacific [AP], 13.9% in Europe, Middle East, and Africa [EMEA], 7.0% in Latin America [LA], and 74.5% in North America [NA]). It contains demographic data, 35,874,039 laboratory, and 140,016,249 treatment observations as well as frequently recorded medication information, and clinical outcomes (e.g., hospitalization and mortality). Several regional differences can be observed using these data, such as age, treatment modality, and treatment time. Conclusion: Creating a robust multinational dialysis database offers vast opportunities to conduct real-world research and data analytics, including the development of artificial intelligence models. These activities hold promise of advancing the understanding of kidney disease and dialysis therapies. It can serve as comparative resource for the nephrology community.