Clinerion launches its Federated Machine Learning Platform, a global, federated machine learning (ML) platform on its Patient Network Explorer infrastructure, incorporating its global network of partner hospitals. This new platform will enable sophisticated analytics and pattern recognition use cases on real-world data (RWD) across the Patient Network Explorer network, ultimately offering better diagnosis of patients.
The Clinerion Federated Machine Learning Platform allows the development of advanced statistics and machine learning (ML) algorithms on data at local site-nodes in Clinerion’s global network of de-identified electronic health records (EHRs). Training and validating machine learning models require large and varied data sets for increased accuracy. Clinerion’s platform, covering 425 M patients in 24 countries, can uniquely provide the required scale of data for such advanced use cases.
Using the Clinerion Federated Machine Learning Platform, users can train models on local data to identify pattern in local patient cohorts, and further optimize models by using data from multiple sites. For RWD, it can be used to develop detailed models based upon all possible heterogenous disease presentations and novel predictive biomarkers, enabling the prediction of high risk in patients before they reach critical state, as well as the generation of disease phenotype sets for the identification of undiagnosed diseases, e.g., to find rare disease patients.
Clinerion’s new platform facilitates federated learning, learning across site-nodes and model deployment to sites, enabling direct, anonymized data metrics and per site / per country patient population insights through advanced analytics. For countries where data access is strictly protected, Clinerion’s Federated Machine Learning Platform enables safe and secure setup of analytics, as the patient data remains under the control of the hospital and never leaves the hospital infrastructure. The platform offers access to multiple dimensions of RWD from hospitals and trial sites around the world, incl. demographics, diagnoses, procedures, medications and laboratory results, longitudinally, through time. This allows actionable insights into the patient journey, as well as access to diverse population data to tackle longstanding health inequities and increase the diversity and inclusion of racial and ethnic minorities in clinical trials.
“This platform builds on Clinerion’s patented technology for distributed query to a hybrid cloud-and-local-node infrastructure,” says Andreas Walter, Clinerion Chief Technology Officer, “which preserves patient privacy at both the local and federated levels. We have years of experience under our belt in how to maintain and leverage a distributed health network, for performance, security and speed – due to this history, the Federated Machine Learning Platform comes out of the gate running.”
“While the Federated Machine Learning Platform supports clear use cases, I am also happy that it also brings innovation to many important initiatives and trends within healthcare,” says Barış Erdoğan, Clinerion CEO. “Personalized medicine, diversity and inclusion, and rare disease treatment will all benefit from this innovative new technology platform.”
Clinerion has developed this technology platform with Innosuisse – the Swiss Innovation Agency, the Bern University of Applied Sciences and the University of Zürich.
