Merck Partners with Atropos Health to Accelerate Real-World Evidence Generation
Atropos Health has announced a collaboration with Merck to utilize its suite of tools for real-world evidence (RWE) generation, including GENEVA OS, Green Button, and the Atropos Evidence Network. The partnership focuses on rapid cohort creation, advanced analytics, and generating publication-grade RWE within 48 hours.
As part of the collaboration, Merck’s data science teams will work with Atropos Health to replicate studies, produce insights from real-world data (RWD), and address critical evidence gaps for life-saving treatments. The federated Atropos Evidence Network includes over 300 million patient records, offering datasets evaluated for quality through its Real World Fitness Score.
Dr. Brigham Hyde, CEO and co-founder of Atropos Health, said:
“Decisions based on RWE are becoming a reality, and the speed of medicine requires quality data to be available for translation into accurate RWE rapidly. This collaboration accelerates the rate at which insights for decision-making are presented and acted upon, improving the pace of innovation that ultimately improves patient care.”
Megan O’Brien, Associate Vice President of Real World Evidence Capabilities at Merck, added:
“We see the importance of real-world data, evidence, and insights to ensure safe and effective use of life-saving treatments. Atropos Health will enable rapid conversion of data into evidence using multiple assets to ensure feasibility and relevance and validate evidence for accuracy and increased confidence.”
The announcement comes at a time when the pharmaceutical industry is grappling with challenges in data acquisition and representation. There is ongoing discussion that the bottleneck in leveraging AI for drug discovery and evidence generation is not primarily algorithmic sophistication but rather the availability, quality, and integration of domain-specific data.
While pharmaceutical companies have access to vast amounts of RWD, issues such as inconsistent data formats, limited interoperability between systems, and incomplete datasets hinder effective utilization. The integration of data from diverse sources, including clinical trials, electronic health records, and registries, remains a big challenge. Even when data is abundant, gaps in representation or alignment between datasets can reduce the utility of AI models and analytic frameworks.
Additionally, the need for more precise data representation is being increasingly highlighted. For instance, variability in how datasets are processed or differences in experimental protocols can make it difficult to integrate results from multiple studies or predict outcomes in real-world settings. Addressing these issues requires not only sophisticated tools but also novel approaches to data collection and curation, which can align disparate datasets into a unified structure suitable for advanced analytics.
