There’s no shortage of healthcare data in the world, and the amount of data being collected increases exponentially by the day.  But at the same time, AI innovators continue to struggle to get the access they need to “fresh” data to train new algorithms.

 The solution, says one stakeholder, is federated learning.

In a recent commentary, Yuval Baror, co-founder and CTO of Rhino Health, an AI platform developer, pointed to three challenges in health IT development overall, and AI development in particular.  

First, the aforementioned insufficient access to “real-world data” is hampering the development of advanced solutions.  Next, the reality of “disparate data sets and systems across multiple sites and systems” is an impediment to the validation of new AI-based solutions.  And, three, the maintenance over time of AI-based solutions requires access to large amounts of data on an ongoing basis, and that access is not readily forthcoming.

Given that the core stumbling block in play, here, revolves around concerns over patient data privacy, Baror points to “federated learning” as the likely best solution.

“It allows connecting data from different data silos while not requiring any movement of patient data,” he explains. “No huge files to transfer and store. And, most importantly, no risk to patient privacy.”

Digging in a bit further, Baror notes that “(d)ata remains with the hospital where it was created. Copies of the AI model are sent to each hospital, and training is performed at each hospital with its local data. Only aggregate information is shared back and used to create a global model. This makes it possible for healthcare AI developers to utilize data across hospitals and health systems without those care providers ever moving data, transferring ownership, or risking patient privacy.”

Moreover, he says, federated learning “makes it much easier for AI developers to access larger, more diverse data sets that better represent today’s real-world patient population. This will make models more robust for different patient populations, enabling AI-based healthcare solutions to scale globally at an unprecedented pace.”

Of course, federated learning by itself doesn’t constitute a silver bullet.  

Indeed, says Baror, “training a model once on a diverse data set is not enough. The world is constantly changing, and AI models must be able to adapt and evolve with these changes. AI developers need access to a continuously updating data set, reflecting the changes to patient populations and treatment protocols. They also need a way to evolve their AI models over time to adapt to these changes and to deploy their updated models to production environments in a safe way, in line with regulations.”

The starting point, however, is to “break down” those data silos, and for that, says Baror, federated learning is the way forward.

Photo by Peter Howell/Getty Images