Matching patients to clinical trials they are eligible for has been a major challenge for everyone from pharmaceutical companies to hospitals.

THE PROBLEM

Finding the right matches based on certain health records and demographic eligibility is difficult and time-consuming. It costs a lot of money, can slow clinical trials down and is one of the major reasons trials fail. This slows down progress in advancing new medical treatments and advancing health outcomes.

"On the flip side, patients want to be involved in trials, but often can't find the right ones," said Sean T. Manion, chief scientific officer at New York City-based blockchain and machine learning vendor ConsenSys Health.

"Nearly 50% of cancer patients want to be involved in clinical trials, yet only 5% are able to do so," he said. "This limited access to groundbreaking treatments and being part of the system of science to improve treatments for future patients is the other side of the problem."

One of the major reasons clinical trial-matching and recruitment is so slow is that it's largely limited by laws and regulations related to patient privacy.

PROPOSAL

"At ConsenSys Health, we are focused on a new way of approaching these concerns using a federated approach, where the data stays safely and privately at rest and only answers to questions about that data are shared," said Manion. "This is made possible by the combination of three families of technology: blockchain, decentralized AI, and privacy-preserving software and hardware.

"We are working with Intel to align their next generation of Intel SGX hardware for confidential computing with our blockchain-orchestrated federated learning system, Elevated Compute, to solve a variety of problems in healthcare and life sciences."

For clinical trial-matching, the ability to identify sites and individuals that might be good choices for a clinical trial is made possible through this approach, including a validated audit trail of the data via blockchain and privacy preservation with Intel SGX hardware.

"We are working with Intel to align their next generation of Intel SGX hardware for confidential computing with our blockchain-orchestrated federated learning system, Elevated Compute, to solve a variety of problems in healthcare and life sciences."

Sean T. Manion, ConsenSys Health

This will enable faster and less-expensive trials, with more access and inclusion for patients, all resulting in a more robust advancement of new treatments and improved health outcomes, he noted.

"We were using Intel SGX hardware to simulate the privacy-preserving attributes for comparing to eligibility for patients for clinical trials, allowing deployed algorithms and analytics sent to the data to do the necessary analysis of the data in the trusted enclave the Intel SGX hardware creates, without ever moving and exposing the data," he explained.

"This served to demonstrate the viability of this privacy preservation for future use in safely applying our blockchain-orchestrated federated learning solution to clinical trial-matching and other future-state use cases like data analysis for multisite trials without centralizing the data," he added.

MEETING THE CHALLENGE

ConsenSys Health worked directly with the Intel SGX team using synthetic health data to run this proof of concept.

"We compared their latest version of Intel SGX in Ice Lake to their earlier version," Manion said. "There has been a significant increase in the enclave size – the amount of data that can be processed at once – in the new version, allowing for faster and more data-processing in a privacy-preserving way. We also worked with a large pharmaceutical company to validate the problem area and solution design."

RESULTS

ConsenSys Health learned that this privacy-preserving approach using the next generation of Intel SGX hardware was much faster, while maintaining the necessary privacy preservation.

This confirmed that it can be used not only for significant improvement in clinical trial-matching, but also a host of other potential use cases relating to clinical trials that can result in increased speed and decreased cost, without sacrificing patient privacy, by using a federated approach.

ADVICE FOR OTHERS

"I would recommend looking at blockchain, federated learning and other decentralized technologies, as they are quickly becoming the future trend for clinical trials and more," Manion advised. "The current pandemic has moved up the timeline for interest and adoption.

"Organizations like the newly founded 125-member Decentralized Trials & Research Alliance are quickly advancing the dialogue of an overall decentralized approach to research," he continued. "Not only are these technologies going to play a role, and building capacity and knowledge now will help organizations prepare, but the solutions will be built around the standards and workflow of those doing the early work."

Getting to the table now is the alternative to adjusting to someone else's standards in a few years to avoid falling behind or failing, he concluded.

Twitter: @SiwickiHealthIT
Email the writer: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.