Certara’s regulatory and medical consultancy, Synchrogenix, introduced the artificial intelligence-enabled solution to meet new data transparency requirements released in January 2015 by the European Medicines Agency (EMA).
The new transparency and disclosure rules, Policy 70, requires clinical study report publication for all successful marketing authorization applications submitted on or after that date.
On March 2, 2016 the EMA published clarifications to Policy 70, which, as Synchrogenix President, Kelley Kendle, told us, expands on the type of clinical trial data to be published to include not just clinical study reports, but also patient narratives and other regulatory documents.
However, in order to be compliant with Policy 70, sponsor companies will be required to redact patient information and confidential company information throughout these documents before publication.
An AI solution
In response to the new regulations, Synchrogenix developed a knowledge-based technology system, which applies learned protocol to the redaction of such confidential information. The AI solution “applies machine learning that can’t be duplicated by humans,” explained Kendle. “Its processing capabilities can handle thousands of pages quickly and with 99% accuracy.”
The system is built on a comprehensive AI engine that requires statistical models to determine the data to be redacted, much as a human would if taught the process, according to Kendle. According to the company, Synchrogenix’s technology is the only AI-enabled solution in the biopharmaceutical industry.
“We provide the system with general rules, as identified in the sponsor’s redaction requirements document. The system is then trained on those specific, redaction requirements and learns specifically what needs to be redacted,” she explained.
However, unlike human redaction, the system never needs to be retrained or replaced. Additionally, “the system is not limited in its capabilities and is able to redact hundreds of studies per day, consistently, accurately, and efficiently,” said Kendle.
Yet the system’s development has not been without its challenges, the greatest of which has been “balancing the application of the engine to the forthcoming regulatory requirements, while meeting the ongoing needs of trailblazing sponsor companies,” explained Kendle. “We recognized that there would be pent-up demand as companies that had a post-January market approval would need to play ‘catch up’ to meet the new timelines.”
According to the company, the technology has already been applied at both large pharma and smaller biotech organizations.