AI catches fire in clinical trials

Artificial intelligence (AI) is having a moment. From writing wedding toasts and even news articles to generating (often unsettling) images of hands, individuals and companies have increasingly embraced the new technology since ChatGPT became a household name.

The healthcare industry also wants in, particularly to improve the clinical trials process.

New AI-enabled software like eClinical Solutions’s Elluminate cloud platform help clients better manage clinical data, monitor trial participants, and analyze results. Some companies, including cancer health tech firm Flatiron Health, have taken it a step further by using AI-enabled software to review electronic medical records (EMRs) and streamline clinical research.

Those capabilities offer a preview of how AI-enabled technologies could shape the future of clinical trials, health tech executives told Healthcare Brew at a November conference in Boston.

“The generative AI movement now—it’s epic technology. It’s unprecedented in terms of what we are seeing. It’s going to transform industries,” eClinical CEO Raj Indupuri said. “Three, four years from now, we’ll look back and will think this moment [is] almost equivalent to how we see the internet in the early 2000s, how we see the mobile revolution—or the iPhone revolution—after the financial crisis, how we see the cloud [in the] last decade.”

AI not only allows for automation but also has the ability to take human-machine interactions “to the next level,” opening up “unbelievable opportunities for [the] life sciences industry,” Indupuri said.

Software companies, he added, will have to embed AI capabilities to stay competitive in the biopharma space. And the adoption of such technologies could come quicker than in the past thanks to industry pressures, like costs.

“Two years from now, we’re not going to be talking about ‘We have AI’—it’s a given. That shift is already happening; it’s kind of expected,” Indupuri said.

That shift comes as the FDA mulls over new rules to streamline and embrace clinical trial-specific digital health technologies, which often cost an average of $700 million to $1 billion.

Flatiron Clinical Research General Manager Alex Deyle offered that AI has the potential to change the way clinical trials are designed so that researchers can focus on “what’s absolutely needed to answer the scientific question.”

“There are so many clinical trials that it’s hard to run all of the clinical trials that the industry is trying to run to generate the evidence we need to know what treatments work for what patients,” he said. “If we can lower the burden of running these clinical trials […] we can just learn a lot faster for more patients in a more integrated way.”

For example, AI-powered software can be used to comb through EHRs to find patients who may be eligible for a clinical trial and prompt doctors to offer participation as part of a regular visit. That process has historically relied on clinical research coordinators manually reviewing records to find patients—something Deyle compared to finding a “needle in a haystack”—or doctors just knowing about clinical trials that are recruiting patients.

Despite the promises of AI, Deyle cautioned that such technology can only be as good as the data underpinning it. Executives should therefore ensure there is some sort of human element to the collection and oversight of health data, he said.

“It has to be a combination of data technology and truly fit-for-purpose human operations,” Deyle said, noting that Flatiron has spent 10 years curating EHR data. “I think a lot of companies are underestimating the need and the hard work that goes into creating good data to train models.”