“We’re in a Third Wave”: How Tech Is Innovating Clinical Trials
Clinical trials are the heart of the healthcare industry, the portal to new drugs and therapies. But they’re heavily regulated, lengthy and expensive. That means they’re ripe for disruption, with high hopes for artificial intelligence.
- Medical trials lie at the heart of the healthcare industry, with novel drugs forced to go through three stages of testing.
Take a peek into your medicine cabinet and you’ll find boxes of tablets, syrups or balms. All of these drugs — every single one — began in the same place: a clinical trial.
As drug incubators, clinical trials are highly regulated and conducted in three separate phases, seeking to identify effective and safe treatments for patient care. Globally, over 430,000 trials were recorded in 2022 alone.
Trials are expensive, costing tens of millions of dollars to shepherd a drug through all three phases, and lengthy, lasting up to 10 or even 15 years. Fewer than 10 percent of candidates make it through all the stages. Trials are also complex to manage and the risk of human error hangs large. It’s hit or miss stuff.
That’s where new technology comes in. New tools are working to improve clinical trials to greenlight them faster. The space is increasingly buzzy, with startups like Paradigm raising over $200m. There are also entire funds now dedicated to pharmaceutical research and development (R&D), like Switzerland’s Debiopharm.
For some, the ultimate vision is for clinical trials to be totally simulated. Instead of testing on humans, predictive machines may one day be able to safeguard drugs based on enormous patient data sets.
While that’s still some way off, Mark Braganza, a partner in EQT’s Healthcare team, says we are entering “a third wave” of clinical trial optimization. The initial waves focused on moving from paper records to digital data capture and then digital data collection; now, we’re seeing tools that can compute and analyze that data.
“Pretty much everything now is in some electronic format. So there’s a huge revolution going on with people taking those data sets and working with them,” explains Braganza.
The previous waves bolstered reliability, but failed to make a real dent in drug approval times; arguably, they just moved the bottleneck. The hope is this latest tech wave will be different, by cleaning and reviewing the data, a process that is still very manual and error-prone.
Even shaving three months off the length of a medical trial could be a huge advantage, Braganza adds. Switzerland’s Risklick goes further, predicting that they could take a year off the drug-to-market timeline.
Challenge and opportunity
Big pharma can reap huge profits once drugs are approved and give patients faster access to life-saving care, so speed is everything — making this a lucrative market.
There are four key areas for streamlining clinical trials:
- Analysis of participants’ data, helping quickly identify anomalies.
- Patient recruitment and matching.
- Patient monitoring, including management platforms.
- Regulatory compliance, such as automating paperwork and standardization.
Startups in this space, therefore, have a huge opportunity, but industry scepticism is a challenge, says Morgan Hanger, who heads up the U.S. Clinical Trials Transformation Initiative.
“The biggest hurdle here is that human and governance factor, and that’s exacerbated at some of the big pharma companies… It can be onerous to test all these interventions,” she says. One pharma company she knows is “test tired” after going through 100 new initiatives to optimize trials.
While the potential impact of AI in clinical trials is enormous, real adoption is still in its infancy, adds Kirk Lepke, another EQT partner.
“Everyone is talking about AI in clinical trials. And while the long-term trend is very promising, it will be a journey because we are dealing with live patients in a regulated environment,” he says, noting that — for now — advanced machine learning models are what we are seeing applied at scale in real-world trials.
“It will likely be harder for a brand new player to suddenly break-in,” argues Lepke. “The companies who have already built the trust, have access to trial data and a deep understanding of the vertical, will likely be best positioned to leverage AI.”
The CluePoints approach
CluePoints is one such firm. The EQT-backed company has de-risked over 1,600 studies in the last decade, testing for trial fraud and data anomalies. Its tech is even used by the U.S. Food and Drug Administration to proof-test trial submissions. CluePoints is going further now with machine learning tools that can understand medical terms, and ultimately identify adverse patient outcomes.
“We’re not making bold claims about how much we can reduce timelines. But what we do do is make sure that the data you capture is of the highest quality so that when you submit it to the drug regulator, you don’t get questions back,” says CluePoints CEO Andy Cooper.
“The last thing you want is for the regulator to find the problems you should have found before you submitted it.”
It’s not just in the trial phase that this technology offers promise. AI is also being tested to accelerate the development, discovery and design process, before a drug gets to trial. Companies like Norstella, for instance, use AI to quickly sift through old trials to inform protocol design. It’s also led to scientific breakthroughs, like predicting how molecules and proteins will interact, resulting in AI-engineered drug candidates.
Founders say the ultimate metrics of success will be clinical success rates climbing and the pace of getting drugs to market accelerating.
“In this regulated space, people don’t trust the technology yet to do everything,” says Braganza. “But what they trust it to do is find everything that’s relevant so you can make an assessment. It’s helping you assess that first piece.”
ThinQ is the must-bookmark publication for the thinking investor.
