Clinical trial data company Phesi’s new analysis found that fewer than one in three trial protocols are connected to documented patient data and outcomes, highlighting the risk that flawed decision-making may be repeated in clinical development.
In its assessment of 600,000 clinical trial protocols, Phesi found that 29.3% of protocols are linked to patient data and outcomes, according to a May 13 release. Without a connection between trial design and what happens to patients, trial design mistakes may be repeated.
Oncology trials, in general, and breast cancer trials specifically showed similar results. The report found that just 30.9% of 116,746 oncology trials are linked to usable patient data. Only 31.2% of 15,977 trials for patients with breast cancer, the world’s most studied disease, are linked to actual patient data.
Phesi President and Founder Dr. Gen Li noted that many protocols are designed based on existing trials, so when current trials are disconnected from their target patient populations, issues are more likely to be repeated.
He pointed to the use of AI in clinical trial design as a contributing factor. “In the past, protocol writers could be selective about the protocols they used and apply human judgment to the connection between design and the target patient population,” he said in a statement. “AI can ingest far more historical templates, but without the right logic or judgment, it may fail to make that connection… In essence, flaws are being scaled, not solved."
Trials fail to link protocols to patient data for numerous reasons, according to the report. Some trial designs restrict the ability to enroll patients and collect data, so results are never reported. When there is no connection between the intent of the protocol and what happens to patients, it is harder to understand which designs work, which fail and why.
The lack of connection puts trial sponsors that rely too heavily on AI at risk of reusing trial designs that faced recruitment issues or failed studies, Phesi says. Trial failures can also have a significant negative financial impact on sponsors.
Li emphasized that datasets must account for the full patient population rather than a late-stage subset. “There is [a] huge opportunity for AI to optimize clinical development, but only when the platforms being used as the basis for AI can identify protocols with patients and outcomes reported,” he said.