Clinical trials are drowning in data. In a TransCelerate/Tufts analysis published in late 2025, phase 3 protocols averaged more than 5.9 million data points, up 67% since 2020. The researchers found that nearly one-third of procedures and associated data were non-core or non-essential. In other words, a good portion of this data does not directly support primary objectives or key secondary endpoints.
That complexity matters because every extra field carries a downstream cost: more time for site staff, more queries, more monitoring, more protocol amendments and more burden for participants.
Unwinding that degree of complexity is no easy feat, and too big of a task for any single organization, said Janice Chang, CEO of TransCelerate. “The system is built with friction, fragmentation and systemic barriers.”
TransCelerate, founded in 2012, brings biopharma sponsors together to work on shared clinical trial pain points that single companies struggle to solve alone, Chang said. The nonprofit’s goal is to bring more patients into clinical research by reducing friction in how trials are designed and run.
“My North Star is about bringing clinical research to the patients and meeting them where they are,” Chang said.
“My dream would be that there is no more ‘recruitment.’ Every patient who needs or chooses to participate would be readily enrolled in a trial as part of their care,” she said.
Chang points first to geography. In the U.S., a significant share of counties do not have access to clinical research, she said. County-level analyses in oncology show how sharp that gradient can be. A 2023 ASCO policy analysis, for example, found that about 70% of U.S. counties had no active cancer treatment clinical trial, with availability concentrated around metro areas and academic centers.
Access is only one barrier. “Many people can’t reach trials, or don’t believe trials are designed for them,” Chang said. Sites are also strained, she added, citing statistics suggesting approximately 50% report being overburdened by administrative work such as training, paperwork, and contracts.
In one breast cancer program, Chang said the protocol initially required a second biopsy for screening. After study teams gathered early patient feedback, they learned the procedure was too burdensome and logistically difficult. The protocol was amended to allow archived diagnostic samples instead, reducing burden and improving feasibility.
Technology is often pitched as the answer: decentralized trials, remote monitoring, AI-powered patient matching. But the industry’s ability to deploy those tools lags behind the marketing.
A recent McKinsey report found that while most companies are investing in AI, only 1% describe their gen AI rollouts as mature. “We’re far from mastering it,” Chang said.
Chang framed the technology as disruptive. “Companies face a fairly binary decision: either you lead, or you get disrupted.”
Yet for now, she sees AI mostly supporting existing processes rather than transforming them, but the near-term gains are real. One member company reduced clinical study report drafting from approximately 180 hours to 80 hours using generative AI, she said. She also said some member companies report approximately 10% to 20% recruitment improvement from better patient matching.
But without shared frameworks, those wins stay siloed. “There’s no aligned regulatory guidance, and everyone is experimenting differently,” Chang said. TransCelerate is positioning itself as a neutral convener, bringing sponsors together for vendor-agnostic conversations with regulators. “One regulator told us this is a good space to ‘co-learn,’” she said. “I love that.”
Filed Under: clinical trials, Drug Discovery
