New analysis reveals billions risked on overly complex trials that recruit from oversaturated investigator networks

Sponsors face mounting risks of late-stage trial failure across multiple indications by continuing to design overly complex studies without incorporating real-world data, according to new insights from Phesi.

The company’s latest analysis, using its AI-powered Trial Accelerator platform, warns that Phase 3 programs are under threat as investigator sites become increasingly oversaturated and patient burden climbs.

The report zeroes in on atopic dermatitis (AD) as a case in point. As of June 2024, AD ranked second only to non-small cell lung cancer in number of active Phase 3 trials, with 18 industry-sponsored studies underway. Despite AD affecting more than 200 million people globally, the number of active recruiting sites has surged 205% in four years — from 1,944 in July 2021 to 3,981 in June 2025 — stretching global capacity and placing pressure on a limited number of qualified investigators.

“Phase 3 trials are high-stakes, resource-heavy investments. Sponsors are under pressure to deliver therapies faster, but without real-world data and AI-driven trial design, many are setting themselves up for failure,” said Dr Gen Li, CEO and Founder of Phesi. “In a competitive space like AD, being first to market is critical. That means designing smarter trials — minimizing recruitment friction and placing patients at the center of study design.”

One striking example from the report highlights a single investigator in Fountain Valley, California — a city of just 55,000 — currently involved in 14 active AD trials, raising clear concerns over feasibility and data quality.

To support more efficient design, Phesi also unveiled a Digital Patient Profile (DPP) for adult AD, highlighting a near-perfect correlation between EASI (Eczema Area and Severity Index) and %BSA (body surface area affected). While EASI is widely used, it is a complex, multi-factor metric — whereas %BSA offers a simpler, patient-friendly measure that could ease trial burden without compromising scientific rigor.

“This is not just an AD issue — it’s endemic across indications, including rare diseases,” added Jonathan Peachey, Phesi’s COO.

“Sponsors need to rethink how they’re approaching feasibility and site selection. The potential of AI — from predictive modelling to digital twins — is enormous, but underused. Now’s the time to evolve.”