Study Identifies Strategic Decisions and Poor Recruitment as Key Causes of HNSCC Trial Failures

A study published in JAMA Otolaryngology – Head & Neck Surgery identified 346 trial failures for the head and neck squamous cell carcinoma (HNSCC) indication. Most clinical trial failures are strategic, with the majority of phase 1, industry-sponsored, and immune/targeted therapy trials subject to early termination. Recruitment challenges emerged as another leading cause for trial termination, encompassing many later-phase, non-industry-sponsored trials.

The leading reason for failure in the early-terminated trial group were strategic decisions—non-scientific sponsor-driven decisions to end the trial prematurely. Strategic decisions were more frequently the reason for failure in phase 1, industry-sponsored, and trials investigating immunotherapy and targeted therapy. Poor recruitment was a more common reason for failure in later phase trials, non-industry-sponsored trials, and trials investigating chemotherapy, radiation, chemoradiation, combination treatments, and supportive care. The study also identified an increasing rate of trial termination/withdrawal. A key protective factor was increased log-transformed actual enrollment, and industry funding is a risk factor for trial termination/withdrawal.

Overly restrictive eligibility criteria may be one explanation for the trial failures attributed to poor recruitment. It is possible to incorporate safeguards to reduce the frequency of operational failures through stronger site selection, better data monitoring practices, and other methodological considerations.

Clinical trial delays often originate in early site selection decisions, where misalignment between protocol demands and site capabilities undermines startup, enrollment, and data quality despite later efforts to correct course. Capable, experienced sites are completing feasibility, expressing interest, and then hearing nothing. At the same time, studies move forward with sites that struggle to activate, enroll, or maintain consistency. As protocols become more complex, the margin for misalignment narrows. Studies now frequently involve biomarker requirements, complex dosing schedules, hybrid or decentralized elements, and increased data collection demands. Feasibility often functions as a high-volume, time-constrained process that does not fully capture how a site operates. Sites are asked to complete detailed questionnaires under tight timelines, often across multiple studies at once. The questions are standardized, designed for scalability rather than nuance. Feasibility is rarely a two-way exchange. Sites provide information but receive limited feedback. There is little visibility into how responses are interpreted or why decisions are made.

The gap between feasibility responses and real-world execution is where many studies begin to drift off course. Protocol expectations are often developed without a clear understanding of how they will translate into site workflows. Feasibility responses may not reflect a site’s true capacity. When feasibility and site selection are misaligned, startup timelines lengthen, enrollment becomes unpredictable, protocol deviations increase, and monitoring and oversight demands grow, adding cost and pressure across the study. Data quality can also be affected.

Many trials are already struggling long before enrollment begins to fail. Operational cracks often appear quietly at the beginning. A protocol may be scientifically strong but operationally unrealistic. Study teams inherit timelines disconnected from clinical workflow realities. Communication becomes fragmented across sponsors, contract research organizations, investigators, coordinators, and regulatory teams. Small inefficiencies accumulate long before recruitment numbers visibly decline. By the time enrollment visibly struggles, deeper structural problems may have existed for months. Modern clinical research systems often rely on unsustainable human adaptation to compensate for operational complexity. Much of that invisible labor never appears in publications, dashboards, or executive summaries.