New research has shown widespread conflicts of interest (COI) in oncology clinical trials, particularly in US-led studies and among leading authors, raising issues about the influence of biopharma companies on investigators and scientific independence.

The investigators said their findings should prompt “…serious reflection not only by journals and regulators but also by researchers and academic institutions, who have a responsibility to move beyond simple disclosure toward more proactive COI prevention strategies.”

The study, published in Ethics in Cancer Care [link here], used AI to analyse more than 2,000 clinical trials published in the Journal of Oncology between January 2010 and January 2025.

The analysis was further divided into three periods based on date of publication of trials: January 2010-January 2015, January 2015-January 2020, and January 2020-January 2025, in order to ascertain COI trends over time.

The data extracted was the main medical products under investigation, the responsible biopharmaceutical companies, author lists, their affiliations and disclosures.

Bevacizumab in 2010-2015 then pembrolizumab across the next two time periods were the most studied medical products.

The analysis found that overall 72.6% of clinical trials had at least one author with a company-related COI – 70.0% in the earliest time period, 77.0% in the middle time period and 72.0% in the most recent time period.

Specific COIs followed similar trends:

• company employment was reported in 31.3%, 38.3%, and 36.2% of studies across the three periods
• consulting or advisory roles increased from 52.5% to 61.5%, and then slightly decreased to 59.1%.
• honoraria recipients rose from 43.8% to 49.4%, before a slight decline to 48.8%.

The study said the temporal trends in COI prevalence, plateauing then declining, may indicate that industry-academia collaboration has reached a steady-state dynamic, and that the initial impact of transparency initiatives has reached a point of saturation.

Speakers’ bureau roles, not disclosed before March 2015, were reported in 26.0% (2015-2020) and 30.9% (2020-2025), indicating increased disclosure over time.

Among studies with COIs, a median of 35.3% of authors in each study had COIs linked to the company, with period-specific values of 31.2%, 36.4%, and 37.5%.

It also found that across the three study periods, 75.2%, 82.8%, and 76.0% of US-led studies reported COIs.

“Compared with the rest of the world, US-led studies were significantly more likely to have COIs (77.6% v 67.3%;P < 0.001), particularly those related to employment (38.9% v 31.0%; P < 0.001) and advisory roles (61.2% v 52.8%; P < 0.001),” it said.

“In contrast, studies from the rest of the world more frequently had authors who received honoraria from the company (50.9% v 43.0%; P < 0.001,”.

It also found that among the first or last authors 61.9% had a COI with the associated biopharma company, with increasing rates of 60.4%, 62.5%, and 63.3% across the three periods.

Company employment among first or last authors declined from 28.9% to 4.6% and 4.4% over time; COIs from advisory or consulting roles increased from 18.4% to 41.3% and 46.4%; and COIs from honoraria were 5.3%, 25.0%, and 22.8%, for the three periods.

The investigators, from Ohio State University in the US, said the higher prevalence of COIs in US-led studies compared with those from other regions may reflect “…differences in funding models, regulatory environments, and cultural norms surrounding industry-academia collaboration.”

They said their findings on disclosed COIs suggest a substantial proportion of contributors to cancer trials may be subject to potential financial or professional influences.

“The observation that a substantial proportion of first or last authors have COIs with associated biopharma companies is particularly significant, especially given that these rates have continued to increase steadily, rather than plateauing or declining as seen with general COIs.”

They said the decline in company employment among leading authors over time, coupled with the increase in advisory/consulting roles and honoraria, may reflect a shift toward less direct, but still potentially influential, forms of financial relationships.

“These findings underscore the need for future research examining whether financial COIs among first and last authors are associated with trial outcomes or patterns of data interpretation, particularly given their critical role in shaping the interpretation and dissemination of clinical trial findings,” they said.

“Ensuring robust academic oversight remains essential to preserve scientific independence and credibility.”