The use of physicians’ or investigators’ choice of treatment in oncology randomized clinical trials (RCTs) may mask the limited choices actually available in control groups, according to research published in JAMA Network Open.

The researchers noted that RCTs may be limited if patients in control groups do not receive standard care. Allowing physicians to choose the control therapy may overcome this limitation. However, there is concern that the set of physicians’ choices may be too restricted to yield an overall optimal control group.

For the current study, researchers evaluated RCTs with physicians’ choice control groups to determine whether the choice in the relevant trial was restricted or unrestricted. All data used were publicly available.

Continue Reading

The researchers analyzed data from 92 RCTs published between 2007 and 2021. The most common disease subtypes in these trials were breast cancer (23.9%), non-small cell lung cancer (8.7%), and ovarian cancer (8.7%).

Of the 92 RCTs, 78 had restrictions on physicians’ choice of control therapy, and 14 had no restrictions.

A majority of the RCTs (n=82) were sponsored by industry. Industry sponsorship was observed in 91% of studies with a restricted physicians’ choice and 78.6% of studies with unrestricted physicians’ choice.

“Through imprecise wording, potentially masking substandard control group, treating physicians may inaccurately think that the reported results can be generalized to their patients, whereas this may not be true,” the study authors wrote. “Our findings suggest that editors and regulators should demand clarification in the use of these terms within RCTs protocols and reports.”

Disclosures: This research was supported by Arnold Ventures, LLC. Some study authors declared affiliations with various companies. Please see the original reference for a full list of disclosures. 


Olivier T, Haslam A, Prasad V. Reporting of physicians’ or investigators’ choice of treatment in oncology randomized clinical trials. JAMA Netw Open. 2022;5(1):e2144770. doi:10.1001/jamanetworkopen.2021.44770