Study criticizes journals for leaving out trial details

JOSEPH S. BAILES, MD

Editor

Randomized controlled trials published in major oncology journals do not consistently report essential therapeutic details necessary for translation to clinical practice, according to a new study by Florida researchers. Based on their findings, the authors are calling for improved reporting from all therapy-based trials to include specific details about performing adjustments for hematologic and organ-specific toxicity.

"We're just taking it to the next logical step, which is, 'how do we apply these results to the masses of patients who need to benefit from scientific progress?"

— THOMAS GEORGE, JR., MD

"We were frustrated that we were trying to adjust the therapeutics to the side effects that were occurring in the care of patients and we realized everyone had different styles," said lead author Thomas George, Jr., MD, director of the University of Florida's gastrointestinal oncology program. "For example, a clinician might reduce the dose of chemotherapy if neutropenia was detected, or add growth factors or do some combination thereof—or do nothing. So we would go to the article that established the original regimen to get an answer. We found there was so much variability among providers we could not consistently find the answers."

He said randomized clinical trials improve clinical care via evidence-based results. He noted that guidelines exist for reporting results from randomized clinical trials, but often lack specific details needed for clinical translation. Dr. George and his colleagues identified 10 essential elements of randomized clinical trial reporting, including drug name, dose, route, cycle length, maximum number of cycles, premedication growth factor support, patient monitoring parameters, and dosing adjustments for various toxicities.

The investigators analyzed all therapy-based oncology randomized clinical trials published between 2005 and 2008 in five journals (see Table).

They found that premedication, growth factor support and dose adjustments for toxicities were reported in less than half of the articles and only 30 articles (11%) met the main objective of complete data reporting (all 10 essential elements).

The highest percentages of complete data reporting were found in JNCI (25% of the time; 5/20 articles), followed by Cancer (18%; 5/27 articles), JCO (11%, 18/165 articles), Blood (5%; 1/19 articles); and NEJM (3%; 1/31 articles).

"It just boils down to willpower on the part of the journal editors to agree that this is an important issue," Dr. George said. "I think the scientific community, the publishers, the editors, and even the investigators who conduct the studies have been appropriately focused on justifying the scientific methods and merit of the study. We're just taking it to the next logical step, which is, 'how do we apply these results to the masses of patients who need to benefit from scientific progress?'"

The researchers found that the dose of the drug was almost always reported. However, only 43% of the papers reported what kind of premedication was necessary and only 42% reported how they adjusted dosages if the therapy proved toxic (J Natl Cancer Inst 102:702-705, 2010).