COLUMBUS, Ohio, March 27 (UPI) -- Researchers reviewing 75 group-randomized U.S. cancer trials found fewer than half used appropriate statistical methods to analyze the results.
The Ohio State University-led review suggests some trials conducted during the five-year period studied might have reported interventions to prevent disease or reduce cancer risks were effective when, in fact, they might not have been.
Professor David Murral, the lead author, said more than a third of the cancer trials contained statistical analysis the reviewers considered inappropriate to assess the effects of an intervention being studied. Eight-eight percent of those studies reported statistically significant intervention effects that, because of analysis flaws, could be misleading to scientists and policymakers, researchers said.
"We cannot say any specific studies are wrong," said Murray. "We can say that the analysis used in many of the papers suggests that some of them probably were overstating the significance of their findings."
Murray and his colleagues call for investigators to collaborate with statisticians familiar with group-randomized study methods and for funding agencies and journal editors to ensure such studies show evidence of proper design planning and data analysis.
The review appears online in the Journal of the National Cancer Institute.
