Better with time: Randomized controlled trials on the rise in rad research, but still room for improvement

Both the quality and quantity of radiologic randomized controlled trials has improved dramatically over the past 20 years, according to results of a new study published online in the American Journal of Roentgenology.

Randomized controlled trials, which significantly limit the possibility of bias in reported results, are considered to be among the most reliable methods to research new medical interventions and treatments.

That’s not necessarily the case when it comes to radiology research, said lead author Su Jin Hong, MD, and his colleagues from Kangdong Seong-Sim Hospital in Seoul, South Korea.

“Unlike other clinical specialties, radiologic [randomized controlled trials] involving diagnostic imaging face additional challenges with regard to feasibility, acceptability, and standardization,” they wrote. “Furthermore, methodologic and financial concerns often make it difficult to conduct [randomized controlled trials] in the field of radiology.”

Hong and his team conducted a bibliometric study to assess the characteristics and quality of radiologic randomized controlled trials over the past 20 years. To do so, they performed a search of the PubMed database to identify all radiologic randomized controlled trials published from 1995-2014 and extracted information including the specific journal, radiologic subspecialty, imaging technique, number of subjects, study result, funding, number of authors, number of institutions and country of origin.

Each randomized trial was scored for methodologic quality by applying the Jadad scale, a scoring tool ranging from 0-5 according to levels of randomization, double-blinding, and withdrawals and dropouts. When a trial’s total score was 3-5 points, the study was categorized as high quality; when the total score was 0-2 points, it was categorized as low quality.

Their results showed significant increases in the number of radiologic randomized controlled trials published during the study time period, from only 43 such trials conducted between 1995-1999 to 172 conducted from 2010–2014. Randomization was appropriate in 46 percent of all trials, double-blinding was adequate in 20 percent of articles, and appropriate reasons for withdrawals and dropouts were described in 64 percent of studies overall.

“The quantity and quality of radiologic [randomized trials] have significantly increased over the past two decades; however, the methodologic quality remains suboptimal,” the authors concluded. “There were significant improvements during the study periods in the reporting of appropriate randomization methods, reporting of withdrawals and dropouts, number of studies using inappropriate randomization methods, and overall quality.”