Study Finds Fabricated Data in Up to One Quarter of Clinical Trials

A 2020 analysis by John Carlisle examined more than 500 randomized controlled trials submitted to the journal Anaesthesia. Carlisle, an anaesthetist with England’s National Health Service and an editor at the journal, reviewed individual participant data for 150 of the studies.

He determined that 26 percent of those trials contained problems so extensive that the results could not be trusted. Carlisle labeled these studies “zombie” trials because they appeared legitimate but lacked reliable data.

Carlisle lacked access to raw participant data, he examined only summary tables. In those cases, just 1 percent of trials were judged untrustworthy. The difference in detection rates prompted Carlisle to recommend that journal editors request individual participant data before publishing randomized controlled trials.

Other researchers have reported similar issues in fields including women’s health, bone health, and COVID-19 studies. Ian Roberts, an epidemiologist at the London School of Hygiene & Tropical Medicine, stated that about one third of randomized trials on some topics may be fabricated.

Ben Mol of Monash University estimated that 20 to 30 percent of randomized trials included in systematic reviews of women’s health are suspect. A 2022 Cochrane review found that more than 40 percent of randomized trials on ivermectin for COVID-19 were untrustworthy.

Unreliable trials can be incorporated into systematic reviews and clinical guidelines that physicians use to decide treatments. Alison Avenell of the University of Aberdeen noted that work by one retracted researcher had been cited in 88 systematic reviews and guidelines.

Žarko Alfirević of the University of Liverpool said additional studies are underway this year to measure the extent of the problem more precisely.