Fewer than half of clinical trials that receive funding from the National Institutes of Health (NIH) are published 30 months after completion, new analysis shows.

Researchers from Yale School of Medicine sought to investigate the prevalence of scientific evidence remaining unpublished in a peer-reviewed, MEDLINE-indexed biomedical journal over two years after the results have been obtained.

Using data provided by the trial registry ClinicalTrials.gov, the team cross referenced NIH funded studies registered after September 30, 2005 and completed by December 31, 2008.

They found fewer than half appeared in a journal 30 months after completion, while a further third were not published 51 months after they concluded.

Joseph Ross, M.D., first author of the study and a Yale assistant professor of medicine, said: "When research findings are not disseminated, the scientific process is disrupted and leads to redundant efforts and misconceptions about clinical evidence.

"Such inaction undermines both the trial in question and the evidence available in peer-reviewed medical literature. This has far-reaching implications for policy decisions, and even institutional review board assessments of risks and benefits associated with future research studies."

Dr Ross suggested that while there may be reasons why a number of studies did not appear, for example being rejected by journals, efforts should be made to offer alternative publication platforms. These include the federally mandated results database at ClinicalTrials.gov.

He also highlighted that the issue is not concentrated to NIH funded trials, with similar issues being reported by the pharmaceutical industry and medical device manufacturers.

"More work needs to be done to better understand impediments to publication," Dr Ross concluded.

The research is published in the latest issue of the British Medical Journal, which focuses on unpublished evidence.

Writing for the journal Perspectives on Psychological Science, Marco Bertamini of the University of Liverpool and Marcus Munafò of the University of Bristol recently argued that 'bite-size science' is becoming too prevalent.

The pair argued the trend could see a greater number of false alarms, as smaller samples increase the risk of statistical deviation.