Mike: Hello, I’m Mike Clarke podcast editor for the Cochrane Library and co-ordinating editor of the Cochrane Methodology Review Group. In this podcast, you’ll hear from Marian Showell of the University of Auckland in New Zealand about a significant concern in medical research known as publication bias. The updated Cochrane Methodology Review was published in November 2024, bringing together evidence on the size of this problem, drawing on more than 200 studies.

Marian: Publication bias occurs when the likelihood of a study being published is influenced by the direction or strength of its findings. Typically, studies that report positive results tend to be published more frequently than those yielding null or negative results. In some instances, unfavourable trial outcomes may even be deliberately withheld from publication.
This bias poses critical challenges for the scientific community, as the selective reporting of trial results can obscure the true efficacy and safety profiles of healthcare interventions. If the time from trial completion to publication significantly differs between studies with different results, healthcare professionals may erroneously accept new treatments as effective, based solely on incomplete evidence; and the time to publication might be years shorter for positive findings than for negative ones.
To investigate the extent of publication bias, we conducted a systematic review of the publication rates and time to publication for health-related clinical trials. We aimed to identify whether various factors, including the nature of trial results, sample size, number of participating centres, and whether the trial had commercial or non-commercial funding affect these publication dynamics. We conducted a comprehensive search across multiple literature databases and identified 204 studies published from 1992 to 2023, encompassing more than 165,000 clinical trials.
Our findings revealed several noteworthy trends. First, we discovered that only just over 50% of all trials were ultimately published, and this proportion has not shown significant improvement over time. This stagnation raises concerns about the transparency of clinical research that influences decision making across health care, not least because publication rates were significantly higher for trials with positive results. Trials that reported favourable outcomes were also published more rapidly, averaging 24 months from completion compared to 31 months for those with null or negative results.
Other factors influencing publication rates included trial size and funding source. Larger trials were published more swiftly, and trials funded by non-commercial organizations generally experienced faster publication timelines than those supported by industry funding. Notably, we observed no substantial difference in time to publication between multi-center and single-center trials.
In summary, our investigation underscores the persistent issue of poor publication rates and protracted timelines for disseminating clinical trial findings. These results have significant implications for systematic reviews, particularly when the inclusion of trials in such reviews is linked to their publication status. This is particularly concerning for reviews based on a limited number of studies, which may lead to incomplete or biased conclusions regarding treatment efficacy and safety.
As we continue to navigate the complexities of health research, it is imperative that we advocate for greater transparency and the publication of all trial results, regardless of their outcomes. This commitment will foster a more accurate understanding of the effects of healthcare interventions, ultimately benefiting both practitioners and patients.

Mike: Thanks Marian. If you would like to learn more about the challenges caused by publication bias, you can find the full review online at CochraneLibrary.com. Typing “publication bias” into the search box, will show it near the top of the list.