Background: Publication bias is a highly complex and common problem in the area of clinical research. Its adverse consequences range from the waste of resources to the unnecessary duplication of studies and the distortion of the public perception about the benefits and harms of treatments which can ultimately lead to wrong clinical and health care decisions. Data based on previous research suggest the mean rate of publication of randomized controlled trials (RCTs) is around 50%, although there has been a trend towards improvement in recent years.
Objectives: To assess the publication rate of RCTs in the area of cancer authorized by the AEMPS.
Methods: We performed an observational study consisting of a cohort of RCT records with a 9-year follow-up period. We identified all the RCTs authorized by the AEMPS in 2013 corresponding to the therapeutic area of cancer from the Spanish Registry of Clinical Studies (REec). We identified the corresponding publications from the same registry, clinicaltrials.gov registry, through formal searches in Pubmed and, ultimately, through the Google meta-search engine.
Results: A total of 215 RCTs were identified. Of these, 129 (60.0%) were published and 86 (40.0%) were not. About 87.9% have also been registered at clinicaltrials.gov. The median size was 181 (IQR 82-450). International multicentre (86.5%), phase II (39.5%), or phase III (38.1%) studies, with a commercial sponsor (84.7%), predominated. The main outcome most studied was progression-free survival (27.0%). The multicentric character (mainly international) was positively associated with the publication of the study (p-value = 0.014), as was the size of the study (p-value < 0.014).
Conclusions: Although there has been a slight increase in the publication rate in Spain over the past decade, there is still much room for improvement. It is mandatory that all actors involved in research be more proactive in order to ensure compliance with current legislation and ethical codes that require full publication and dissemination of results. Patient, public, and/or healthcare consumer involvement: A PhD student was involved.