Abstract: Results from clinical trials are an important basis for decision making in clinical medicine. However, trials may be reported late or not at all, resulting in a waste of research resources and a distorted knowledge base for evidence-based practice, which may lead to ineffective or harmful clinical decisions.
We have followed up reporting of clinical trials at medical universities and university hospitals in the Nordic countries. Eligible trials were registered in the EU Clinical Trials Registry and/or ClinicalTrials.gov, completed in 2016-2019, and led by a Nordic medical university or university hospital. We automatically identified summary results posted at the trial registries and we searched manually for matching results publications.
We included 2113 trials. Trials were typically non-industry sponsored, small (less than 100), monocentric, and prospectively registered. Overall, 1098/2113 trials were reported within 2 years of completion (52%, 95%CI 49.8-54.1%). Few trials had summary results posted in a trial registry within one year of completion (42/2113, 2%, 95%CI 1.5-2.7%) or at the end of follow up (250/2113, 11.8%, 95%CI 10.5-13.3%). Any results, whether as summary results or a results publication, were found for 1642/2113 trials (77.7%, 95%CI 75.9-79.4%). For the 1642 reported trials, median time from completion to first results reporting was 506 days (IQR 602 days).
This project is a first step towards following up all trials conducted in the Nordic countries, regardless of registration status/platform and sponsor. The project will enable further meta-research on characteristics, quality, and usefulness of Nordic clinical trials. We are seeking to collaborate with stakeholders to improve the reporting of clinical trials in the Nordic countries.
Bio: Gustav Nilsonne is an associate professor of neuroscience at Karolinska Institutet, Stockholm, Sweden. He works on metascience to assess and improve transparency and reproducibility of research using quantitative methods. His particular interests include multi-analyst designs, quality control and error checking, researchers’ incentives, and evidence synthesis. Gustav is a previous visiting scholar at METRICS and at the QUEST center in Berlin, Germany.
