Public availability of randomized clinical trial protocols: A repeated meta-research study.
Schönenberger CM., Chiaborelli M., Heravi AT., Kübler L., Gandhi P., Kontár Z., Hüllstrung J., Elalfy M., Glasstetter J., Gryaznov D., von Niederhäusern B., Blümle A., Busse JW., Lohner S., Hopewell S., Griessbach A., Matthias Briel None., Speich B.
OBJECTIVE: Making protocols of randomized clinical trials (RCT) publicly available is important for the trustworthiness and quality of medical research. In a previous study assessing 326 RCTs with ethical approval in 2012, only 36% had a publicly available protocol. We aimed to generate current evidence on the availability of RCT protocols and to evaluate changes over time. STUDY DESIGN AND SETTING: Using a representative sample of RCTs approved in 2016 in Switzerland, Canada, Germany, and the UK, we investigated the number of available protocols by searching PubMed, Google Scholar, trial registries, and Google. Up to June 2024, we systematically searched for (i) protocols available as peer-reviewed publications, (ii) protocols attached to trial registries and (iii) protocols shared with result publications of RCTs. We used multivariable logistic regression to examine the association of protocol availability with trial characteristics such as sample size, drug vs. non-drug interventions, multicenter vs. single center status, and RCT approval in 2016 vs. 2012. RESULTS: Of the 347 included RCTs, 228 (66%) had an available protocol. Forty-three percent (150/347) of the protocols were available as files on trial registries, 26% (91/347) as supplementary material to result publication, and 23% (81/347) as peer-reviewed publications. Protocol availability improved over time in industry trials (83.4% in 2016 vs. 34.6% in 2012). Protocol availability for non-industry trials remained low (46.4% 2016 vs. 38.1% 2012). Multicenter trials (206/256; 77.7% vs single-center trials 22/82; 26.8%) and larger sample size (>500 participants 68/77; 88.3%, 100-500 participants 131/191; 68.6%, <100 participants 29/79; 36.7%) showed higher protocol availability. CONCLUSION: The availability of protocols increased in RCTs approved in 2016 compared to RCTs from 2012. This was mainly driven by industry sponsored trials. Efforts to further improve protocol availability should be continued, especially in non-industry sponsored RCTs.