Trial registration as a safeguard against outcome reporting bias and spin? A case study of randomized controlled trials of acupuncture

RESEARCH ARTICLE Trial registration as a safeguard against outcome reporting bias and spin? A case study of randomized controlled trials of acupuncture Jiyoon Won1,2, Seoyeon Kim1,2, Inhu Bae3, Hyangsook Lee ID1,2* a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Korean Medical Science, Graduate School, Kyung Hee University, Dongdaemun-gu, Seoul, Republic of Korea, 2 Acupuncture & Meridian Science Research Centre, College of Korean Medicine, Kyung Hee University, Dongdaemun-gu, Seoul, Republic of Korea, 3 Kyung Hee University Korean Medicine Hospital, Dongdaemun-gu, Seoul, Republic of Korea * Abstract Background and objective OPEN ACCESS Citation: Won J, Kim S, Bae I, Lee H (2019) Trial registration as a safeguard against outcome reporting bias and spin? A case study of randomized controlled trials of acupuncture. PLoS ONE 14(10): e0223305. https://doi.org/10.1371/ journal.pone.0223305 Editor: Spyridon N. Papageorgiou, University of Zurich, SWITZERLAND Received: December 19, 2018 Accepted: September 18, 2019 Published: October 3, 2019 Copyright: © 2019 Won et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are included in the paper and its supporting information files. Funding: This work was supported by the National Research Foundation (NRF) of Korea funded by the Korean government (Ministry of Science and ICT, grant No. NRF-2017R1A2B4006407). Competing interests: The authors have declared that no competing interests exist. Trial registration is widely endorsed as it is considered not only to enhance transparency and quality of reporting but also to help safeguard against outcome reporting bias and probably spin, known as specific reporting that could distort the interpretation of results thus mislead readers. We planned to investigate the current registration status of recently published randomized controlled trials (RCTs) of acupuncture, outcome reporting bias in the prospectively registered trials, and the association between trial registration and presence of spin and methodological factors in acupuncture RCTs. Methods Acupuncture RCTs published in English in recent 5 years (January 2013 to December 2017) were searched in PubMed, Cochrane Central Register of Controlled Trials, and EMBASE. Trial registration records identified in the publications and trial registries were classified into prospectively registered, retrospectively registered, or unregistered. Primary outcomes were identified and the direction of the results was judged as statistically significant (positive) or statistically nonsignificant (negative). We compared registered and published primary outcomes to assess outcome reporting bias and assessed whether discrepancies favored statistically significant outcomes. Frequency and strategies of spin in published reports with statistically nonsignificant results for primary outcomes were then identified. We also analyzed whether the trial registration status was associated with spin and quality of methodological factors. Results Of the 322 included RCTs, 41.9% (n = 135) were prospectively registered. Among 64 studies that were prospectively registered and specified primary outcomes, 25 trials had the PLOS ONE | https://doi.org/10.1371/journal.pone.0223305 October 3, 2019 1 / 19 Association of trial registration with outcome reporting bias and spin in published acupuncture trials discrepancies between the registered and published primary outcomes and 60% of them (15 trials) favored the statistically significant findings. Among 169 studies that specified primary outcomes, trial registration status was not associated with the direction of results, i.e., statistically significant or not. Spin was identified in 56.4% out of 78 studies with statistically nonsignificant primary outcomes and claiming efficacy with no consideration of statistically nonsignificant primary outcomes was the most common strategy for spin. Trial registration status was not statistically different between studies with and without spin. Conclusion While trial registration seemed to have improved over time, primary outcomes in registered records and publications were often inconsistent, tending to favor statistically significant findings and spin was common in studies with statistically nonsignificant primary outcomes. Journal editors and researchers in this field should be alerted to still prevalent reporting bias and spin. I. Introduction Research findings should be reported in complete, transparent, and accurate manners [1]. However, there are various obstacles to such reporting. Among them, outcome reporting bias in a clinical trial, defined as a selective reporting of outcomes influenced by the nature and direction of the results [1–4], has been widely recognized as one of the most substantial biases affecting the results from individual studies [5,6]. Separately from outcome reporting bias, a misrepresentation of study results regardless of intention, that overemphasizes the benefits of the intervention and exaggerates safety compared with those shown by the results, is known as spin [7–9]. Some researchers have recently demonstrated that when spin occurs it may mislead readers to inadequate interpretation and dissemination of research findings [10]. One of the recommended safeguards against outcome reporting bias and spin by securing the transparency of reporting is clinical trial registration, i.e., the systematic public disclosure of key descriptive information about a clinical trial before the commencement of study [11– 13]. After the U.S. Food and Drug Administration (FDA) enacted legal mandates for registration of trials in 1997 [14], endorsement of trial registration before the onset of enrolling participants by the International Committee of Medical Journal Editors (ICMJE) followed [15,16]. Accordingly, the number of trial registration substantially increased [17,18] and in October 2018, ClinicalTrials.gov, the largest and most known trial registry, contains trial registration records of more than 280,000 studies from over 200 countries [19]. Despite requirements that trial registration should be completed before the first participant is enrolled at a publicly accessible trial registry, there have been concerns regarding inadequately registered studies [20]. Trials that were retrospectively registered or not registered have been reported more likely to overestimate the treatment effect than those prospectively registered [21]. Furthermore, regarding the quality of registered information, the discrepancies between information in the registry and publications such as changed, omitted, or newly introduced primary outcomes, have been identified [22,23] and such discrepancies are known to prevail across (...truncated)