Registration of clinical trials has long been identified as best practice [1]. Amongst its various advantages, registration of research reduces the risk of publication bias, as a discrepancy between study registrations and published studies is a clear indicator of such bias. It also reduces the risk of reporting bias, whereby outcomes are reported selectively based on their significance, and not based on a priori definitions [2]. It also reduces the risk of waste in biomedical research by having a clear record of studies that have been, and are being, undertaken [3].
Whilst emphasis on registration has traditionally been placed on clinical trials and systematic reviews, there is an increasing recognition of the importance of registering all research. In 2013, the Declaration of Helsinki was modified to ensure that not only clinical trials were registered, but all studies involving human participants. Despite these efforts, there is increasing acceptance that observational research is seldom registered [4], and that there are limited venues to register research that has already begun recruiting participants. This was the basis for the foundation of the Research Registry® [5]. This provides free and immediate registration of any study type involving human participants. Its development has been outlined previously [4], and we have reported the first 500 registrations; finding that a majority of these to be retrospective cohort studies, case series or first-in-man case reports, fulfilling a previous unmet need for the registration of these study types, that the quality of the registrations improved over time and that ∼16% of studies were incorrectly registered [6].
The aim of this study is to describe the first 2000 registrations received to the Research Registry®, and to separately report the first 200 registrations of systematic reviews received to the Review Registry.
This will be retrospective database analysis, and involves no human participants, so is not registered. No reporting guidelines exist for study types of this design.
Data will be extracted from the Research Registry® (www.researchregistry.com) database when 2000 complete registrations have been received. Characteristics to be extracted include: study design; country of origin; date of registration; number of patients anticipated or included in study; source of funding for study and study phase at which registration was made. Curation reports for these 2000 registrations will be collected, and the process curation of the registry has been described previously; involving a weekly process of screening for inappropriate registrations by trained curators [5]. Any registrations missing these data will be excluded by list-wise deletion, and their characteristics compared to included registrations. Separate analyses will be performed for the Review Registry, the part of the Research Registry® designated for the registration of systematic reviews and meta-analyses.
To establish the quality of the registrations received at the Research Registry®, we have previously described a quality score that was developed based on Sir Austin Bradford Hill’s criteria for what a research study should convey [5]. Each registration is marked by providing an appropriate answer for each of the following: primary investigator, participating institutions, countries recruiting, key questions and objectives, patient population, intervention, control or comparator, primary outcomes (and secondary, if used) and dates of enrolment. Each of these items scores one point, and therefore each registration can score a maximum of 9. If any of these items are inapplicable, such as the use of a control or comparator for first-in-man case reports, then the item is marked as such and the denominator of the quality score for that registration decreases. Each registration will be marked by two researchers independently. Where consensus is not reached on the correct score for a given registration, the decision will be raised to a senior author.
All records will be held in a Microsoft Excel 2007 (Microsoft, Redmond, WA) database. Data will be cleaned and descriptive statistics will be used. Continuous data will be presented as either mean or median (with standard deviations and interquartile ranges respectively), and categorical data as a number (with percentages). The quality score will be compared between different time points (divided by month) with an independent samples Kruskal–Wallis test, with post hoc comparisons to compare quality scores between months. To compare continuous data, we will use student’s t-test if parametric, and a Mann Whitney U test if non-parametric. χ2 tests will be used to compare categorical variables. All statistical tests will be performed within R [7] and the analysis code will be provided.
As no human participants will be included in this study, no ethical approval was required. Findings will be disseminated by the team working on Research Registry® as soon as data are available. The data will be published in a peer reviewed journal and presented at international conferences.
No funding was received for this work.
RAA is the Director and Founder of the Research Registry; AJF is the Assistant Director of the Research Registry; CL is the Director of Quality of the Research Registry, DJJ is the Director of Usability of the Research Registry.
AJJ, DJJ – drafting, study design, RAA – concept, study design and critical appraisal, CL – critical appraisal. All authors approved the final version of the submitted manuscript.
Riaz Agha
This is not a human study.
Sim, I., Chan, A.-W., Gulezoglu, A.M., Evans, T. and Pang, T. (2006). Clinical Trial Registration: transparency is the watchword. Lancet 367(9523): 1631–1633. [PubMed]
Page, M.J., McKenzie, J.E., Kirkham, J., Dwan, K., Kramer, S. and Green, S. (2014). Bias due to selective inclusion and reporting of outcomes and analyses in systematic reviews of randomised trials of healthcare interventions. Cochrane Database Syst Rev,
Moher, D., Glasziou, P., Chalmers, I., Nasser, M., Bossuyt, P.M.M. and Korevaar, D.A. (2016). Increasing value and reducing waste in biomedical research: who’s listening?. Lancet 387(10027): 1573–1586. [PubMed]
Loder, E., Groves, T. and Macauley, D. (2010). Registration of observational studies. BMJ 340: c950. [PubMed]
Agha, R. and Rosin, D. (2015). The Research Registry® – answering the call to register every research study involving human participants. Int J Surg 16(Pt A): 113–115. [PubMed]
Agha, R.A., Fowler, A.J., Limb, C.E., Al Omran, Y., Sagoo, H., Koshy, K., Jafree, D.J., Anwar, M.O., McCulloch, P. and Orgill, D.P. (2016). The first 500 registrations to the Research Registry: advancing registration of under-registered study types. Front. Surg. 3
Core Team, R. (2014). R: A Language and Environment for Statistical Computing In: Vienna, Austria: R Foundation for Statistical Computing. URL http://www.R-project.org/.
Supplementary data associated with this article can be found, in the online version, at https://s3-eu-west-1.amazonaws.com/ubiquity-partner-network/ijs/journal/ijsp/IJSP12.zip.