Publication of Clinical Trials on Medicinal Products: A Follow-Up Study

Kinga Amália Sándor-Bajusz University of Pecs: Pecsi Tudomanyegyetem Andrea Kraut University of Pecs: Pecsi Tudomanyegyetem Odgerel Baasan University of Pecs: Pecsi Tudomanyegyetem Gergely Márovics University of Pecs: Pecsi Tudomanyegyetem Károly Berényi University of Pecs: Pecsi Tudomanyegyetem Szimonetta Lohner (  lohner.szimonetta@pte.hu ) University of Pecs: Pecsi Tudomanyegyetem https://orcid.org/0000-0002-6292-4802


Introduction
Clinical research should provide reliable evidence to clinicians, health policy makers and researchers (1), which is only possible once the results are made publicly available (2). On a national level, published research means that the resources expended are not waisted and the results of the research become part of the international medical knowledge. Published research with a domestic author contributes to the assessment of the scienti c performance of a country.
As of 2014, any trial of any medicinal product conducted in a member state of the European Union (EU) is required to be registered in the European Union Clinical Trials Register (EU CTR), which is administered by the European Medicines Agency (EMA). Following the 2012 European Commission guideline 2012/c302/03, sponsors must ensure that all trials registered on EU CTR disclose their results to the EMA within 12 months of trial completion; phase I trials are exempt unless they are part of a pediatric investigation plan (3). Voluntary initiatives (4) and recommendations (5) started to emphasize the importance of registration of clinical trials and subsequent reporting of results. The EU CTR also tries to increase awareness on mandatory posting of results (6), and recently launched a page with a tutorial to facilitate posting of results on the EU CTR webpage (7).
Beyond mandatory posting in EU CTR, European researchers also often make the decision to register their studies in clinicaltrials.gov, the largest trial register worldwide. However, the extent to which trials conducted in the EU are registered in clinicaltrials.gov and whether this "multiple registration" has bene ted the reporting of trial results or scienti c quality of publications is not well investigated.
Results posted in registries currently have limited impact and awareness in the scienti c community.
Advantage of results reported to a registry is undeniably their standardized format (8,9); however, these results do not undergo rigorous evaluation as do full scienti c publications during the peer review process. Besides, publications and scientometrics are -despite international initiatives to change this (10) -currently an integral part of research evaluation and play a crucial role in decision making for national research policies, funding, promotions, and the careers of scientists (7). Therefore, it should be underlined that results posted in registries do not contribute to the total research output of either the participating researchers nor the participating country.
The aim of this methodological cohort study was to investigate how and to what extent does an authorized medical research conducted in a given country become visible and affects the research output of that country. We investigated publication rates, time until publication and the relationship between posting results in trial registers and publishing them as a full scienti c publication. Further, we aimed to

Methods
In order to examine the publication rates and time of clinical trials conducted in Hungary, we collected detailed data on the human clinical trials on medicinal products authorized in the country back in 2012.

Search strategy
We used the advanced search function of the EU Clinical Trials Register (www.clinicaltrialsregister.eu) to identify clinical trials registered within the date range of January 1, 2012 to December 31, 2012 with Hungary as a participating research center.

Inclusion and exclusion criteria
Clinical trials were eligible for our study if a) Hungary was a site of the clinical trial; b) they were registered in the database by the National Institute of Pharmacy (Hungary) in the year 2012; c) no restrictions were applied to the trial phase, trial status or participant characteristics (e.g. age, gender, disease group).
Identi cation of trials in clinicaltrials.gov and data extraction from registries We tried to identify included trials in the register clinicaltrials.gov by searching the EU CTR identi er or by the use of speci c PICO terms.
We extracted pre-de ned study characteristics from the study registries: full title of the trial, authorization date, trial start and completion dates, information on participating countries, sponsor, funder, trial scope, trial design, blinding, sample size, study phase, therapeutic area and presence of a data monitoring committee (DMC).
We determined whether or not study results were available in the study registries EU Clinical Trial Register and ClinicalTrials.gov. In this current paper, we aim to distinguish results available in the registries ("results in registries") from results published as full scienti c publications ("publication").

Identi cation of corresponding scienti c publications
Full scienti c publications were de ned as publications which were published in a scienti c journal of any type and were reporting study results on pre-de ned outcomes. We excluded methods papers and published protocols, and publications which reported results of a secondary analysis.
The availability of scienti c publications was rst checked in February 2019 and then 16 months later in June 2020. Publications were identi ed in a step by step process for each trial. First, we checked whether publications were already added to the register. In a second step, we searched for publications in the PubMed database with the following identi cation data: a) the trial register number, b) the investigators' names, c) keywords describing the intervention or the condition (PICO elements). As a third step, Google was searched with the same search terms.
All identi ed publications, potentially belonging to the registered study were checked for their content (study design, population characteristics, dates of recruitment, intervention, comparator). Publications which clearly described the results of the originally planned and registered study were included.

Data extraction from scienti c publications
We extracted the following data from the scienti c publications: the presence of author(s) with a Hungarian a liation; the number of Hungarian authors or whether Hungarian participation in the study was mentioned in a way other than author a liation; the journal's name, and date of publication. In cases when there were different forms of publishing (e.g. published electronically ahead of print), we took the rst date when the full text of the nal manuscript was accessible.
To estimate the time to publication, we counted total months between trial end date available in the EU Clinical Trials Register and the publication date. We expressed publication rate as the percentage of clinical trials with a full scienti c publication divided by all clinical trials. The number of all clinical trials was calculated separately for each month after trial completion, by adding trials without any publication to the number of trials with a publication only until the elapsed time since their completion.
Impact factors for each journal were derived from the Journal Citation Report, Clarivate Analytics via www.webofknowledge.com. Scimago journal rank (Q1-Q4) was derived from www.scimagojr.com.
We also wanted to see the public's degree of accessibility of results from published scienti c publications. We therefore investigated whether scienti c publications were published openly or with

Included clinical trials
A total of 614 clinical trials were identi ed in our search. After excluding trials where 2012 did not correspond to the Hungarian registration date, but to the registration date of another participating country of multinational trials, a total of 330 Hungarian national or international clinical trials were eligible to be included in our methodological cohort (see Additional File 1). Eight years after trial authorization (in June 2020), a total of 232 clinical trials were "completed" trials for at least 1 year. Baseline characteristics of these trials are presented in Table 1.
Most of the trials were international, initiated and funded by the industry and assessed both e cacy and safety of a therapeutic intervention. Of the investigated clinical trials, 91.8% were registered both in the EU Clinical Trials Register and the clinicaltrials.gov database.  In uence of the industry on transparency and scienti c impact of a trial The number of authorized clinical trials initiated by the academy was extremely low (2.7% of all authorized trials). Clinical trials initiated by the industry or the academy are shown and compared in Table 3. All industry-initiated trials with accessible information regarding funding were funded by the industry. Investigator-initiated clinical trials were also partly funded by the industry. For one third of investigatorinitiated trials information on funding was unavailable.
A DMC was available to a much larger extent in trials initiated by the industry.
Results for investigator-initiated trials were signi cantly less likely to be posted in a clinical trial register (73.1% vs. 11.1%) and also slightly less likely to be published as a full scienti c publication (62.8% vs. 55.6%).
Both the rate of publications with at least one Hungarian author (33.3% vs. 12.2%), and the rate of publications mentioning Hungarian participation in any form (44.4% vs. 27.5) was higher among investigator-initiated as compared to industry-initiated trials.

Summary of ndings
Our study provides empirical evidence about the publication tendencies of authorized clinical trials in Hungary, the impact of authorized clinical trials on the scienti c reputation of the authorizing country, and the role of multiple registrations in increasing transparency.
A total of 97.3% of authorized and EU CTR -registered clinical trials were initiated by the industry. About 20% of clinical trials were published within one year after trial completion. Trials conducted only within Europe and registered only in the EU CTR register were published signi cantly later.
Universality is a fundamental principle of science (11); open access publications have therefore the largest impact on the scienti c community. In this study, 70.93% of publications were found to be openly accessible to the public. However, we were not able to identify trial characteristics which might in uence the access to scienti c publications.
Publications with Hungarian co-authors occurred in 21.5% of cases only. Clinical trials with study sites within Europe and trials initiated by the academy resulted more often in scienti c publications with Hungarian authors.
Trials registered not only in the EU Clinical Trials Register, but also in the clinicaltrials.gov database were more likely to be identi ed as full publications. The results of clinical trials available as full scienti c papers but absent in registries were in low numbers. However, we found that almost one quarter of results of the investigated clinical trials were available in registries, but not as a full scienti c publication.

Strength and limitations of this methodological research study
Our study analyzed a representative sample of trials authorized in Hungary, the results can therefore be generalized to the country. All trials were registered in study registries, thus basic study information was available for all the trials included in our study. All data extractors were trained and the main outcome data were double-checked and double-extracted.
Our study has limitations. The cohort was composed of trials that were authorized and included in the study registry EU Clinical Trials Register by the national authority. Due to the low number of investigatorinitiated trials it is di cult to draw rm conclusions on their publication tendencies. Nevertheless, our results may indicate a publication trend also for investigator-initiated trials.
Another limitation is, that the information provided for researchers in the EU Clinical Trials Register on trials de ned as completed (when it "has been completed in accordance with the full requirements of the protocol") might be interpreted in different ways by researchers and may have impacted our results. We searched for scienti c publications in 2020 and trials completed within one year before the search date were excluded.

Comparison with other studies
To our knowledge, this is the rst study investigating how and to what extent research authorized and conducted in a given country becomes visible and affects the scienti c performance of that country.
Authorship issues were already discussed is several papers, dealing mainly with gender distribution of authorship (12,13); association between sponsorship and authorship (14,15); under-representation of researchers from speci c regions in papers published from research done in these regions (16, 17); and di culties and possibilities in determining authorship in multicenter clinical trials (18,19). These studies were mainly based on publication data sets. The approach to prospectively follow trials authorized in a given country until publication and investigate authorship in such a cohort of studies is novel.
There was one large cohort study investigating compliance with requirement to report results on the EU Clinical Trials Register up to December 2016 (8). This study found that trials with a commercial sponsor tend to be substantially more likely to post results on the EU CTR than those with a non-commercial sponsor (68.1% v 11.0%)(8). This is in line with the results of the present study: a signi cantly higher posting of results in the EU CTR for industry-initiated trials were found (73.1%) than for investigatorinitiated trials (11.1%). Besides the need for standardized procedures (20), periodic quality control assessments during trial implementation (21), improved reporting about funding (22) Posting trial results in study registries might be the rst step to make study results become openly available for the public; however, study results should be published also as a scienti c publication as soon as possible after trial completion. Systematic reviewers and guideline developers are advised to search clinical trial registers in addition to electronic databases to identify study results, which have not been published as full text publications at the time of the search.
The present study also showed that the participation of Hungarian researchers in industry-initiated studies on medicinal products has only partial measurable scienti c bene ts, as Hungarian researchers appear as authors in only a fraction of scienti c publication derived from these trials. Several publications not even contain the list of countries of trial participants. In line with the Lancet journals, which strongly support the inclusion of authors from local countries on papers reporting studies from those countries, we also would like to "encourage authors to include researchers who originally collected the data, where possible, and to share expertise in analysis and other skills, so that the research capacity of the country from which the data were obtained is strengthened" (24); i.e. to enable local researchers to ful l the criteria for authorship developed by the International Committee for Medical Journal Editors.
The scienti c performance of universities and countries is evaluated and ranked -despite valuable initiatives for change -based on research productivity (i.e. the number of scienti c publications), research impact and research excellence (i.e. the number of scienti c papers in high impact journals).
Considering that slightly over a fth of authorized Hungarian trials result in scienti c publications with a Hungarian co-authorship, we can conclude that the authorized, mainly industry-initiated clinical trials on medicinal products currently result in limited measurable scienti c bene ts to the participating researchers and their countries.

Conclusions
We call researchers of investigator-initiated clinical trials, to register their trials in an openly available clinical trial register. Trial registers have to be considered as an important source of information of clinical trial results, as they may contain results from unpublished trials or trials published with closed access. Domestic scienti c impact of trials on medicinal products has to be further improved; an increase in the number and role of investigator-initiated trials might help to achieve this goal.

Declarations
Ethics approval and consent to participate: Not applicable Consent for publication: Not applicable.
Availability of data and material:The data supporting the conclusions of this article is included within the article (and its additional les).
Competing interests: The authors declare that no competing interests exist.