Article Text

Download PDFPDF

Discrepancies between research advertisements and disclosure of study locations in trial registrations for USA-sponsored research in Russia
  1. D Patrone
  1. Correspondence to Dr Daniel Patrone, Union Graduate College–Mt Sinai School of Medicine Bioethics Program, Philosophy Department, Union College, 80 Nott Terrace, Schenectady, NY 13847, USA; dpatrone{at}gmail.com

Abstract

Background and Aim The full disclosure of all locations at which research is conducted is an important requirement of clinical trial registration. Yet, little is known about how well researchers and sponsors disclose this information in their registrations. The aim of this study is to examine the adequacy of study location disclosure on http://ClinicalTrials.gov/ for recent USA-sponsored research in the Russian Federation.

Methods and Results Information from advertisements for clinical research participation opportunities in Russia was compared with that in the corresponding trial registration in http://ClinicalTrials.gov/. Of 218 USA-sponsored trials advertised to Russian patients and physicians, nearly one-third did not disclose Russia as a study location in the registration. Longer trials were more likely to specify Russia as a study location.

Conclusion This study found that recent USA-sponsored clinical trials advertised to patients and physicians in the Russian Federation often do not specify Russia as a study location in their registrations on http://ClinicalTrials.gov/. The failure to fully disclose the locations at which research takes place undermines patients' ability to access information about potentially beneficial research opportunities, limits access to information, which may be important for identifying or avoiding ethically questionable practices, and may frustrate the ability to interpret the results of research.

  • Trial registration
  • disclosure
  • transparency
  • study locations
  • ethics
  • research on special populations
View Full Text

Statistics from Altmetric.com

Various legal, professional and ethical standards and guidelines require the registration of clinical trials in some publicly accessible databases. The Declaration of Helsinki endorses the requirement to register trials, stating that “every clinical trial must be registered in a publicly accessible database before recruitment of the first participant.”1 Section 113 of the 1997 Food and Drug Administration (FDA) Modernization Act requires that studies of drugs and biological products for the treatment of serious or life-threatening conditions conducted under FDA regulations be subject to mandatory public registration.2 In 2004, the International Committee of Medical Journal Editors resolved that as of September 2005, journals under their purview would no longer publish results from trials that were not registered before enrolling participants.3 In 2007, the FDA Amendments Act expanded the FDA's registration requirements to include all drug, biological and device trials subject to FDA regulation and required additional mandatory data elements in the registration.4

There are a number of important functions served by public trial registration. First, trial registration aims to support the practice of evidence-based medicine, primarily by combating selective publication and selective result reporting. Evidence-based medicine depends on physicians' and others' access to and evaluation of the results of clinical research. However, if researchers and sponsors readily publish or report only research results that appear favourable, then the literature available to those making relevant treatment or policy decisions will be skewed. In principle, trial registration aims to curb the negative effects of selective publication and reporting by creating a permanent public record of research conducted, thereby providing some measure to determine the extent to which the published literature is representative of the body of research actually conducted.5 6

Second, publicly accessible and searchable registries allow for greater patient access to information about research participation opportunities, thereby helping patients and their physicians to identify potentially beneficial treatment alternatives. One of the initial impetuses for creating legally mandated, centralised trial registries came from patient advocacy groups demanding greater access to information about clinical trials, and one of the principal intended benefits of http://ClinicalTrials.gov/ was to help improve patient and physician access to information about research participation opportunities.7 The FDA expressed this understanding of the justification for trial registration in a letter sent after a 2005 audit of http://ClinicalTrials.gov/ to investigators and sponsors who failed to provide information in the required study location field. The letter states that “the purpose of http://ClinicalTrials.gov/ is to inform members of the public about potential trials for which they may be eligible. Some patients are willing to travel anywhere to participate in a study”, and it goes on to instruct the registrants to “include the city, the state and the country for each clinical trial site so that visitors to http://ClinicalTrials.gov/ may search for clinical trials by location. We also ask that you consider including the name of the facility conducting the trial.”8

Third, trial registration adds a further layer of oversight to clinical research. Although trial registration cannot be expected to provide sufficient information to allow comprehensive ethical review of clinical trials, by requiring researchers and sponsors to disclose what they study, on whom and how, the public can access at least some basic information regarding the conduct of human participant research and can help identify questionable research initiatives, existing research gaps, or potential research redundancies that can expose participants to unnecessary risks.7

The full disclosure of study locations is important for each of these functions served by trial registration. Without full transparency regarding all study locations, trial registries provide limited utility for those in search of suitable opportunities to participate in research, particularly for those patients who would be interested in travelling even great distances.8 Furthermore, research conducted in low- and middle-income countries may involve vulnerable populations and thus be subject to special ethical and legal requirements. Transparency is, therefore, particularly crucial in multinational research.9 For example, it is widely accepted and enshrined in the Belmont Report that justice in participant selection requires that vulnerable populations ought not to be used “solely for administrative convenience.”10 In addition, many codes and guidelines governing international research agree that ethical research must be responsive to the health needs of the host community and must provide the host community access to the fruits of successful research.11 Ensuring that such requirements are being met necessitates that investigators and sponsors are open about where all of their research is being conducted. Finally, although the disclosure of study locations may have little impact on selective publication and results reporting, the practice of evidence-based medicine does warrant a serious concern for transparency regarding research locations, especially given the increasingly global nature of the clinical research market. Whereas the unique contextual features of research conducted in international, multisite trials do not invalidate a well-designed study, they may provide information that physicians, other researchers, or policy makers might want to better scrutinise or interpret results. For example, as Glickman et al12 note, the nature and quality of the healthcare system available where research is conducted, the social ecology and genetic make-up of participant population, and the previous treatment history of participants all may affect the extrapolation of results from multinational clinical trials to populations in the developed world. Yet, they found that <5% of articles reporting results published in the New England Journal of Medicine, the Journal of the American Medical Association and the Lancet report recruitment numbers according to individual country, and they point out that “genetic diversity is often not considered in study design and interpretation and reporting of results.” In a 1998 study, Vickers et al13 found that research conducted in certain countries (Russia among them) is more likely to yield favourable results. They also determined that no trial conducted in Russia had found a treatment ineffective.

Despite the importance of complete and accurate study location registration, little is known about how well researchers and sponsors disclose research locations in their trial registrations. Some previous research has suggested that there is near-perfect compliance with requirements to register mandatory data elements on http://ClinicalTrials.gov/.14 15 However, these studies did not examine completeness for the mandatory data field referring to study location. Moreover, these studies are limited to examining whether information has been submitted for the data elements and whether any supplied information is specific enough to be clinically useful. This method of completeness analysis cannot assess the veracity of the information submitted by researchers, and numerous studies have documented discrepancies between information that investigators disclose about their trials and the research that is actually conducted.16–20 The present study, therefore, aims to assess the adequacy of study location disclosures in trial registration for USA-sponsored research advertised and conducted in Russia by comparing information from Russian language advertisements of trial participation opportunities with information contained in the corresponding trial registrations available in http://ClinicalTrials.gov/.

Materials and methods

Trial advertisements were obtained from http://Clinical-Trials.ru/, a Russian-based, Russian language database that allows public searches of clinical trials approved in the Russian Federation. According to the website, http://Clinical-Trials.ru/ collects information about clinical trials conducted in Russia from various sources and provides this information as a service to patients who are potential participants in clinical experiments and also their doctors, relatives and friends searching for research participation opportunities. The website additionally characterises itself as a service for the organizers and sponsors of clinical trials to submit announcements of their trials “in order to recruit patients.” Because the database only aims to advertise trials that are or may be of interest to potential research participants, unlike registries such as http://ClinicalTrials.gov/, which also serve as public depositories of information about clinical trials, Clinical-Trials.ru provides researchers, sponsors and clinical research organisations with a centralised platform to inform physicians and prospective participants in detail about trial participation opportunities without simultaneously creating permanent public records of the trials or of elements of their protocols.

A search was performed to identify all USA-sponsored phase II–IV trials listed on http://Clinical-Trials.ru/. For each Russian trial identified in this initial search, http://ClinicalTrials.gov/ was searched by study ID number (if available), intervention name(s), sponsor's name, key words from the advertisement, or various combinations of these to determine whether the trial was also registered in this registry. For those trials that were advertised in Russia and registered in the USA, data were extracted from http://Clinical-Trials.ru/ on the trial phase and the numbers of participants sought in Russia, and data were collected from http://ClinicalTrials.gov/ on the start and end dates and on the number of participants sought across the entire trial. Results from the http://ClinicalTrials.gov/ search were then compared with the original Russian advertisements to determine whether the US registration contains Russia as a study location.

The adequacy of study location registration for each trial was coded for statistical analysis as follows. A score of 1 was given to trials that were properly registered as conducting research in Russia, that is, listing the Russian Federation as the study location on http://ClinicalTrials.gov/; a score of 2 was given to registrations that contained study sites but did not specify the Russian Federation among them; and a score of 3 was given to trials that did not specify any study locations in their http://ClinicalTrials.gov/ registration. Taking this as an ordinal variable response, the proportional odds model was used to determine whether (a) phase, (b) the total expected enrolment, (c) the total expected Russian enrolment, (d) age of the trial and (e) length of the trial were factors in predicting whether trials were adequately registered. Given that there were only 2 phase IV trials, phase III and IV trials were combined here. Ordinal regression analysis was performed using PROC LOGISTIC available in the statistical software package SAS (V.9.1).

Results

Eight hundred and forty-seven trials were listed on http://Clinical-Trials.ru/ at the time of data collection. Of these, 249 were USA-sponsored research on products produced by US companies. Thirteen of these were phase I trials and were excluded because regulations governing the registration of phase I research is different from those governing phase II–IV research. Of the 249 trials, 18 were not registered in http://ClinicalTrials.gov/. Therefore, 218 trials met the criteria of being USA-sponsored phase II–IV trials advertised on http://Clinical-Trials.ru/ and registered on http://ClinicalTrials.gov/. Of the 218 trials meeting these criteria, 79 were phase II trials, 137 were phase III trials, and 2 were phase IV trials. The number of participants sought in Russia ranged from 5 to 2300, with a mean of 108 participants sought per trial. Russian participants sought constituted a mean of 39.3% of the total expected enrolment per trial. The combined number of participants sought in all Russian-based research was 23 544. The mean trial length was 27.8 months, with an average start date of April, 2008 and an average end date of July, 2010. Of the trials, 77% (168) were enrolling participants or were active at the time of data collection.

Most (92.8%) of the USA-sponsored phase II–IV trials advertised on http://Clinical-Trials.ru/ were also registered on http://ClinicalTrials.gov/. Of the 218 trials meeting all of the criteria for this study, 69 trials did not disclose Russia as a study location on the US registry. Of these trials, 41 did not have any entry in the study locations data element, and 28 listed some study locations on the registry but did not specify Russia among them.

Statistical analysis

Of the 218 trials meeting the study criteria, two registered an expected total enrolment of 20 000. These two outliers were determined to have a biasing effect on the modelling and were, therefore, excluded from analysis. An additional 19 trials could not be included in the modelling because of missing data in the registry or inconsistencies between the advertisement and the registration. Of these, 7 trials did not provide start and/or end dates on http://ClinicalTrials.gov/. For the other 12, the advertisement on http://Clinical-Trials.ru/ provided larger numbers for the expected enrolment in the Russian component than http://ClinicalTrials.gov/ listed for the entire trial's expected enrolment. Therefore, 197 trials were used in the statistical analysis.

Of the various factors compared with the adequacy of study location registration (phase, total expected enrolment, total expected Russian enrolment, trial age and trial length), only trial length was significant at 95% (p=0.016). Longer trials were more likely to specify Russia as a study location. An increase of 30 days (roughly one month) in trial length increased the chance that an advertised trial would specify Russia as a study location in its http://ClinicalTrials.gov/ registration with an odds ratio of 1.028 (confidence interval, 1.005 to 1.051).

Discussion

The key finding was that nearly one-third (32%) of the trials listed in the Russian database and registered on http://ClinicalTrials.gov/ did not disclose in their registration that research was being conducted in the Russian Federation. This contrasts sharply with previous research on the completeness of mandatory data submissions to http://ClinicalTrials.gov/. In 2005, Zarin et al14 looked at patterns of completion for the mandatory field intervention name and the (then) optional field primary outcome. They found that 100% of the registrations examined had completed the required element and that clinically uninformative or non-specific entries in this field dropped from 10% to 2% with the implementation of the International Committee of Medical Journal Editors policy. Ross et al15 examined the registration for the mandatory data elements identification number, title, primary sponsor, study official, design, type, phase (if interventional), intervention, condition and population studied, along with the following optional data elements: enrolment, trial start and end dates, primary and secondary outcome measures and publication for all phase 2, 3 and 4 trials registered between December 1, 1999 and June 8, 2007. Again, they found that nearly 100% of the trials completed all mandatory fields, although they noted that entries for study official often varied in specificity. Both studies found problems, but insofar as they provide analysis regarding the completeness of mandatory data fields in trial registrations, both suggest that investigators are meeting registration requirements. A recent study by Mathieu et al20 reflects somewhat less favourably on the completeness of trial registrations generally, finding that 39 of 234 (12.1%) of the 234 registrations they examined had no or unclear description of the primary outcome or had stated inconsistencies regarding the timing of assessment, although this study was not limited to registrations in http://ClinicalTrials.gov/. Although methodological considerations prevent straightforwardly comparing the present finding with those of these earlier studies, the apparent differences do suggest that inadequate registration of study locations may occur at a higher rate than does inadequate registration of other mandatory data elements. One possible explanation is that the previous studies focused only on whether useful data were provided in a registration data field rather than also comparing information supplied to the registry with those in external data sources that could help identify inaccurate or incomplete submissions to the registry.

The use of such external comparative data here, however, cannot alone explain the apparent higher rate of non-compliance with study location registration requirements than other mandatory registration requirements suggested by the present study. By even restricting the analysis to whether information was entered in the trial's registration, 19% of the USA-sponsored trials advertised in Russia can be identified as inadequately registered for study location. This still indicates a much higher rate of non-compliance for study location registration than the near perfect rates that were found by the previous studies for other mandatory data elements submitted to http://ClinicalTrials.gov/. This suggests that researchers and sponsors might be less willing to disclose study locations than other data regarding their research, although this is tentative and highly speculative and the present study was not designed to confirm this. More research needs to be done to determine why researchers and sponsors do not register study locations.

Another significant finding was that the length of a trial was predictive of whether Russia was registered as a study location. Longer trials were more likely to have specified Russia as a study location. This was particularly surprising because the age of the trial was an insignificant variable. Thus, the most obvious explanation for the significance of trial length to adequate registration of the Russian study location (ie, that longer trials are more likely to be adequately registered because they provide lax registrants more time to complete their registrations) is unlikely to account for this result. That the age of the trial was not significant suggests that retrospective registration of study locations probably does not occur frequently. However, because this study used research advertisements to identify trials in Russia, it is limited to examining ongoing or recently completed research. Because researchers and sponsors have the opportunity to go back and update the registrations at later times, an examination of older trials than those maintained by http://Clinical-Trials.ru might find that age is a significant factor in study location registration and that adequate study location overall does occur at higher rates than it does for trials that are currently advertised. In any case, even if retrospective registration of study locations is a prevalent practice, given the distinctive functions of study location registrations, there can be no justification for researchers and sponsors failing to register study locations before beginning the recruitment of participants.

Conclusion

Although the full and adequate registration of study locations for clinical trials is an important legal, ethical and professional requirement, there has been little research examining the quality of study location disclosure in trial registrations. This study found that current and recent USA-sponsored trials being advertised to patients and physicians in the Russian Federation often do not disclose Russia as a study location in their registrations on http://ClinicalTrials.gov/. Such a tendency undermines patients' ability to access information about potentially beneficial research opportunities, limits access to information, which may be important for identifying or avoiding ethically questionable practices, and may frustrate the ability to interpret the results of research.

Acknowledgments

The authors thank Daria Eremina for her assistance with the statistical analysis and David Resnik, Sean Philpott, Kwinn Doran and Alexander Zelikin for their comments on the manuscript.

References

View Abstract

Footnotes

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.