Science and Technology CommitteeWritten evidence from PLOS

Executive Summary

1. Clinical trials on humans cannot be considered as private undertakings, since they require the participation of human volunteers. However, it has become the norm for information generated from commercially sponsored trials to be held as private by default with the release of much of this information happening only according to commercial needs and in way that is dictated by the sponsors. There is now substantial evidence of harm as a result of this withholding of clinical trial data.

2. Mechanisms are in place or under discussion that could ensure all clinical trials are tracked and data from them are made available. However, these mechanisms are not currently complied with, legally mandated nor sufficiently enforced.

3. This is an international problem, but one in which the UK could usefully show leadership by mandating registration of all trials in a World Health Organization (WHO)- or International Committee of Medical Journal Editors (ICMJE)-approved registry, prospectively requiring reporting of all data from ongoing and future trials within a specific time frame after completion, as well as requiring release of data from previously completed trials.

Background on PLOS

4. PLOS is a not-for-profit organisation headquartered in San Francisco, USA with an office in Cambridge, UK, with a mission to transform scholarly communication. PLOS publishes peer-reviewed research papers in the science domain with a focus on biomedical sciences.

5. PLOS publishes seven journals: two highly selective journals, PLOS Medicine and PLOS Biology; four community journals, which are based on a publishing model that is similar to journals published by scholarly societies and focus on specific domains of Computational Biology, Genetics, Neglected Tropical Diseases, and Pathogens; and PLOS ONE, the world’s largest journal, which publishes across science and medicine.

6. PLOS applies the Creative Commons attribution license (CC-BY[1]) by default to all the content it publishes. Under this Open Access (OA) agreement, all the papers it publishes are freely and immediately available to read and, crucially, to reuse.

7. Since publishing its first articles in 2003, PLOS has seen phenomenal growth. In 2012, PLOS ONE published more articles registered in the PubMed bibliographic database as being funded by the Wellcome Trust, Medical Research Council (MRC), Biotechnology and Biological Sciences Research Council (BBSRC), Cancer Research UK (CRUK), Canadian Institutes of Health Research, and the US National Institutes of Health (NIH), than any other journal. PLOS published 11% of all Wellcome Trust research in 2012, 9.3% of CRUK, 8.5% of BBSRC, and 6.9% of MRC. PLOS has published many hundreds of trials since it began publishing in 2003.

8. PLOS is a leader in high-quality peer-review processes. All of the journals are peer reviewed, and PLOS applies the highest ethical standards to the peer review and publishing process. For instance, unlike most other biomedical publishers, PLOS journals do not accept advertising for drugs or devices. PLOS has also taken a leadership role in promoting higher reporting standards and reproducibility of research.

Declaration of Interests

9. PLOS publishes clinical trials in several of its journals. We have specifically stated we are interested in publishing clinical trials regardless of outcome; this and any mandate for clinical trial reporting are likely to lead to increased numbers of papers being submitted to our journals and potentially more income. To provide Open Access, PLOS journals use a business model in which expenses—including those of peer review, journal production, and online hosting and archiving—are recovered in part by charging a publication fee to the authors or research sponsors for each article they publish. The fees vary by journal. PLOS offers to waive or reduce the payment required of authors who cannot pay the full amount. Editors and reviewers have no access to information on authors’ ability to pay; decisions to publish are based only on editorial criteria.[2]

10. PLOS has also publicly supported the AllTrials campaign,[3] which calls for all trials to be registered and all results reported. Our position in this area is therefore well known.

The Committee Sought Submissions on the Following Questions

(a)Do the European Commission’s proposed revisions to the Clinical Trials Directive address the main barriers to conducting clinical trials in the UK and EU?

(b)What is the role of the Health Research Authority (HRA) in relation to clinical trials and how effective has it been to date?

(c)What evidence is there that pharmaceutical companies withhold clinical data and what impact does this have on public health?

(d)How could the occurrence and results of clinical trials be made more open to scrutiny? Who should be responsible?

(e)Can lessons about transparency and disclosure of clinical data be learned from other countries?

11. We address points c-e in this submission.

What evidence is there that pharmaceutical companies withhold clinical data and what impact does this have on public health?

12. There is substantial evidence that systematic bias exists within the published medical literature due to companies withholding clinical data.[4] This bias begins very early in the planning stages, in which trials and publications are planned in order for drugs to appear in as good a light as possible. “Positive” (eg, reporting results favourable to the drug) trials are targeted at specific high-profile journals, and “negative” trials are either not published or are targeted at lower-profile journals.[5],[6],[7].

13. The lack of negative trials in the literature was acknowledged by medical journals when in 2004 the ICMJE adopted a policy[8] that all trials should be registered in an approved registry before the first participant was enrolled, and unregistered trials would not be considered for publication in these journals. This policy was applied to all trials initiated after 1 July 2005. Since then many other journals, including all the PLOS journals, have adopted this policy. Unregistered trials can have their results alone submitted to ClinicalTrials.gov.[9] However, a 2012 study showed that registration is not universal, and that, even if registered, the study’s registration number is not always included in the journal report of a trial.[10] In addition, the quality of data included in the registry is highly variable and often not complete.[11] PLOS regularly receives submissions of unregistered trials; we reject these submissions along with a suggestion that the authors submit the results to ClinicalTrials.gov. We do not know the ultimate fate of these trials.

14. Many trials are not submitted for publication. A 2009 study showed that trials primarily sponsored by industry (40%, 144 of 357) were less likely to be published when compared with non-industry/non-government-sponsored trials (56%, 110 of 198).[12] A study of a national registry—the Netherlands Trial Register—showed that 48% of trials registered in the NTR had not been published at least two years after completion.[10]

15. Even if a trial is published, the results in a journal article are often a biased subset of the full dataset generated by the trial, and often skewed towards framing the intervention in a light favourable to the sponsor—an effect known as outcome reporting bias.[13]

16. Currently, journals—especially high-profile journals—are more likely to receive for publication trials which are “positive”. A study of six high-profile general journals showed that industry-supported trials were more frequently cited than trials with other types of support, and publication of industry-supported trials was associated with an increase in journal impact factors, a favourable outcome for the journals.[14] These journals obtain a large proportion of their income from selling reprints of industry-sponsored trials[14]—a potential source of bias.

17. A downstream effect of the lack of negative trials in the published literature is that secondary analyses, such as systematic reviews and meta-analyses, will not include all relevant studies and may therefore inadvertently conclude overall evidence of benefit that would not have been found had all the trials, regardless of outcome, been included. This effect has been known and demonstrated statistically for many years,[15] and has more recently been conclusively shown in a systematic review.[13]

18. There are numerous examples of harms to the public that have resulted from the withholding of trial data. For example, harms to patients treated in routine practice may arise if the clinical approach is informed by incomplete data or leads to treatment with drugs for which the harms may have been evident from prior trials, but which have not been made public; harms may also occur if patients are treated with drugs that are ineffective or less effective than other treatments, information which could have been evident from trial data, but which had not been made public. Two prominent examples are the increased risk of death from myocardial infarction from the use of Avandia (rosiglitazone),[16] which was not apparent in initially published studies; and the increased risk of breast cancer from the use of menopausal hormone replacement therapy Prempro.[17] In these cases, the existence of much data came to light only after court cases in which many drug company documents were released. Many other examples exist.

How could the occurrence and results of clinical trials be made more open to scrutiny? Who should be responsible?

19. To prevent the abovementioned sources of bias and potential patient harms, all steps in the information chain of clinical trials need to be as transparent and accessible as possible. This chain starts from the point of planning of trials and extends through to publication of trial reports in journals or elsewhere, and to the reports and data submitted to regulators.

20. Everyone involved in clinical trials therefore has a responsibility to examine the processes for which they are responsible to ensure that their part of the chain is transparent and the associated data are available. Thus, funders and sponsors of trials, those who run and report trials, those who publish trials and those who oversee trial registration and drug regulation all have a role. Currently there is insufficient oversight and regulation to ensure complete transparency for some parts of this chain.

21. Clinical trial registration for all trials, even early phase, if mandated by funders, regulators and journals, and if properly followed up and enforced, will ensure that all trials that are started are adequately tracked.

22. Trial registration needs to be coupled with a mechanism for the reporting of all trial results. This reporting can occur in the form of journal articles, but need not exclusively to be done this way. Alternative mechanisms, which are adequately resourced and overseen, should be put in place to allow reporting of clinical trials into a database.

23. Journals should publish articles reporting clinical trials based on the value of the question asked and the soundness of the methodology, not the direction of the results. Journals should require that trials are well reported, according to the accepted CONSORT criteria,[18] should require submission and publication of protocols alongside articles reporting on trials and should have polices that require access to the data underpinning trials and provide a mechanism to link to such data.

24. In addition, the UK government should support the initiative of the European Medicines Agency (EMA),[19] which has stated that by 2014 it will release publicly clinical-trial data submitted as part of drug regulatory approval. This availability should, however, apply to drugs already submitted for approval, not just future drugs, and should include drugs for which approval was not granted.

Can lessons about transparency and disclosure of clinical data be learned from other countries?

25. Clinical trials are increasingly an international business. Although examples of good practices are available from other countries, it is also clear that regulatory differences among countries have in the past led to lack of transparency. Ideally, any new initiatives will not be limited to just one country.

26. One specific example is the requirement in the US, under Food and Drug Administration Amendments Act (FDAAA) legislation, for mandatory reporting on ClinicalTrials.gov of specified trials of summary clinical trial results within one year of completion of the trial. A 2012 study showed that only 163/738 (22%) had reported results in this way. Enactment of legislation without enforcement is obviously ineffective.[20]

Recommendations for Government Action

27. Require prospective registration of all clinical trials of all phases in a WHO- or ICMJE-approved registry[21] with specific penalties for non-compliance.

28. Require reporting of all clinical trial results within a specific time frame after trial completion in a properly resourced public site with enforced penalties for non-compliance.

29. Support EMA initiatives to require release of clinical trial data related to all drugs submitted for marketing approval, both approved and unapproved, and for historical as well as forthcoming trials.

February 2013

References/Notes

1. PLOS (2013). Open-access license. Available: http://www.plosmedicine.org/static/license. Accessed 20 February 2013.

2. PLOS (2013). Publication fees. Available: http://www.plos.org/publish/pricing-policy/publication-fees. Accessed 20 February 2013.

3. AllTrials (2013). All trials registered—all results reported campaign. Available: http://www.alltrials.net. Accessed 17 February 2013.

4. Turner E H, Matthews A M, Linardatos E, Tell R A, Rosenthal R (2008). Selective publication of antidepressant trials and its influence on apparent efficacy.N Engl J Med 358:252–260. doi:10.1056/NEJMsa065779

5. Vedula S S, Li T, Dickersin K (2013). Differences in reporting of analyses in internal company documents versus published trial reports: Comparisons in industry-sponsored trials in off-label uses of gabapentin. PLoS Med 10(1): e1001378. doi:10.1371/journal.pmed.1001378

6. Sismondo S (2007). Ghost management: How much of the medical literature is shaped behind the scenes by the pharmaceutical industry? PLoS Med 4(9): e286. doi:10.1371/journal.pmed.0040286

7. Melander H, Ahlqvist-Rastad J, Meijer G, Beermann B (2003). Evidence b(i)ased medicine—Selective reporting from studies sponsored by pharmaceutical industry: Review of studies in new drug applications. BMJ 326: 1171. doi:10.1136/bmj.326.7400.1171

8. De Angelis C, Drazen J M, Frizelle F A, Haug C, Hoey J, et al (2004). Clinical trial registration: A statement from the International Committee of Medical Journal Editors. N Engl J Med 351: 1250–1251.

9. ClinicalTrials.gov (2013) Trial registry. Available: http://www.clinicaltrials.gov. Accessed 20 February 2013.

10. van de Wetering F T, Scholten R J P M, Haring T, Clarke M, Hooft L (2012). Trial registration numbers are underreported in biomedical publications. PLoS ONE 7(11): e49599. doi:10.1371/journal.pone.0049599

11. Viergever R F, Ghersi D (2011). The quality of registration of clinical trials. PLoS ONE 6(2): e14701. doi:10.1371/journal.pone.0014701

12. Ross J S, Mulvey G K, Hines E M, Nissen S E, Krumholz H M (2009). Trial publication after registration in ClinicalTrials.gov: A cross-sectional analysis. PLoS Med 6(9): e1000144. doi:10.1371/journal.pmed.1000144

13. Dwan K, Altman D G, Arnalz J A, Bloom J, Chan A-W, Cronin E, Decullier E, Easterbrook P J, von Elm E, Gamble C, Ghersi D, Ioannidis J P A, Simes J, Williamson P R (2008). Systematic review of the empirical evidence of study publication bias and outcome reporting bias. PLoS ONE 3(8): e3081. doi:10.1371/journal.pone.0003081

14. Lundh A, Barbateskovic M, Hróbjartsson A, Gøtzsche PC (2010). Conflicts of interest at medical journals: The influence of industry-supported randomised trials on journal impact factors and revenue—Cohort study. PLoS Med 7(10): e1000354. doi:10.1371/journal.pmed.1000354

15. Egger M, Smith G D, Schneider M, Minder C (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ 315: 629. doi:10.1136/bmj.315.7109.629

16. Nissen S E, Wolski K (2007). Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med 356: 2457–2471. doi:10.1056/NEJMoa072761

17. Fugh-Berman A J (2010). The Haunting of medical journals: How ghostwriting sold “HRT”. PLoS Med 7(9): e1000335. doi:10.1371/journal.pmed.1000335

18. CONSORT (2013). Home page. Available: http://www.consort-statement.org. Accessed 17 February 2013

19. European Medicines Agency (2013) Release of data from clinical trials. Available: http://www.emea.europa.eu/ema/index.jsp?curl=pages/special_topics/general/general_content_000555.jsp&mid=WC0b01ac0580607bfa. Accessed 27 February 2013.

20. Prayle A P, Hurley M N, Smyth A R (2012). Compliance with mandatory reporting of clinical trial results on ClinicalTrials.gov: Cross sectional study. BMJ 344: d7373. doi:10.1136/bmj.d7373

21. WHO (2013). International Clinical Trials Registry Platform. Available: http://www.who.int/ictrp/network/primary/en/index.html. Accessed 20 February 2013.

Prepared 16th September 2013