Science and Technology CommitteeWritten evidence submitted by the UK Research Integrity Office

Scope of Submission

1. The UK Research Integrity Office (UKRIO) is submitting evidence on two particular questions:

Question 3: What evidence is there that pharmaceutical companies withhold clinical trial data and what impact does this have on public health?

Question 4: How could the occurrence and results of clinical trials be made more open to scrutiny? Who should be responsible?

2. Question 3: What evidence is there that pharmaceutical companies withhold clinical trial data and what impact does this have on public health?

3. There is a very substantial body of evidence that many clinical trials, about half, are not published. This evidence is well summarised in Ben Goldacre’s book(1) and a recent editorial in the BMJ.(2) UKRIO has always stressed—through its practical advisory service, its education and training, and its publications—that all research should be conducted to the highest standards of honesty, accuracy, integrity and accountability. We would expect research organisations to share our ambition to uphold high standards and we find the examples of bad behaviour in Bad Pharma all the more disappointing. These are neither historical problems nor isolated examples. Far more should have been done to address the non-publication of clinical trials and the selective reporting of data and results. Regulatory, professional, funding and other organisations now have an opportunity to take a clear stand on these issues and bring about real change.

4. It is important to note that it is not only the results of clinical trials that are not published but also the results of all scientific studies. This is an important problem for all of science because it means that conclusions are being reached on only part of the evidence—and almost certainly a biased part.

5. There is evidence, presented in the BMJ editorial,(2) that trials funded by pharmaceutical companies are less likely than trials funded by, for example, Government to be published, but, as we have said, it is a substantial problem for most research, although it is interesting to note that the UK Health Technology Assessment programme (HTA) has very high rates of publication. This may be related to its policy of withholding part of the funding until the work is published and also of producing its own reports rather than requiring publication in peer-reviewed journals.(2)

6. The emphasis in the current debate is on the publication of analysed results but equally important is the publication of the full (raw) data of the trial or study. It has not been normal in science to publish full data, but the arrival of the internet has made it possible to publish full data sets. Research funders are now beginning to require this, partly to allow full examination of studies but more to allow reuse of the data, which can bring substantial scientific, social, and economic benefits. A systematic review of trials of a drug that uses individual patient data will be much superior to a systematic review that uses only summary data. So it is important to push not just for the publication of results but for the publication of full anonymised data.

7. We have known for a long time that trials funded by pharmaceutical companies are more likely to have results favourable to the company than publicly funded trials.(3,4) We know too that studies with negative results are less likely to be published than studies with positive results and that studies with positive results are likely to be published more than once.

8. The consequence of trials not being published combined with a bias in those that are published is that patients and clinicians are misinformed about the balance of benefit and harm that might be expected from a drug. The usual distortion is that drugs will seem to be more effective and less harmful than they are in reality. The difference between the evidence and the actuality may in some cases be substantial.

9. The strongest evidence we have on this difference is with antidepressants, which are prescribed on a massive scale in Britain. As Ben Goldacre describes in his book, systematic reviews of some antidepressants that use all trials reveal that they have little or no effectiveness and substantial side effects.(1)

10. Because many factors go into the prescribing of drugs it is hard to know how much harm to public health results from the distorted information. It may well be that drugs are overprescribed: benefits may be less and harms greater than expected.

11. Question 4: How could the occurrence and results of clinical trials be made more open to scrutiny? Who should be responsible?

12. It seems relatively uncontroversial to insist that the results and full data of all trials and, indeed, all scientific studies, particularly those funded with public money, should be published. Publication does not have to be in journals, and as publication of full data becomes the norm it will make more sense to publish on large databases rather than in journals. Indeed, current publishing practices mean that trials are often published in journals that allow access only to those with subscriptions or who pay. Pharmaceutical companies are keen to have their most positive results published in high profile journals that reach many prescribing doctors and carry great prestige. We have growing evidence that exciting results published in high profile journals are more likely to be misleading than the results published in less prestigious journals.(5,6) This is a further source of bias in prescribing information for clinicians, who receive much of the information from these journals.

13. We support the registering of all trials on publicly available databases, and we support the move to publish the protocols of all trials. There is considerable evidence showing important changes between protocols and published studies, and some of these changes are designed to make drugs being trialled appear to be more effective than they actually are.(7) Clinical trial registration has a critical role to play. It is the start of a process that should ensure publication of trial results, and in a complete and unbiased way. Currently trial registration is an incomplete process and inconsistently enforced. If a more uniform approach to trial progress tracking could be achieved and those failing to deliver could be subject to some form of penalty, then we might make progress.

14. Many parties—authors, employers, funders, ethics committees, regulators—might potentially play a role in ensuring that all trials and studies are published, but there is clearly a danger that something that is everybody’s responsibility becomes nobody’s responsibility. We suggest that the prime responsibility should lie with the funders of the research, and they should develop and implement processes for ensuring that they: register all studies that they fund; follow up the studies; and insist that they are published. The mechanism for achieving the latter could be withholding the last part of the funding until publication has taken place, following the example of the HTA.

15. Although prime responsibility should rest with funders, there needs to be regulatory oversight. As the main concern is with drug trials and because much research is international, the European Medicines Agency is probably best to provide that oversight. It should almost certainly work with the US Food and Drug Administration (and other international drug regulatory bodies) to fulfil this oversight role. Any requirement to publish a clinical trial within a certain period following its completion would be worthless unless it was actively enforced. A new rule on its own would not be enough; its implementation would need to be monitored and action taken against those who choose not to comply.

16. In addition to these statutory roles it will be useful to write into codes of scientific conduct and employment contracts the duty to publish the results of all scientific studies. Publishing within 12 months of the completion of a study might be the aim but could be quite tough to achieve in practice, as the publication processes of some journals are slow. Setting a deadline of publication within 18 months of study completion may be more realistic.

17. Ethics committees might also develop mechanisms to ensure that all the research studies they approve are published, though they would need to be better resourced in order to carry out such a role without their other work suffering. Decisions would need to be made on a case-by-case basis and all research should be scrutinised in this way, whether it originated from a commercial organisation or elsewhere. Some argue that ethics committees should be able to refuse to approve research projects proposed by researchers who have not published results from earlier studies or trials.

18. In UKRIO’s experience, policies, systems and even contracts are not sufficient on their own to effect real change. If left unsupported they can lead to a “tick box” exercise, rather than becoming an integral part of the research practices of an organisation. Instead, they must be:

supported with appropriate resources, training, dissemination activities and sources of help;

monitored for effectiveness and periodically reviewed, informed by feedback from researchers, participants and others; and

promoted by senior research and managerial staff in institutions, encouraging researchers to engage critically with standards for good practice in the publication and dissemination of research and with other issues of research integrity.

Declaration of Interests

19. This submission draws upon the views of the Trustees, Advisory Board and staff of UKRIO. These include persons who have: undertaken medical/scientific research, including clinical trials; worked as editors of academic journals which publish medical/scientific research, including clinical trials; and/or held senior roles in institutions such as universities which undertake medical/scientific research, including clinical trials.

20. UKRIO is funded by subscriptions from UK public sector or charitable research organisations, including over 30 universities. It has received funding from bodies that fund or undertake medical/scientific research, including clinical trials. None of the bodies which fund or support UKRIO had any input into the content of this submission.

21. UKRIO has never received any funding from private sector organisations which conduct pharmaceutical/medical research or clinical trials. During our first phase, which ran from 2006 until mid-December 2010, UKRIO received £10,000 in funding from the Association of the British Pharmaceutical Industry (ABPI).

22. UKRIO is a signatory of the All Trials petition for the publication of clinical trials results. Further information on the petition can be found at www.alltrials.net.

About the UK Research Integrity Office

23. The UK Research Integrity Office (UKRIO) is an independent charity, offering support to the public, researchers and organisations to further good practice in academic, scientific and medical research. We promote integrity and high ethical standards in research, as well as robust and fair methods to address poor practice and misconduct. We pursue these aims through our publications on research practice, the support and services we provide to organisations, our education and training activities, and by providing expert guidance in response to requests for assistance.

24. Since 2006, UKRIO has provided independent, expert and confidential support across all disciplines of research, from the arts and humanities to the life sciences. We help all involved in research: researchers, research organisations and members of the public, including patients and research participants. UKRIO covers all research sectors: higher education, the NHS, private sector organisations and charities—wherever the research affects the public good. No other organisation in the UK has comparable expertise in providing such support in the field of research integrity. We welcome enquiries on any issues relating to the conduct of research, whether promoting good research practice, seeking help with a particular research project or investigating cases of alleged fraud and misconduct.

25. We are not a regulatory body and have no formal legal powers. UKRIO fills gaps between jurisdictions, where no overall regulation might apply, and helps to direct researchers, organisations and the public to regulatory bodies when issues fall within their jurisdiction. We help institutions achieve high standards when they have to manage challenges to research integrity and support individuals faced with bad practice. Our advice and guidance emphasises the good practice that runs across all research disciplines and all regulatory remits. In this way our role complements that of regulatory bodies for research and supports the work of Government and research funders.

February 2013

References

1. Goldacre B. Bad Pharma. Fourth Estate, 2012.

2. Chalmers I, Glasziou P, Godlee F. All trials must be registered and the results published. BMJ: British Medical Journal. 2013 Jan;346. Available from: http://dx.doi.org/10.1136/bmj.f105.

3. Lexchin J, Bero L A, Djulbegovic B, Clark O. Pharmaceutical industry sponsorship and research outcome and quality: systematic review. BMJ. 2003 May;326(7400):1167–1170. Available from: http://dx.doi.org/10.1136/bmj.326.7400.1167.

4. Lundh A, Sismondo S, Lexchin J, Busuioc O A, Bero L. Industry sponsorship and research outcome. Cochrane database of systematic reviews (Online). 1996;12. Available from: http://dx.doi.org/10.1002/14651858.MR000033.pub2.

5. Young N S, Ioannidis J P A, Al-Ubaydli O. Why Current Publication Practices May Distort Science. PLoS Med. 2008 Oct;5(10):e201+. Available from: http://dx.doi.org/10.1371/journal.pmed.0050201.

6. Ioannidis J A. Contradicted and Initially Stronger Effects in Highly Cited Clinical Research. JAMA. 2005;294(2):218–228. doi:10.1001/jama.294.2.218.

7. Lundh A, Krogsboll L, Gotzsche P. Sponsors’ participation in conduct and reporting of industry trials: a descriptive study. Trials. 2012 Aug;13(1):146+. Available from: http://dx.doi.org/10.1186/1745–6215–13–146.

Prepared 16th September 2013