Science and Technology CommitteeJoint written evidence submitted by Cardiff University School of Medicine and Cardiff and Vale University Health Board

1. 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?

1.1 Anything that makes the setting up and undertaking of clinical research should be applauded. The changes may marginally help low risk drugs or biosimilars, but we have concerns that they do not address the fundamental problem of extremely costly overregulation. Every small change still has to go through a four week review process and the paper trail will be as immense. The plan to introduce a “lighter regimen” for low risk studies is welcome but the proposals could have gone further. The restriction that “... investigational medicinal products are used in accordance with the terms of marketing ...” could have been more generous. We have a phenomenal amount of knowledge of drugs by the time they have gone off patent and a less regulated way of using a drug in a slightly different way to the “terms of marketing” should be accepted, especially as such a change would still have to be approved by a REC.

1.2 For already licensed medications with many years of usage in clinical populations there is still a level of approvals and monitoring burden in excess of that needed for safety purposes. This hampers our ability to ensure that what is used in practice is evidence based. We are still in the situation where a clinician can chose between treatments in a relatively arbitrary way (possibly influenced by drug representative), but if they were to randomise between two treatments they would have to go through approximately six months of approvals. The example given in Goldacre’s book is of comparing two statins, but there are similar prescribing choices being made every day on a poor evidence base and we cannot address this unless we find a new governance approach for this type of study.

1.3. There remains a view that a fundamental re-think and re-engineering of the approval and monitoring process is required and until this is done, opportunities for improved patient care and economic income will continue to be lost. One colleague favours a model of approval of Clinical Trials and monitoring by a “Responsible Officer” (like a building of fire safety inspector) as is done for many if not most other high risk activities (radiation protection, data protection, hygiene etc) but there may be other options. A major concern is the delay at individual NHS Organisation R&D offices. Monitoring and publishing individual institution performance would be likely to help.

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

2.1 The HRA is positively perceived but felt to be moderately effective with no major impact in Wales to date. It is felt that some studies that have been adopted—certainly in the cancer field—have been poorly scrutinised and are virtually impossible to deliver in large parts of the NHS.

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

3.1 It is felt that this goes on, more in smaller companies than major Pharma, but that it also happens in the academic sector. The danger to future patients is that if a trial fizzles out because of toxicity and it is not recorded a repeat could happen (eg Northwick Park). The option not to submit a final report should be removed. One local investigator reports having seen areas where the reporting clinician of a certain adverse event wishes to register a particular AE as “possibly or probably” related to a given drug but the sponsor hasn’t been completely in agreement.

3.2 There are a number of systematic reviews published which show that there is a significant difference in the proportion of trials published with pharma involvement that show a positive finding compared to those trials published with no pharma involvement. This has been taken to be evidence that pharma must be avoiding publishing studies. However there is little distinction made between early and late phase trials in these discussions and it is possible that more phase 3 trials are positive if there is a better decision making process made at phase 2. However if those early phase trials are not published then this still distorts the picture of what products are successful when we come to have an overview of the evidence in a systematic review.

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

4.1 Publication bias in all its guises is a big issue. The more benign kind, nevertheless devastating to the corpus of published research, is where “statistically significant” findings are preferred for publication. The more blatant kind is an organisation prohibiting publication of research that disfavours its product. The present research governance system should be helping a little but changes in the publication system so far haven’t really addressed this issue at all.

4.2 The obvious solution is a system in which research governance approval entails a mandatory commitment by the researchers that a basic report on the study—consisting of the final protocol and a brief summary of how the study progressed (such as the CONSORT flowchart) and of the main findings—would be placed in the public domain within a mutually agreed, appropriate timescale. The appropriate placing would be a freely accessible website, a single one to cover all the approved research in the health domain in the UK (?HRA). The final protocol would be published at the start of the study, the summary of progress and findings would be published later and would have a link back to the protocol. Both would have the same reference code which would identify the year, the competent authority, and a serial number.

4.3 Researchers would normally be expected to seek publication of a definitive article in a peer-reviewed publication. When this occurs, clear links should be made in both directions between the publication(s) and the material for the study on the database described above. The R&D authority would follow up the investigators at six-monthly intervals from a pre-agreed date a few months after the study close for say three years, to ascertain progress towards this objective. If this has not been achieved within this time period, the study would then be flagged on the database as “not published within three years”—which, if left as it stands without any explanation, would be construed by the research community as a black mark.

4.4 An exception, where one would not expect to see results published, are studies that collapsed for some valid, unforeseen reason such as withdrawal of a medication from approved use, actual recruitment rate grossly below what was anticipated, or death or serious incapacity of the main researcher to complete the project. Here, we would expect a summary on the database, linked to the approved protocol—this summary to be as informative as possible, including regarding lessons to be learnt from the failure of the study where this is applicable.

4.5 The scrutiny of a particular study largely resides with the Sponsoring company, MHRA and in the end Referees of the particular journal the Chief Investigator chooses to submit the manuscript to. Sponsoring companies seem to vary in their particular vigilance but the MHRA are very sound in their work. Following publication some journals allow quick, free access but many do not. Whether it would be possible to only License drugs if the results are published in free to access journals is open to debate but this would certainly broaden access.

4.6 The sponsor should be responsible for ensuring that trials are registered and that the results are made available (preferably open access) and that there is a mechanism for access to original datasets for independent verification of results and conclusion. While publication of results is laudable, the strength of the evidence base would be increased if there was sufficient scrutiny of the methodology used.

4.7 To take a concrete example, in trials of cholinesterase inhibitors in Alzheimer’s disease, many of the original trials on which decisions were made failed to follow-up patients who ceased therapy. They then imputed results using last observation carried forward; ie the last result while on treatment was used following withdrawal. In a declining disease this will artificially bias results in favour of the treatment with greater toxicity; and the active treatment in these trials suffered greater early dropout. If the methodology is flawed in this way, journals (say those which find it difficult to get access to high quality statistical review) will generally accept the paper, meaning that clinicians, who should not be expected to understand this issue in detail may gain an inflated opinion of a treatment’s efficacy; alternatively, the journal will reject on the grounds of scientific value, meaning that the trial is lost forever.

4.8 This is not a simple problem to solve; this sort of issue would probably be caught at the funding stage by one of the larger funding councils; but in trials that do not go along that route, there may be a need for greater scientific review at some point. Conversely, data should be made available following publication as a safeguard—but again one needs to guard against the production of “results” of spurious validity obtained by looking at subgroups of data, or using the sort of analytical methods given above. Clearly independent scientific review of proposals is required—but resources would need to be found; and an arbitration system introduced to stop people with a vested interest, faced with a negative trial, being allowed to massage or dredge the data for spurious signs of hope.

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

5.1 We are not aware that any given country is particularly superior to the UK in this respect. We consider this to be a global issue, not least because evidence doesn’t stop at our country’s borders.

6. Conflict of interest

6.1 We have no conflict of interest to declare.

February 2013

Prepared 16th September 2013