Select Committee on Health Minutes of Evidence

Examination of Witnesses (Questions 420 - 439)



  Q420  Dr Naysmith: We will come to some questions later on about the role of the industry and marketing the product.

  Professor Vallance: Can I make a comment on this? It seems to me that the job of saying which of the trials are missing and what the patients want that are not drug-related trials is a matter for the NHS to set the agenda and to say, "where are the gaps and what should we be addressing in a clinical trial?" It is unrealistic to expect industry to go out and want to do that.

  Q421  Dr Naysmith: I think we will come to interesting questions later on related to what the role of the Department of Health should be versus the Department of Trade and Industry and whether it should operate through the Department of Health. I want to move on to something else. I have been very impressed in this inquiry by witnesses who have been able to tell us about the practice of ghost writing, someone else writing papers, sticking their names on a paper that they know very little about. I am sure that Richard, as a scientist, will not think that is a very good idea, but does it happen widely in the industry? I am really talking about clinical trials rather than original research.

  Sir Richard Sykes: I am sure it happens because academics are very, very busy people, and they prefer to do research than spend a lot of time writing papers. If the industry puts forward a method of relieving them of that chore, then I am sure that that does happen throughout the industry. That would be true generally. Is it a good idea? I think it can be, as long as everybody is in agreement with what is written at the end of the day, the results and what they are.

  Q422  Dr Naysmith: You would never have done it as a scientist, would you?

  Sir Richard Sykes: It would never have been within anything that I did as a scientist to do that. I am not a clinical researcher, so I was never informed of that situation.

  Q423  Dr Naysmith: How much will the chief executive of a big British pharmaceutical company know about the details of the sort of things we are talking about now, and the kind of research that is being pursued? Is it seen in terms of one, two or three new drugs progressing, or would you be interested in things like incentives and disincentives for sales representatives and the pressures they face to meet targets; or would you be really interested in producing a drug that benefits humanity?

  Sir Richard Sykes: I must make it clear to everyone here that people go to work in the pharmaceutical industry generally—and particularly the researchers—because they really have a mission. They want to discover and develop something that is really going to benefit people, and the young people that go into industry today go with that desire. I have met very few people who did not want to achieve that end. The people working in industry work very closely with academia in the big companies, and they all have one objective at the end of the day. In my opinion, these are not people that are driven by greed, money or avarice; they are driven by the desire to be successful and to do something that is going to benefit people. When you get higher up in the organisation, those objectives change to some extent, but not the people doing the work. As a scientist and a chief executive, I obviously took a lot of interest in what was going on, because that was my desire as well; to make sure that we could develop things that really benefited people.

  Sir Iain Chalmers: May I add something on ghost writing and publication practices? One of the developments under Sir Richard Sykes's regime at GlaxoWellcome was that there was a publications unit headed by Dr Elizabeth Wager, which worked with some people in other companies developing good publication practices. One of the things that came up was ghost writing and gift authorship and so on. They developed good publication practice guidelines for the pharmaceutical industry; they are on the Web. When I wrote to the successor of Sir Richard Sykes, Jean-Pierre Garnier, congratulating him on supporting this activity through the successor company GSK not only did I not get a reply but Dr Elizabeth Wager and her team were sacked. The last time I looked only six companies world-wide had endorsed the good publication practice guidelines. I think it is a very, very serious situation. If companies are really genuine about wishing to improve publication practices then they should sign up to those guidelines.

  Q424  Mr Amess: Gentlemen, the Committee would be very interested in hearing what you think of the idea of setting up a public register of all clinical trials. If you think it is a good idea, can you tell us at what point trials should be registered? Should this register be operated by an independent authority, if there is such a thing as an independent authority—Richard over there calls himself an independent, but he is up to his neck in the grubby world of politics; but if you could find an independent authority why do you think it could be operated by the pharmaceutical industry?

   Sir Iain Chalmers: I have already said that I have been wittering on about this for nearly two decades.

  Q425  Mr Amess: I am sorry if I am repeating things which you have answered already.

  Sir Iain Chalmers: I apologise I have not yet given a detailed answer on this. Registering trials at inception is one of the steps needed to tackle this problem of publication bias. It has been proposed at least since 1986 as a way of addressing this problem. Very good progress has been made in some respects. For example, the trials that are funded by the NHS or by the MRC are now all registered on the meta Register of Controlled Trials set up by the publishing company, Current Controlled Trials, which has done an extremely good job. In addition there is a need to uniquely identify trials, in the way that books are identified uniquely. There is an ISBN scheme, which was introduced initially by two publishers, and then became the responsibility of the International Standards Institute. WHO has recently assumed responsibility for developing the International Standard RCTN number. It is a very important initiative which has now has international backing from WHO, and from the International Committee of Medical Journal Editors and from others. In terms of which trials should be added to that register, and indeed for which trials it should be required, I think it is reasonable for industry to say that in the early days of drug development, there are issues of commercial confidentiality that they will want to keep under wraps. However, at the time when studies start to be likely to influence patient care, those studies really do need to be registered prospectively and reported fully.

  Q426  Mr Amess: Who would judge that?

  Sir Iain Chalmers: One way of judging is to see if the study is randomised. If it is, people are really taking a serious interest in making sure that they are not misled by bias. I would say randomisation would be a good trigger for registration, although it has to be said that drugs are still being licensed on the basis of rather inadequate evidence, looked at from my perspective of uncontrolled case series, particularly in cancer.

  Q427  Mr Amess: You mentioned the need for registration of trials at inception. Can you explain to the Committee why you feel that is so important?

  Sir Iain Chalmers: Yes. If you wanted to address uncertainties about treatments you want to look at all of the relevant evidence that has been generated, and not just a sample of it, and certainly not a biased sample. Unless you identify, at inception, the studies that are addressing a particular question, you cannot know that you have identified all of the relevant evidence. There are examples of people being harmed, and certainly resources being wasted because disappointing studies have not been published. Trying to identify those studies in retrospect—in the mid eighties we wrote to over 40,000 clinicians around the world to try and flush out unpublished studies—does not work. You need proper registration of these studies so that you have a handle on them right from the beginning.

  Q428  Mr Amess: Gentlemen, do you have anything to add to your colleague's remarks?

  Professor Vallance: I agree with what Iain said. The cut-off point in a sense is when a trial is trying to establish a therapeutic end-point, then that should be in the public domain and should be registered. I think that is true for academic and industrial-sponsored trials.

  Sir Richard Sykes: I agree.

  Q429  Mr Amess: Gentlemen, do you think the current draft regulatory system is adequate? If it is not, what do you think is wrong with it? Might the MHRA take a greater role in addressing the problems of adverse drug events or post marketing clinical trials or, again, do you think it would be better to be handled by some sort of independent authority, if you can think of such an independent authority?

  Professor Vallance: There are a number of problems. MHRA, when it licenses a drug, is looking for the effect of that drug over and above placebo very often. I understand why that happens as a regulatory process, and that is in harmony with other places in the world, but it is not a question you want answered in terms of whether you want the drug available in the NHS. That is where NICE comes in. There are some boundary issues that need to be explored, and whether NICE has said in advance the criteria it wants to put in place to say something is a true advance over existing therapies.

  Q430  Mr Amess: What is your view of how it operates at the moment?

  Professor Vallance: NICE has done a lot of very good things, and in many ways it has worked tremendously well. It is often on the back foot when a trial shows an advance, because it is then left with asking how big an advance it is and whether it means something. The problem is that very often it does not mean very much, but they are always seen as denying—

  Q431  Mr Amess: So this links up; it is all written up.

  Professor Vallance: Yes, there is an expectation by patients and doctors and others. I would prefer to see at least as a pilot NICE looking at the idea of setting criteria as to what is an advance in a given therapeutic area.

  Q432  John Austin: In your evidence you have suggested there is a lot of information on the effects of drugs which is collected in the routine of clinical care. Could that be used as a base for evaluation?

  Professor Vallance: There is an immensely valuable database called the GP Research Database, which is routine clinical data collected on five million patients through computerised systems, and that will increase in the NHS. It is an under-utilised resource for looking at drug safety. There is good evidence it is under-utilised, and one of the reasons is that it has been incredibly expensive to access, and just about the only groups that have accessed it in the past have been in hospitals in the USA. UK data of direct relevance to safety in this country has been used mainly by groups in Boston. That database is now owned by the MRHA and it extremely valuable for drug safety, and there will be more of that coming along as NHS computer systems get put into place in hospitals.

  Q433  John Austin: What is the barrier now?

  Professor Vallance: Cost and to some extent expertise.

  Q434  Mr Amess: I do not think you addressed the point about the drugs.

  Professor Vallance: The database is a way to get that. You can get a lot of safety information and you can monitor safety much more accurately. I suspect that some of the issues around cost might have been picked up using that database.

  Sir Richard Sykes: I think you have to recognise that you can never win at this game. The MHRA or any regulatory agency gets a set of data. Usually you do this through clinical trials with from anywhere between 3,000 and 10,000 patients. If an adverse event is one in 20,000 or 30,000 or 50,000, you are never going to see it. There has got to be a process of making sure you have enough information to give an approval to have the drug into the clinic, but then there have got to be very clear monitoring processes for seeing that drug operate in a true market place, where now you are not selecting the patient who receives the drug but patients of a great genetic diversity are now receiving that drug. That, by definition, will produce adverse events. Remember that these drugs, at the end of the day, must be poisons, otherwise they would not be working. The body is a very complicated organism. It is all connected. If you inhibit one bit, you are going to inhibit something else somewhere else, and that is an adverse event, and that is the risk/benefit relationship at the end of the day. The more information we can get about this, the more we understand about the underlying mechanisms of disease and how those mechanisms are connected to other mechanisms. This will change over the next 20 years absolutely dramatically. That is where we have to make sure agencies like the MHRA are tied in to this modern technology so they can get better, more valuable information, to make better decisions.

  Q435  Mr Amess: You think they are the appropriate body.

  Sir Richard Sykes: As long as they change and develop with progress, then they should be fine.

  Q436  Dr Naysmith: Why should it be better in the future because it has not been all that good in the past? I know it will change with the different scenarios, but you are still going to have to regulate.

  Sir Iain Chalmers: I have a slightly different take on your question. It seems to me that unless you establish that a new treatment is either better or cheaper than what is already available, then worrying about side effects is displacement activity. We need better information, complete information on the effects of treatments and their benefits, in particular if better information about what difference drugs make to outcomes that matter to patients. This is very important because often they are evaluated for licensing in terms of outcomes that are completely meaningless to patients. If we had evidence that a drug is beneficial in these terms, and we can be reassured that there is no evidence that there is no publication bias, then that drug becomes interesting in terms of trying to make sure it does not have any unexpected side effects. The initial problem is the worst one; that we do not yet have good mechanisms for identifying drugs that are useful to patients, partly because of publication bias, but partly because the outcomes studied are either not sufficiently important to patients. I was listening to the Chair of the Psoriasis Association speaking yesterday, who is a lay person. He was saying that the studies of psoriasis do not measure the things that he and other sufferers of psoriasis rate as important, such as itchiness and pain. There are the fundamental earlier problems that we need to address before looking for side effects.

  Q437  Mr Amess: To help the Committee in reaching its findings, can you think of a good mechanism?

  Sir Iain Chalmers: I do not know how radical this is, but I would propose that there should be a provisional licence given to drugs on the basis of the kind of outcome measures and the type of follow-up that has been done in studies submitted to the MHRA; and that a decision about whether this new drug represented good value for the NHS should wait until there is more evidence about whether or not it affected outcomes that matter to patients.

  Q438  Chairman: Is that a model that applies anywhere else in the world?

  Sir Iain Chalmers: I do not know.

  Q439  Mr Burns: In parallel with your criticism of the industry you have said "denial of access to information held by the MHRA puts the interests of the pharmaceutical industry ahead of those of patients and prescribers", and you categorise it as "deplorable". Can you explain to the Committee what exactly you mean by "deplorable"?

  Sir Iain Chalmers: Let me give you two examples, both of them relating to pharmaceutical products. One relates to a treatment that has been used ever since the Japanese attack on Pearl Harbour for treating people who were severely burned or otherwise critically ill. The treatment involves giving a transfusion of human albumin solution. Ever since that time, until very recently, there has been no adequate assessment of whether the claim that this was a way of reducing the chances of those individuals dying was substantiated in good evidence. Human albumin solution has been used repeatedly in this country, and it was re-licensed in 1993. Yet a systematic review of all of the studies that provided information about death done in the mid 90s showed no evidence that albumin was helpful, and some worrying evidence that it might be harmful. Furthermore, there are mechanisms through which this harmful effect might be mediated. The reaction of the Medicines Control Agency to this news was to slightly modify the labelling, but to keep confidential the evidence upon which the drug had been re-licensed in 1993. It was not considered in the public interest to allow that to be made public, and I think this is indefensible. If you have a question about a widely-used medical product, which has been used for decades, you need to know what criteria the licensing authority, which is meant to be looking after our interests, has used to decide that the product but it was left to Australia and New Zealand to address the concern. Their study has failed to detect any advantage of this human albumin solution compared to salt water. As you can imagine, salt water is a good deal cheaper! There is some evidence that in a certain sub-group of patients albumin may be harmful, but the worry from the systematic review was not substantiated. However, we have been paying millions and millions of pounds for a way of resuscitating patients which is no better than salt water. That is one example. Another example is so-called evening primrose oil for eczema. This was given a drug licence, but in the early 1990s the Department of Health commissioned a systematic review of the evidence relating to the effects of this drug in eczema. The reviewers did an extremely good job, and included both unpublished data provided by industry and published data . They were unable to find any evidence that the drug was useful in eczema, except "in doctor-assessed itch"—"not patient-assessed itch", which raises an interesting question! The drug continued to be sold to the NHS, at a cost of about £7 million a year at the time the study was done. However, industry put pressure on the Department not to release the results of the systematic review. Two years ago however the MCA withdrew the licence on the grounds of lack of efficacy; but it did not make available the evidence that had led it first to license the drug, and then to withdraw it. Those are examples of the sort of secrecy that those outside the system have to put up with. That secrecy is not in the interests of the public.

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2005
Prepared 26 April 2005