Genomic Medicine - Science and Technology Committee Contents

Examination of Witnesses (Questions 500 - 519)


Professor Rory Collins, Professor Andrew Morris, Professor David Porteous and Professor Julian Sampson

  Q500  Lord Broers: Can we backtrack to a simple matter of sequencing. The data we are hearing—at least that I am hearing—are not consistent. There is a US website offering to sequence your DNA for $999.90; is that a complete sequencing? And is complete sequencing required? We have also learned in our papers that you can sequence the entire coding regions of BRCA 1 and 2 for perhaps £10 soon. So could you talk about that? Do we need a complete sequence or is much of the sequence irrelevant? And what is the price anyway?

  Professor Porteous: The price is tumbling—we have said that and I think we need to restate that. In a sense that is less important than why you are sequencing and what you are looking for. I would just like to echo Rory's comments by emphasising the value that perhaps we have not yet teased out yet adequately about the biobanks. You have heard in your evidence previously about the really remarkable success of the Wellcome Trust case control consortia, where large groups of investigators have come together, amalgamated their samples—collected some 2,000 or so cases of Type 2 diabetes or Crohn's disease or one or other disease—then scanned the whole genome and looked for signatures from genes that suggest that there might be a risk factor there. That has worked extremely well but it still only captured between 10 and 20 per cent of the total genetic liability that we know lies somewhere in the genome that has yet to be found. Therefore we have valuable but still partial information. So we need to do two things; we need to look for the remaining information, where it comes from, and we need to seek evidence of whether or not those findings made in one study are replicated in a second, independent study. Moreover—and this is where I think the biobanks come into play—it is very important that you validate those studies in the general population rather than simply in populations that have been acquired because they are a case of inflammatory bowel disease or a case of diabetes. What we want to know is how much do those genetic factors play in the risk of an individual later in life developing diabetes or later in life developing Crohn's disease or depression or any of the other psychiatric disorders that you have also taken evidence on. So I think that there is an attempt by a number of speculative companies to get rich quick. There will be people gullible enough to take up their offers of genetic testing, but I would have said that the chance of the results being meaningful is extremely low because the evidence base is far from complete.

  Q501  Baroness O'Neill of Bengarve: I would like to come on to the question of the information that you need. We have heard several times about differences between the way that centralised healthcare records are maintained in England, Scotland and Wales—so you are the perfect group of people to enlighten us on that. We have also heard that the healthcare informatics framework in Scotland works better as a basis for research than that in England. Do you think that that is the case and what are the differences and what makes one system better or worse than another for research purposes? A large set of questions there.

  Professor Morris: Shall I kick off as it is my area of interest? An initial comment would be that we should not ignore the phenotype because there is so much emphasis placed on genotype; but arguably, I would say, that phenotype is actually of equal if not more importance. An example is that Rory says that UK Biobank is powered by incident or new cases of disease. What we absolutely require is robust mechanisms to ensure that we pick up all new cases of disease in these very large cohort studies. To do that efficiently I think we have to look at the domain of health and informatics to support these research programmes because if that efficiency is not built into the system, the costs associated with these studies will spiral. If you look globally, the healthcare economies that do this well have a certain size—about five million people. I think that size is in Scotland's favour because it is much easier to do this work with a population denominator of five million people; allied to a low rate of migration, allied to a single health provider, NHS Scotland, with less penetrance of the private sector; also allied to a collaborative inter-disciplinary approach to answering these questions. Scotland arguably is ahead of the game because since the 1970s we have used a single patient identifier for every episode of healthcare. This is not rocket science, but in general practice we have used the so-called community health index number to track all episodes of healthcare, and this is now being implemented throughout the health service and is reaching 93 per cent implementation in hospitals. I think that single fact has given us a competitive advantage; it allows us to track people through the fragmented journey of care from their general practice or laboratory to the pharmacist and the hospital. What is the research benefit? I think a great example—again going back to statin drugs, comes from one of the biggest studies of statin drugs. This was led by Ian Ford in the West of Scotland. He in the mid-eighties tracked people in the context of a clinical trial for four years, and he showed that taking a statin reduced the risk of heart disease and stroke. That study probably cost £25 million to £30 million. What he published in the New England Journal at the end of last year was the ability to track these people; they were no longer part of the study but he could track them for vital events by linking into hospital admissions, and he published the 15-year follow up of the initial study. The key here was that the cost of the 15-year follow-up study was only £60,000 because of the efficiency of the linkage into vital health events. What he showed was two things: firstly, the protracted and prolonged benefit of statins in terms of cardiovascular risk; secondly, he put to bed the suggestion that statin use was associated with cancer development—so they are safe. So drug safety and drug efficacy. Scotland, because of its integrated records, is able to run such studies. It is not without its problems but perhaps I will let others comment before I come back to the problems.

  Q502  Chairman: Is there a comment from either the Wales or England side?

  Professor Sampson: I can comment a little on the situation in Wales. In Wales there is a developing centralised NHS healthcare informatics system and this is termed Informing Healthcare. It is certainly behind the state of development of the Scottish system. It was initiated in 2003. It is a single system across Wales and it is being developed by evolution rather than revolution. In addition to the NHS-based health informatics system, there is in parallel a research system which is under development at HIRU (the Health Information Research Unit), in Swansea University. That unit is developing algorithms to assimilate anonymised data for research purposes from the various health databases that exist in primary and secondary care. At the moment neither of these health informatics systems—the NHS one nor the anonymised research system—are actually linked into genetic data and will not be until concerns about the confidentiality of genetic data are resolved in relation to these systems. But there certainly would be potential for these integrated clinical health information systems to be linked to genetic data in those circumstances where consent has been given, for example in Biobank. Indeed, there are significant developments now linking clinical health information systems through to, for instance, the Wales Cancer Bank, which again is a fully consented biobank.

  Q503  Baroness O'Neill of Bengarve: Could I just ask for clarification whether you mean by anonymised they are irreversibly anonymised and de-linked or whether you mean that it is anonymised but a key is held.

  Professor Sampson: A key is held but it is double encrypted.

  Q504  Lord Warner: I am slightly puzzled by some of this—and I say this as the guy who had responsibility for Connecting for Health in England as a minister. The custodians of these records are essentially clinicians and the biggest obstacle to integrated use in England was the clinicians, who are deeply suspicious of integration and who are very protective about individual records. Is there something different about the clinicians in Scotland and Wales? Is it something about population size which actually makes the clinicians behave differently because they have been the biggest resistance or integration of records in England?

  Professor Sampson: I think there probably is a difference. Wales is a much smaller community. There is, I think, a stronger sense of collective ownership, much less private medicine and a stronger socialist tradition in Wales. These cultural differences may help the development of integrated electronic health records as well just the logistics of scale. Databases such as the screening databases in Wales have been developed with a great deal of clinician involvement from the onset.

  Professor Morris: I think you have put your finger on a key issue. The databases I alluded to are the national NHS Scotland administrative datasets and they may be rich in clinical content—for example we have the national screening systems, the national hospital admissions database, outpatients' database, GRO, and they are coded according to the international classification of Disease 10 or operating codes. So they are rich in clinical functionality, yet they are administrative databases hosted by NHS Scotland.

  Q505  Lord Warner: So you bypass the clinician cooperation issue in a sense because they are there.

  Professor Morris: I think "bypass" perhaps is a strong word! The governance of the system includes clinicians but the data sources and governance are under the aegis of NHS Scotland. Just to follow up, the real prize for the future is the linkage of clinical datasets, as you say, and that is where the challenge lies. We, for example, have a health care economy budget of a billion pounds a year. We probably have about 150 clinical systems, each with its own jurisdiction, and if we are really going to yield the value of these large cohorts we actually need to get into the granular detail of these clinical information sets. I think that is where the challenge lies because of lack of clarity of governance, data protection, confidentiality and there is a lot of work that needs to be done in that domain if we are really going to see the value from clinical datasets.

  Q506  Baroness O'Neill of Bengarve: I think we would like to hear from Professor Collins about England in this respect. How does it look for UK Biobank? I should say that I enrolled and I know the process and the consent seems to be fine, but of course you have to link to subsequently developing clinical information. Is Connecting for Health to provide that? How is it to work?

  Professor Collins: I think personally that this is the most critical issue for many of these studies. It is not the technology—that is developing. It is building these resources and making them as informative as possible. If you look back at the similar studies that have been done in the past what they were able to do was to look chiefly at the diseases that killed you and at cancers because there were registries of death and cancer. Why did Richard Doll do the study in British doctors? Because British doctors got paid if they were registered, so he could follow them. It is the ability to follow people and to find out what happens to them that allows you to really increase the value of these studies. The key issue that we have now is not being able to follow so readily the diseases that do not kill, but that disable, that maim, that cause misery. Dementia, joint disease, why are these not as well studied as cancers and heart disease? It is not because they are less common, it is because we cannot follow them in many of these studies. So the ability to link participants in these studies to their full health record is absolutely critical. Scotland is great but we are now seeing increasingly the ability, in principle, to link to more and more records elsewhere in the UK. So, for example, not just death from cancer but to hospital statistics so that you can find out at least the diseases that get people into hospital. In principle, you should be able to link to pharmacy record details so you could find out what drugs people are getting and look at pharmaco-vigilance in terms of safety of drugs. I do not think the problem is the ability in principle to link, it is actually more a problem about being allowed to link. It is the bureaucratic obstacles to linkage that are the concerns. They were highlighted in the Academy of Medical Sciences' report on personal data. They have been, I think, understood by the people in Connecting for Health who are developing the Research Capability Programme, which is looking at ways to facilitate research. But actually what is required is a desire, a willingness to understand that research is at the heart of healthcare. If we do not have research then we do not good healthcare, and it needs to be seen as part of the healthcare system.

  Q507  Baroness O'Neill of Bengarve: Just one very small point there. Do you think that it is fundamentally a bureaucratic or a legislative problem? Is it the data protection legislation with its many intricacies, if not incoherences, or is it the administrative framework that has grown up around its attempted implementation?

  Professor Collins: It is both. This is the summary of this report by the Academy of Medical Sciences on the working group. The legislation is not clear; it can be interpreted in a variety of different ways, and the consequence of that is that different people with different agendas interpret it in different ways. So I think it is a combination of legislation which is unclear and so leads to differences of opinion, and the consequent bureaucratic obstacles.

  Professor Morris: Can I build on that because I think this is an absolutely pivotal issue? The Department of Health guidance suggests that this domain is affected by 43 relevant pieces of legislation. There were 12 sets of relevant standards and eight professional codes of conduct. What this has bred is a culture of caution, confusion, uncertainty and inconsistency. Even Lord Phillips suggested it was like weaving his way through a thicket. So for us to interpret it and to have consistent interpretation from legal bodies who have data protection responsibilities is absolutely key. Currently this is a the major issue in terms of the ability to safely link data in a way which is in the public good with appropriate security. This was a major focus of the Walport Report, which was broadly welcomed.

  Q508  Baroness Perry of Southwark: I am puzzled as to how you are going to get the kind of follow-up data that you want. Is it not your intention to keep in touch with the people who were in your cohort and revisit them every seven years so that you do keep a record? And if you do not have that facility then collecting the biobank is going to be severely damaged in its usefulness, is it not? If you are looking for the sample of your subset that develop heart disease or whatever you are looking for, how are you going to find them unless you keep bringing the subject back? I speak as an enrolled person in the Whitehall Study, the stress and health study, which has been going for over 20 years, and we are brought back every three years or so to find out which of us have died off or gone off in the meantime. That is the essential part of the value of the study, is it not?

  Professor Morris: With respect to UK Biobank certainly one would ideally wish to keep in touch with participants regularly but the thing that one is interested in is what happens to them primarily in terms of health outcome, and that linkage can actually be done in principle through health record systems. As I have mentioned, for the whole of the UK, death, cancer, hospital episodes can be linked. Already in Scotland one can link beyond that to primary care.

  Q509  Chairman: To clarify, how does the UK Biobank currently link with databases, GPs, hospital episodes and death records?

  Professor Collins: At the moment we are not linking but we have approval from what was the ONS to link to death and cancer registration and you can then link to the HES system -Hospital Episode Statistics System. We have piloted that in collaboration with Professor Beral in Oxford, who has linked the 1.4 million women in the Million Women Study to Hospital Episode Statistics data. As the ability to link to more record systems, with the consent of the participants which we have, becomes easier, then we will be able to do so beyond Scotland (where you can already do it in primary as well as secondary care) to those other record linkage systems. In terms of seeing participants again, we will be seeing a representative sub-sample of the participants in UK Biobank, not to find out what happened to them in terms of health but in order to assess the variation, the changes over time, to the risk factors related to the way they live which will inform us about the whole cohort. To be efficient, we do not have to see all half a million again; we will see 20,000 or 30,000 every few years in order to track variation in the ways in which they live; obviously their genes will stay the same. But it is these linkage systems which allow us to look at their health outcomes. I should also add that such linkage could allow us to look at other health-related records. We can look back in time, at participants health before they went into the study and at other things, like occupation, which will tell us about risk factors, exposures, as well as looking at subsequent health in the long term.

  Q510  Baroness Perry of Southwark: If you are only looking at the health records, the hospital records and so on, and the records of death, are you not missing out on the people who may have the same genetic variation but who do not get that particular illness? Is it not just as important to know that 10,000 people who have a genetic variation all developed bowel cancer or something, but another 10,000 or 5000 or so who had the same genetic variation did not. Is that not just as important? You will not take that up from the health records, will you?

  Professor Collins: Yes, because we are linking to the records, and the people who have the disease would be in there and the people in the Biobank project who do not get the disease will be linked as not having had that condition, at least diagnosed, and we then compare the people with and without the diagnosis.

  Q511  Lord Warner: I am trying to understand this. This is quite a critical issue. Are you saying that all the people who are in Biobank have given consent to you going in, through the connecting for health, presumably information through the spine, from the GPs to give them the information that you request from the GP? I am trying to understand this, because this seems to be quite a critical issue.

  Professor Collins: We could ask one of the participants what they felt they had consented to, but I will answer, first, for them. Yes, there is explicit consent from participants for linkage to health-related records. That is defined as not just medical records, but other records—and it gives examples—including, for example, employment records both in the past (that is, before they went into the study, so that tells us about exposures before they joined) and also their health-related records in the long term through their medical records systems.

  Baroness O'Neill of Bengarve: If I remember—it is quite recent, and I should—there are ways in which, as you consent, you can exclude certain types of things like attendance at STD clinics and perhaps abortions—I cannot remember other things but there were things like that—but otherwise it is consent. I would add that I happily gave my consent to the lot. Having always been treated on the basis of knowledge that was obtained by treating other people, I do not feel this information, if duly anonymised, is in any way private or mine. But I may be one of the few people who thinks that. The law does not.

  Q512  Lord Warner: We have had a good go over, in a sense, whether the present IT systems are sufficient for joining up information from different databases I think the sense in the answers we have had so far is that there is still some way to go. Could we explore a bit more how we might improve the joining up because it is not just about technology? Can we explore some of the other issues that you think need to be tackled if we are to join these databases up?

  Professor Morris: We have touched upon the clarity of governance and the legal status of linkage. Currently the situation is that the clinical data is very much silo'd and it sits under different jurisdictions. I think, therefore, we need to look at the recommendations of the Walport Review and come up with a response to that which is clearly in the public good as good governance, and legislate if required, because there are differences between England and Scotland. Currently, I understand that PIAG has legal status; whereas, in Scotland, the Privacy Advisory Committee of ISD which oversees the governance of these national datasets is advisory and has no legal status. I think we need clarity around that. Second, I think we need to get on the front foot in terms of public and professional education. This is something which the Information Commissioner is very strong on. He says we should be much more forthright, informing people about how we use data, why we use data, the benefits of using data, and the safeguards which are in the system to ensure confidentiality and security. Certainly in Scotland, in 2002 a very good report, called the CSAGS report, recommended a public consultation campaign to demonstrate the clear value that research has in informing and improving health services. That campaign has never happened, so I feel there is work to be done about informing the public about how we use information and the benefits of using information. The third issue is around trusted models of linkage. Currently we have these multiple data silos with huge variance in the way that data are linked or what is allowed to be linked. I think we should look beyond health at other examples of things that work. A good example that is work which has come out of the ONS called the Virtual Microdata Laboratory. They have built a very resilient and secure system, with "thin client" secure and safe and audited access from multiple sites from the UK. They have been able to link very confidential person-specific datasets for the benefits of research. Researchers go into a protected environment. They are protected, they are overseen, but then given "thin client" access, which means that they are not allowed to download any data. They are given access to a central data repository for research purposes. The virtual microdata laboratory is seen as a huge success. It is now a model which is being adopted throughout Europe. I think there are three broad areas where we need to focus our attention so that this linkage is seen to be in the public good and also has industrial strength in terms of its security and confidentiality.

  Q513  Chairman: In the long term, all these databases will need to be linked to genetic databases.

  Professor Morris: There are various ways of approaching that. On a case-by-case basis, yes. Currently, the way that we think genetic data is that, following ethical committee approval, Caldicott approval, NHS R&D approval, we are allowed to link for specific research questions and that process has a very distinct audit trail.

  Professor Porteous: Clearly, the discussion we have had at the moment has been about linking health records to genomic data, but we should not lose sight of the fact that we need also to have IT solutions to linking and securing and auditing the genetic data. It should not be taken as a simple "press of a button" and you get an accurate answer out with certain technologies. You have to have laboratory information management systems in place—which we have in Generation Scotland and which I know the UK Biobank has—that can accurately track and audit all the samples that are received, and do that linkage too. If you have inaccuracies, it is always the weakest link that is going to be where the problems arise. It has to be said that much research in the past that has been done in this area has been done very much on a short-term basis. I think one has to be cautious about, for instance, using legacy data (samples have been collected in the past, stored in some clinical researcher's fridge or laboratory and then taken out subsequently), to be sure that you are matching up with modern technology to material that was stored in the past without those kinds of secure ways of coding the information, coding the samples, and then pulling it all together. There is a laboratory side to this IT technology that has to be addressed as well. But that is easier to do, because you have control over it, but you have to be very careful when you pull that data together from different biobanks to make sure that all the biobank information you are getting from around the world is set at an appropriate standard. That is where organisations like P3G and HUGO and others come into play to try to set those international standards.

  Q514  Lord Warner: We have spent a lot of time painting a picture of a world in which there is better linkage between big databases. There is a burgeoning volume of information available. Do you have any concerns about whether the informatics skills capability is going to be up to the job? What do we need to do if it is to secure benefits from this kind of linking of the huge amount of information?

  Professor Porteous: I think this is a very exciting opportunity. The physicists who have developed IT solutions for large number calculations have skills that we are barely aware of. I can give you a good example using a problem that we had. We were frustrated by the inability of a well-used statistical algorithm that is used to try to glean information out of genetic data: it is well-used in a certain setting, but when you try to use it to look at larger segments of the genome, it just falls flat, it cannot do it. We took the problem to our local colleagues at the National e-Science Centre in Edinburgh, and they developed, within a matter of a month, a grid-based solution to this that cut the computational time a thousand fold and gave us a working solution. At that level, I do not think we have a problem. The only problem will be if we think that our home-made solutions are adequate. We need to talk to people who are real experts in the field at that level. They are there; they just need to be alerted to the opportunity and the need that we have and I think they will be able to respond.

  Professor Collins: I want to agree with that. I think that we are generating resources that will generate huge amounts of data—not just the genetic data, but the protein metabonamic data. It will be working out what is real in those data that will be the most difficult part of it; it will not be the technology of measuring it. Encouraging people with bio-informatic skillls to work in the biology area is a critical one, and building a career structure for those individuals. The Wellcome Trust had a meeting in January at which they discussed this. One of the key things which came out of that was that we need to stop thinking about these people as information technologists and start thinking about them as information scientists, because they will be the ones who help us to tease out what is real in these resources. Maybe we should see the banking crisis as an advantage, because we can recruit all these people into biology. Once they come in, they will not go back because it is a lot more fun!

  Q515  Lord Warner: If a bit less well-paid.

  Professor Porteous: A steady, low income

  Q516  Baroness Perry of Southwark: One of the issues we have constantly talked about and had evidence about in the field is what you tell the subject of the investigation after they have given their samples. Particularly as knowledge advances, it will become apparent that some people have a very high risk of a disease which may be curable—and that is one reason for wondering whether to tell them—or may be incurable—and that raises different ethical issues. Are there any plans, if these do emerge, to inform the subjects and their families?

  Professor Collins: I could speak for UK Biobank. This was discussed at great length and included wide consultation with patient groups and the general public, as well as ethicists, lawyers, doctors, and researchers. The conclusion we came to—and this was signed off by the Independent Ethics and Governance Council for UK Biobank—was that no feedback of individual results would be provided to participants. The invitation to take part and the consent is explicit in saying that people understand there would be no feedback of their individual results. There are a number of reasons for that. Partly, we are not doing any assays of the samples upfront—we are doing them in five, 10, 15 years—and we do not know what we are going to measure, so it is impossible to counsel people on what the implications will be of things we might do in 15 or 20 years time, or, indeed, what the relevance of the things that we find might be. Going back to previous comments about these websites where you can find out your risk and it is widely divergent, we do not yet know when we get research results what their full impact is on disease, and we do not know everything about that patient in terms of their health like their own doctor does. We will, however, be feeding back the findings of the research that is done on the study. Information will be made available to all the participants and to other people as to what findings have emerged. If findings emerge that should change health care, which might include screening tests that are validated and really strongly predictive of disease, then those should become available in an appropriate way through normal healthcare systems. But we get explicit consent that there will not be individual feedback of results.

  Q517  Baroness Perry of Southwark: Do you see this as being permanent, or might it change in 10 or 15 years, when both the technology has changed and the findings that will emerge over that time may make prediction much more accurate?

  Professor Collins: The idea that the funders had of setting up the independent Ethics and Governance Council for UK Biobank was that things may change. Their role is an oversight role and to provide guidance. I cannot say that things will not change. They are there to guide us on that. Participants also have the opportunity to voice their concerns, because they can withdraw at any time. They can let us know if they have concerns about the ways in which the resources are used. There was, however, very detailed discussion and consultation for UK Biobank about the pros and cons of feedback. I cannot see the rationale on which that was based changing.

  Q518  Lord Colwyn: We have covered some of the problems of incidental findings that are already causing, for instance, dilemmas such as this discovery of abnormalities in brain scans. Genome studies are likely to accumulate much more information and have an effect on health, so privacy and confidentiality are very important. What are the risks and how are you managing the risks to privacy and confidentiality for participants and their families in biobank collections? What role do patients and volunteers play in deciding what will happen to their sample and information supplied as part of biobank studies?

  Professor Porteous: Perhaps I could reflect on the previous question first, because I think it does follow on from the one you have raised. The situation with Generation Scotland is very similar. There was a broad view taken and we have very much followed and, indeed, been part of the discussions that led to the formation of the UK Biobank. Generation Scotland has taken a very similar view and operates in a very similar principle; that is, if something clear and important and overt of a clinical nature emerges at the point of the clinical examination, then that information is fed back, but thereafter not. We are very keen that we emphasise the fact that we are doing studies that are looking for predisposing factors. To go back to the points made earlier by Andrew, the process of taking that evidence through to clinical application is one that requires replication and validation and several steps in that process before one would have, for example, a health technology assessment view as to whether or not something had diagnostic value and should be implemented. For all of those reasons, the notion of giving feedback on an individual basis we think is inappropriate; but like UK Biobank, we too have an advisory board, set up at arm's length from Generation Scotland by the Chief Scientist Office, chaired by Lord Sutherland, with advisers who also link between the UK Biobank and Generation Scotland. Graham Laurie is a member of both. Thus we have, I think, a good way in which we can see at a UK level, both UK Biobank and Generation Scotland, issues as they emerge and to get good advice from external groups as to how those should be handled. So far as risks to privacy and confidentiality are concerned, that is clearly of great import, because without that confidence in Generation Scotland or UK Biobank, we will neither gain nor retain participants. Of course, the retention of goodwill is particularly important for a family-based study, so we have been given this great thought and, we believe, care and attention, to ensure that we build in, through the IT structures, the appropriate firewalls that prevent information flow in anything other than a fully controlled fashion. I am sure Andrew will go into that in more detail as to how that works. I think you are right to raise the question, but it is one that we have spent a lot of time deliberating on. But, particularly for a family-based study, it is paramount that those entering the study feel confident that they wish to be part of it, feel confident in the reason for the research being done, and feel confident that the measures that we have in place secure that information and that it will be used for approved research only.

  Q519  Chairman: Professor Collins, would that apply also to Biobank?

  Professor Collins: Yes, we have the same security measures. There are two parts to your question. The first part relates to key code anonymising data, so separating identifiers from the data and keeping the identifiers separately and the codes separately and secure; encrypting data, particularly when it is being transferred. We spent about a year getting permission to get names and addresses to invite people into the study from the NHS, and it took us a little bit longer then to persuade the NHS to stop sending them in unencrypted data files. So it is doing things which make sense: as I say, protecting the identifiers and keeping them separate; having firewalls; paying people to try to hack into your system—we employ the National Computing Centre, to try to hack into our systems and identify potential areas of failure; and only allowing, a limited number of individuals to have access to identifiers. In addition—and I think it goes to one of the concerns that was raised in your question about being able to identify people from genetic information—in terms of what researchers will have access to, it will be anonymised or key code anonymised data, but they will also have to sign agreements that they will not use the data to try to identify individuals, so there will be contracts with researchers in that respect. On the second part of your question: What control do participants have over the users? They can withdraw at any time. We tell them about the records we are linking to, we tell them about the uses of the resource, and if they feel that there is a problem with that, then they have the right to withdraw at any time and pull all of their data out of the study. They also have the Ethics and Governance Council to raise their concerns with. So they really have control.

previous page contents next page

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

© Parliamentary copyright 2009