 Ond oeddwn i'n dweud o'r tawg. Felly dyna'n mynd yn ymddangos i'r ysgrifennu. Dyna'n mynd yn ymddangos i'r colegau Carlotta ar y FFW, felly mae'r dweud o'r dweud o'r dweudio'r dweud, ac yn ffwrdd Christine Woods ar yr Ysgrifennu dweud. A ond byddwn ni'n dweud o'r dweud i'r dweud o'r dweud o'r dweud o'r dweud o'r dweud. ac we now have a what we call a final professional development framework for people accessing data in safe settings. So obviously we're generally most I'm assuming most of your dins here are researchers or people who want to use data. Is that anybody else in the room that might be from a data supplier or government department? Can't? Oh, okay, so where are you from? From the Office of National Statistics. Oh, right, okay, so I have to be a little bit careful what I say. You don't look like the average typical civil servant I say with a t-shirt on but that's nice. Well of course, we all need data and that cancer research UK along with other research institutes.. Data is really the heart of what we do. of what we do, and more and more as researchers and a research community we're interested in much more detailed information about data, so for example we're interested in pulling up data about education, health records, tax and benefits, our shopping habits, what music we listen to, what we eat, what we drink, and information about our families as well, and this is really just the type of the iceberg, and not only do we want to analyse these data separately, we actually want to link these data together because by linking these sources of data together we can explain a lot more about how we behave as individuals and as a society. And of course once you start using data like these, and especially when you begin linking these sources of data together, then the data become much more sensitive because there's a risk that somebody could be identified from the data and that would reach all kinds of confidentiality and data protection issues. And that's why we have these safe settings like Danby was just talking about to make sure that research can occur with these data but within a safe environment. And we use these data because we want to produce statistics which can then go on to inform public policy and hopefully make our lives and our society better. So a little bit about the good, the bad and the ugly of data access. So I've been in this field for about 11 years now and things have improved immensely during that time. I actually started my career at the Office for National Statistics in their virtual micro data laboratory and since then a huge number of administrative data sources have now become available. Many more research is accessing these types of data than ever before and that's in down parts thanks to services like the UK data service, the administrative data research network and NHS digital. Things can still be a bit slow sometimes but at least it is possible to access data. And UK data service and people like ADRN do a fantastic job as well as just providing access to data but also providing information about data. So good standards for metadata and data documentation. Those are really key elements which make it easier and more effective to do research but not everybody is like UK data service or ADRN. And there's quite often as well as delays, often data come poorly documented and quite often we spend quite a lot of time working out what the data are showing before we can even start using the data properly for analysis. So that's a bad and ugly. Sometimes data access is never going to happen or work so slowly that we end up abandoning projects and my team have abandoned one project now that we applied for nearly 12 months ago. It took so long but actually it's not feasible for us in our current work plan to carry on with that work. That was going to be a nice bit of work about the variation of treatment for lung cancer which is like any project it's quite important and we've had to drop that one now because we couldn't get the data in time. Despite spending quite a lot of time trying to apply to access for the data. And generally this kind of situation will just lead to poor outcomes because there's a lot less we know about society as a result. In general though I think that most data owners, government agencies or government departments are generally willing to make data available and certainly at the top of government I think that willingness is there. So for example if we speak to the chief executive of Public Health England Duncan Selby he's very committed to making data available for research. The problem is it never quite gets off the ground and that's what I want to kind of talk about today. So what are the practicalities of making data available and who's going to make this happen? Because quite often we're talking about relatively junior civil servants within government departments who are going to be charged with making this happen. Do they have the support and the skills needed to actually enable access to data? There's definitely the buying at the top, the question is is there buying and the capability at the bottom. So what's the problem? Well first of all those staff have just mentioned take on quite a lot of risk for making data available to researchers and they often receive very little reward. So that's a bit of a problem. If you're a relatively junior civil service manager and you make access to data available and it goes wrong you're going to be in a lot of trouble. So the safest thing you can do is make it as difficult as possible or not make data available at all. And I think that's for my experience that happens quite regularly. So that's coupled with the fact that often staff aren't confident to make the right decisions when they're providing access to data. And so then we've got to think about why are they lacking in confidence. Quite often there will be data access teams where civil servants have been drafted in. They might have been moved around from other parts of that department. They might not have any expertise or background in data or data access. They've just been put in this role and said this is what you're going to do. So they're going to be lacking confidence. And that coupled with the fact that they're taking on a certain amount of risk means that they're generally going to be quite risk averse individuals. And in addition to that there's a lack of investment I think in the staff who are making data access possible which is really kind of a central theme of this presentation. As researchers and people applying to access data what do we expect? So we expect that when we liaise with these agencies to access data that they understand why we're applying to access their data. That they understand what research is and why research is important. We kind of generally expect that I think. We would hope that they would understand the data that we're applying to access that they understand their own data. We would hope that they can produce metadata and be accompanying documentation that comes with the data. We would also hope that they can provide us with legally sound advice which isn't necessarily that the law is there to protect data and stop access. Often the law is there to actually enable access to data. And for all of those of you who have been involved in the GDPR you've probably sort of thought of GDPR perhaps as compliance. GDPR has actually got a lot of stuff in there about making available access to data for scientific research. We also expect then to provide us with support and we would like to have an efficient fast service as well. I don't think anyone's kind of expecting data to come within a week of applying. Wonderful as that would be. But you definitely don't want to get to a stage where it takes a third or a half of somebody's PhD programme for example before they even get to see the data. You want things to be a little bit quicker than that. So myself Carl Otter and Christine and a group I'm going to talk about in a moment started thinking about this and started thinking, well is this field, is the data access and data support field, the things that we do for researchers and analysts, is it really seen as a profession or is it just something that we end up falling into by accident and perhaps some of us stay in this field because we quite like it and find it interesting to work with researchers. And is there any professional development that we can follow to enable us to provide an effective service for researchers and also for data suppliers? And we thought well actually there's very little. If you're within government departments you might be a member of one of the professions like the government statistical service or the government economic service where you might start to handle data a little bit. There's now a strand of government statistical service to deal with data science, but I'm not really sure to what extent that's actually developing a staff with skills in managing access to data. So, the working group for secure data access professionals that I now chair and that Tanbi chaired previously to me was established in 2011 and it's basically a group of people who are providing secure access to sensitive data to meet, facilitate development of standards to share views and experience etc. We've got a number of members now including myself from Cancer Research UK, Office for National Statistics, NHS Scotland, HMRC, UK Data Service Health Foundation and recently connected health cities and University of Edinburgh. And I hope that members of the ADR will still keep a part of that stay in that network. And essentially we started thinking about the problems of providing access to secure data and identified as a bunch of people doing this work as I say what is there for us in terms of career development and our ability to progress ourselves. Because if that's not in place then we are likely to move away from these positions and then for the researchers who are trying to apply to access these data they're just going to be presented with somebody who doesn't understand data, doesn't understand research etc. So, we began putting together what we call a competency framework and the competency framework looks very much like a civil service progression routes. It covers a number of areas which we think are important for staff who are in contact with researchers accessing data. So, for example, we think that staff should be aware of what analysis is all about. It should have an understanding of data legislation, be it statistics and registration services act, digital economy act or GDPR or health and social care act etc. They should have experience of handling and managing data including adding value to data as well because there's nothing worse from my experience of research, of getting a research data set where variable names are confusing and there aren't any labels. So, we would expect that staff can actually kind of add that metadata and they should be aware of data settings, data management. They should have experience hopefully of supporting people and managing users as well. This is really important. Statistical disclosure control is an important skill if you're working within a safe setting and they should be able to train staff and users as well. What we did was we put this competency framework together with these skills and we set targets for beginners, people with perhaps one to two years of experience and people with a little bit more experience as well. The idea of this is that, as members of staff, they've actually got a framework by which they can develop themselves and continue working in this field and develop themselves as data access professionals. What we think this will provide is expertise which is really, really important. It will also provide a career development pathway. Hopefully it will mean that staff retention is much better. I know of one government department where staff come and go roughly every three months in the middle of applications. It's really frustrating actually but I can understand actually why they're moving on. This makes the whole process of applying to access data and the secure data, administrative data, much more efficient. I think it's also worth saying that there are going to be more of these settings around in the future. Yesterday I read the government response to the Lord's report on artificial intelligence. There's a lot of stuff in there about data settings and more of these coming about. We obviously need more staff to actually man these settings and therefore we need people with the right skills, the skills like I mentioned previously. We need some kind of framework. I think this is what the secure data access professionals working group have put together which you can now find on our website. If you've got any questions you'd like to talk to me about this a bit further, the email address is at the bottom as well. You can now go to our group's website, go to guides and resources, download the competency framework. I found it really useful for myself. I've actually just recruited somebody to assist me with some of my data access work. It's been really neat to be able to put together a programme of work for this person based on this framework. These are kind of niche skills. It's very difficult to recruit people into these roles. When we recruited this person, we maybe had maybe three or four feasible applicants and we're just down the road from Tech City and you would have thought maybe we might have had a few more. It is important that we get more of these people trained up and skilled to do this work because unless researchers actually want to do all of that themselves and not do research, I think we're going to need many more of these staff working in this field. Thank you.