 Felly, rwy'n gweithio, wrth gwrs, o gweithio i'r rhaglenu i'r rhys-grifennu i'r First Australian, o'r rhaglenu ar y llan sydd yn ymlaen i ddweud, ac mae'n ddweud i'r Cydnod, rwy'n gweithio i'r Eora nesaf. Rhywbeth Rosy Hicks, rwy'n gweithio ar y ddau cyfnod ddataeth waith Australia. ymgyrchu ar gyfer y Fawr Ffordd, i'w gwirionedd ar gyfer y Ff locations, am gynhyrchu ar gyfer y Ffordd. Yn y pethau ar y gyfer, mae'n cael ei gweithio a bwyd i gweithio gael eu lleol sy'n gwneud a'r gweithio i gweithio ymgyrchu cyfnodol i gwybod, i'r fawr i gweithio ar gweithio. Felly, mae'r fawr i gweithio ymgyrchu ar gyfer y Ffordd. Ffurf hwnnw, d систем, ar amserion, mae cyfnodd, yr amserion a pharwyr ddamer, ond mae'r iddyn nhw fyddwch ar gyhoeddwch gael panaldus iechyd. Cyn bod ymddirionedd yw'r cadw iawn i gyd-igion o'r cadw iawn, ddechrau, a'r cyfrannu, ac mae'r ddechrau i'r cyfrannu a chael ddataeth o wneud yn fy neid o'r amserion o greff stryd. Access to sensitive and identifiable human data is essential for researchers to gain new insights to understand human health and well-being. By analysing and integrating these datasets, researchers can unlock a wealth of knowledge to drive innovation and positive outcomes for the community. The ARDC is heavily invested in enabling researchers to share sensitive and identifiable human data through a number of initiatives such as our Health Studies Australian National Data Asset, or HASANDA. This has an initial focus on clinical trials and we have 72 health research organisations across Australia collaborating to build a national infrastructure to enable researchers to access and share data from health studies. The initiative aims to stimulate new data-driven research ideas, increase the impact of health research and ultimately improve health and well-being of Australians. No doubt there are many challenges facing researchers working with sensitive data but the potential benefits of harnessing this data to generate innovative research solutions is enormous. It's timely and an important topic and I look forward to hearing the various insights of our panel members and I am very grateful for all of their participation and efforts in joining us here for this afternoon's discussion. I'm going to hand over to Professor Joe Shepter who will lead us through the rest of the afternoon. Thank you Joe. Great, thanks Rosie. So I would also like to acknowledge the traditional owners of the lands, all the lands of course that we're meeting on today both here in Sydney and of course around the country. My role is as the Pro Vice Chancellor for Research Infrastructure at the University of Queensland where of course this is a particular challenge for us and in a little while you'll hear from a UQ researcher who's facing these challenges in a project and that'll be the case study that you'll hear. I also then sit on a Queensland-wide organization and we talk about how we're going to share this data and in fact I'm also on the board of the ARDC where we have this national conversation about solving these data challenges. So this is an incredibly important issue that we as a country really have to figure out good ways to solve. So we really would like to I guess have a very wide-ranging conversation today but in particular we really want to think about a few things, one of which Rosie has already highlighted and that's one simply of access. The good news is unlike in lots of cases there is lots of data. The challenge now is to make sure it's in the right hands of the right people. The second challenge then of course is reusing that data. So just because we collected the data for a certain thing or a certain application doesn't of course mean that it can't be used for many many other things. And of course if you solve the first problem where lots of the right people have access to the data now you can do new and exciting things with it. Not only do you get new people with new perspectives looking at the data but now you get all kinds of machine learning opportunities, AI opportunities. And so again access opens up that reuse possibility and it would be great to hear people's thoughts about that. Then of course I'm going to mention the thing that probably not allowed to mention of course and that is sort of quote unquote who owns the data. Okay we're going to share the data, who gets to decide what can be shared and what can't be shared. If it does get shared who owns the data now that somebody did something with it. So again those are all challenges that we really need to solve of course as we go forward to make sure that people understand what they can and can't do with the data and what the outcomes are going to be. We don't want all this work to happen and then it not to have some impact in the real world. So that's going to be really important. So that access reuse and understanding who owns the data are some of the key elements that we'd really like to talk about. So the way today is going to work is the first we're very lucky to have some fantastic panelists and you'll hear from them in just a second. And then after that Professor Peter Sawyer who is online from Adelaide is going to give you a case study about a project that he's running and it will highlight a lot of the challenges that exist in this space and that we have to solve. And then after that we should have about 45 minutes for some Q&A and we'll certainly do that with people in the room. And of course we've got somebody monitoring online and if you have any questions or comments in the chat function please put them in and of course we'll make sure they're part of the discussion. So with that I'm going to turn us over to the panelists. A panelist has been asked to sort of tell you a little bit about who they are and maybe give you some thoughts in the sensitive data area. So our first speaker is Philip. So hello everyone my name is Philip Gould. I'm the first assistant secretary for the health economics and research division at the department of health. So I've got a really interesting role where I have a lot of people who are actually working right now on fascinating data sets generating insights and learning things that we didn't know beforehand. But I've also got a lot of people working on data governance and in the department data governance often means working out how we can better share the really important information that we have with the people who need to use it. So I'm going to I've got a few minutes I think give you a bit of a kind of narrative arc I suppose from when I started in the public service 10 years ago to where we are now. Prior to joining the public service I worked in banking and finance and that meant using a lot of data. A lot of interest. I'm still really keen to know what the RBA does with rates today. You kind of can't lose that interest in data. But there the sensitivity around data was actually paywalls. So we had a lot of providers who had fantastic information that could give you a market edge but you had to pay for it. And there was a lot of work and a lot of money spent on getting access to data via providers like Bloomberg and Lehman Brothers. Many of you would have heard what happened to Lehman Brothers in the end but data was very central to all of that work and finding the right staff gave you an edge. I then worked in road safety where I did my PhD doing econometric analysis of road safety measures and again data was really central to that work. And that was the first time when I came across this idea that information that I knew was clearly in the public benefit was too sensitive to be shared. So I wasn't actually able to get hold of information that as a researcher I knew would help my research because it might have potentially led to the identification of an individual being involved in a road accident where for example drink driving or drug use may have been involved. I found that incredibly frustrating but I found my way around it and continued on with my work. Fast forward a few years and I decided to move to work in the government sector and I sort of remembered I knew data was really important. I went to work for the ABS because I knew that I wanted to help people get access to better data because it's an extremely valuable resource for researchers and I'll be really honest here. I was absolutely horrified when I arrived at the mentality towards data sharing that we had in government. We had people asking us to integrate data sets and we were saying no we wouldn't do that to have access to it. No you can't have access to it and then lecturing people about the fact that legislation didn't allow data to be shared with researchers. So it was a really backward view of the world and I think over the last 10 years governments come a long way. I still think there's a long way to go but I think that we should actually say as a group here of researchers and government people that a lot has actually been achieved in the last 10 years. We see now the ABS has given people access to a lot of really valuable potentially reidentifiable information often quite sensitive information via their data lab. And we've seen ideas like the five safes approach really actually give government comfort that if they are to provide better access to data there are proper frameworks for doing that. So we've seen this big shift but there's room for a lot more and we've got to work together as a group to help drive that shift. We need the research community to be able to demonstrate the value of what it can offer through better use of data and we need a government that's willing to listen and understand what that value is as well. The last thing I'll say is as part of that shift I think we saw with COVID a really interesting change in government mentality. And it's this term that the former interim data commissioner my boss Deb Anton talked about and I really liked it which was this idea that government needs to have a sense of a duty to share. So that moves away from a duty only to protect but also thinking about the potential that data has to do good things and that we as public servants need to think about what we need to do to enable the right people to use data safely and effectively. So that puts the pressure back on us that duty to share it's not just about protecting and saying no. And I think if we can keep building on that and I know that the Department of Health by our work is adopting that sort of posture to use a very management consulting word I think that that's going to lead to really great things going ahead. And I ask you as a community research community to keep pushing us to do better but also work with us and be constructive. So I guess that's kind of my few words. Thanks Phillip. My name is Marion Hemphill and I'm the general counsel and the chief privacy officer for Australian Red Cross lifeblood. So if you don't know what lifeblood does we used to be called the Australian Red Cross blood service so we provide Australia's blood supply to hospitals. We collect it from donors, we test it and process it and then distribute it out to Australian patients. We collect about one and a half million donations per year and the donation itself, the blood itself has information within it that's very useful. We also test all the blood that we collect and there's a lot of sensitive information that we need to get from the donors that goes along with the donation which we need to keep for safety purposes. And one thing that we've learnt that while we collect information for safety both for the donor and for the patient the way we collect it and the way we protect it is really important because we run on a social licence. And if that licence relies on trust and if we breach the trust of the public and our donors that may mean that people no longer wish to donate. And that has real consequences not just for our research programs but really importantly for Australian patients. It might mean that there aren't enough blood donors and there isn't enough blood available. And given that one in three Australians are going to need donation in their lifetime that could have really a real world consequence. So we've had experience around a bump because we did have a major data breach about seven years ago and we were really aware that public trust is hard to get and easy to lose. And if you do lose it, it's hard to get it back. So when we did have a data breach, we really focused on transparency and trying to regain that trust. And a big learning that we had when our donors spoke to us about their information being involved in a data breach was there was a lot of concern about who could access their data. They'd given us data to save a patient's life not for any other purposes. Even though our consent forms had talked about research and other things as well, it sort of opened people's eyes as to what we might do with that information. And it was really clear to us that our donors were very worried about third parties, particularly things like insurance. If there was something, because we test blood, was there something in their test results around their iron levels, for example, that might prevent them from getting insurance coverage a bit further down the track. Some people were worried about police getting hold of that information. As Phillip was saying about alcohol levels and that sort of stuff, even though we don't test for it, that seemed to be a real concern for people. So we're really focused on trust and making sure that we collect, we're really clear about why we collect it, what we're going to do with it and how long we're going to hold it for. But the thing is that data is so rich. We could do so much more with it. And we do have quite wide consents where we talk about research and making of reagents. And it's in the forms that people read when they come to donate blood. But how many really get that far with their reading? They've answered lots of questions. They're focused on the stuff about fainting. We're not really sure how much people take in about that research part of it. But we want to use that data because it's so rich. It could benefit the donors as well as the patients. Because we're doing tests, we could start to tell people a bit about their own health. We could go down the road to precision medicine. It could also help society at large. So one of the challenges for our business is to find a way through. And I get asked a lot by our internal teams. Is it legal? Is there an exemption in the Privacy Act that we can use to use this, even though it might be a little bit outside the realms of what we collected it for? Is this within reasonable expectation? And I think my teams find me a little frustrating because as a lawyer, I think they expect me to say, there's no law against it, go right ahead. But because I'm an ethical lawyer, I think the question that we need to be asking my organisation needs to ask, and I'd like us to ask today, is not what can you do, but what should you do? And put an ethical framework over it. And it might mean that we push out to more accessibility because that's where the scale tips. I could keep going, but I'll pass on. Thank you, that's fantastic. Thanks. OK, well I'm Maren Smith. I'm the chief executive of the Population Health Research Network, or PHRN for short, and I'm based at the University of Western Australia on the banks of the beautiful Blue Swan River. I'm particularly interested in sensitive data. It's something that PHRN works with every day, and I put together a couple of slides just by way of background that I think might help the discussion. So as I said, PHRN works with linking sensitive data. Now we're all on life's journey from cradle to grave, and information about the health services we receive and other human services we receive is collected at each step along that pathway. And of course we live in a federation, so some of that data is collected by states and territories, and other data is collected by the Commonwealth. And PHRN is able to work with groups around the country to link this data in privacy-preserving ways and making it available for a range of research. And the sort of data we link at the individual person level includes birth data, hospital records, emergency department attendances, cancer registry data, and of course death records. And this data can then be linked to information held by the Commonwealth, such as the medical benefit schedule data or pharmaceutical benefits data, and aged care veterans affairs data, and of course immunisations data, which has been enormously important during the pandemic. Most of that linkage is done without informed consent, but it is lawful. It's allowable under a whole range of legislation, which unfortunately is not standard across the whole country. And we're also able to link this data with other clinical trials data, other clinical registries data, and other research data sets. And generally when that happens, that linkage is with informed consent. In terms of the PHRN infrastructure, I won't go into too much detail, but to say that linkage is done at the jurisdictional level. So every state has got a data linkage unit, and they do that linkage based on fully identified data. So they take the name address, date of birth and gender of the person to build the linkage map. They don't rely on any particular identifiers, because of course some of this data is health data, and may not have a Medicare number on it, but other data such as death data doesn't have a Medicare number on it. So it's not reliant on a particular identifier. The other thing we do is provide a secure access environment. So once data has been linked and put together in a linked data set at the individual person level, generally without the identifiers on it, then it can be interrogated within a secure environment, and then only the summary data is taken out, and the data is actually checked as it leaves that environment. So I think that's quite a helpful mechanism as we're thinking about how you manage sensitive data in the discussion today. And probably in our experience, just to say, we think it's just enormously important that whatever is done with the data is lawful, ethical, and acceptable to the community. As you say, community trust is enormously important, and it's easily lost and very hard to win back. An informed consent is probably a helpful mechanism. It doesn't have to apply all the time, but where it can apply, it's helpful. Another helpful construct is the five safes, and some of you will have heard about that. So it's safe projects, safe people, safe data, safe settings, and safe outputs. I've just mentioned a little bit about safe settings and safe outputs in what I said previously about those secure environments. And last but not least, I think accessing health information in Australia in the 21st century is complex, but it can be done. And there's a couple of examples down the bottom of some uses. The first one is around the human papillomavirus, vaccination, HPV vaccination. Most people probably know that the in-phrase from the University of Queensland was instrumental in the development of this HPV vaccine. A linked data was used to link the vaccine information with the cervical cytology data, and that clearly demonstrated in the population that that vaccine was enormously helpful in preventing cervical cancer. And the second example is a vitamin D trial that's currently underway in Queensland, and there's over 21,000 patients enrolled in that trial, and linked data is being used for five-year follow-up, in this case with consent. So that's probably what I wanted to say just to provide that sort of framework because I think it will probably be relevant to the later discussion today. Absolutely, thanks. And our last panelist couldn't be here in the room with us, but I'm hoping Satyrus is online and can tell us a little bit about his background and his interest. Thank you very much. Thank you for having me in this panel. I hope you can hear me all right. I could join you in person, but I'm delighted to be part of this discussion. Just a big introduction. I'm Satyrus Bartlaikis. I'm a Professor of Global Environmental Health at the Australian National University. And I direct the new NHMRC network called HEAL, Healthy Environmental Wives, which focuses on environmental health, climate change, environmental change, environmental pollution and other environmental and socio-economic stressors and health in Australia. We try to develop the evidence base and tools that will support decision-makers aiming to protect the general population about all sorts of other groups from the effects of climate change, extreme events like bushfires, floods that we've seen recently and also issues related to contamination, to unsustainable housing or in relation to food security and water security and availability. The HEAL network has been founded by NHMRC for five years and we are starting now. We are starting this month formally. It's based very strongly on data systems and data and work with close data providers from across the country. If we can move to the next slide, I'll give you an overview of that. I would like to acknowledge the traditional custodians of the lands we are based at in the country from here in Canera, but across the network we have many Aboriginal and Torres Strait Islander investigators and we are very grateful for their contribution to the network, which of course is very strongly based on indigenous knowledge and tradition. Next slide, please. Going back to the aims of the HEAL network, as I said, it aims to catalyse environmental research and generate the evidence, the tools and the data which are needed to make decisions which will protect the population and particularly at risk groups from climate change and environmental change in the Aboriginal and Torres Strait Islander and in all the other countries. If we move to the next slide, please. So this mark gives an overview of the HEAL network landscape. As I said, we rely heavily on data providers and collaborators from across the country like the Public Health Research Network or the Australian Health and Welfare and other data providers. The intention is to link available data, environmental and health data to better identify risk factors and characterize the risk of environmental exposures to human health and, most importantly, identify solutions that can reduce exposure and improve health and wellbeing. If we move to the next slide, please. In terms of health data and the access to environmental health data and how we share that, there are many challenges and many opportunities there. So I would like to highlight that big data are increasingly used in environmental health research. There is clearly a need for good quality data and for data which are findable, accessible, interoperable and reusable. This is the third framework for data and data sharing. The ability of using data and analyzing data has increased immensely in recent years. Currently, there are very powerful techniques, computational techniques based on artificial intelligence and machine learning for analyzing data and deriving exposure response relationships which help us to make decisions in terms of protecting the population for environmental exposures. This large availability of data and computational tools provides opportunities to gain knowledge into the causal relationship between exposures and health outcomes. But it also poses a number of risks and these are related to private issues. We have seen in studies, particularly coming from the US that it has been possible to re-identify data which have increased exposure in environmental health studies and of course this is this potential very damaging. We can undermine confidence and trust in research and of course you can expose individuals, you can cause some stigma and even have financial impacts for people involved in studies. Therefore, it's very important to ensure that sensitive information is treated as sensitive and we ensure that there is no that we protect the data in a way that they cannot be re-identified. I think it's worth saying that this kind of risk can be investigated for a long time in the more traditional genetic and medical research fields but much less so in the environmental health research and it's increasingly important because the availability of data from sensors from exposure studies is becoming much larger so we have access now to data from wearable devices from mobile monitors from biobondering studies and of course this data potentially linked to housing characteristics and the process as I said a risk for re-identification. So I will say that of course it's important for the data to be open and shared and fully for the purpose of improving research and gaining knowledge environmental health data are expensive to collect they are expensive to analyse so it's important to make the best use of them and add value to the research into the research studies which have collected this data but of course we need to be conscious of the potential risk of re-identification and make sure that there are no unintended consequences for participants in environmental health studies so I will post here and before we will discuss anything in this plan thank you very much terrific thank you looks like a fantastic initiative so now I would like you to turn you over to Peter Sawyer Peter is joining us from Adelaide and he is going to present our case study and Peter will tell you a little bit about the project that they are working on then highlight some of the issues and challenges that they have got hi Peter yes thank you so much, Joe and thank you for inviting us and giving me the opportunity to present our research project this is an ACIF for Sailing Cancer Research Foundation funded 10 million grant where we got the infrastructure for 15 3D photopoly imaging systems and it's a collaboration with the University of Queensland University of Sydney and Monash and just that I'm a dematologist which is all this time here in Australia since 2007 with a lifelong interest in the early diagnosis of melanoma and this is also the reason why I decided to come from Australia to Australia because you guys are leading the world in regard to the incidence of melanoma this funding basically is based on a on a technology which is basically 3D total body imaging looks like a spaceship 92 cameras a machine in one second or 92 cameras are shooting and within 12 to 13 minutes with the current computer the 3D total body other types built up I think it's fair to say that these are sensitive data and they are identifiable also so I don't think this needs to be explained by the way this this is one of my collaborators who is also part of our study project and I have his ex written and verbal consent to show his to show him and the beauty of this system is of course that we can study longitudinal data which is very important actually for the early detection of melanoma to understand the biologic ecosystem but it's also important for inflammatory skin diseases and we feel very strongly that this will have a place in the future in dermatology particularly for regional and in Australia I think we are going now into the nitty gritty if you think we have quite a few boxes here the overarching idea is to have a basically a nationwide research data bank which is obviously an image data bank with total body imaging and also with individual demoscopy images but of course we will need clinical data we have done quite a bit of work already with general and genetic data so it's not just imaging data but the specific aspect of our data set is absolutely identifiable and then you see three other boxes which basically at the moment represent the three states we are working in having said this one of our big goals is to develop a roadmap for nationwide screening program and then we will have more states and more territories involved and as you can and you guys probably know it better than I that there are different rules in Victoria and in U.S. Wales and in Queensland for example but I understand U.S. Wales has one bug system in the public system Queensland has probably 15 bug systems but on the other hand we have just one EMR so it's challenging in its own light in each of this and actually the data which we are collecting we are collecting which in the public system but also some of our systems are in the private system like in the Mananoma Institute of Australia which is a private non-for-profit hospital and of course the data will need to be identified in the conventional way but the conventional way is not enough if you are dealing with the body imaging so there are quite a few problems as you can imagine how we deal within the various hospital system and each HSS is absolutely independent and then how we deal then with a research data bank originally we thought there would be one research data bank now I understand that there will be each of the states will have a separate one and I mean Ryan Sullivan is the contact person from Sydney and Paul Bonington as who I understand is in the audience today because really about all this at Monarch and it's actually a special situation because Monarch has a close collaboration at Monarch University with the health system at the Alfred which is not always the case in UQ where it's basically a sort of a different setting in David Albarham a person from UQ in this context and I understand that Monarch and UQ are contemplating about the physical storage whereas University of Sydney is thinking on the cloud anyhow this is basically and I come now to my last slide which will show up in a second and basically this is the missing statement of our research basically to transform a melanoma early detection using total body surveillance to enhance individual lesion management and we have three major research streams one about diagnostic intelligence where of course AI will play a major role and my part will mostly be the annotation of the images and link it together from the clinical point of view of course the health service evaluation is a major point and having said this we have a huge group of CIs and we have health economists, pathologists we have computer experts we have a very broad group of expertise under our research and then of course health informatics which is a big issue standardisation of images dichom standards and the idea is obviously to fund then basically to build up a network a 3D network in the various places in the 15 sites we should have at the moment but at the end of the day it should be of course extended to Australia and and finally and this is actually my last part of my slide here is the outcomes I think we will be able to have the largest most comprehensive skin imaging database not just with individual teleimages but also with scope of body images and I mean this is very bold and there are a lot of questions and I really hope that as an outcome of this discussion today we will get some specific feedback specific health. Obviously we are looking for a live solution for melanoma early detection so that they get the melanoma at the stage where they are not melanoma but of course this is quite challenging from the clinical point of view because this means that we excise all moles and all lesions through present on the body of the person and of course we will facilitate artificial intelligence and yeah and as I already mentioned the big idea is obviously to pave the road for a nationwide screening program where there is a lot of discussion about should there even be a nationwide screening program because there are a lot of opportunistic screenings already going on but I think from a societal point of view it is at least worth contemplating a nationwide program. So thank you so much again for inviting us and chair our project with us and I am looking forward to the discussion thank you. Great thank you Peter and it's a really exciting project I have to say great things are going to come from it so now we're going to open things up for a Q&A or comments from anybody both in the room and as I said we do have Keith monitoring what's happening online so if there's somebody online and you'd like to put something in the chat we'd certainly like to throw that into the into the mix so to start the discussion I'm going to start if people will let me do that because I was struck by some of the comments that were made and Peter's very last comment reinforced it for me and that is we've spent a lot of time I think thinking about the technical but how could we make this happen but have we done the hearts and minds work to demonstrate that it should happen so a couple of you at least said said that and I'd like to I guess ask that question have we done it and if we haven't done it how are we going to do it I don't know if somebody would like to tackle that major problem but I think we haven't done that work I think that and as a lawyer I touched my microphone I take my part in that because I think that in explaining to people why information is collected what's going to happen for it and what the benefits might be I think that we haven't actually I know in the health sector haven't been that transparent we give people documents which are very densely written by lawyers and a lot of lawyers write documents that aren't really written to be read they're not in plain English they're quite dense and people are often reading things in a situation where they're maybe haven't got their eye on the ball when it comes to privacy and data I know when I'm asked for medical forms is there something wrong with me and that's my main thought and so it's not duress but I don't really care what the back information says we did have this conversation within Lifeblood recently because we have a antibody register in WA that we're improving and we want to rather than just the blood service have access to it let hospitals have access to it so if someone comes in and they need blood they can check the antibody register to make sure that there's going to be a really good match for the patient that they've got like if someone has unusual antibodies we'll know and there's a real feeling of you know can we just take this information because if we ask donors and patients if we can have it they might not see the benefits and they might not consent and that's a dumb decision if you decide not to have your information on the antibody register and you're in an accident and you've got antibodies you might have a very some standard blood transfusion if you need one it's a very technical example of a person but my feeling there is that it's not about avoiding taking the choice of someone we haven't done a good enough job in explaining to donors and to patients why it is a fantastic idea to be included in this register so I do think there's a lot of work that needs to be done in the hearts and minds area public trust of the health sector not practitioners but the infrastructure particularly government there's a low trust level there so I think and every time there's a breach and it's approached in a political rather than a transparent way that adds to that lack of trust so I do think there's a lot of work to be done around hearts and minds in this sector and in others my personal sort of feeling around things like facebook and other social media if they had come forward there's a lot of very low trust with facebook and meta if that company had said we're going to give you a tool and it's going to be for free other than having to look at a few ads you can look at your friends' photos you never have to sit through a holiday slide show again because you can do it at home and you can keep in contact with people around the world we'll give that for free what we're going to do is take all your data I think people would have signed up and they wouldn't feel betrayed when later on information was being used in a way that they hadn't that weren't expecting I don't think that we'll ever have achieved public trust it's always going to be a work in progress because expectations will change and people will I guess something constantly needs to be renewed but what I think is really challenging about where we're at at the moment is if you talk to a group as I've done before or talk to a focus group essentially about views on sharing data the views will change very rapidly within a 5 minute discussion so if you have someone who's quite contrary to data sharing in the room that can turn a whole room like in sort of 15 seconds against data sharing the right advocate in the same room can actually turn it all the way back so everyone's saying why aren't you sharing more data so that's a really challenging environment to operate in and I think what Marianne said about holding ourselves to being more transparent is the only thing is one of the only things we can do to settle those views so that they don't change so quickly and I think the other area with transparency is around clarity of language so we start using terms like identifiable data and that confuses a lot of people what do you actually mean by that well it's not kind of identified because it doesn't have a name and address on it but you could kind of work it out and you know with Peter's vision there you could maybe looking at the back of that person maybe we are identifiable the tattoo might have helped you turn it round but these are difficult concepts for people to understand and until we have a better public understanding of language we'll continue to face these issues where people get turned on and off data sharing very quickly Thanks Could I just say that I don't think we should underestimate the common sense of a general public I think that people generally speaking when they're faced with for Peter's example like melanoma I think the community understands that melanoma is a major cause of death and that something like this would be more systematic in assessing moles is likely to benefit them so I think the community will understand that but the community will want to be reassured about what's happening to their data and how it's being kept safe and I think as you've said Phillip I've set in on focus groups as well and focus groups can move quite quickly with strongly opinionated people within a small group but there are other ways of assessing community views and I've been involved with some citizens' juries which are a sort of a slightly longer process which start by giving a group of people some information and then enabling that to be in a non-threating environment discussed and out of that generally comes common sense something where the community can actually see some benefit as long as they have some reassurance about the risk so it's possible that Peter might want to actually test that with the community and it's not too difficult, you could set up a focus group or maybe given the drawbacks of focus groups that we've talked about sometimes a survey can help because that is then just an individual fills in the survey and it's perhaps less prone to that influence or some other process so there Peter is something for Peter to take on board so it's probably worth doing Actually we have already done this we have done several consumer forums and just recently we have done a survey where we asked very specific questions and what we understand that people are quite happy that the data shared in Australia but definitely not overseas I make sense this and we have not defined overseas well enough but it's a work in progress and we have actually quite proud to say that we have quite a bit of consumer involvement but as you say the consumer has really a lot of common sense and our consumers are mostly also people who have suffered from melanoma had advanced melanoma so of course they are flying the flag and this is probably also we should also look for absolutely independent consumers and this makes sense Thanks Peter I fully agree with the point about the general public becoming more familiar to this kind of service and sharing and sharing of data so we are increasingly used to where the device is smart phones, smart watches we get this notification all the time to share our location and we make decisions every day about sharing our personal data with companies I think in the case of medical research, environmental health research and public health research it's important to highlight the benefits to the public health, to the population of sharing this data and be able to use and reduce high quality data which are expensive and difficult to collect and analyze so I think the public is increasingly becoming familiar with these kind of processes and it's obviously our responsibility to explain and highlight the benefits of storing the data and sharing data Great thanks Questions from folks in the room? Greg So to the guys many heart and ear fill of content comments that there's a duty to share and that's coming from almost from the data commissioner who's putting that forward I'll take it online So yeah really heartened by the duty to share but we live in a working environment of seven jurisdictions and Peter fully is aware of what that means for his study across a national area and we are still struggling to get our state governments and our state health departments to share data even following the sorts of user forums that Peter's had to say that the people contributing are quite happy for their data to be shared So how do we spend a lot of time in the research space addressing this issue I'd spend hours every week facing data issues and talking through what are the processes what are the governance issues and every research has to jump through those groups How are we going to improve that system and how can ARDC start to build a national story around this So one of the toughest areas is actually dealing with the federated government in Australia so that adds a whole new area of complexity because not only dealing with one layer of government who's got legislation and concerns you dealing actually with two every time I think I don't have a simple answer to that otherwise we'd be in a much better position than we are but I do think one of the things that where government works well and we probably need to together work on how we can leverage this better is government works well on precedent you establish a precedent and quite quickly things become the norm So I think what we're seeing at the moment are some really interesting projects which are targeted to particular areas of policy interest New South Wales is doing some wonderful work with the Commonwealth at the moment on a national disability data asset so that's actually focusing on a particular area of need establishing some of those connections and then hopefully turning those into the norms so I think that's one of the things that we have to do is actually start showing groups that you can achieve something initially modest and build on it and that's how the Commonwealth will matter the multi agency data integration project that was starting with a small number of linkage linked data sets and actually showing yes you can do this let's do more but the challenge now I think is to integrate state and territory data with Commonwealth data and what tends to be the sticking point there is this kind of first mover idea of well I'll share my data with you if you share yours with me everyone agrees that we need to do it that happens in principle very quickly and then to characterise it really glibly is we agree let's share data and then there's the email from the state person it says to me so when you're sending me the data and I send back an email I say I thought you were sending me your data and it's kind of sad that's the impasse that we've been in for a while but I do think specific project to create a precedent of better sharing is probably a really practical way of moving forward Paul Yes hello Paul Wellington Monash University it's going to make some additional thoughts on trust and thank you Mary and for what you're saying around the hearts and minds I'm first and foremost a technologist and I've learnt that the key to success of any technical ambitious project like ASMID is to appreciate that it's primarily about people culture and policy and it's not about technology there are obviously broad challenges for those projects like ASMID trust and awareness of a system facility or service that from people that it's okay to use I'd like to say that groundbreaking innovation and technology projects can only move at the pace of trust there's another element to trust though I'll be kind of keen to hear from the panel your thoughts on so we've talked about public trust but I also think it's about stakeholder institutional trust particularly for those complex multi institutional programs that need to acquire and maintain the trust of the hosting institutions because at the end of the day they're brand as well in their reputation and that institutional trust needs to be earned but it also needs to be nurtured continuously so I'd be keen to hear from the panel your thoughts on that Who would like to tackle that? I can start because it's an issue close to my heart it's a really good point you make about the technology can move but the trust needs to be there and technology isn't moving as well so it's a vicious circle because when the technology fails that hurts the trust but often with sharing if we're looking to share with someone or they're looking to share with us there is this fear of the other side not having as good a standard as we do and I think part of that is that a lot of organisations talk about privacy but isn't really at the heart of their governance in their corporate hygiene privacy training for everyone at the organisation or just a few privacy officers scattered amongst various departments I think businesses need to take privacy and cyber security and their data security as seriously as they take financial security Data is an asset it's worth a lot and so they should be putting sufficient resources and money in to making sure that their practices are ahead of the game and it's not enough to buy something in and to create a new department it's something around embedding good privacy practices throughout the whole organisation and good ethical practices throughout the organisation I think that's a mammoth task and when everyone's already got quite a big to-do list sometimes privacy is last and it's a little bit ad hoc when really it should be for private organisations for their boards or for the governing bodies of research organisations and government organisations do they have privacy specialists on the board how much time do they spend talking about privacy I think until the governing bodies spend as much time and then that that theme and that care around privacy and data security is shared as to when you should share and how careful you should be and I'm still saying when you should share because that's the word that I come from I'm all for sharing but only if we're confident that it's within the expectation and the best interests of the person who I think owns the information which is the person who gave it to us in the first place who's the person that it comes from, the individual so I'd like to see a lot more time and money thrown creating good practice and I think the trust will come out of that but that is a really really hard ask in the current environment Could I just back that up? I really quite agree and I think the problem with trust is it's sort of like faith and belief and it gets on to sort of quite difficult ground quite quickly whereas I think if you can bring it back to sort of information governance and the processes and the practices that are in place to manage this sensitive data then you're on stronger ground and I think then trust will flow out of that I think that comes at the end rather than at the beginning Thanks Mia Got a question from online? Yes, okay here we are So we have a number of questions from online I'll get far off with one of them and this is a question to Marion and Peter You both seem to have to some extent a similar question about how to enable expanded use of the data you've collected Do you think a shared or common approach will be possible to deal with this or will we need specific solutions for each situation? Peter, I'll throw to you first maybe if you'd like to Yes look I mean this is obviously a very difficult question because obviously we will have the total body images we have the individual images we will collect pathology data where we also want to develop AI based on the histopathology and then of course the background each of us have a genetic maker and our maker obviously defines what is happening with our pathologic system and we have researchers who are working very much in this field on the polygenic risk core so we will have different qualities of data each of them actually can be identified if you have a lot of genetic data you can triangulate this not really worry with the pathologic data with the clinical data this is obvious and I mean one of the tough questions which I have been asked in the interview they were asking me Peter in all fairness the Australian AI researchers are dope but don't you think that there are even more better guys in the States or in Europe you have not to mention China would you share your data with these experts because as you have mentioned so there is I mean we put literally millions of dollars into the collection of data and we collect them and we will annotate them well and then of course I think in an ideal world we have been to researchers of working in this field but then of course all the researchers or many of them have been industry connections so it's really a very complex question which I can't answer I mean we have various committees in place we have a strategic advice of the team we have then obviously our major universities and having said this we are also part and all knows it of the SERP project so I am not so much concerned about the security of the data but more than the custodian needs to decide who will allow to have access to the data to the researcher in Canberra yes to the researcher in Paris no I mean these are really tough questions and I do not have an answer to these questions so can I add maybe add to that question is are we happy to allow a researcher who wants to in your case Peter do BMI work as opposed to melanoma work I mean I personally very keen about this because the strengths of the system are the longitude and the data so it may well be if we start at the moment three years of follow up we will get I mean we have an enegymasico hordcran ddorima from Monarch with 3.2 million ideally we will get follow up grants and we will collect over 10 years and of course the BMI will change I mean we are all humans and I think these are important data because the BMI is strongly correlated to the Cardio Metabolic Rift and I think these are information where we can do research and of course it has to be feedback to the person to the research participant to in some way will be often also a patient so I think the data needs to be feedback also for study participants I think it is a really hard question and I feel like I am giving a fudgy answer but I think you can start with general principles like de-identify things as much as possible only ask and use what you need for the outcome that you are generating so you can have some broad principles and guardrails but then I think you are going to have to descend into the particular for a tool like the one that Peter was just demonstrating that would have a very different approach from something which is more static information around answers to questions and things like that so I think that we might have to work out because also we won't know what the potential uses are so it's hard to anticipate them coming but we could put guardrails in such as if something was for a publicly funded health outcome was reproved by ethical researchers maybe that could be within the guardrails but then what happens when it's something around a commercial product maybe something like this is where I'm going to show I'm not a scientist or have any medical expertise but in this maybe it's the manufacturer of sunscreen wants access to this data so that they can make a better product and okay we benefit from that product but it's a commercial product so is that outside the lines so I think it's a really excellent question that the person has asked and I think we could have some guardrails but they'll never be good enough and every time I think it will be up to the custodians to make a call if I just can say indeed we are able to measure and eventually also automatically to assess the degree of sun damage and the area of sun damage actually this can be easily expected and this example which you bring with the sunscreen industry is just one example but of course it goes even into military or industry where people are working outdoor so you could then I mean brother is watching you if you image someone over 10 years and then you can even on the pattern of the sun exposure you can then define if this is related to his leisure activities because of serving or because of the professional exposure so there are quite a few significant ethical questions interesting another question from the room so question in the university research sector we've been talking a lot about sharing government data previously sharing data sharing moving focusing on data custodians and their role in determining who has access to data in the university sector we see a lot of examples of where health consumers researchers data custodians have agreement on a process for data sharing on the policies for data sharing and the systems are in place and they work well however in those examples where the process is falling down and falls short is where the data custodians and the people in the research involved in the research know what they're doing and all agree to it I think it's a good idea to share data and then it gets handed over to the lawyers for the different universities and so research ideas that can take 40 40 days for new researchers that have never met before to discuss a new project idea working over shared data can then slow down and take months or years for their lawyers at their institutions to come to terms to drafting a contract or drafting an executed data sharing agreement so the research can effectively stop and not go ahead for what is otherwise everyone agrees is a good idea how could that be addressed get a new lawyer I think a short comment we have to deal not just with university lawyers which are usually they understand the research questions quite well we have to deal with the lawyers obviously in the public health system and at least in Greenland each health and hospital service district is completely independent so a lot of time is involved with discussion with lawyers and so on the data obviously is another one because one or the other the service district says is done with the research project within our hospitals so we own the data not some universities so it's a very complex issue and I recall well ACIF was telling us we want the bold project sometimes I think it's the project is too bold so Tories could I ask you for a comment because I suspect this is a major issue for you thank you absolutely I think it's a very good point and I would like to point out that a lot of the research we are doing is international research so quite often we collaborate with institutions in Asia and the Pacific in Europe of course in America and it's very challenging to navigate these data issues and electoral property issues when we work across different jurisdictions that can discourage research in some cases and and I would like to say that obviously it takes time, it takes a lot of effort to get the legal teams to agree on the sharing of data across different countries especially in countries with different very different legal systems but it is obviously very important and some of the countries low income countries low income countries will benefit even more from data sharing and high income countries so it's very important to make this extra effort and to self-convent the obstacles and hurdles to get the agreement and do the data sharing of course now with your attention to and privacy issues and so I think that's very important and also it's important to take into account cultural ideas as properly in my introduction about the importance of working and involving and respectfully engaging with indigenous populations in Australia and overseas so I think it's very important to to take explicit to account the cultural issues human data or cultural knowledge from traditional custodians of this information and data and of course this complicates the process but it's absolutely essential to do that and and of course to take advantage of the knowledge that these cultures and different systems of knowledge can offer right thanks I'm just going to say it's fundamentally important to listen to the lawyers because what you do with the data must be lawful and the law is complex and it differs between jurisdictions and it differs between countries but there is some law for example last time we went to the GP and the GP wrote some notes about you whose data was that and you might know better than I did but I think the law says that that data is actually the GP's data his professional opinion about what you've told him and he's written it down and maybe prescribed you something so it's actually his but then do you have some rights so the patient may have rights and legislation patients do have rights to change things they think somebody's written down something about me and it's wrong and I'm going to tell them it's wrong I don't think the person that owns the data has to actually change the data but they need to keep a record saying we think it's wrong so it's sort of complex and you need to engage that complexity rather than hope it'll go away because I don't think it'll go away and just while we're talking about that there is an additional complexity which I think Satyrus was adhering to which is Indigenous people and data sovereignty so that's something that's come on the scene relatively recently but it's certainly an issue for Aboriginal Torres Strait Islanders it's certainly an issue for our colleagues in New Zealand and in other jurisdictions so it's sort of an international movement about data sovereignty of Indigenous people and it's a challenge to manage in that context because it's possible that those ideas will spread beyond just the Indigenous population and that will be another degree of difficulty for us I think I just want to make one addition it's kind of like this Mia Cooper something I've been really guilty of in the past particularly when I started in the public service was asking the lawyers the wrong questions and the wrong question is can I do this using BLAR whereas the question it's a lot more helpful if you ask the question how can I do this legally because they might then look for different ways of achieving the result that you're looking for and I actually I reckon I've wasted a year in the ABS when I was working there asking the wrong questions of lawyers so that can really help there's some new legislation which I have to talk about which I used to work on the data availability and transparency act no longer the DAT bill which at its best will also help override a lot of the complex Commonwealth legislation which prevents sharing so I'd encourage people to think about how that might actually be able to help them as well and if I could just add to that I think it is really important to engage with your internal lawyers and the views I'm going to express in this bit are not lifebloods abuse but my personal view as a lawyer I think a lot of lawyers are not well set up to assist you doing what you're wanting to do we're taught to advocate for one party and that's our job and so a lot of lawyers tend to look things through blinkers and if you ask a very tunneled question a lawyer who doesn't perhaps have time or experience will stay in the tunnel rather than working with you to try and pull the question apart a little bit I don't think that all lawyers are actually that well suited to work in this area because it's not about you're not negotiating with the other side to get a win which is often how some lawyers approach things so I think I would encourage people to engage early with their legal teams don't they leave it to a 450 on a Friday when the law has already got plenty to do and for a lot of organisations it's a small legal team it's a big organisation your request is not going to be a priority so it's easier to just say no and then to leave it dwindling that's not good practice so I think as organisations what we could do is engage early with legal teams but I also think organisations more broadly need to it's going back to the question we were making before about building trust it's actually equipped people in privacy to get a couple of in-house lawyers to manage a large organisation and hope they cope or use external lawyers invest in privacy professionals who can assist different parts of the business because you're going to get a quicker and more nuanced solution if you're dealing with a specialist and that's a specialist who doesn't just know privacy but they know you and they know your business and they know what you're working through so the earlier you can plan for it the better and what I would say about just to add about that's only just on the internal on your side it is incredibly frustrating if you were trying to be very practical on your side and you're dealing with say a North American firm who are representing and again it might be someone very small it's a small piece of their organisation you're dealing with a lawyer and it's just not a priority for them it's impossible to get some lawyers to move away from templates and templates say no so the more investment we can make in privacy and data the better outcomes I think you'll get Graham, is your comment related to this? A short comment I really like both of you and Philip I think we have to ask the lawyer the right question and really as you we have to equip the lawyer in privacy and this is certainly a process and I think our our example is quite quite a challenging one but I think we need to address it thank you So is your comment related to this, Graham? Well it's related to the indigenous Graham Galloway University of Queensland we've been talking for the last probably 10 years around fair internationally now and a lot of discussion last year's international councill research infrastructures that always needs to be associated with care and I think in Australia and I think it's good for our DC to start adopting let's be fair but care approach to data and that involves talking to indigenous populations at the beginning not coming to them at the end and saying oh we've got this great project and you want us to want to join we need to be talking to them at the beginning how can you contribute to this project and how can we work with you so for those of you who haven't looked up care there's a lot of work being done international around care collective benefit authority to control responsibility and ethics we can all repeat fair acronym now we need to also be using the care acronym equally just as much thanks Graham so I think we'll go online and go Gith okay so there's a whole bunch of questions online quite an engaged discussion also in the chat I think so there's a number of questions that have already been covered around changing the discussion maybe from data sovereignty about owning the data versus ideas about trust and getting sense of the trust and building trust around the data one question that sort of also came up in different forms and I think is an interesting one is imagine a situation in which you are gathering more data than you need for your direct use and with actual minimal effort you can collect that data and bring it together might be of interest and use for others and for other research projects are we ready to aim for that added potential or do we still really need to focus on the hearts and minds and making sure that there's trust in just the basic research that can happen with the data I think a really interesting example again was the establishment of the multi agency data integration project by the ABS so that really kind of moved the commonwealth from a single specific purpose use of a data set so there was this kind of idea that you would actually integrate a data set and then it would be used for quite a specific domain specific project and then it couldn't be reused again and that was actually baked into some of the commonwealth principles around data integration and Peter Harris who was the chair of the Productivity Commission a few years ago took a very active role in a Productivity Commission report into data use and he said that was basically equivalent to burning books only using that information once the way the ABS and others in the commonwealth, AIHW included have made gains in this era is through establishing very clear governance arrangements around these multi use data sets and it's not always particularly streamlined but it's getting quicker over time but it took a year's worth of really hard work to actually establish arrangements that would allow for that safe reuse of information for different purposes and again I get to that point about precedent once you start doing a few of these things a few times and working out what works and then putting solid documentation and governance behind it that's really your friend I think in terms of being able to use data sets for more than we ever intended them to be useful and that's so true of administrative data that we use in the commonwealth you know all of that MBS and PBS data that we talk about that wasn't collected for research purposes that's administrative data for administrative purposes and now that's become the backbone of such a huge amount of research and there are opportunities to keep building on those sorts of successes Thanks Phil Can you hear me? Hi, Sally Pearson from UNSW I guess I just wanted to first focus back on the purpose of the panel discussion today make a comment and then ask a question of the panel So I think we've been sharing sensitive and identifiable data for a very long time there are mechanisms to do that there's safe ways to do it sure we would like it to happen faster quicker more expediently but I think we do need to acknowledge that there's great things in place that allow that to happen already and there's some really wonderful research that has resulted from that in my area of expertise that the world has shifted greatly things that weren't possible five years ago are now possible so I think it's really important we focus on some of the good things that have happened but I guess what breaks my heart is the waste of public money and what I mean by that is how do we uplift this area so we are focused on generating high quality output for the public good rather than spending all of our time building the enabling data infrastructure so I guess what I would like to ask the panel is if you had a magic wand what would we be putting in place right here and now to uplift that so we start investing in outcomes rather than the inputs so we've just got a couple of minutes so I'm going to give the panellists each 30 seconds to answer that maybe with one idea I should have asked that earlier there's a real challenge so I guess we'll run through this order I think actually to get maximum impact from the data that we have we need to have I hate to use this term again better data storytellers because we have all of this amazing information that's kind of hidden and inaccessible to policy makers because they're talking about the Bayesian ridge regression algorithm that they used and focused on that and a lot of the time a simple cross tab or a simple graph presented to the right person in a timely fashion makes a huge difference so if we had more people that knew how to talk with that kind of cut through it would make a big difference Laws are introduced and frameworks are introduced because people don't go of themselves and do the right thing so if I had a magic wand everyone have had appropriate awareness of how to do the right thing and they actually do it if people respected other people's data and how to use it and how to look after it you wouldn't need rules and frameworks or the bureaucracy that goes with it My thinking was we have a virtuous cycle and we're sort of moving around that cycle but we're not at the end yet where we don't get the benefit until all that data is actually used for the sorts of things that Sally's been talking about and if I could mix a metaphor and use a Melbourne example in us to a Sydney audience I sort of feel that if we're in the Melbourne Cup and I don't think we're in the final straight I think we're sort of in the back and sort of turning the corner towards the final straight of getting that virtuous cycle in place so I think we need to collect the data collect it once you use often and generate the value Is it a tourist? I think there's a lot of value in large scale biomedical databases CLDK Biobank many other examples in the last week I think this is a very valuable investment for research and for the public group and of course it should be regularly updated should be documented with more focused research data and of course protocols and procedures are important can be time-consuming but it's important to ensure the quality of the data and the correct use of data so I think it's an important investment for the public Peter Furthermore, thank you for having me quite a few things that are resonating with me what's resonating most is what Philip was saying to create a precedent in this congenital landscape and also to tell the data story even better Thank you I'm just mindful of the time in our now zoom world of course everybody online probably has a meeting in three minutes so they're going to close this meeting and go to another one so apologies for those people so it really just leaves me to sum up I mean I was quite struck a fantastic conversation I'm sure we could go on for another hour in fact a really great conversation so I was really struck by a couple of things I guess that one of them comes down to the fact that I think the last point that was made and the answers and it went back to the very first question we must be able to demonstrate the value of doing this and if we can't we will lose the fight it's as simple as that and I have absolutely no doubt the battle has been won in certain areas I would hesitate to suggest that battle has been won across the sector and I have to say I'm absolutely delighted to hear that government is now sort of I guess pushing and saying well actually we have to be in this space and we have to push forward and we have to not only help make it happen but demonstrate that it can happen and see all the great things that come from it so you know I leave with I guess really hopeful it looks like a fantastic landscape of some great things I was struck by the hearts and minds is an ongoing activity and also struck by the fact that you must resource that so you just can't you know I'm a researcher so I'm allowed to say this you cannot leave that to the researchers okay we must find people who are good at messaging and get that message out there effectively a couple of other things that came up I think the comments about how do we just go down the path and get it to happen how do we make sure that we've got all the right agreements in place and we can share data effectively we must get better at doing that part of that is solving the jurisdiction problem that Peter and Graham talked about it's you know there are seven states we just must have we must have national solutions if we're going to solve these problems seven times or even worse 50 or 60 times because of all the different health networks that we've had here a very long time and so I think one of the great challenges is to find the national solutions we didn't actually talk much about who owns the data I have to say but I did think the example about commercial outcomes commercial outcomes are not bad things in a lot of instances so we should be quite open to that happening and if it happens and you know lots of Australians benefit from a health sense because a new company grows and a thousand people get employed that's a good outcome and so we should be ready to embrace that so so it really just leaves me now first of all to thank Ian and Ash and all the ARDC team for putting this together a fantastic conversation I suspect we could do this every six months for the next five years and it would be a really rich conversation maybe that's a great idea but there will be lots of other of these leadership series conversations and I would encourage you all to get involved and finally I'd like to thank all the panelists and Peter for giving up their time and their fantastic insights and maybe I can ask the people in the room to show your appreciation and the people online can clap too if you want so please stay engaged there will be lots of stuff happening in this space and we look forward to keeping the conversation going so thanks everybody take care