Tim Chaston 0:03 Well, if we're all ready to go, we might start with the Welcome to Country. So as we gather for this meeting physically dispersed and virtually constructed, let us take a moment to reflect on the meaning of place, and doing so recognise the various Traditional Lands on which we do our business today. I live and work on the Wurundjeri Woi Wurrung lands of the Kulin nation. Perhaps we could all acknowledge the Lands that we work on and live on in the chat, while we acknowledge the Elders past and present and emerging of all the Lands we work on and live on, and their ancestral spirits with gratitude, and respect. So today, we are here to make the first steps towards the task of supporting the entire HEAL consortium in its range of disparate themes, with data and decision support systems. To this end, it's time for us all to have the first conversation about how to consolidate various data and computational resources that we hold across the country. And how to best facilitate the research that is crucial for adaptation or mitigation of environmental change. You're all likely to have come across FAIR data principles: findable, Accessible, Interoperable, and reproducible data. I'll just share my screen. So I think we're all probably familiar with FAIR data principles. Perhaps we can focus this hour and a half on the first word in this acronym, findable and set the scene for assembling a central source of information. First, we'll hear from our speakers who will introduce four exemplary research infrastructure projects. Our first speaker is Dr. Ivan Hannigan. He's the new senior lecturer in Health Impacts of Climate Change at Curtin University School of Population Health in Perth. Congratulations, Ivan. Starting in 2022., Ivan will be the director of the World Health Organisation Collaborating Centre for Environmental Health Impact Assessments based at Curtin. Ivan completed in NHMRC postdoctoral fellowship at the Centre for Air Pollution, Energy and Health Research at the University of Sydney focused on environmental epidemiology and data science. Ivan applies a multidisciplinary approach and his research seeks to understand the health impacts of air pollution, socio-economic disadvantage, extreme weather events such as bushfire smoke plumes, dust storms, droughts and heat waves. Ivan has experience in the integration of health, social and environmental data, which involves reconciling a large and complex ecological datasets with information from health and socio-economical population domains. I hand over to you, Ivan. Ivan Hanigan 2:48 Thanks, Tim. I'll start sharing my screen. And thanks for the introduction. The Centre for Air Pollution, Energy in house Research or CAR. What is an NHMRC Centre for Research Excellence? Can you see my screen? And it's hopefully going to turn into presentation? (Yep, we can see that.) Great. Yes, I got my postdoc here at CAR back in 2017. And I set out to do two things: One was to do environmental exposures and epidemiology, and the other one was to consolidate all of the data that the first iteration of the CRE had amassed, and coordinated into a data platform that we called CAR's Data and Analysis Tools: CARDAT. And what it needed to do was to pull together all the different researchers from different institutes and overseas, and make something that was able to collect the data, share the data, and ultimately to use the data. So we designed the system to make primarily our researchers, but also their close companions, their collaborators; able to discover and acquire the data, do the data analysis and to manage the outputs. And I'll step through exactly how we're doing this. We have a data inventory. And that has everything that we have, plus things that we know about and even things that we wished existed but haven't been done yet. If you can find it in there, then it exists; then you can ask to access it, if it's restricted. If the answer is no, then we work with you to be able to get it. But if the answer is yes, you can just go and get it. We have both online cloud platform, called CoESRA leveraging off the terrestrial ecosystem Research Network where I worked previously, we built this system called CoESRA. Or you might just want to download the data and do the analysis on your personal computer. Either way, we encourage you publish the results in form policy and feed the whole system back into itself. And so those outputs can be managed by CARDAT and shared as appropriate. So this is how it all works. I'm going to step through an end to end procedure. From the start, where the research and the collaborator in this case, epidemiologists that I was working with, have been collecting to contact tracing the locations of infection from a virus. It's very topical. This is not about air pollution, but I thought this is about environmental impacts on virus, which is very interesting and its geographic spread. Now, with geography, the hypothesis emerged it was something to do with the proximity to inland waterways. And so GIS analysis was constructed. To do this, we helped data discovery and acquisition in the form of waterways, urban areas, coastline... you can see that these dots are actually locations of health outcomes. This is private and de-identified and able to use and intersected with the environmental data. So we follow the five safes, which is a very rigorous approach to confidential health data, which means that we can then link it with environmental data, and also socio-demographics. In the next stage of the analysis, what you do is you can go into the system and get the data with a GIS that's already there. Or if you've got a GIS, you might want to take it away and do some stuff. What actually happened in this example, was a team of a geographer, myself, and epidemiologists, statistician, and mathematician, and an entomologist all got together and pursued the different strands of a research project. And we discovered that while proximity to waterway was an influence on the risk of infection, it was different between urban and rural populations, indicating something to do with susceptibility and vulnerability, and other aspects of transmission. The output was very easily put through the general traditional peer-review process. And this is the paper that came out of it. But as I said, we put it back in. Meaning that it can be reused and repurposed. And I use it for training. It's a really, really good example, for training on how to manage health data in a spatial environment. It's findable because we have this data imagery I mentioned, and this is a standard metadata language that is machine readable. And this Ross River Virus status that is there able to be found and requested. Once you get access to it, we can go through a stage two analysis and secondary analysis, I'm going to show you something that we did in the cloud. This is CoESRA. You can launch a virtual desktop in the cloud, ask for as many CPUs and RAM and days of access that you want. This is in a browser. So you operate in your web browser. And it shows you a desktop in the cloud, and it has the data access right there. You can take a copy of this data set, put it in your own workspace, which is on the right hand side, and do the analysis deed or do something new. In fact, that's what we did. We were very interested in this particular dataset, we extracted the risk effect estimate, which is the proximity to waterways estimate. And we could verify it was accurate and maybe do something different. But what we ended up doing was to actually hopefully get a new output. We linked it in New South Wales, which is the other side of the country. The data was from Western Australia. In New South Wales, we linked small area-level disease rates with the distance to water in that location and also population density, population type. And we estimated the spatial indicator of the excess risk, which is on the right hand there. They are only for coastal waterways by the way. So this is currently published as a New South Wales DPIE report, which indicates that, for example, the mid-north coast around Byron Bay has some small pockets of high risk, and they should be targeted for interventions, especially if there's coastal inundation risk that this might spread and that could be modelled as well. So in conclusion, the CARDAT system allows us to find and share the data and tools across the consortium. There's no need to download the large datasets unless you desire to have the datasets, and you can synchronise your changes locally to the cloud. These cloud computers can be completely pre-configured with tools that are sometimes hard to get and scaled up to get more compute power and grunt. But beyond that, it actually is just an easy way to work collaboratively on the same projects. Even across these big distances between Western Australia and the east, or our collaborators overseas. I did a project with collaborators in the US EPA and it was great. So in summary CARDAT makes data and analyses more fair, findable, Accessible, Interoperable, and reusable. Thanks. I think we're gonna have questions and comments all at once at the end of the session. Is that right, Tim? Tim Chaston 10:15 Yeah, I think that's right. I think we've got to move along pretty quickly. So thanks a lot, Ivan. That's terrific and well done with sticking to time or even actually getting us ahead. Now we're going to hear from Dr Aiden price. Aiden is a research associate at the Centre for Data Science, where he manages the AusEnHealth project, a National Environmental Health Strategic Planning Digital Twin. Aiden research is currently focused on spatial and temporal analysis of environmental and population health data, identifying the impact of bushfires on human health and conservation focus work through the lens of aesthetics in the Antarctic Peninsula. I hand over you, Aiden. Aiden Price 11:01 Thanks very much. So we're just going to go through I guess the platform. I'll give you a very brief demonstration and talk about I guess, where we're at, where we are, and where we're going. So I'll start by saying, I'll start by going to the next slide. So a bit of background information. We're a multi agency funded projects focused, I guess, in environmental health as the name suggests, being AusEnHealth. You can see some of our supporting organisations there. It's led by FrontierSI and QUT. We've also got support on the software side for NGIS, who are developing the platform for us. Our aim is to provide access to environmental health data for visualisation analysis, and provide tools to support adaptation planning, vulnerability, assessment and decision making. At the moment, we've explored Australian data and consulted a range of experts from research and environmental health organisations; and we're in the process of building the AusEnHealth demonstrator. And that's what I'm going to show you a bit of today. So I'm showing the Technology Readiness, I guess, to give you an indication of where we're at as far as the capabilities of the platform. In our initial, I guess, planning, we set out to deliver an experimental proof of concepts so that we'll be aiming to demonstrate the technological capabilities of such a platform, but with inflexible design or inflexible capabilities that we'd have to tailor towards specific demonstration. And what we ended up achieving was a technology readiness level of six, which really means that it's a functional platform as far as users being able to navigate through different components and produce information outside of what we would typically demonstrate. So it's, it's coming further, I guess, than we thought it would. So today we'll talk a little bit about, I guess, the core UI components of the platform and its capabilities, starting with a landing page and some information about the data. And then I'll talk through three of the platform's demonstrators. And then I'll finish with some of our next steps, contacting and all the rest. So this is the AusEnHealth projects landing page. You can see, I'll direct you very quickly to the six panels along the bottom, starting with heat risk assessment, climate change, assessment, etc. So these are our demonstrators for the platform. The idea is that we'd designed the demonstrator in a very modular way. So our two focuses from the onset were extreme climates and air quality, and their associated health outcomes. And so we've got here a set of six demonstrators related to heat and air quality focused on the different impacts, I guess, that we could give to different end users to really motivate the future, I guess, what AusEnHealth might become. So these are six, I guess, tailored different demonstrators, which show different facets of extreme climates and extreme air pollutant levels. Great. The idea, I guess, of these demonstrators is that we can take these either two different designs based on how well AusEnHealth goes in the future. Or we could introduce future modular panels, such as vector borne disease, mosquito borne disease, water quality, food, etc. While we're here, I wanted to highlight, I guess, the Environmental Health variables that were displaying. So I've got there there are 158 different environmental health variables, those span across climate, air quality, demographics, things like land cover, water cover, and come from a pretty wide range of data custodians. So in the first phase of AusEnhealth we focused on national data, and that's to really display the value or, I guess, promote the value of such a national platform in the future. And so that comes with it the set of, I guess, typical national data custodians for environmental health, so Bureau of Meteorology, ABS, fire risk, etc. Excellent. And we've used those data, the Environmental Health variables to generate indicators for extreme climates and air pollutants, and also for heat waves and cold waves. And we combine those with demographics and hospital counts, etc., to come up with population vulnerability indices relating to heat and air quality as well. So now we'll jump into the first demonstrator. This is our heat risk assessment demonstrator, focused, I guess, on extreme climates generally. You can see on the left, maybe I can mouse over, we've got a cold section as well, where we can, I guess, flip the extremes and come up with some, some really good data and insights using that data for extreme cold as well as extreme heat. And I guess the motivation is, is pretty well drawn out today that, you know, extreme climates and the future of climate change is a really important problem to start tackling. You know, on the on the range of climate adaptation intervention strategies, research has been able to draw out trends and research that indicates the risk of the health and also just for the general public to raise awareness. So I suppose the first demonstrator really highlights that while there are platforms dedicated to displaying environmental data, they don't really focus on, I guess, the perceived risks that, you know, alongside that data. So for example the Bureau of Meteorology, which is a host for climate data, displays daily temperature. But what we do here is a little different in taking that temperature and turning it into an indication of risk. So for example, the one that's displayed at the moment is our heat high percent variable, which displays a region based comparative percentile, I guess. So per region, we compare to its historical temperature, and then determine how hot it is relative to that historical temperature. Cool. So this is SA 3 level, we have the opportunity to go to SA 2, we have the opportunity to change years, to change temporal resolution, etc. And we've got different representations of the data. So we've got a choropleth map, which is I guess, front and centre. And we combine that with tables and charts as well. And you can see there's an option to turn that temporal as well down the bottom. So say, we're a government official and we're interested in a component or a specific region in Australia, one of the SA 3 regions here. A user can search in the top right, we've got this search panel over here, or they can just click on the panel or click on a row in the table and everything works kind of together in the platform. So you can see here, if I want, say heat vulnerability index, I would select that from our filter at the top, I want, say 2016 data, and I want to choose Alice Springs. Then in the next slide here, you can see what would have happened, I'm not using a live demo, because they're potentially fraught with disaster. But you can see here that the platform would zoom into that it will show the location in the audit histogram down the bottom, and it will highlight the row in the table as well to really focus the user on what information they want in particular. So this talks about, I guess, data that we have now, and data that's historical. So I guess the next question, and what we've been talking about a lot this morning is, what will happen with future climate. And so our next demonstrator is focused exactly on that. So our climate change assessment demonstrator. So we've got here on the left hand side, our 2000-2010-2020, and then our forecast data, which comes from the recent IPCC reports that came out earlier this year. And this really demonstrates a very quick inside view into what's happening across time for each of the states in Australia. And we've even got a time slider down the bottom right here, which shows I guess those changes over time live if you were in the demonstrator. While it opens at the state level, we can also go into a state, and we can see the data breakdown there at the SA 2 level. I've also got, and I didn't mention this before, the opportunity to download the data because the team has put in a great deal of effort collecting, cleaning, processing data, and we'd like that to be usable by others in the community rather than just being stored in some USB somewhere and not seeing future use out of, I guess, the sustainability push of the platform as well. Great. And so I guess the focus here is on high heat days. So these are days that pass the 85th percentile of historical temperature. And you can see I'm on 2020 at the moment, and the the, I guess, highlight of this climate change assessment is changing that timescale. So just going to 2060 there shows the extreme difference in the number of days that we're seeing above that threshold, and therefore shows the value of having a tool like this immediate disposal. So this is, I guess, again, where we're moving forward through the time. We've talked about data. We've got data that might be, but what about data that we don't have? Well, the third demonstrator that we are using, or the we're showing today, is our heat health data analytics demonstrator. The idea of this demonstrator is to connect to a range of open data repositories using a query system that allows a user to bring the data, or bring alternative data or separate data into the AusEnHealth demonstrator. So this, I guess, demonstrated uses the latest multi cloud technology, I think, based on Google, and represents more or less the future of data visualisation. We have our own data that we've processed, and we've gone through a lot of time to reduce this platform, or this demonstrator takes data from elsewhere and makes it very seamless to bring in efficiently, I guess, ready for extracting value. And you can see on the left, I've selected hurricanes from NOAA, I selected just some of the parameters, I think, the storm identifier and the season, and it generates the query automatically. And all I'd have to do at the bottom is press Run, and it's generated straightaway into the table there. And this is under development. In future you'll be able to visualise this visualise it with other parameters, etc. But just showing the capability there. And for now, I guess, that wraps up the demonstration. So moving forward, we're looking to finish the demonstrator. There are a couple of UI components and capabilities like I just mentioned before, and continuing to explore other environmental health projects and data sources. We're also implementing a roadmap for future phases of AusEnHealth right through to operationalization. And we're exploring future collaborations and are excited to show more individuals or organisations the current and future of AusEnHealth administrator. If you've got any feedback on any of the I guess, slides that I've just shown, about application value, additional functionality or anything else, please feel free to email our project director, Paula. And also, if you want to book a demo session, we're happy to arrange anything like that as well. So please, if you want to know any more information or you're open to collaborate, you've got a great data source that you'd like to share, please feel free to make contacts. We're happy to chat. Thanks. Tim Chaston 22:22 Thanks a lot, Aiden. That's terrific, terrific development of user-interface and it certainly seems like there's much common ground and good opportunity when we consider CARDAT and AusEnHealth in the same half hour. I now invite Dr Yuming Guo to present the Global Health Research. Yuming is a professor of Global Environmental health and Biostatistics and head of the Monash Climate and Air Quality research unit. He is supported by an NHMRC career development fellowship. His research group focuses on environmental epidemiology, biostatistics, global environmental change, air pollution, climate change, urban design, residential environment, remote sensing modelling, and infectious disease modelling. He's developed and participated in several large international collaborations to assess the impact of air pollution, residential environment and climate change on human health. His research is primarily supported by the National Health and Medical Research Council, the ARC and China National Science Foundation. He's also appointed as an Adjunct Professor by the University of Melbourne, University of Queensland, University of Oulu in Finland, and he was awarded a visiting fellowship at the London School of Hygiene and Tropical Medicine. Take it away, Yuming. Yuming Guo 23:40 Thanks, Tim, for your introduction, and thanks for inviting me to give the talk on the discussion about multi-country collaborative study. You know, first I want to talk about, you know why you want to perform a multi-country collaborative studies. Now, the most important thing of that is, you know, to perform innovative studies, for example, we can develop a new method, new technology and new tools in this area to promote health risk assessment or climate change in our health of global environmental change in our health. And also, with the use of big data to solve the problems which can't be solved by the single cities' data or single country's data have seen already data. If we have, you know, the big data enough, we can do a lot of work. And if you do, too great, the new knowledge, new evidence, and the evidence will be, you know, will be reliable because we have a big data and we use reliable method. And you know, because this data is from multi-country across the world, so the evidence can be generalised to different areas and the policymakers can use evidence to develop some policies and to fight a climate change, to make policies to mitigate climate change and to make policies for climate change deputation that is very helpful in this area. And also, you know, from developing the multi-country collaborations, we can develop a relationship between keeping countries people and allow, you know, you to develop a relationship with the local government, national government and international agencies, and we can provide the evidence and the data to them, and then we can use this data to do something. And also it is good opportunity to communicate with different people from local to international, and you know, big data and reliable evidence can make a big difference in the world. And if we just have data from a single city, you know, the evidence might not be good enough to persuade the local government, to persuade the government, to persuade the generators, we need to make the data and you know, more reliable and evidence more reliable. Currently, this is the first study I wanted to introduce. Now for this one that use a multi-country study to look at short-term effects of environmental data like, air pollution, temperature, flooding, wildfire smoke, and also hurricanes, cyclones, and other you know, extreme weathers, or natural disasters. And for this dataset, for this collaboration, we have collaborators from 43 countries. Currently, we have already collected data from 750 cities. And this collaboration is new, exciting, and many people have expressed their interest to join our network. For this dataset, we have daily analysis data for mortality, and we have expanded to include the hospitalisations and emergency vulnerabilities and other health outcomes. And for the environmental data, we have weather conditions, air pollution data, and the way we link it to other environmental factors. You know, this network was developed in 2013. This is initiated by me and Professor Antonio Gasperidi, from the School of Hygine and Tropical Medicine and let's tie him we two just graduated from PhD study. And so we discussed how to make a joint collaboration. And now, at that time, I developed a network including Australia, Brazil, Thailand, China, and the Philippines and only development another network in Europe. So we discussed how to make this as a joint network. So we now successfully, jointly, were successfully joined together to make this as a truly global study. So we have done a lot of work in this field. And another one that is in the Asia Pacific cohort consortium with the long term effects of evironmental factors. This is just in each city and for this study, we want to look at the long term effects of air pollution, weather conditions, greenness, residential environmental, noise on the human health. This is study, this network, including 22 cohort studies, about 6 million participants from 10 countries or regions. We have already gone to great lengths from other PIs or the principal investigators from this cohort studies and they are happy to share the data with me, with my team, and we can get to the data for the mortality, morbidity, lung function, blood pressure, and other biomarkers like the DNA methylation and other biomarkers. And also for this study, we have already done some estimation for the air pollution, greenness, noise and the weather conditions then we can link this estimated data with each process, we can assess the long term effect of these environmental factors on the human health. Here's some examples from the network and is the first one that is you know the temperature and the mortality. You can see if we have big data, we have different countries data, we can compare the results from different countries and we can get the percentage without, for example, for the minimum temperature. We can see for most of the countries, the minimum mortality temperature is located at 75% higher in other countries. And also based on these results, we have estimated how many deaths can be achieved while in cold and heat in these countries. Also now because we have many status data about 750 cities data. So, we can get to the effect estimate between temperature and mortality in this database and we can develop a meta-regression to estimate the relationship as each grade of the world and then we can use this relationship, to estimate how many deaths attributable to temperature in each grade, combining the temperature you get grade and population and the mortality rate you that's great. So, that's that means, you know this is truly global and that association has already been reflected by the predictors like socio-economic, population density and the local temperature and also represent rate. So, basically, you know, big data can provide the evidence and also this is the result for the air pollution and mortality in this database. We can say, you know, there is no threshold on the PM2.5 on mortality risks, that means at any level of PM2.5 or PM10 there is a risk of about anything, because you know, this data is from many countries and many cities, the evidence is reliable. And also here, you know, because we have estimated that the data for the wildfire smoke, wildfire PM2.5 and we link this predicted wildfire PM2.5 in this multi country network, and we estimated the relationship between wildfire and the mortality risk, we can say for total mortality, cardiovascular mortality and maternal mortality, all of them are related to the widlfire PM2.5. And also we can get the leg effects and also the cumulative effect. Okay, but you know, we can do a lot of work using the multi-country collaborative studies, but there are also challenges for collaboration. For example, data collection, because, you know, different countries have different policies or that data sharing or data cleaning, and also the data linking. Or, you know, we need to know to discuss or argue with data providers how to share and how to make the data matching to each country's policy and also because we know the data, the level of the data are different in the different countries. For example, in Thailand, they provided the province level data, but in Australia we provided an SA4 level data and then how to match the spatial level that is also important factor. And also funding support, because, you know, this is a larger international collaboration and at the very beginning, we just develop this collaboration by a cost effective model that means, you know, if they already have the data, we can come together to work on this work, but, uh, you know, with you know, the programme growing, we needed to find some funding to support this collaboration. So, you know, in Australia, we did not have enough support or we needed to seek support from international agencies or from different countries, as the funding agency. So, currently, we deal with effort to get support from different level of support, and also the communication sometimes you know, communication is not effective, because you know, we have a time lag in different area, we want to find a good time to communicate that is really hard and also you use different language sometimes, you know, some people are very good at English and we sometimes it is harder to get understand what you are doing or what we are doing. We need to find a way to solve the problem. And also, you know, the vision. At the very beginning of the race is very simple, just that we want to develop some new knowledge and provide some sort of evidence, but always programme growing we needed to provide in the vision and the location of our data our work. You know, with the International interest and interest from different people... Tim Chaston 35:08 Yuming I'm sorry to interrupt. We must move on now but, um, thanks so much. They're really impressive that you've managed to get around all those challenges to collaborating with so many people from overseas. But now we best hear from Dr Nasser Bagheri, who will introduce very different environmental health research infrastructure one focused on mental health. Nasser is a Senior Research Fellow and spatial epidemiologist at the ANU and a recipient of an ARC DECRA Fellowship. He leads the visual and Decision Analytics laboratory at this in the Centre for Mental Health Research. His research has focused on the healthcare ecosystem and chronic diseases risk management, particularly mental health, cardiovascular disease, diabetes, dementia and multiple sclerosis. His primary research focuses on the application of spatial risk assessment methods and population based intervention studies, to provide epidemiological evidence for more rational implementation of strategies for the control of chronic diseases. I'll let you take it away now, Nasser. thank you very much. Nasser Bagheri 36:21 So today, I'm going basically introduce the aim and the area of the VIDEA lab. And thank you, Tim, for that introduction. We just need to, like, emphasise that we just moved to University of Canberra, our whole team. So now the VIDEA lab is now sitting at University of Canberra. So VIDEA lab emphasise the healthcare ecosystem combining visual and decision analysis for optimising communication, mainly the policymaker, in public and private sectors. We have different expertise in VIDEA Lab: epidemiologists, computer science, and people working in AI and machine learning. So, and our partners at the moment, we are focusing on mental health care ecosystem, and we are ongoing collaboration with Department of Health, the ACT health, with Community Mental Health Australia and office of mental health and wellbeing and other collaborators. What we are doing in VIDEA lab, basically, we are modelling the healthcare ecosystem. And you see here, the future drivers of health ecosystem focused on the future drivers of outcomes. So the healthcare ecosystem has different dimensions. One dimension is only, this is the clinical care, which unfortunately, we are spending too much money on clinical care, but there are other dimension in health care ecosystems such as health behaviour, physical environment, as lots of colleagues mentioned today about the relationship between air quality and health outcomes, social and demographic. So, in VIDEA lab, we are trying to quantify the interaction between these different dimensions and visualise in a way which is understandable for policymakers. And even we are focusing on the context for translating evidence to implementation, context is important. So we are applying a system thinking approach in VIDEA lab. And we are using different tools and expertise, such as artificial neural network, machine learning, GIS, social network analysis, Bayesian network analysis in this VIDEA lab. For example, here, this is one example of social network analysis. It shows that the connection and relationship between different providers of mental health in ACT as you see those dots the bigger size, they are main hub of mental health care ACT, which is near which other mental health provider refer the patient for these main hubs. And we can do the advanced analysis to see the central degree of similarity between nodes and the weighting and other types of analysis of social network analysis. This is just example how the pattern of mental health provision looks like in ACT. And other aspect we are focusing on Bayesian network analysis, this is basically funded by NHMRC Idea Grant leading by Marc Daniel, but I am a CI in that project as well. So, in this project, we are doing, how the neighbourhood indicators or neighbourhood environment factors can affect or have impact on cardio metabolic. So, they are generating, for example, indicator for healthy food index, for community resource index, for public health transport, or public open spaces and different types of indicators; they're called 20 minutes indicators and using the Bayesian network analysis to visualise the relationship between the different indicators and their impact on the cardio metabolic and then applying this knowledge to the another area in ACT to look at, to predict how the environmental can shape the outcome of the health outcome mainly caregiver topic. And other areas. We are focusing on VIDEA lab using the machine learning. This is the type of like, a self organising map which for example, we have mental health. We have a different dimension for service utilisation in terms of like, say for example, one domain, one variable for utilisation would be for example, using admitted patient data, for example, discharge or one layer shows here length of stay and, for example, another layer shows the readmission prevalence incidence and frequency of mental disorder using the patient admitted data. And then we can apply the self organising method to come up with a kind of a heatmap here to show with the utilisation services that are problematic. And definitely GIS is another arm of the VIDEA lab, which we are collaborating closely with another colleague team which they are focusing on Australian Dual Health lab. For example, in this study, I developed a social fragmentation index for whole Australia at SA1 level using the publicly available data from some from the Australian census. So and then, the idea was that those people living in communities with highly, socially fragmented, might have high level of mental disorders such as depression. I wanted to test this hypothesis. So firstly, I visualised the pattern of the index I developed for social composite index for social fragmentation in western Adelaide. As you can see here, those suburbs locating in the near the ocean they are less fragmented compared to the communities in Central City. And similarly for depression hotspots, you can see the various hotspots of depression using the GP practice data, because the GPs data they have information for depression and anxiety. Then use the multi level modelling to see how the contextual factor and the social fragmentation has the impact on depression, for example, one standard deviation change in social fragmentation the corresponding change in the in depression would be for example 6%. And here we visualise how the convergence between social fragmentation and depression looks like. And so, in mental health unit policy, we are analysing this mental healthcare ecosystem in three main levels: microsystem, mezasystem and macrosystem. For example, they are looking on the hospitals or like, say, they are then be gathering information from macrosystem, which is the hospital's data, and then if you look at the PHN level or state level, we are looking for gathering information from meza-level and macro-level. Then classifying that information based on the algorithm you're using because there's the algorithm or description for mental health, different types of mental health care, that could be daycare, social care, residential care, social care, psychological care, acute care, or clinical care. And then then that taxonomy done, then we are ending, for example, the codes similar like, for example A it shows that this coding is for adult population, and F0 to F99 is the ICD codes for mental health and our shows that for example, this is for only residential care. Then when we have this information, we're able to visualise and map then again for another type of the care we can do this type of analysis. And then we can compare or doing the benchmarking, for example, the pattern of mental health care in a city, any like say other PHN in Australia or even other countries, as long as we have the basic unit of analysis similar in all catchments. Here, as you can see, the compared the mental health pattern of care for adults in ACT and in Central Eastern Sydney PHN as you see, for example, accessibility is higher in ACT compared to Central Eastern Sydney, or the problem for example, for daycare for mental health, or in ACT and in Central Eastern Sydney is the problematic. This approach has been used in 34 countries: North America, mainly Europe, European countries, and Australia. Now we are doing this kind of integrated Atlas of mental health in Middle East and in New Zealand. So here's the recent published paper using this approach as I explained. For example, in the table, the paper I just published in the Journal of Alzheimer disease, it shows the impact of built and social environment factors, they can diagnose and estimate risk of dementia. So we use the satellite data. My colleague, Luke, he provided the satellite data at city level and then we can model from that point and then use the multi-level modelling to investigate the relationship between for example, NO2 and dementia risk. And dementia, we had already diagnosed cases of dementia plus, we estimated the future risk of dementia using GP practice data. And another phase of use the rule of visual analysis in supporting mental healthcare system, which is publishing in Journal of information management, using how the visualisation helps the policymaker to make a better decision. And also other papers on self management or some net using those hexagon heat map type-papers. In brief, these are focusing on at explaining the activities we are doing at the visual and decision analytic at University of Canberra. And happy to get comments from and question from all collectors. Thank you very much. Tim Chaston 48:45 Thank you Nasser. We'll move very quickly now. We've gone over time a little bit. So thanks to all of our speakers in the session, we've seen quite a few different types of data and some terrific systems for analysing them and for making them available. I'll hand over now to Kerrie Mengersen who will lead the conversation about how we can make all these resources findable. Kerrie Mengersen 49:16 Can you hear me? (Perfect) Alright, so I'm going to... Thank you very much for that. I'm going to just share some slides. And we're going to go into a panel session. So stay with me. All right. So thank you very much for the opportunity to have this panel session and acknowledging that I'm in Yugara and Turrbal land here. And I'm proud to be part of the very long tradition of learning and teaching that's happened on these lands here. We want to talk in this panel session about moving towards or proposing and having a discussion about a trusted online and Environmental Health Decision Support Platform. And the this is because we feel that one of the major challenges that we all face is the way that we can access and accessing data about exposures and health outcomes, and also the analysis and reports about those data. So when we can canvased in previous discussions about the challenges that we face if we're going to move forward, we need to know where we are now. And if we want to know about where we are now, it's important that we know what kind of data we have, and what kind of studies we have and how we can access those in order to understand where we're at, and then be able to move forward. Now, we've heard some great talks about the type of analysis and information that's around and it's setting a really good framework for what we're going to be talking about here. So what I want to do is to have a talk about the desired outcomes of this session, I want to propose a vision and straw man proposal, and then invite some people who know about this area to talk, to take some questions about this particular proposal. And then I'm going to ask the audience to provide some input about this as well because we really want to know whether this is something that we should take forward as part of this programme in HEAL. So if the objectives for this session are to understand if there is an appetite for a trusted online environmental health decision support platform, to consider the challenges and the benefits of such platform, and to consider what a prioritised list of activities might look like to address these identified challenges. So if we look at where we are now, then we can think about this challenge that we have about actually obtaining data and reports. And if we could put it in the context of FAIR data, then we might think about this as being unfair, you know, can we actually find the data and the reports about a particular aspect of health and exposure that we might be an environment that we might be interested in? Are those data accessible? Are the data interoperable or the system's interoperable? And are they reusable? And most of the time, I would argue, the answer is no without a lot of effort. We also have an unintegrated knowledge base. So we have these data that are siloed, we also have analytics then that are written for a particular challenge or a particular issue, and then put down again. Similarly with the data, we collect those data for a particular challenge or the issue, and we drag them from the different places, and then we put it, we finished that particular study and then we put the data down again. And similarly with platforms. So we have this, we, we have this issue of, you know, having to maintain software, having to keep data, having to keep models, having to keep the system, these diverse systems, and up to date, and also then thinking about how they interrelate, and then also, we tend to do all of the analysis and then we throw it away and then we do it all again when the next challenge comes up. So what we have now is this sort of uncoordinated activity. So we have this in data, we have it in analytics, we have it in platforms, and we have it in our ecosystems. So if we want to move forward as HEAL, then can we do better than this? So there's two things: can we do better in the operation of this? And also, if we want to know where we are in terms of the science, how do we find that out without a lot of effort? If we're going to move forward, we need to know where we are now. So what we'd like to do, as part of our group the Data and Decision Sciences group, is to think about what's the solution to this issue. And of course, there are many solutions and we're going to be working together to address the many challenges that we all face. But one of those is what we've been talking about, is just this, where is the data and the where are the data in the reports now. So if we want to envisage a solution for that, then what we'd like to think about is, as one of the first discussion points for the group is the construction of this interactive online resource that enables users to find the data, find reports, link existing or emerging platforms that access expertise. And then, as we, as this resource grows, access gold standard methods, be able to run analysis, be able to gain insight from these data. Now, this is not a proposal about building the one database to rule them all and aggregating the data into it. Where data are available, that would be really useful. And something that number of us are attempting to do in a bit through the various platforms as we've talked about in the last session. But in this case, even if it's about understanding where those data are, where those studies are, who's, who are the experts in the space, and whereas the information available for particular issues and challenges. So even if we have that at some metadata level, then that would be really useful. So if we think about the principles for this, it must be inclusive and collaborative. It must be a coordinated effort across the different groups, we must have careful stakeholder consultation, it must be a whole enterprise supported initiative, it must have continued support and currency, and we must ensure that whatever we create actually delivers in early usable outputs. So if we think about where we are now, and where we could go with such a platform, then we can see that we could benefit from an infrastructure perspective, from a data perspective, from reporting, from insights, from governance, and from decision-making. So going from those silos with undisciplined, unfindable data, scattered reports, partial insights, immature governance, and fractured decision-making to something that's more connected, audible, ordered, findable, federated, collated, cohesive, mature, and cointegrated. So that's the proposal. And if we think about what we might gain from that, then here's some potential benefits and then I'll leave you to to read them. But I'll point out that this could be a coordinated resource that could benefit all of us. We can have, we can move towards the need that we all identify for some sort of standardisation of tracking and reporting. Hopefully, this will reduce duplication and redundancy of effort and it will help us to have make informed decisions. So we can see here that if we do this, then this information source will not only be good for researchers, but it would also be good for managers and policymakers as well when they're fast, when they're after fast information, access to information and making fast decisions. So we could really use this resource as a way to be able to to bridge the evidence to policy gap as well. So we were contemplating like, what would happen if we don't do this, if we start HEAL and the steps to the future for HEAL without having this kind of resource. Then what we think is that we would actually have health and planning, adverse health and planning outcomes, we would also have adverse economic outcomes, and we would have adverse research outcomes. So that's the proposal and what I'd like to do is to turn this over now to a panel and our panel members here, who have kindly agreed to discuss this topic. Dr Claire sparks, who's from AIHW, we have Paula Fievez from FrontierSI, we have Mark Taylor from EPA Victoria, we have Karen Bleicher from the Sax Institute, and Stuart Barr from AURIN. Now, these people we've deliberately selected because they are experts in their field. They work closely in this area of needing to use data and to access data and information arising from the data and also the are outside the academic sphere. So we really wanted to hear from people who are really at the coalface with this kind of issue. So thank you very much all of you for agreeing to participate in this. I'm going to ask the questions in open-ended manner and I'd like all of you to feel free to jump in with a response. And then to talk to each other as well, if you'd like and then we'll move to the next question. So these are the questions that I'd like to ask you. And the first one is: What's your opinion of this straw man proposal, to have this trusted online environmental health decision support platform? So first off, I'll just ask Paula, would you like to start the discussion? Paula Fievez 1:00:34 Yeah. And I think you know the answer from me, Kerrie. I think it's an absolute essential initiative really. At the end of the day, I guess this has to happen. All the things that you spoke about earlier, about the duplication, we just see that so often, and you see, you know, new studies coming together and people doing exactly the same task over and over and over: where do I find the data? Who can I go to let us apply for the data? Let's see if we can access the data and then let's process the data and then then analyse it. And then somebody else wants to do a similar study using similar data, they've got to start right back from scratch. So for me the biggest benefit of something like that is just getting rid of that duplicated effort. So absolutely supported by me. Kerrie Mengersen 1:01:21 Thanks, Paula. So coming to any of the other panel members who would like to make a comment on this? Claire Sparke 1:01:29 Kerrie, it's Claire here. I agree with Paula and you, we would strongly support this. From our organisation perspective, we get asked for similar data for similar projects all the time. And we want to support people getting that data to do the good work that the research community does in this space. So I think it's a no brainer if that can be done. I think it's worth obviously looking at what is already out there. Maybe not starting from scratch on this. But yeah, as Paula said, to reduce that duplication I think is really important. And from our perspective, I guess it's governance that gets in the way for a lot of this work. Not that we're suggesting that... Not that I'm suggesting that this platform would be all about governance, but if there was some effort put into that, thinking about that, you know, you extracting once for multiple use use cases, that would be useful. Kerrie Mengersen 1:02:51 Thanks, Claire. That's great. And I know that there are people who are listening to this, participating in this session, who also have some suggestions about platforms that we could, should consider to be integrated. And as we talked about, you know, this could be a way of networking between those platforms that we've heard about, but certainly we don't need to start from scratch. We've got some great foundations. We heard about some of them, CARDAT and AusEnHealth, and so on earlier. Okay. Mark, would you like to comment on this at all from your perspective from EPA Victoria? Martine Dennekamp 1:03:33 Hi Kerrie. I'm not sure if Mark is here, but I was originally gonna be on this anyway. So yeah, I thought, you know, I might as well butt-in. I've just messaged him: do you know that you're actually with your face on the screen? So, as you know, you know, I've obviously been in the steering committee and I'm also across CARDAT. And in principle, I think it's a great idea. There's obviously quite a few things that I think we're going to have to sort out. And one of them is actually something that I even put in the chat there as well, which is it'd be great to get small spatial scale health data but there's obviously ethical considerations with this. So that will require a bit of thinking, I think, you know, ideally, we want SA1 data if we could. And the other thing is, in terms of when data comes from academia, there's often an IP associated with that. So, you know, there's also a few issues there to sort out but I mean, in principle, yes, it'd be great to have a dataset like this. Thhe one thing that we do need to also think about, I think, is that there is the metadata that goes with it. So it's really clear what the strength or the limitations are, and just have that background so that when that data is being used that, you know, it's used appropriately. Kerrie Mengersen 1:05:09 Yes, thank you, Martine. I think they're very good points. Certainly we need to be aware of the provenance of data of all kinds, the sovereignty of those data, and also what constraints there are about using them. And I totally agree about the metadata. And really, I think that's probably one of the most important things if we can find out what data are there even and the metadata, I mean, then, and the permissions to use it and how to access that or who to ask. And that would be just such a bonus, and I think something we don't have at the moment. So thank you. Okay, so, Stuart I'm going to ask you about this second question. Perhaps you'd like to talk to the first question as well. You're intimately involved with, with AURIN and building platforms like this. What sort of activities do you think would be required to create such a platform? And and also, in your response here, what's your opinion of having such a platform? Stuart Barr 1:06:16 Thanks, Kerrie. Well, first, second question first, and like everybody on the panel, I suspect everybody in in the audience, you know, I think the argument for having such a capability or platform moving forward is very strong and I would actually extend it beyond just environmental health and actually say that, you know, if we are going to be able to make positive interventions today and into the future. This is clearly in terms of population, wellbeing, health, livability of cities, etc., this is very much a sort of multi-dimensional, multi-objective problem. And it's not just solving this for environmental health data, which is challenging itself, but across all dimensions that are impacting our populace in Australia. So all the way from the engineer, environmental, and the social dimensions to the built environments. And so we've got a big, big challenge on our hands if we want to tackle this and do it from this sort of multi-dimensional perspective. In terms of how, I think the previous speakers have actually touched on some of the key things that we would need to start with, and that is coming together as a community and starting to develop first principles: that governance framework, the standards, the protocols, and the procedures that we would use to develop such a federated, geographically distributed, environmental health data analytics and decision support framework. And we can't, as you've correctly pointed out, ever hoped to achieve this by having a single, all singing, all dancing, monolithic database source of all knowledge. This is ultimately always going to be federated and geographically distributed. And the way you tackle that is coming up with the metadata framework and governance structure, and process, and protocols because that's what you can do that's new and that's what you can do to reach out and give those end points to where all the data repositories are, where the analytics is sitting. And also critically, where the high value outputs of research are being generated. And stored based on that data and the analytics. I think there's a really important component here about remembering the outputs from the research as being a key component of any decision support system or capability that you might want to build. So I would start very much and look at this actually, as a metadata, meta-information challenge rather than one of, you know, how do we bring 2530 homes and platforms all together? That that just won't work. Kerrie Mengersen 1:09:30 Yes, thank you, Stuart. I think we're on the same page there. So that's excellent. Karrin, how would this affect you? And how would this, you know, as from the Sax, from the perspective of the Sax Institute, you know, you obviously are involved in data and the 45&Up studies is a great national asset. So a platform like this, would this be useful to you? Would it be something that the Sax Institute would be, would find benefit in and what activities would you recommend for a us if we were considering doing this as part of the HEAL Network? Kerrin Bleicher 1:10:11 Thanks, Kerrie. Can you hear me? 'Cause I'm having my network dropping in and out. Oh great. Um, so I guess they're even within my team, for example, the 45&Up study, and we're dealing with many researchers. We've had over 800 researchers user study and people presenting here, Ivan and Geoff and many others have used the 45&Up study. But even then, with that sort of level of knowledge, we still find that there are current challenges in sourcing out, sourcing what data is available for Environmental Health Research. And you know, who to contact to get that data, even if it's sort of government administrative data. So I think, when we're kind of quite expert in the field and had that difficulty and challenge, it's just unimaginable for younger researchers in this space, or newer researchers in this space. So I guess that's one point that I want to make. So because I think this is essential, it's just has to happen, to have that metadata repository. And I think the other thing to think about is that it's not just about the data, it's really a backbone to knowledge management. Without knowing where data is, of course, people might not even get involved in the first place. And you do want that volume of people using and accessing the metadata and getting involved in environmental and health research. But you want inclusivity and you want diversity. And you also, that's not just across different types of researchers who are in or people who are impacted communities, but also across different fields, whether that's academic or policy, or planning people. So I think, you know, it's an absolute must have that repository, that you and you need to have that broad support. And I think one of the things we've talked about quite a bit of what are the outputs. And I think we need to think about outputs and be really clear that outputs are well beyond academia and they really need to sort of go into that policy space. And so I would really hope that the activities that are undertaken, I'll just mention that in a the moment, but really, actually have a focus on what the impact can be of the Environmental Health Research. So it's not yet so it does have an impact and impact beyond academia. And I think that is was Alistair who was talking about that sort of frustration around the last decade: massive scientific progress and understanding, but actually policies gone. Has it progressed, shall we say, to be kind possibly? Well, maybe recent recent times. So I think you know, that the activity, certainly around a broad engagement it's making sure that we're moving from sort of making sure it's not just about data availability and mapping that, but it is about optimising knowledge. And I would hope, maybe identifying opportunities, and the know-how to impact policy and practice. I think it's around resources and scoping, and that sort of leadership, and a broad based which to draw those ideas from. Kerrie Mengersen 1:13:38 Yes, thank you. That's very helpful. And so, Paula I'm going to come back to you then. So, I mean, you've thought quite a lot about this. And would you I'm going to ask each of you actually, what your what the main activities that you would see that we would need to be thinking about to go forward. So some of you have touched on this, but that you might either reiterate that or say something different. So I'm going to start Paula. Paula Fievez 1:14:09 Yeah, and I just wanted to note, a couple of messages in the chat about, you know, looking at a whole new platform will just be duplication in itself. And I think, you know, when we've been talking about this idea about coming up with a complex platform, it was definitely acknowledging what's already available, ensuring that this just isn't another one platform. But I think the the challenge that we have is that if everybody's wants to make their one platform, the platform, we're going to run into trouble. That's why I think the first activity has to be building that consortium of stakeholders who all want the same thing who may already be involved in building platforms and come together, because if we compete, there are going to be many platforms, and we're going to keep buying into the very same problem. So for me one of the very first thing is bringing people together with a common vision, building that stakeholder group, and then moving forward. And then I think the second thing, very importantly, is coming up with common and agreed national environmental health indicators, because at the moment, people are using different indexes and indicators for different studies. And I don't think we have a common understanding of which are the nationally agreed upon ones for Australia. So for me, that's another big activity that would need to be done as a fundamental, I guess, a knowledge base to underpin anything that we do going forward. Kerrie Mengersen 1:15:41 Okay, thanks, Paula. Claire, what about you? Claire Sparke 1:15:47 Yeah, I would echo that this shouldn't replace platforms, I guess, maybe that maybe the word doesn't work. We have a number of specific websites at the AIHE we're calling clearing houses. So it's where you get your reports and access to data, links to other websites or platforms that where you can access data, that kind of thing. One example is our aged care Clearing House, the other one is the indigenous closing the gap one. So I think it's, yeah, maybe there's some sort of discussion to be had, if this would be moved forward by the HEAL networker to sort of name and not not cause confusion, if that makes any helpful. Yeah. I mean, other things, getting the right stakeholders on board. I would suggest talking to, there's a lot going on in government about trying to get a lot of data together as well, not just in academia. So we have the newly formed Australian climate service, which is really focused on the environmental data at the moment, but I know they're going to be moving to the health data as well. The Australian Digital Atlas that Geoscience Australia, so it's, I would be, that's just federal government, I'm sure there's a lot more going on at local levels as well. So I think it would be good to engage broadly there as well as with the research community. And some sort of framework for high level monitoring environmental health moving forward, I guess it's very similar to what Paula said in the indicator space would be really helpful. I know there's, that Victoria has quite a few. There's some in the state of the environment report. There's lots out there. But yeah, trying to work out the the main things that people are trying to do research and analysis on, because, you know, equally that would lead to what goes on this website, or what goes on this platform Clearing house, whatever you want to call it. Kerrie Mengersen 1:18:05 Great. Thank you. I think we've got resource at the moment. So yes, I agree, thank you. Thanks, Claire, that's really helpful. Okay, in the interest of time, I'm going to move to just the last question, and this is a question of (get it up here) on what would be the benefits of such a platform? So I'm going to ask Stuart, for you. One, one primary benefit of such a platform. Stuart Barr 1:18:31 It ould allow us to generate a consistent and objective set of indicators and improved understanding and knowledge. I think knowledge was one of the terms that was used earlier on around the current and potentially the future of population health of the Australian nation. It's as simple as that: without it, it's very difficult to think how you would make decisions over resourcing and social infrastructure into the future. Kerrie Mengersen 1:19:07 Okay, thank you. Um, Mark or Martine? Martine Dennekamp 1:19:15 Apologies on behalf of Mark, he was unable to make it. Kerrie Mengersen 1:19:20 Thank you for stepping in Martine, I really appreciate it. Martine Dennekamp 1:19:23 That's all good. I was basically gonna echo what Stuart said, which is around consistency and quality. There's one more thing that I think would be good is the fact that it will be easier to compare. I don't know, when you compare states, for example, because the states seem to still work a little bit in silos sometimes. So this would be really beneficial. That, you know, we compare the same and then we're able to compare between states. Kerrie Mengersen 1:19:55 Okay, perfect. Thank you. Kerrin? Kerrin Bleicher 1:20:00 Um, ditto the comments so far, you know, consistentcy in quality and comparability. I think also its capability. It's about building that capability in health environmental research. And I think beyond that is the capability to translate that research. I would hope that that would be an additional part, even though it's not normal metadata platform, those ideas around data visualisation tools, to use all that kind of background could be maybe a second stage. But it's certainly an important stage as part of that storytelling and translation and getting policymakers and planenrs is to listen to the evidence. (Thank you. Very good points). I do have to say the Australian quality for safety and health care commission from the safety and quality and health care actually had a session yesterday around building a platform, pretty similar, actually, in a way. So a one stop shop, searchable database on all this the health research that's going on. So definitely that environmental scan is very important (Yeah, thank you) in the what we're trying to do. Kerrie Mengersen 1:21:06 That's right, there's so much activity in this area that you're right, and others have touched on it as well. So we really do need to make sure that we connect rather than reproduce. Claire? Kerrin Bleicher 1:21:17 Just wait, just one more thing on something Claire said and it was around, so that framework, and I think was clear around indicators and framework. And that's used very much in sort of the health service performance space. And I think that would be a great opportunity to bring that together along with, you know, consistent definitions, like Meteor where you got same formatting and it makes it makes it much more usable, the data sources and that get that cross jurisdictional agreements. I think there, there are benefits that would flow as well. Kerrie Mengersen 1:21:50 Yeah, thank you. That's great. Claire, what's your primary benefit, do you think? Claire Sparke 1:21:57 Um, I think just being able to find things Kerrie. I known that's very basic. You know, and reduce that duplication. But yeah, at the end of the day, as what Kerrin just said, knowledge, translation to policy. That's what is of most, you know, always research is most useful for and I've just noticed, somebody talked about our health metadata resource Meteor in the chat, that's certainly something that could be expanded. I agree. Just to note, also, with Kerrin talking about the health performance indicators, we actually have environmental indicators in the Australian Health performance framework that aren't all developed yet. So there's a really strong use case there for us to to get those developed. They are currently empty. We do need to develop them. Kerrie Mengersen 1:22:58 Okay, great, thank you. Well, you're talking to the right people, so I think we've got some great people who can, who can, who will be very interested in that and also potentially want to talk to you about it.