Ann Mossop: Welcome to the UNSW Centre for Ideas podcast, a place to hear ideas from the world's leading thinkers and UNSW Sydney’s brightest minds. I'm Ann Mossop, Director of the UNSW Centre for Ideas. The conversation you're about to hear, Future Medicine, features Anand Deva with journalist Tegan Taylor, and was recorded live. I hope you enjoy the conversation.

Tegan Taylor: The land I'm coming from now is that of the Yuggera and Turrbul People. I acknowledge this land and Elder's past, present and emerging. It's a familiar situation for many of us, you get a diagnosis, say, of a skin cancer. It's a scary situation. But what makes it even more discombobulating is when you realise that you need to run all over town to see the dermatologist, the cancer doctor, get your blood tests, report back to your GP. Surely, you think, there must be a better way? Well, to discuss what that might look like. He's Anand Deva program, head of plastic and reconstructive surgery at the Faculty of Health and Medical Science at Macquarie University. Welcome, Anand.

Anand Deva: Thank you.

Tegan Taylor: So you're an architect, or, involved in integrated care models. What are you talking about when we're talking about integrated care models?

Anand Deva: Well, integrated care means many different things to many different people. But essentially, what we're trying to do is to simplify the system for patients. So as you mentioned, the diagnosis of something like skin cancer could certainly be quite scary, particularly, then, if you add confusion, cost, waiting times, efficiencies of going from one doctor to another, from one place to another. And so in its simplest form, an integrated care model around skin cancer will put all the elements that will be required to treat that patient in the one place, at the one time. And that's exactly what we've done.

Tegan Taylor: It's not just about putting people in the same place. It's also just those people talking to each other. How do you facilitate a centre like this?

Anand Deva: Well, that's a very good question. I think to get a system that's naturally fragmented and at times adversarial, to work in a collaborative fashion is really, really difficult. It takes, I think, a collaborative mindset, an open mindset to start with. And I think to put that into perspective, if you look at the health system in Australia, it's kind of grown, almost, in many different directions, over the decades. The biggest change, of course, came when Medicare was introduced in the 70s. And I think that I'm a firm believer in universal access to health care, I mean, if all developed nations put health of their citizens first. And so I think that allowed us then to provide health care, at some level, to all all patients and all Australians. But the problem is that since that time, we've had private versus public sector, we've had, you know, specialties versus GP’s, we've had health funds versus doctors, we've had industry versus private hospitals, for example. And so each of these components don't actually necessarily like playing or working together. So to start with, I think you need to find people that are open to collaboration, and that's not easy. And then you need to pick a cause. And ultimately, if you have a patient sitting in front of you with a problem, there's nothing like that to actually make you united, as a system, in order to help that particular patient.

Tegan Taylor: Can you talk about the centre that you're involved in? How was it formed, and what has it taken to get it to work?

Anand Deva: Well, you want to summarise nine years of toiling, in a few minutes! I'll do my best! It started essentially, with support from a grant from New South Wales Health. And I think for that, I'm truly grateful. The grant was, kind of, written almost in a fugue state, because having been a surgeon for many, many years, I've seen firsthand the effects of that disengagement, the fragmentation. Because patients present to me having been through a very, very fragmented and discontinuous pathway, should I say. And the ultimate effect is actually a poor outcome for the patient. So the call came out from the Ministry at that time for any clinician working within the public sector to come up with ideas on how we can better deliver care. And knowing about skin cancer, I thought, well, why not we put in a grant that actually brings together all elements so by that I mean, GPs a lot of GPs are involved in skin cancer care, and some of them do an excellent job. Dermatologists, specialists in skin that also treat a lot of skin cancer and people like myself, plastic surgeons. So I wrote it all in a paper, with some wonderful diagrams and, kind of, sent it off and forgot about it. And a few months later actually got the grant. And that's when I actually had a moment of panic, to think oh my god, I've got to convert this idea from paper into reality. It was through, actually, engagement and connections that I've had with my colleagues over decades of practice, I think that's made the difference. Having I guess that open and collaborative mindset and finding people with that, like mindset. The centre was built partly through public funds, but also through partnership funds. And I have to acknowledge the help of Ramsey Healthcare, Sonic healthcare, those big corporates that actually are not easy to deal with, but can sometimes, with a persuasive argument, give a little bit to help support a model like this. Nine years on, we have now three centres delivering this model, we've treated over 30,000 patients. And the idea with this clinic is a simple one, but very difficult to actually establish and then to maintain and to grow. But in effect, a patient can walk in off the street, concerned about their skin cancer or skin cancer risk, have access to a really well trained general practitioner, we call them a GP plus, because they've done more training in this area, and we vouch for their skills, and we work collaboratively with them. And on that day, they have access to specialists, if they need them, and also access to great facilities for the treatment of that particular cancer. So, you ask how it's done. It's done through I think, a little bit of support and funding, I think that's crucial. It's done through making other people within various aspects of the healthcare system believe in the model and the vision. And then the hard work is actually persuading people to be part of it. So yeah, that's, I guess, a short, relatively short answer. I know professors tend to be very verbose, but relatively short answer in terms of how we established one of, now, many models that we're delivering.

Tegan Taylor: That's so interesting, because I think from a patient… like, the petty jealousies… or I won't say petty… the jealousies between medical specialists seem a little bit inside baseball, when you're the patient being all like, I got a melanoma, like, help me. 

Anand Deva: Yes. 

Tegan Taylor: The patient benefit seems very clear.

Anand Deva: Well, that's the one thing, that's the one secret – well, not so not so secret – weapon that I have up my sleeve, because I think if we are truly going to build the healthcare system of the future, and if we are truly going to to be true to our words, that is a patient centred model, that patients are empowered, that they're given true information that's transparent, and actually for their benefit, rather than for the system's benefit, or for the practices benefit, or for the specialist benefit. Well, that's the sort of mindset that we need to develop. And that's not an easy shift from a system which is naturally pitted against each other where there is competition for work, where there are commercial drivers here that are pushing, perhaps, healthcare in directions that are perhaps not in the best interests of the patient.

Tegan Taylor: So you obviously see this as a model that could be applied across different disciplines or different diseases, I suppose. Skin cancer, plastic surgery, dermatology, those things cluster together quite nicely. But how do you decide who's part of one of these hubs when patients' needs can vary so much?

Anand Deva: The real struggle with some of these other models is funding. So whilst Medicare was designed with the best of intentions, it's very much a transactional system. By that I mean, if I was the doctor, for example, and I was treating you as a patient, what I get paid will be dependent upon what particular intervention, what item number, and what packet of funding I could unlock. And that has repercussions across the whole system. Why is it for example, someone like me, a procedural specialist, is paid huge amounts of money to deal with a crisis problem, when in fact, it's the GP right at the beginning of this problem a decade ago that actually should be incentivised to solve this before it becomes a crisis. So it's, kind of, Medicare's kind of introduced a funding based system that favours urgent hospital crisis management, rather than actually what we should be doing, which is getting people early before they become very sick and unwell. The funding for some of these chronic models falls short because Medicare doesn't cover allied health intervention. And so what I've learned over the last five years developing some of these models is that a lot of chronic disease management really requires input from not just doctors, but other healthcare professionals. And that is difficult, because we can't access that in the same, sort of, level as we can say, for medical intervention. Our chronic wound clinic, for example, has been running for some time, and essentially we can't fund it because nurses, the nurse specialists that deal with a lot of these wounds, with dressings, and pick up problems and help us to heal these wounds simply do not, are not funded. So we need to find the money from somewhere. And so part of these models have taught me that yes, it's lovely to have collaboration, it's lovely to find like minded individuals within the healthcare system, it's lovely to find administrators and universities that actually believe in their heart that they want to do this. It's not so easy to fight some of the, you know, the established thinking and also to be a disrupter to some extent. But what's even more difficult now is the challenge of how do we sustain these models to provide ongoing care, value for money and tackle these difficult problems, but actually sustain the funding and keep them going.

Tegan Taylor: Right. So I mean, educating students from when they're students, sort of, throughout their career, obviously one of the steps along the way, but it seems to me that a much bigger problem or challenge is just the whole healthcare system as a whole. Do you see a need to completely redesign Australia's health care system and funding?

Anand Deva: Oh gosh. Well, I think, I look, I have, I've worked overseas, I've worked in other health care systems through my career, I think we have, we should really, we should be quite proud of the system that we have, you know, it's a good system. You know, and time and time again, you know, when I've compared, say, access to treatments, for people in other countries, I think Australia has done remarkably well. We are essentially a wealthy country, we're an educated country, and we have amazing resources. So I don't think we need to design a new system, we just need to try and get this system to work a bit better, to extend the value of the Medicare dollar to build these models, to ensure that every dollar spent on health care is spent efficiently and in an integrated way that allows patients to access the right treatment at the shortest possible time route at the most affordable price.

Tegan Taylor: And just finally, can we talk a bit about another field of work that you're in, which is wound care. What you're looking at and what you're hoping to achieve there?

Anand Deva: Well, this is a flow on from the integrated care work that we did. So shortly after we established the skin cancer model we looked at the integrated wound, chronic wound model. So humans are unique, in some ways. So animals, you can injure them, and they generally will heal. Humans for some reason, as we get older, perhaps it's our, it's the fact that we live longer, and perhaps lifestyle issues. But there is difficulty with healing wounds. And if a wound doesn't heal by itself, let's say in six weeks, it becomes by definition, a chronic wound. This is a hidden problem. And you know, the focus currently on aged care and the standards of aged care have been called into question. There is a hidden sea of these chronic wounds sitting in aged care facilities right around Australia. It may not be anything that grabs your attention, but it is a drain on the system, drain on the patients. And so I've, as part of my research interests, and actually my clinical interest, I've always been interested in why these wounds? Why is it humans develop chronic wounds? And the answer is probably a mixture of getting old, our ability to repair and regenerate degrades over time. Chronic Disease, once again, like diabetes, smoking vascular disease, so the things we do to our bodies is not great. And then infection is another really big problem. Because once you've got an open wound bacteria colonise on the surface and the body can't get rid of them. So there's been a lot of developments, once again, in technologies and wound healing. And I've been involved in some of the research on that. But how do you then deliver those benefits to these hidden patients? How do you find them? In fact. So we set up our chronic wound care model based in the Sutherland Shire in Sydney. Once again, you start small and develop and optimise the model and then hope to scale. And the model, essentially, it sought to do two things. It sought to get into some of these aged care facilities, and we've used telehealth for this. So that's been a remarkable, rapid uptake of telehealth. And we've seen the value of it and it's not, I wouldn't say it's the answer to everything. But it's allowed us to, sort of, lift the cover, so to speak and actually deal with some of these patients, but more importantly, the staff looking after them. And being involved now, in peering into some of these aged care facilities, the variation in staffing, their knowledge, even their level of English, sometimes, you can sort of see how these chronic wounds are left to fester and kind of ignored to the point where they then end up in hospital acutely and you know, 1000s and 1000s of dollars and time spent to fix these wounds. So identifying them, getting into these homes using technology and then educating their staff to recognise signs that a patient is either at risk of developing a chronic wound or has a chronic wound, was the first step. The second was to build an infrastructure so that we could diagnose what the problem was, and so we built a bricks and mortar clinic. We armed it once again with our GP plus, who I can say is an absolute find. She's so dedicated and has actually now started to spread her knowledge and educate some of the GP registrar's that come through our training centre. And of course, she's then matched with nurses who are committed to take care of chronic words. So the model is involved… education, empowerment, identification, building the infrastructure to deal with some of these wounds. We've diagnosed wounds that have had foreign bodies in them for two years and then pull this stuff out, and the wound heals. Or we've picked up vascular disease, for example, that's been missed. Some of these chronic wounds are actually skin cancers. So, can you believe that? They've been dressed for months, and oh, maybe this is a skin cancer. So simple things like that, where you can literally solve the problem and get the wound healed. The wound model is now ready to scale. So that's really exciting. So we're now starting to talk actually to New South Wales Health, and we're starting to work with other districts that might want to try and replicate this model. The challenge always is to find, you know, GPS that are interested, and obviously, there's specialists that then help the clinic are vascular surgeons, plastic surgeons, endocrinologist to treat diabetes, aged care. So the links are very strong in the Sutherland Shire, the challenge is can we develop these links and find these individuals and other parts of the state?

Tegan Taylor: Okday, so there is a lot of stuff that you're working on, Anand. But if you could only pick one thing to focus on in the future, what would it be?

Anand Deva: Well, I think if I was to pick one thing, I'd love to see the system move away from acute crisis management and more into prevention. You know, it doesn't make sense to me that we wait for the wheels to fall off, and then we spend a lot of money. And so I think it's challenging, but I think, you know, these integrated care models, the change in the culture, you know, ultimately influencing the policy and the decision makers and the fundees of healthcare. That would be what I'd like to see. So rather than wait for the patient to present to the emergency department on a Friday night, the patient's picked up 10 years before, interventions are put in place, such that the patient never has to be seen in hospital. That's what I'd like to do.

Tegan Taylor: You’re still going to deal with those crises when they come up, though, how do you balance the two?

Anand Deva: So I think it's a slow process. So as we get more towards prevention, the hope is that the crises slowly come down. So it's like a slow balancing act.

Tegan Taylor: Again, ambitious, but I like it. Anand Deva, thank you so much for joining us. 

Anand Deva: My pleasure.

Ann Mossop: This event was presented by the UNSW Centre for Ideas and UNSW Medicine and Health, and was supported by Inspiring Australia as part of National Science Week. Thanks for listening. For more information, visit centreforideas.com, and don't forget to subscribe wherever you get your podcasts.

Ann Mossop: Welcome to the UNSW Centre for Ideas podcast, a place to hear ideas from the world's leading thinkers and UNSW Sydney’s brightest minds. I'm Ann Mossop, Director of the UNSW Centre for Ideas. The conversation you're about to hear, Future Medicine, features Louisa Jorm with journalist Tegan Taylor, and was recorded live. I hope you enjoy the conversation.

Tegan Taylor: The land I'm coming from now is that of the Yuggera and Turrbul People. I acknowledge this land and Elder's past, present and emerging. Well, hardly a day goes by when we don't hear about how machine learning or artificial intelligence is disrupting an industry. It's usually framed as a good thing. And it often is, but there are ethical considerations, especially when the data that's being crunched is people's medical information. Healthcare settings are flush with data. In fact, there's almost too much. So, how do you sift through this? How do you protect patients? And how do you train the people in the system to make the most of the power of AI when technology is developing so quickly? Luckily, we have brains like Louisa Jorm’s on the case. Professor Louisa Jorm is the foundation Director of the Centre for Big Data Research in Health at UNSW Sydney. Thanks for joining us Louisa.

Louisa Jorm: Thanks so much Tegan, and good evening to everyone out there.

Tegan Taylor: Can you give us an example of how data could be harnessed in say, an intensive care ward?

Louisa Jorm: An Intensive Care Unit is one of the biggest producers of data within a hospital. I mean, you all know about the machines that go beep. If any of you have been in an intensive care unit, you will have seen the amount of equipment that's there. And all of those bits of technology are, sort of, producing data, continuous streams of data about physiological parameters. You know, things like heart rate, things like speed of respiration, blood pressure, blood glucose. But in the past, that data has mainly been used just on the spot to monitor that individual patient, what we can now do is actually bring together the data that's generated through ICUs from multiple patients, potentially 1000s of patients, and actually then apply these artificial intelligence or machine learning techniques to them to try to produce much more personalised approaches to intensive care. So based on the experience of 1000s of other patients like you, what is likely to be the best, sort of, course of treatment for you in your ICU stay. And we can also look at things like predicting for an individual patient, what might happen. If we change, for example, their artificial ventilation rate, or we administer insulin to them to try to improve their blood glucose levels. So it's all about using large amounts of data from many, many patients to then produce a greater degree of personalisation for one patient. 

Tegan Taylor: Is this happening in ICUs in Australia at the moment?

Louisa Jorm: Really, it's still in its infancy. There's some great examples, sort of, emerging from the research sphere, including in our own work, that do relate to sort of automated blood glucose control and automatic control of mechanical ventilation. But even in the ICU as, like, all other parts of the hospital, there's actually quite a big implementation gap between what is possible using data and using technology, and what actually works in the clinical setting of a hospital, which as you can imagine, there's huge amounts of human factors. And in particular, there's many of the current, sort of, generation of doctors and other clinicians, are not necessarily particularly familiar with, or comfortable with all of these technologies. Rightly they have concerns about who's making decisions, and are they good decisions? Are there possible ethical and legal implications if decision making is done in an automated fashion? So I think the big challenge is how to actually make it, sort of, AI or machine learning, assisted decision making, but with the clinician still feeling in control, and also still being able to involve patients and carers in some of those decisions. It's not always the machine making the best decision.

Tegan Taylor: Right. I want to come to that. But just on the ICU as an example, there's such complex environments, the people literally, lives hanging in the balance, that's why they're there. The people who work in those settings are highly specialised, the machines are multiple. Just… can you talk about the challenges of implementing something like you're talking about, which would be really useful, into such a complex environment?

Louisa Jorm: There's a lot of challenges. One is actually even extracting the data to do the analysis that I'm talking about. Many of the electronic medical record systems have been, sort of, set up as fairly standalone systems as I said, for the point of care, bedside care of patients. Actually getting the data out from the back end of those machines, and then integrating them across the various machines, and integrating with other information that's really important about the patient, like their age, their sex, the health conditions that they have, it actually poses quite considerable technical challenges. And in particular, because proprietary software is often being used, that may not be compatible with other sorts of software. So that's one of the biggest challenges we face, is actually extracting data in a form that is then able to have these techniques applied to them. So that's one big challenge.

Tegan Taylor: So there are examples in other industries, or in other machine learning scenarios where the machine has learned to be racist, for example, because of the inputs that were put into it, the biases that go into it. How do you account for that?

Louisa Jorm: Well, there were examples for medicine as well. And a great example is IBM Watson Oncology, which was developed to try to set up new AI driven treatment recommendations for cancer patients. And the problem that they ran into, not so much that the algorithms weren't good, but rather, the data that they got wasn't sufficient. They didn't have enough patients, they didn't have enough depth, they didn't have enough multimodal data about patients to build a robust algorithm. So what they decided to do was to work with some expert clinicians to develop some sort of pseudo patient journeys. And they incorporated these into the dataset that were being used to train the algorithm, and found that as a result, the algorithm actually was producing some inappropriate treatment decisions. So yeah, basically, the algorithms and their results are only as good as the data on which they're built. And as I mentioned earlier, that's one of the biggest challenges that we have, is actually accessing the really large numbers of patients and the very detailed data that we need to actually use things like deep learning, which is one of the most successful, sort of, AI techniques that currently underpins so much of what you see on the internet at the moment, all of the facial recognition techniques and so on.

Tegan Taylor: When you're designing these AI machines, you're really… the risk is that you're baking in biases that you don't even realise are there…

Louisa Jorm: That's right. And, again, I come back to this, there's a huge amount of data being generated. But harvesting that data and being able to access it for these types of purposes is a big problem. And, in fact, around 40%, of all, sort of, published, machine learning research in health actually uses one single data set, which is called mimic three, and comes from basically, Boston Massachusetts, one hospital there. And they have managed to create a version of that data set that meets, sort of, privacy, privacy protection principles, and make it quite generally available. And so the machine learning community has basically pounced upon that data set and used it. And you know, it has some good features because it enables them to, sort of, benchmark how well their algorithms perform against one another. But it has some pretty potentially negative downsides as well, because if we look at that population, it's really not a very similar population to what we might get in a Sydney Suburban Hospital, for example. Very, very different sort of mixture of ethnicities and backgrounds, and, for example, obviously, no Aboriginal Australians are present in that dataset. And so you do run a risk of basically excluding certain groups and not being able to be sure that the algorithm is performing as expected for all parts of the population.

Tegan Taylor: Right. So it feels like there's a lot of possibilities. A lot of challenges, though, are they as big as they seem? Or is it just, are there small tweaks that need to be made? Or are they being, kind of systemic… 

Louisa Jorm: No, they’re pretty big. 

Tegan Taylor: Ok. 

Louisa Jorm: They’re pretty big. And they do relate to, you know, as I said, accessing data, they also relate to things like making sure that you bring the community along, people have to feel safe and secure in the way that their data are being used, that their privacy is being protected. And, you know, many members of the community don't have a strong understanding of these technologies, because they are developing so fast. But equally, many clinicians don't have a strong understanding of the technologies either. And so I think one of the big challenges is this really rapid upskilling of the health and medical workforce, and probably a, sort of, a different health and medical workforce that includes additional clinicians, doctors, nurses, and so on. But also data scientists and engineers working alongside them. You can't wrap up every skill that you need in a single individual here because you need that clinical insight and understanding of clinical workflows, understanding of patients, as well as really, sort of, highly advanced technical skills in data analytics and data management.

Tegan Taylor: Do you see data scientists, data engineers, like you just said, being embedded in hospital teams in the future? Say in the ICU example we were talking about before?

Louisa Jorm: Yeah, absolutely, it is starting to happen. But when I first started at UNSW, five or more years ago, this was something that people in the health system, health departments and hospitals were coming and talking to us about, is, we don't have people embedded within our systems who can do these things. And we could sometimes hire people who come from engineering or computer science backgrounds, but those people often take quite a long time to actually become useful within the health system, because they don't actually have any background in health, understanding of biology, or, you know, understanding of how one would integrate these sorts of technologies into clinical scenarios and clinical workflows. And that's, in fact, we set up our own master's program in a health data science at that stage, which is now starting to pump out the graduates. And many of those, in fact, I think all of the graduates so far have actually taken up employment within various health settings or research settings. And many of them as part of their training have actually done research projects, basically sitting embedded in clinical settings. So only yesterday, we had the final presentations from our latest round of students. And it's really, sort of, gratifying to see that there are people who are now really quite ready, workplace ready, to go and apply data science skills in health and medical settings.

Tegan Taylor: So that's the Master of Data Health Science program that you just said at UNSW. Can you give some examples of what those people are currently working on or what, sort of, projects seem to be in the works?

Louisa Jorm: They're very diverse. But an example, one that was presented yesterday, was using very complex link data from FACS here, so from the community services sector, to evaluate the sort of longitudinal journeys of children who have contact with Community Services and with Child Protection, and to try to look at whether or not very specific programs that are operated FACS here to try to improve outcomes for these children and prevent, for example, out of home care, placements are being effective. And so that project brought together data from I think, around 12 different, sort of, data systems, and then had to apply quite advanced analytic techniques, because the data were observational, it wasn't like a randomised control trial, these were real children, and there were many, many, sort of, external factors and other factors that had to try to be controlled for, to look for this. 

Tegan Taylor: So it's really exciting that you have this master's program that's developing those specialists. But what need is there for embedding this sort of content into the courses that doctors are learning when they're at medical school, for example?

Louisa Jorm: I think there's a total need, and there's something that I've actually been trying to advocate for since I've been at UNSW. And we've over the years managed to get a little bit of content, and we have a digital doctor content for the year three medical students. But excitingly, as of next year, we are actually going to be running a new medical honours coursework in clinical AI. And that's very much, you know, obviously, the clinicians need to have an understanding of the underlying principles and methods, but most of them are probably not going to be themselves crunching data or developing algorithms, but they need to have an understanding of how they are implemented in health, and how they can critically appraise their use within health. How do they know that a particular AI tool is a good AI tool? When do you trust the tool? How do you, sort of, work with it? How do you integrate it into the way that you work? And excitingly a fairly substantial number of the UNSW medical students excitedly put up their hands to participate in this first round of the clinical AI Honors Program. 

Tegan Taylor: That's so exciting. So it must be a real challenge to be teaching something that by the time someone is actually out in the workforce and perhaps being exposed to these things. The landscape might have changed, because it's such a developing field. What are the, sort of, core principles that you can teach that are going to keep someone in good stead over the next five or 10 years?

Louisa Jorm: Well, as I said, it is largely about how do I assess and critically appraise whether or not this is a good tool? And there are, as we're, sort of, developing some sort of checklist type of approaches for them to use. But I think really importantly, it's going to be the young clinicians who know so much more about this than the, you know, the traditional model where the, you know, the senior clinician imparted wisdom, in this case, the young clinician is the one who's going to be educating the older ones. And it's something that struck me ever since I've been at UNSW, is so many of our medical students have done really high level mathematics, you know, while they're at school, and then they enter the Medical program, and they don't do any mathematics or statistics anymore. So in some ways, those capabilities haven't been nurtured in the way that they could be. And as I said, there's quite a lot of excitement amongst the medical students in now being able to do this. And I guess, you know, it's going to be those, the older clinicians are going to be accepting of the fact that the younger guys are the ones who are going to be teaching them.

Tegan Taylor: So if you're looking forward, the next 5, 10 years –- let's be optimistic – what do you see as the main challenges? And what are the main things that you're excited about that you think might be on the horizon?

Louisa Jorm: Well, obviously, I'm excited about the burgeoning availability of Electronic Medical Record Data in Australia. And I think, you know, encouraging signs that it is going to become more readily available. I'm really excited about the, you know, some of the new algorithmic techniques that are starting to be applied in health and medicine. And I mentioned deep learning, it had many applications in the area of image processing, but is now moving, you know, it's really overtaking most other forms of machine learning, for a whole range of, sort of, predictive tasks. And it's really very exciting to see how many publications are coming out reporting those techniques. The challenge, as I've already mentioned, is how do we actually put all this good use to actually result in improvements in health and healthcare. And people like me, who are a bit funny, who may be interested in the algorithms, do need that interaction with people who are implementation scientists to make things work. And I think we haven't got that quite right yet. The other thing that is, you know, remains a challenge is ensuring that we do maintain public trust and that we, the privacy and confidentiality of individuals' information is maintained. Another thing that I'm quite excited about, which I thought I might mention is a thing called the Join Us Register. Which is a new development being led by the George Institute and UNSW. But with 33 different university and Medical Research Institute partners from across Australia. And the idea behind this is to increase the ability of Australians to participate in Health and Medical Research. Basically, people are asked to join the Join Us Register, and what they're then doing is agreeing to be contacted about medical research studies that may be of interest to them. And there's a big issue at the moment that most Australians are quite interested in participating in clinical research, particularly if they do have a health condition. But only a very, very small proportion of them actually ever do get enrolled in a clinical trial. So I'm really hopeful that Join Us will help to address that problem and, sort of, make clinical research much more accessible to the whole community.

Tegan Taylor: It's a fascinating space to watch the laser. Thank you so much.

Tegan Taylor: Thank you, thanks Tegan.

Ann Mossop: This event was presented by the UNSW Centre for Ideas and UNSW Medicine and Health, and was supported by Inspiring Australia as part of National Science Week. Thanks for listening. For more information, visit centreforideas.com, and don't forget to subscribe wherever you get your podcasts.

Ann Mossop: Welcome to the UNSW Centre for Ideas podcast, a place to hear ideas from the world's leading thinkers and UNSW Sydney’s brightest minds. I'm Ann Mossop, Director of the UNSW Centre for Ideas. The conversation you're about to hear, Future Medicine, features Joseph Powell, Louisa Jorm, Anand Deva and Anushka Patel, speaking with Tegan Taylor, and was recorded live. I hope you enjoy the conversation.

Tegan Taylor: The land I'm coming from now is that of the Yuggera and Turrbul People. I acknowledge this land and Elder's past, present and emerging. We're exploring the future of medicine. Tonight, we're going to hear about how integrating machine learning into healthcare settings can make for better care, as long as it's done ethically, having diverse specialists in one place could streamline care for patients, and applying the idea of precision medicine to public health. But first, imagine if you could understand undiagnosed cancers before they were observed, detect cancer clones that were going to be resistant to treatment ahead of time. Now we're really talking about the future of medicine. Underpinning this future is cellular genomics. Joining us now is a man who's helping translate the technology into clinical practice. Associate Professor Joseph Powell is a biomedical researcher and statistical geneticist and head of the Garvan-Weizmann Centre for Cellular Genomics. Welcome Joseph.

Joseph Powell: Hi Tegan, thanks for having me.

Tegan Taylor: Let's start with a definition. What is cellular genomics? It sounds very fancy.

Joseph Powell: Yes, cellular genomics is essentially a technology type, which allows us to generate sequencing data. So, information on our genomes, but at the level of individual cells. And the reason why this has been so revolutionary is that genomics and generating sequencing data has been around for quite a while, it's made a huge impact already in medicine. But it's traditionally been done at the level of what we call bulk sequencing, nowadays. And that's really where you would take a sample from a patient, a cancer sample, and you would sequence all of the content from millions and millions of cells. Which is fantastic, it can be used for some really important outcomes. But it doesn't give us any information about what's the difference between one cancer cell and another cancer cell, for example. And cellular genomics as a technology type gives us that information, and then an analysis of that data allows us to understand why the differences in cells in a cancer, for example, or immune cells that circulate around your body, why do those genetic differences between them impact response to treatments, or, you know, why we respond to an infection, or indeed, why do we even develop disease in the first place?

Tegan Taylor: Right. So the moment the way you’re sampling things, it's really diluted in the size of the sample? Is that what you're saying?

Joseph Powell: We think that it is the current, sort of, approaches are really lacking the resolution of getting down to that cellular level. And that's not to say that there's anything wrong with them, they work brilliantly, but you miss a lot of information that you get with cellular genomics.

Tegan Taylor: So I've heard cellular genomics talked about as, like, being the original, kind of, histopathology process is like a 2D process, and this is like a 1000D process. Can you talk through what you're looking at? What is the resolution that you're getting? What's the extra information that you're seeing?

Joseph Powell: Yeah, so to talk in those dimensions, as you've just alluded to, I mean, that's a good starting point. Because you’re right, histopathology is 2D, physically. You know, you take a section of a cell, or sorry, a tissue sample, but you then also generate probably information on a handful of, for example, proteins, maybe 5,6,7,8, maximum. But the information encoded in that tissue sample is probably somewhere between 20 and 30,000 parameters. And so cellular genomics not only gives you the 3D stack within that tissue sample, so you generate this information from all the physical coordinates, but you then generate for every single one of those cells, 20 or 30,000 parameters of information. And so you have, you know, an explosion in the inner dimensions in both the amount of information you generate for an individual section of a tumour, for example, but also the whole, the whole aspect of that information across a tumour sample, rather than just a very, very limited histopathology slice, as it happens.

Tegan Taylor: So you've got all of this information and how do you even process it? Like, there's the ability to do this and then you've got to actually be able to crunch that data to make something useful out of it.

Joseph Powell: Yeah, this is, I would say this is probably the biggest challenge that we have. So, as you, sort of, mentioned in the introduction, my background is statistical genetics, which I think nowadays is, you know, nowadays probably more more frequently referred to as machine learning or even artificial intelligence. And this is really where a huge amount of the research at UNSW and the Garvan-Weizmann Centre is focused on is the application of algorithms, the development of algorithms and the development of statistical approaches to analyse that really, really high dimension data to figure out right, well, here is the tumour flow which is resistant to treatment, here is the the immune cell that is causing that pathogenic effect when you get infected with COVID, for example, or so on so forth. And so we develop and work really extensively on the big data analytics of this hugely, you know, high parameter sets of information, the scale of it is, you know, is genuinely quite phenomenal.

Tegan Taylor: So, it's such a powerful tool, you can look at cancer cells, you can look at healthy cells, how do you know when to apply it? Like it sort of sounds like you could get a wealth of information from any sample on any part of any body? So then how do you know, okay, when is this an appropriate tool to use?

Joseph Powell: Yeah, so this is, I guess, the story arc that that we work on in the research, we do a lot of what we call discovery or fundamental research, which is trying to unpick these mechanisms, trying to understand what's, you know, what's happening, you know, this new discovery of knowledge that essentially, you know, wasn't there before. And that is really foundational for us, that's incredibly important. And but in doing that foundational research, one of the most important things that we continually keep in mind is, what can we learn from that, that has got applications into a clinical setting? You know, discovery knowledge is great, you know, that's to be, frankly, a huge motivation for me, but, you know, we want to use that information to make, you know, make an impact on patients and the population and the community. And so, one of the, sort of, the settings that we do is work very closely with clinicians and clinician researchers to think and identify the practical problems that they have, you know, if they have lung cancer patient that, you know, 30% of them respond to a, you know, immune checkpoint inhibitor, what can we do to understand the molecular underpinnings of the 70% that don't respond? Or potentially, what can we do to figure out what's really responding in the 30%? And what can we learn from that to inform the treatments for the remaining 70%. So, as we go through that discovery research, we focus increasingly on this translational component and picking out the examples that we can make an impact. And then in parallel with that, making sure that we're able to think about the practical aspects of putting that, kind of, new knowledge into practice. And that's important, because there's a lot of very careful considerations about the way you change treatment, or you suggest a new approach, or you use this sort of stuff to underpin new drug developments. We have to work within, you know, existing frameworks, regulatory frameworks, ethical frameworks, and so on.

Tegan Taylor: How do you facilitate those conversations? Like, you're in your lab doing your research, I know that it's more complicated than that, and doctors out there with their patients, who's approaching who  in these conversations?

Joseph Powell: Yeah, so we, when we established both the Garvan-Weizmann Centre, which I head an entity between the Garvan Institute and UNSW, called the UNSW Cellular Genomics Futures Institute, we made a very deliberate decision to build multidisciplinary teams consisting of what we call, basic scientists – not basic but how we refer to ourselves – 

Tegan Taylor: I’m sure they love being called that. 

Joseph Powell: And clinicians and clinician scientists, so the teams consist already, you know, of clinicians working at that interface. And that, you know, that part is really important, because I think if you just have, you know, a bunch of academic researchers, a bunch of clinicians and a facilitated conversation, because that works well, you can often have, you know, a nice conversation, but you need people that really understand the nuances of both sides of things, to work at that interface. And so we embed within all of our teams, clinician researchers, you know, as part of the research programs, as part of the translation programs, and then you are able to help leverage their understanding about the as I said, the, sort of, the practical nature, by the way that you put these things into clinical practice, what are the real problems that they see in their fields and, and really supercharge that that pipeline of translation.

Tegan Taylor: How has this work been applied so far, in real people?

Joseph Powell: So we have, I would say, roughly, sort of, two major programs of work. One is in immunology, and one is in the cancer space. And the cancer space is, you know, as I've just been describing. So we've focused very much on what can we learn about patients that respond to current drugs on the market? And so why, you know, there's drugs like checkpoint inhibitors, fantastic, brilliant, but they only work in a certain percentage of patients. And so we've been focusing really specifically on working, trying to understand what's the cellular landscape of patient samples that are not responding, and using that to basically inform new approaches in clinical trials. And where possible, putting it into practice, already using existing drugs on the market. In the immunology setting, it's a little bit different, because one of the things that we, you know, are quite aware of is that the immune system does really diverse things, actually, you know, does lots of very positive things, does exactly what you wanted to be doing fights infections, and so forth. It also does things that you don't want it to do, like causing autoimmune disease. And so you want to be quite careful about the way that we understand how the genetics of people, that our genetic differences that exist between us, control that really fine balance between super healthy function and disease function. And so the work that we've been focusing on in that space is really what I call, you know, this foundational drug discovery space where we are looking at how genetic variation between individuals, how you know, are what we call our differences in our DNA makeup that exists, kind of, you know, between everyone, how does that functionally manifest itself in immune cells? And can we then use that to develop new drugs that target the genetic backgrounds of patients that are developing autoimmune disease, but not target people that have healthy immune systems? So that's a much longer program of work before that's translated. So this is really, you know, what I call, drug discovery work, which is, you know, has a long horizon for it to be translated, and a lot of difficult hurdles to overcome. But it's very distinct in the way that we would think about the way you translate that, compared to the cancer program.

Tegan Taylor: If someone has a cancer, like a solid cancer, you can take a biopsy of that tumour and study it. How do you know what to sample in someone who has an autoimmune disease?

Joseph Powell: Yeah, that's a fantastic question. In some instances, we don't know, at all. And so that's where that discovery, fundamental basic research, is incredibly important. And so, we typically take a blood sample, and from that can extract all the immune cells – that's a very standard approach – and we sequence the transcriptome, or the genomes of all of those immune cells, and we do that for very, very large numbers of patients. And indeed, people without autoimmune disease. And so we then start using statistical genetics, or our machine learning, or AI, or whatever your password is for this day, approaches to figuring out what are the differences in individual cells between the autoimmune patients and the healthy individuals? And once you understand that, and we've, we've done a lot of this work, then you can start really untangling what is mechanistically happening in the cells in the autoimmune disease patients versus the other. So we have this, you know, this discovery component to this where we just need to understand what are the differences in the cells across these 20,000 parameters of information. And five years ago, you know, that was almost intractable. But the technology has moved at such a rate that now this is, you know, we sequenced 20 million cells last year, for example, it's phenomenal the change in the technology, and how that's enabled our research to continue.

Tegan Taylor: I know that you've described this as a field that is relevant to almost any condition that someone will want to see their clinician about. The applications in things like cancer and immunology, you've explained, but what other things do you see this being applied in the future?

Joseph Powell: Yeah, so yeah, you're right, I think about this a lot. I'm incredibly passionate about the translation of genomics into clinical practice, for lots of reasons. And, you know, Australia, in particular, but many countries, has been fantastic at doing this for, you know, for some cancers, you know, for things like, you know, non invasive prenatal testing, and increasingly for what we call rare conditions, where there's often a really important diagnostic odyssey for patients. But genomics has largely, I think it's fair to say, been unused for the wide range of conditions and diseases that, you know, humans suffer from and that's for a whole variety of reasons. But the, I would say, you know, an important generality to that is that it's because those things act at the level of individual cells, and cell to cell differences are massive. And so this is why cellular genomics in my opinion, is able to make such a transformative impact because we now can unlock the understanding of all these genomic processes, which, you know, we do know how to work, we do know how to translate them, but we can do so for all of these cell and tissue systems for almost any different disease and condition.

Tegan Taylor: It feels like a really exciting area, that’s really, kind of, on the edge of what's possible. What do you see as the biggest challenges that are lying ahead for this field?

Joseph Powell: That's a good question. I don't know if I'm overly optimistic, I don't see… Well, I see challenges. I don't know… they're all ones that I feel can be overcome. So I don't know if it's, yeah, I don't know how big they are. No, I think I would, sort of, flip that and think less about what are the challenges and more about what are the problems that we need to solve and, you know, for me, as I've, sort of, described, we have a route for average genomic technologies that are being put into clinical practice, we do know how they can be impactful, and they can be transformative to patients and, you know, make it make a fantastic contribution to society. But cellular genomics has this, you know, this other level of complexity because we generate, you know, 20,000 parameters of information for a single cell, we do that for 10,000 cells for a patient. So that one of the things that we need to think about is the way that all of that rich data can be transferred into a clinical setting where the relevant information is transferred, but the irrelevant information, all of these, you know, as you alluded to the potential things that might be alarming for a patient but in reality, they don't need to worry about it, are not transferred. And so that data transfer, that, you know, that gap between the generation of massive amounts of data in a, you know, in a diagnostic lab or in a research setting, any information that gets put on a GPS form, you know, that that is an important piece for us to focus on, and I think, you know, it's a problem that needs to be overcome, you know, with some finesse and not without its own challenges, I suppose.

Tegan Taylor: It's been such a fascinating journey to talk to you, Joseph Powell. Thanks for joining us.

Joseph Powell: Pleasure. Thank you very much.

Ann Mossop: This event was presented by the UNSW Centre for Ideas and UNSW Medicine and Health, and was supported by Inspiring Australia as part of National Science Week. Thanks for listening. For more information, visit centreforideas.com, and don't forget to subscribe wherever you get your podcasts.

Ann Mossop: Welcome to the UNSW Centre for Ideas podcast, a place to hear ideas from the world's leading thinkers and UNSW Sydney’s brightest minds. I'm Ann Mossop, Director of the UNSW Centre for Ideas. The conversation you're about to hear, Future Medicine, features Anushka Patel, with journalist Tegan Taylor, and was recorded live. I hope you enjoy the discussion.

Tegan Taylor: The land I'm coming from now is that of the Yuggera and Turrbul People. I acknowledge this land and Elder's past, present and emerging. Scientia Professor Anushka Patel is professor of medicine at UNSW Sydney, and a practising cardiologist. As the Vice Principal director and chief scientist of the George Institute for Global Health, she has a keen focus on making healthcare both affordable and effective. Anushka, welcome.

Anushka Patel: Thank you. 

Tegan Taylor: Who's most at risk of being left behind when we're, sort of, looking at this future of medicine?

Anushka Patel: So I think if we look at Australia, for example, you know, we we sit in the bottom half of the OECD rankings in terms of health equity, when looked at, for example, the ratio of life expectancy among those of us who are at least educated compared to those who are most educated. So there are some major disparities within the population. Of course, we all know about the major gaps in life expectancy between Indigenous and non-Indigenous Australians. So there's quite a few gaps. There's many social drivers of those gaps that sit outside the healthcare system, and outside of medicine, and those, of course, need to be addressed. But really, I think transformation about how we deliver health care is another approach that really needs to address these inequities.

Tegan Taylor: I mean, we have Medicare, we have social health care in Australia that other developed countries, like for example, the US doesn't have. I think a lot of Australians probably think of us as being, like, pretty good. Where are the gaps?

Anushka Patel: So, I think we are pretty good, Tegan. I think, you know, I think we've got a health system that we can be really proud of. But the health system we have today has really been developed and established for problems of the past. And most of the health inequities that we see today are problems of the present and of the future. So perhaps I can give you a couple of examples. So, you know, when I graduated from medicine a little over 30 years ago, and since that time, the mortality rates, the age adjusted mortality rates from cardiovascular diseases, which is an area I work in, have dropped dramatically, the death rates have reduced by 50%. And that's in no small part due to major advances in public health measures such as reductions in smoking levels, but it's also been introductions of major innovations, new drugs, new devices, and new therapeutic approaches, such that a person who, for example, presents to hospital today has a much better chance of surviving from a heart attack and surviving in a healthy state than, say, 30 years ago. But those improvements have been inequitable, they're not fairly shared across society. So there are many segments of society where smoking rates remain high, and where outcomes from a hospitalisation from an acute event don't do as well. But because of the improvements in general, what we're also faced with is an ageing population. So you know, in the year 2000, about 12% of Australians were aged 65 years or above. Now it's more than 16%, and that's increasing by about a percent per year. We've also got the major problem of chronic disease multimorbidity. And that's where an individual might have more than one chronic disease, so they might have heart disease, they might also have diabetes, chronic kidney disease, mental health issues, such as depression and anxiety. But one in five adult Australians have that, but if you're over the age of 65, it's above 50%. Some major problems and again, it's inequitably distributed in the population. So those who sit in the sort of bottom quintile of socioeconomic status have high levels of multimorbidity. And our health system is really not equipped to deal with an ageing population and the growing problem of multimorbidity. It really needs transformation into the future.

Tegan Taylor: What does that transformation look like?

Anushka Patel: I think it could look like a number of things. But you know, perhaps, what I'd like to do is focus on three principles around changes that might not only result in better equity, but also a concept that's gaining a lot of credence around prevention. That is, those called, sometimes. the three P's. So it's around innovations or changes to medicine or health care that not only prevent disease, but also at the same time help promote equity and protect the planet, which is another dimension I'd like to talk about. So the first of that is moving the health system, or transforming the health system towards patient centred health care, with a much greater focus on prevention than cure. What I mean by patient centred care is that, you know, for each person, we're delivering the right care at the right time, in the right place, with a really strong emphasis on shared decision making. It’s care that is fundamentally customised. So it's individualised for a person. It’s collaborative between healthcare providers and patients. And it's coordinated between healthcare providers, particularly in the context of multimorbidity. It's also accessible to meet the needs of the patient, not really the convenience of systems and processes that our traditional bricks and mortar health systems are used to. And these aren't new concepts Tegan, but they are happening very incrementally when they need to be transformative. But I think, you know, for that transformation to patient centred care to happen, it will require really major shifts in vision, values, leadership, drivers of quality improvement, which include funding models, but also new workforce strategies. But I've also no doubt that maybe some of the other innovations we've talked about today around data and technology are going to be critical enablers for any transformation, particularly transformation that's going to promote equity.

Tegan Taylor: So you mentioned personalised care. And I often think of something that's bespoke as being like a premium product, but you're talking about equity at the same time and targeting this to that… the people who are most in need of it, which is, you say, that lowest fifth of the population in terms of socio economic status. How do you do both of those things at the same time?

Anushka Patel: So that's a good question. I think, you know, precision medicine, or personalised medicine, which are similar related concepts. The goal is to really identify the optimal care for an individual, you know, based on their unique profile, rather than that of the average population. And, you know, we do have tools to do that for chronic diseases, an area that I work, but they're relatively crude. And I think, you know, the power of precision medicine, in data, has a big role to play, even in chronic diseases. Where I think you can bring in the dimension of equity is to talk about the concept of Precision Public Health, which is a little bit controversial. But it's… a way to describe Precision Public Health is around delivering the right public health measure to the right public population at the right time. So using again, data, often big data to identify those populations that potentially have the greatest burden of disease, but also, signal detection, to identify those communities that are at risk of developing future burden of disease, and targeting public health measures to those populations. It's an approach that's been using not only health data, but non health data. So for example, looking at how people purchase food, what types of food they're purchasing in their areas, what's the neighbourhood walkability, like? What's the safety? What's the ability to exercise? Those sorts of aspects can be incorporated into a precision public health approach.

Tegan Taylor: I don't think the public has ever been as aware of public health, in the past, as they have been this past 18 months. We've seen public health interventions in practice. Do you think that that's something that's going to help with buy-in from the public?

Anushka Patel: Yes, I think it is. I think COVID-19 has given us great examples of where you can have fantastic innovations in medicine, like the development of vaccines and absolute record time, that none of us expected. But then we've fallen over in some countries, perhaps you can say most countries, including in Australia, on the public health aspects. And it shows us that these two areas, technological advances in medicine, and improving public health are not two sides of the coin, they're intricately linked, and they need to be complementary to really deliver health outcomes equitably across the population.

Tegan Taylor: So you've talked about your work as being around designing around patients and not systems. What do you mean by that?

Anushka Patel: So maybe I could give you some examples of what I mean by that. 

Tegan Taylor: We love examples. 

Anushka Patel: So, you know, I don't see, like, patients like this everyday, but it's not it would be not unusual for me to see a patient, maybe a 60 year old woman who is retired, she's 60, so, her average life expectancy having gotten this far, so far in life, is probably another 20 to 25 years. She's had a heart valve replacement, maybe 10 years ago. She has a history of long standing high blood pressure and diabetes. As a result of the diabetes, she's got chronic kidney disease. So she sees me regularly as a cardiologist, she also sees a nephrologist, a kidney specialist, she sees an endocrinologist to manage her diabetes, she's developed early Parkinson's disease, so she sees a neurologist. And she's got a very good GP who tries as best as she can  to coordinate the care of this patient. So I see her every six months. But she comes in a bit early to see me because her GP’s worried that she's a bit short of breath, and she's not sure what's going on. I see her and I think, you know, this might be a bit of heart failure, but I know she's also got chronic lung disease, I want to get some blood tests. And she said, I just saw my endocrinologist a week ago, and I've just had blood tests, I don't want more blood tests. And I say, well, I want to change your medications. I'm on 10 medications, I don't want an 11th medication, I can't handle this. And so, you know, the moment, the way the system is positioned, this patient needs to go and see specialists in their rooms, coordinated by her GP. They send her for various tests, they don't coordinate those tests between them, we don't see each other's results in a timely manner, we all adjust medications independently, we try and find out what's going on, but it's not easy to do. So she's rattling along with multiple medications, which she's finding it difficult to adhere to. So there are many ways that you can imagine we can change a future health system, so it's much more patient centric for such an individual. You know, could it be, for example, that she has a technology platform at home, one that is suitable for a 60 year old person, who may not be as digitally aware as others, but enables all her care providers to know what's happening at the same time. She could have some simple but smart sensors that recognise changes that need attention. She could have, you know, we could have technologies that ensure that if she does need, for example, additional tests, these are coordinated by a digital healthcare assistant that coordinates what she needs from her various providers. Now, these sort of patients are going to become the norm in the future. And so we need to really transform the system to meet their needs, rather than ours.

Tegan Taylor: The other thing you mentioned before was farm polypharmacy. So when people are on a bunch of different pills. You've talked about the idea of polypills as a potential innovation in the future. What's that?

Anushka Patel: So this is an area of research that, well, it's not researched, it's actually in reality now. So we know that with the problem of ageing, and multimorbidity, that many patients are on many medications. And because these drugs are effective, they're highly effective, but it's very difficult to remain… for doctors to prescribe and for patients to remain adherent long term to many medications. So the concept of a polypill is really around simplicity. Now, if we combine multiple medications into a single pill, will that make it easier to prescribe for a healthcare provider and easier for a patient to take and to take long term. And indeed, we and others around the world have done quite a bit of research that’s shown that that's the case, if you can do that, it does actually help. Now there's an extra hurdle to get these polypills into practice, at various levels from a regulatory perspective. From an implementation perspective, it's, you know, it's difficult to change healthcare provider practice, people are used to prescribing individual drugs and fiddling with the doses, even though the trials are showing that this is as good if not better than that traditional approach. So that's where implementation science comes in. And that's also an area where we're working.

Tegan Taylor: And climate change disproportionately affects people at the lower end of the socioeconomic spectrum as well. It's all interconnected.

Anushka Patel: Absolutely. It is, absolutely, absolutely. 

Tegan Taylor: So if you only had one intervention that you could implement to give the biggest bang for your buck, like, what one intervention would most benefit the people who experienced the worst health inequity?

Anushka Patel: That's a very challenging question, partly because I don't think there is a single solution but if you know, to be honest, if I was to identify one area, broad area, I would say it's actually outside of health and medicine, I think it's the social determinants of health. You know, we have in this country, and many other countries, we spend a lot of money within healthcare systems to try and address the needs of disadvantaged populations without major gains. And part of the reason is because that's not the driver to the inequity. Its other social determinants, its around education, around housing, it's around employment, etc. And so, you know, I think much more attention needs to be paid to the social determinants of health.

Tegan Taylor: So, some big challenges ahead, but also some really exciting opportunities. Anushka, thank you.

Anushka Patel: Thank you very much.

Ann Mossop: This event was presented by the UNSW Centre for Ideas and UNSW Medicine and Health, and was supported by Inspiring Australia as part of National Science Week. Thanks for listening. For more information, visit centreforideas.com, and don't forget to subscribe wherever you get your podcasts.