 Okay, so this is working, yeah. So thank you very much for the introduction. So I'm going to give you today an overview of how AI can be used for person-centered care and especially what we mean by it, okay. So why we want to, we think that this theme is important and why we think that we need to do research on it. So there is a strong concern within the European Union about the sustainability of healthcare services in the next 20 to 30 years. So there is an increase in healthcare expenditure, there is the aging of population and also very, very much relevant is the high cost of managing chronic diseases that goes of course higher and higher with the aging of the population. So especially in Ireland we have a huge percentage of Irish population that is age 65 and over and as you could see in the previous talk, actually this percentage is increasing and increasing. We have 21% of these older adults that are frail and then often live with multiple chronic conditions. So actually there are some studies about how this will affect the capacity required in primary care and those numbers are actually quite scary. So we need to do something about how to make the healthcare more efficient, how we can improve health outcomes and bring sustainability to the overall system. So how we can do it? So essentially we have a need for a person-centered approach to care versus the disease-centered approach that we see now. And especially in this framework what is very relevant is understanding human behavior and how this can be influenced to improve health outcomes. And in this framework we are focusing our attention on the interactions of multiple chronic conditions and how this can be tackled using behavior change and the person-centered approach. So I'm going to present to you two projects that we are doing in the lab. So one is around human behavior modeling. The other one is around multiple chronic conditions management and how we can use the human behavior modeling to tackle multiple chronic conditions. So the first project that I wanted to present to you is the human behavior change project. So this is a research project so it's pretty scientific and essentially the collaboration is between IBM research and those universities on the left and it's funded by Wellcome Trust. So why we want to focus on behavior change. So healthy behaviors, when I talk about healthy behaviors I talk about, for example, exercising and dieting, keeping healthy body weight, not smoking, not drinking, those are some examples of behaviors that are very important to keep on a population level. Because for example, lifestyle habits can increase life expectancy up to 14 years or 15 years actually as well. So it's very important that we actually include the healthy behaviors within the healthcare system and within the model that we want to build around the person. And actually this is very important, so we are focusing on aging and multiple chronic conditions but this is also very important for example, people in the workforce because in that case we are trying to increase the life expectancies but also the quality of their life. And this of course have a direct impact on the healthcare expenditure and quality. So when we talk about the behavioral scientists and the behavioral research, we are trying to let's say answer to this question. So when it comes to behavior change interventions, what are the type of interventions that works, when they work, how well actually they work and for how long, for example, which quarter population with which characteristics and how we can deliver these interventions. So those are the questions that we are trying to answer with this effort. So essentially what we are building here is an AI system for behavior change intervention prediction. So we are working on two different levels. So one is that we are building an NLP system to extract those information from literature, from documents, from paper and we are processing this information to build a reasoning system to generate new knowledge on one side and to recommend the most effective intervention for a specific quarts and population on the other side. So in this framework, the human behavior change project is a very research oriented project. So how we can use these results on a real life setting. So I will introduce you now to the other project that we have. There is essentially a bit more practical and will let us see how we can leverage behavior change in a real life setting. So we are focusing here on multiple chronic conditions. There are a lot of use cases that you can think about. We are focusing on this because there is a huge need for managing multiple chronic conditions in healthcare systems and there is so scarcity of research in this space. There is a lot of research when you think about single disease. So there is a lot of research on diabetes, a lot of research on COPD, a lot of research on heart conditions, pretty scarce research and effort on managing multiple conditions. And this is actually quite relevant because it happens quite often that when you have multiple chronic conditions, the care plan that you might have for each one of your conditions can actually contradict between each other. So it's important to understand how the person is doing on the overall when they have more than one chronic conditions. So PROC is a research project and we have a trial of 120 patients over 65 years old that have more than one chronic conditions in the space of diabetes, COPD, and heart diseases, so CHF and CHD, they are equipped with wearable environmental sensors and we are monitoring them and try to see how they are doing and how we can promote them to actually self-manage their conditions. So this is a high-level platform data flow. So on the left, you can see the person. So the person is equipped with a set of devices. You can think about blood pressure cuff, you can think about blood glucose monitoring depending on the conditions that they have. They also have a tablet where we put a care up where they can actually answer questions about their mood levels, about their overall state, about their social interactions, about their activity levels. We collect all of those data. Those data are collected into an IoT platform, these CABISIMS, and after that they are actually pushed, they are de-identified, anonymized, and pushed through interact. There is a cloud-based service where we have built a bundle of care analytics to process this data and get insights from the data that we have. So interact is our cloud-based platform. So it is essentially a platform for multi-morbidity management. It's built on top of IBM cloud and uses IBM cloud for storing the data. So what is very relevant about this platform is that it lets you to run a sort of care analytics that will give you insights on your population and your patient. So I'm going to present now to you just one of those care analytics, the health and wellness profile. So we have built a tool to essentially represent the state of the person when they are, of course, depending on the conditions that they are affected to. So the state of the person is actually scattered over several dimensions, including behaviors. Behavior is a very important one. What we would like to see is how all of those variables actually correlate between each other and what is the effect that they have on the overall health of the patient. So we have used the tool for a number of use cases, including risk identification and prediction, and next best actions. So what is very important here is that when we think about people affected by multiple chronic conditions, we have to think that those people have a network of care workers. Okay? These care workers need to be, need to decide about the next best action for each one of this person and needs also to know how is that person doing compared to what they are expected to do, given their state, given their conditions. So for this reason, we have also used this tool for resource allocation. So how can I, so how the schedule of the care workers can be optimized to serve the needs of the pool of people that they have to care about, to care after. So I don't have enough time to show you the demo, but please feel free to reach out to me if you want to know more about it. But essentially we have a dashboard where you have a lot of variables describing your person. So it can be demographics, conditions, vitals and symptoms and behaviors. And you can actually click on each one of those tabs and the system will give you, will update all the probabilities and will give you the probability, for example, of a person having, for example, it will give you the probability of the social participation level for a person having hypertension and diabetes or it will give you the probability of being obese for a person having CHF and hypertension. So please feel free to reach me if you want to know more about it. I don't have enough time to actually to show you the demo now. So I'm actually done. So thank you for your attention. Thank you very much, Alessandra. That was really interesting. I think it really brings to the fore the importance of collecting data and understanding, I suppose, what is happening in the day-to-day lives of patients to drive better outcomes within our healthcare system. We've alluded to already the spend within healthcare that is, you know, quite really an incredible chunk of our GDP and to be able to drive any efficiencies within that. One of the challenges that I'd just like to quickly ask you about, collecting this data, you're looking at, let's say, potentially elderly, infirm, chronically ill patients, their interaction with these collection devices, granted wearables, could be fairly innocuous and they wouldn't have to interact too much, but something, you know, like the tablet computer or if there's an app on their phone and interacting with that, how did you manage those challenges? Yeah, so actually, that was quite surprising. So we had a team of researchers and actually triage nurses as well devoted to take care of these pool of people to teach them how to use the care app that they had, the tablet, the wearables and so on. So we had, let's say, a period at the beginning where people were actually a bit struggling in using the devices and the tablet, but then after, let's say, a month or so, they got used to them and they were actually so much engaged that some of them actually brought the tablet and the sensors with them in the hospital when they were hospitalized. And they were going on sending data to us and getting our recommendations on activity levels and everything, because they felt that they were in power of their condition and of their, let's say, health state. And that was actually quite surprising. Very good. I suppose the old adage of you can't teach an old dog new tricks, really doesn't hold true that they're certainly there and if they can be engaged and empowered, that's certainly something that they're well capable of. Excellent. Alessandra, thank you very much. Most interesting talk. Thank you. Thank you.