[00:00:05]
>> All right thank you so we I want to have Also 1st you know our something so since she is the director of the center and then I'll take it over and give the welcome I just well they will come it's really really wonderful to have all of you here with us today especially on this very very cold day going to call for Atlanta old I know some of you lived in other places and maybe you think this is not too bad but for those of us who live in Atlanta for a long time it's probably one of the worst days of the winter so we are a bit spoiled but thank you for joining us really had a very exciting program ahead of.

[00:00:38]
The panel discussion presentations and posters there so I hope you enjoyed make some new connections. Meet people that you didn't connect with people that maybe you knew before or learn some new things generate ideas and hopefully build some new collaborations with us while he hopes ha but without further ado I'll turn it back over to my colleague Dr knows all and looking forward to interacting with all of you during the day thank you thank you for being Thanks in art.

[00:01:15]
Ok So thank you very much for joining like welcome you were very excited so I should hear all right no that's Ok. So I'd like to welcome each and every one of you for joining us today we're very excited to share all of the research work that has been done at Georgia Tech at Emory and other places I want to welcome all of our panelists and I would rapid fire presenters and I want to welcome our students who took time out of this kind of the last the end of the semester where I know you're all just trying to keep your head above the water at this point.

[00:01:55]
And without further ado so let me get started with. First let's see if it works. So just kind of a quick reminder I would center the center for House and humanitarian systems for my vision and for my mission perspective I want to highlight on the vision that our goal our vision is to develop tools and I'm collaboration's in order to transform the decision making because we want to improve the health care operations health care delivery and humanitarian systems operations we accomplish this through multiple pillars education outreach projects and innovative research so when I was thinking about how to do the introduction before I do let me say welcome also the rest of the director so I mean to do that you might you know art my co-researcher that actors are.

[00:03:00]
Dr Ira I don't know if they're here now but I know that a lot of their students have poster presentation and I when I thank all the support staff. And the sort of beneath Taylor and I began Thank you all for your support for putting this event together it's not easy saying it's really logistically very challenging.

[00:03:25]
And we appreciate all of their work for putting this together so let me laugh. Ok so I would when I was thinking about this talk and kind of just the introductory I was curious to know what is the history of health care in the u.s. where it's every day we hear about health care you open the t.v. And that's something about how scary open very you know you listen to Fox There's always something about health care so I wanted to kind of look at it when did it start how did it evolve and you know I'm not do not walk you through every step here it starts in the 1800 where it was started for the manufacturing workers where companies.

[00:04:10]
Try to protect some of their workforce against sickness there was no structure it was a trial and error type of system and then it's all just you know the American Medical Association been. Grew tremendously during the specter decades. That was organized medicine and then hospitals in $1000.00 and American Association of labor legislation drafted insurance.

[00:04:40]
And as you can see as we walked through this. There was a few themes that emerged out of this one of them. That refuted Lee there was an attempt to have government sponsored health care coverage that repeatedly failed in the us and there were some incremental kind of changes the health care coverage system over over the years over different administrations.

[00:05:12]
And that was kind of where it started us and the Europe started to kind of separate and their health care coverage journey another theme that one would see. A lot of the health care initiatives were quite reactionary right so the. Post World War 2 employers were not allowed to kind of offer to raise to give to kind of attract more workforce so the one way that they were able to kind of gets more people is to offer them some Has care covers so kind of it was always kind of holding little by little in order to.

[00:05:52]
The current problems and the decades. In the 1960 s. that's when the health care expenditures started to get dropped and in the u.s. they were 5 percent of the g.d.p.. Over the years it started to grow little by little from $8.00 to $12.00 in the ninety's $13.00 in the 2000 and now it's $17.00 So this is the health expenditures.

[00:06:26]
As percent of the g.d.p.. So I'm I'm sure everybody has kind of heard or seen something like this which is. We would expect wealthier countries to how. Much higher per person expenditure on health care that number as the country is but even if we kind of zoom in on the on the vote the valid countries we know that the United States health spending per person versus g.d.p. is kind of an outlier compared to other developed countries in fact so anybody can guess who this is which countries this one and I guess a tiny tiny tiny tiny tiny country in Europe Luxembourg this is Luxembourg and this is Switzerland and then over the years we can see there is a specific decade where the u.s. the rate at which the u.s. expenditures started to grow at a faster rate than comparable countries so in the $9090.00 s. when the United States g.d.p. sorry and actually x. percent of the g.d.p. started to grow at a higher rate and kind of the difference started to Barbour's another interesting thing is when we look at public versus private spending.

[00:08:07]
Among across these countries we see that United States in line with other countries when it comes to public spending. The difference is really private spending on health care and so while these countries kind of the average about 2 to 3 percent the u.s.. At a much higher. However right well let's say a couple of more interesting statistics I was looking kind of you know let's dig deeper into these costs and one of the I found those statistics to be interesting is most of the caught a big chunk of the cost comes from administrative costs in the health care system so he was.

[00:08:54]
Mistaken and. Another interesting statistic that's related to kind of see is that we work and was disintegrating is that 50 percent of all chronic diseases are linked to preventable problems and then the last one that I found also interesting missed appointments how much they cost the u.s. health care system every year so $150000000000.00 per year are the costs.

[00:09:20]
Missed appointments just people not showing up to an appointment and this one doesn't mates with me as someone who does a lot of work with senior with the capstone So you know design projects just about every health care clinic project that is something about you know shows and the no show right and how different clinics have to deal with that by you know buffering by double booking triple booking appointments just to kind of hedge against that uncertainty now let's look at some good news for the u.s. though right yes where the highest cost but in terms of innovation I we come and.

[00:09:59]
Relatively top lists right and we've actually climbed a few spots from 20229000 this innovation and if you're curious about it it takes many many metrics and it combines them both on their input to the end of ation the infrastructure the market. Sort of this is them on the output how much technology technological advances are generated by that time tree and when they put that together and they do the different weights the United States comes 3rd and the world.

[00:10:40]
Another thing that was interesting in that report they lay out a lot of promising fields for medical innovation and one of the things that I'm interested that nor which of them is the process organizational and process innovation and this is most fun it's not meant to be read even though if you read it that's that's quite impressive.

[00:10:58]
The one thing that out of the few things that I'd like to highlight are you know things related to draw on the livery of medications telemedicine virtual medicine and artificial intelligence and these types of times right they resonate with us people and in the logistics field whether you're working with health care or just regular general supply chains.

[00:11:21]
These are common technology is now right in distribution centers and supply chain systems so it's going to be very interesting to see how some of the research that's already been established and. Their supply chains retail and commercial supply chains how that can be translated into the health care systems when it comes to drawing delivery or artificial intelligence and machine learning and they to mining so now I want to kind of that that was a good segue into talking about.

[00:11:57]
Our systems engineering approach to looking at these extremely complex system so when we talk about health care systems the health care system is quite complex Why are they complex there's multiple stakeholders there's multiple and there to pendent subsystems that you cannot model you cannot design one system subsystem on its own and we all know what happens if we try to optimize one aspect of that subsystem and there's an unintended consequence to another subsystem.

[00:12:29]
There are supports physical spaces and limited resources and multiple organizations there is government there are people there's insurance. And of course these resources are shared right so you have a lot of these limited resources that are also shared across multiple subsystems that makes these systems quite complex Another aspect is that doing objectives or conflicting objectives so we have you know if you try to minimize the cost it's quite risky because then you're looking at sacrificing the service or the patient care and so where is that balance between the $2.00 and $1.00 of the things I was reading on the differences between the u.s. and other developed countries.

[00:13:15]
The reason why one of the high level reasons of why the costs are so high for the us are 2 things 1st people come visit the high utilization of their resources so u.s. patients visit doctors or seek medical help more often than others nations per person and of course the cost per service is much higher and $1.00 of the things that got highlighted and.

[00:13:43]
From multiple resources is the idea about. The defensive medicine that we have in the u.s. where you know you will always you know get the test that you think you should get it right if you tell the doctor I need you know this particular test most likely you would get it while this is not usually the case and many of the other countries around the world where there is more kind of.

[00:14:11]
Not every test is available for every patient. And then the other aspect of these health care systems is of course the data complexity and the uncertainty with the data and all of the gaps that we have we have in this so all of these need to. Having these health care systems being extremely complex and the need for systems engineering type approach to design or to operate these systems are in fact the White House.

[00:14:47]
And that's the president's Council of adviser on science and technology they prepared the report in which they outlined some of these and here are some kind of selected quotes from that report highlighting the importance of systems engineering in order to address some of the challenges in the healthcare delivery systems but also how systems engineering concepts should be embedded in education and training for people involved in health care.

[00:15:19]
So what we believe is that you know if you look at the different subsystems and you take them apart they are fragmented and then through systems in your generic you can have a more coordinated and integrated health delivery system and we have many many tools that can be applied some systems engineering you know you sometimes or traditionally that were used for you know designing airplanes and now there are more to weren't designed in supply chains and distribution centers but many of those tools are applicable to the health care system and we're always looking I am personally always looking for kind of seeing how we can make those tools even more.

[00:16:03]
They consider more the practical aspects of the health care delivery system because it has got driven. In the manufacturing floor lie there is a car that's going through this line while in this case you have patients and patients behavior is unpredictable but also the person who is working on that mission is also a human rights a kind of all these human working on the human makes these systems even much more challenging than that's traditionally.

[00:16:30]
The flow system so I'm going to end with a challenge to you I hope that you would benefit tremendously from this. But really it's kind of to have some concrete output what we would like for everyone to kind of think about these questions What knowledge can I take away for myself from my colleagues from my administration and the Constitution.

[00:16:57]
What action can I take and what action can I must've 8 others take so this is your homework right for this forum and hopefully by the end of this you would have an answer for each one of them. Thank you. For raising the money. You know everyone. Good to see you this is certainly a very warmed up crowd.

[00:18:01]
So I'm going to get started and. Then struction for 3 to 5 so I interpreted that as 6 or 7 so bear with me. So I'm a city planner and when I think about what what the main contributions are to health I think about this iceberg it's interesting that it's freezing city and we're using an iceberg.

[00:18:26]
We start at the bottom of the iceberg where we have the social determinants of health and so that's really based around where we live work and play we're born the systematic structure of society so think about this 1st column here to the right that shows public transit that's moving along quite well and efficiently 1st is what we have in Atlanta next to that.

[00:18:52]
The system itself leads to certain behaviors those behaviors lead as you can see here indicated by people engaging with one another walking being active physical activity a human connection 1st is lack there and a fair amount of stress and perhaps many of us have seen have experienced being in that kind of traffic here in the city and then finally the events the events are at the top of the pyramid and often these events of the things that we focus so incredibly on it's the hardest tack it's the back pain it's the headache it's the broken leg from being in a car accident or it's the joy and relationship that we have when where connecting with other people and where in communities where we really can thrive and so it's in this frank work that I think about what the true source of health is so and so.

[00:19:50]
The World Health Organization indicates that health status is determined the majority comes from the social conditions and behavior so 55 percent and 30 percent in health care here from the w.h.o. is not it to contribute 10 percent and this is not to say that health care is not an important component of seeing a healthy population but I see that this is one piece and we have totally ignored in many respects when we think about health healing the bottom of the pyramid and the behaviors that lead to these events and these outcomes and then when we think about the health impact pyramid this was authored a couple years ago by Dr Friedman that the c.d.c. the bottom of the pyramid indicates their socio economic factors and the context in which.

[00:20:38]
The individual's default decisions can be healthy it's at the bottom of the pyramid where these interventions can have a population level impact. But when you have a population level impact where you're changing the socioeconomic factors how much for people being paid are they living in safe and appropriate housing Do they have access to good quality education and they get good jobs later so they can pay for the resources that they need to thrive in society.

[00:21:06]
Do they have solid connected sidewalks so they can walk and get to places safely that's lost a lot of money at the top of the pyramid are individual and group level interventions that don't cost as much and so for changing the contacts to for dealing with socioeconomic factors we really have to make a commitment to resources.

[00:21:29]
And so in just a few slides I have left I'm just going to give you an overview of the kind of work that I've been doing around health equity engagement and so this is a project that I started a couple years ago it's called Measuring the dream and the primary question here is since Dr King's assassination how far have we come in realizing his dream how far have we come in realizing his dream and so we developed a resource that essentially quantified.

[00:22:00]
From 1054 to 2014. Different components that we find in the dock and I have a Dream speech and you can see here with regards to health we start at point 73 and we end up point 91 worry quality as we measure it as one and you may be thinking that's great he's made great strides but the strides that we've made here as it relates to getting to point 91 is because there's a there's been a significant increase in mortality of white males between 30 and 45 because of the opioid epidemic so we don't want to see any improvement in this index because one group is doing poorly we want to see improvement because all groups are improving overall.

[00:22:46]
I've also done work on dashboards you saw in the circles and so the 1st on the last is the neighborhood quality of life and health dashboard where we actually combine quality of life for Built Environment mesh metrics with what health data and this is organized based on the neighborhood planning unit system which is the decision making system here in the city of Atlanta so communities can truly engage in what's happening around health and design in the structure of their communities then we moved on with some c.d.c. funding to build the Fulton County health environment livability platform and now we're working on the house a 2 year old data collective and this is a database in collaboration with the folks in the College of computing to look at affordable housing in the city of Atlanta.

[00:23:36]
Current project that the New York ng on is the using gates right now action for health project this is engaging with chits So we want to understand how training kids to be advocates in their communities leads to improved physical activity out. Individual agency and wellbeing in those kids so we've worked with kids all over the u.s. about $400.00 kids and Boys and Girls Clubs afterschool programs and so on and you can see pictures from Hawaii which is perhaps the best trip to here in Georgia and Virginia.

[00:24:11]
We created a fantastic curriculum and we've been sharing that with folks cross the country we have a paper that just came out or we are examining the intermediate steps in what we refer to as the ladder of citizen participation to understand how kids can truly partner with adults to have an impact on their built environment and this latter kind of short hand moves from the bottom where communities are receiving information.

[00:24:39]
To planning processes community processes health processes they are giving informations their feedback where there may be consulted or there on a some sort of board to actually sharing our partnering with the information and the decision making and having power and what happens and that's where you see True true power when we think about kids in this process.

[00:25:05]
The Using an action for health program has kids engaging in. And in understanding how to assess the environments in which they live whether it's their school environment their nearby park their home environment or neighborhood and then they present to their adult leader he's running their club we want to work on water fountains or we want to work on improving our park and based on those presentations they pitch to a decision maker based on evidence and they follow that through to some decision and so it's a waterspout example that was our kids in south Florida they were able to get a new.

[00:25:47]
Protocol established to have the water fountains cleaned on a regular basis and you may think that's not a big deal but when you're 12 years old and you're in south Florida and it's blazing hot and the water fountain is nasty having a clean water sound is a big deal and with regards to the park cleanup that was in Hawaii in the hooey the Boys and Girls Club there and they presented their county council make up 80 $1000.00 to improve their part and so these are place based changing the context of the environment so the healthy choice is the easy choice is the default choice it's the attractive choice.

[00:26:29]
I'll leave you with 2 final slides to final size this is somewhere that has been done in collaboration with my colleague Ted Russell and environmental engineering our student Abi law all I don't know if he's here today. And some colleagues in Minnesota took a look at demographic inequalities in health in air exposure in metro Atlanta What do you think is the case is truly the case but you can see the black lines show kind of a clear demarcation in different health outcomes so the darker the color the worse the health outcome the slides the image the right shows red majority black neighborhoods blue majority white neighborhoods and so you can see that there truly is a disparity when it comes to these health outcomes with regards to air quality and then finally we're doing some really cool stuff with Google Street View we're rippling images from Google Street View we just finished a study in Atlanta we're working on this one that compares Atlanta to San Francisco where we coated the images from Google Street View based on air.

[00:27:42]
Road and sidewalk based on our algorithm we're able to determine how walkable or code areas based on how walkable they are and we compare that walkability index to walk score and we found our measure truly is more representative where people indicate they were walking relative to where walks for what walk score indicates walking environment and so this is helpful in that we can direct resources to areas that need to improve their walkability especially we know that there are vulnerable populations and that in that community.

[00:28:18]
And so I'll just end by turn emphasizing that when you think about moving from step care to health care we want to consider broader on slope or a broader umbrella of what the medium by which that is. Should be operated through and so yes I do believe very dearly that if you need to go see a doctor go see a doctor and you should be able to find it find that health care.

[00:28:46]
Resource in your community and it should be accessible but they are prevented as things we can do kind of the routine stems idea they were prevented if things we can do so you don't see it so we don't have the problems of heart disease and diabetes and all the things that we know we have in this country that do not show up in other places Thank you.

[00:30:24]
So so. So this. Hired me one thanks for having me here this afternoon what I want to talk about is some of the things we're doing in pediatrics around Georgia Tech and I just it's Ok. Startle them so you may or may not know that Georgia Tech has something called the pediatric Technology Center it was stood up in 2012 in conjunction with Children's Healthcare of Atlanta and it started with a $10000000.00 gift and so we were able to fund a lot of research in that relationship and we did it in research areas all across campus so everything from 3 d. printing and medical devices to NATO technology and special coatings were 2 beings were generally of medicine diagnostics and imaging maybe even applying machine learning to those d. can see how it just changed over time mobile health is huge as you can imagine health data analytics we did a lot of that work here and I asked why a drug delivery methods and we also have programs for students so there's a lot of opportunities to work with Children's Healthcare of Atlanta and other pediatric hospitals that we work with character to check one thing we didn't though is we only work on clinician driven problems and that's really important because academia in general tends to like to study something and then figure out what could you do with it after you understand it what we're doing is we're saying we have a problem in the health care or the clinical setting and we want to fix it so what can your engineers how much was to do.

[00:32:16]
We have several different funding mechanisms to make those ideas come to reality so everything's for. An innovative program called Quick when's the idea there is very much applied research so what what can you do in a matter of like maybe 12 months to 2 years what can he bring to the clinic that can help them already start solving their problems we John and I blame research so more of the basic kind of research maybe looking at virus like particles or whatever it might be child impact grants are also along those lines kind of the bigger brain with a longer period of looking at and and now we also have an innovation fund so the employee Foundation decided to give Georgia Tech and Children's a $5000000.00 endowment and that we get to fund a couple of projects every year through that innovation fund and again we work in all areas of campus so it doesn't matter we've done things in policy and.

[00:33:11]
Modeling and whatever else it might be we work all across campus so anybody can be included it doesn't mean that you have to have even a health background at all a lot of times it's really just the engineering or science skill set that we're looking for and maybe we're applying it to something you typically apply to again we have lots of students options for students as well there's one program in particular called the petite Scholars Program it's run out of Id be and the idea is to bring an undergraduate and was a faculty member and and post-doc or graduate student to work on a project during the course of the calendar year and the goal is and to be able to have that student present a paper to really help further whatever research area they're working on and I will say with that students from all over the the city can participate so you could be at Morehouse or Spelman or Agnes Scott or Emory it doesn't matter what school in the local area that you are Georgia State kind of.

[00:34:12]
It's we also have another resource on campus that you may or may not know about it's called the Center for the global center for medical innovation and the idea there is to help small companies with their start up they do small batch manufacturing they can help you with f.d.a. regulations and all sorts of things like that so they're a fabulous resort resource you can also do animal trials or cadaver studies so it's another great thing that if you're interested in place feel free to reach out after this talk like I said to start with Children's Healthcare of Atlanta but we've been working with a lot of other hospitals as well one is children's we're seeing Kansas City Shriners Hospitals as a new one and they're interesting because there are actually 20 hospitals around the u.s. and one in Canada and one in Mexico and so it's they're going to a really great partner they do a lot of a mobility studies and understanding how how a person moves and Charles national in d.c. there are great too and we just had them the campus and also Cincinnati Children's a significant new one that we're going to start in the New Year One thing I wanted to convey is that as a part of this just like we're working with all sorts of other children's hospitals in pediatrics the resources are really small and so you have to use them wisely and one of the things that happens in pediatric research that is not as common and other types of research is it's very open everybody wants to do what's right for the kids and so if it means that the 2 of us work together and I would normally work with you that's Ok because it's all about making it healthy and so there are several organizations that we participate and one is the International Society for pediatric and of ation this started about 4 years ago it is all about what is health care going to look like in the next 40 or 50 years and the vice president of the organization is sure if the regime is a faculty member here on campus and we've both been involved in that organization since the beginning I sit on their board.

[00:36:11]
And then we also work with the International Children's advisory network so I'm the president of this organization and Sherry sits on this board so it's really great to have both of us here kind of working towards the same goals and kind of bringing these organizations together that in the very beginning they didn't know each other existed I can't it's also been around for about 5 years.

[00:36:32]
And what it's all about is giving the pediatric patient voice and health care and so what Michelle was just talking about were kids would go and talk and get money for a park this is the same kind of thing search today for example we have 7 kids presenting at the f.d.a. and I just received a text with pictures of them talking on stage but it's all about getting them opportunities to tell their stories.

[00:36:55]
They are the experts and so I'm going to go off a teeny bit off topic here for just a 2nd because I'm presenting a little bit for somebody else who couldn't be here today because she got sick. But what's really important here is it's really important to include kids when you're looking at health care it's really important to understand their perspective they are the experts at living with their condition they are the experts at knowing what it's like day to day and what issues that they have.

[00:37:22]
And so even a doctor who's been trained and maybe is an expert in that condition from a medical standpoint they're not living with it every day it's not their life every day and so it's just really great and important to include kids in research and one of the things that I think is really interesting and I know this is the story that Megan would have told she.

[00:37:40]
Didn't get sick is she's working on a project that we have with children's healthcare Atlanta it's called cancer care for kids by kids and the whole idea is to help kids kind of more seamlessly get through. Doctors' appointments when they have to go and for chemotherapy or whatever and those news are long days for them and they're not very fun for these kids and so it really started out as a what could we do to make this a little bit easier on the kids you know maybe less time in a certain waiting room maybe giving them harsh reality video games all the things that you might kind of think of in your head but when they asked when the girls they said you know what would make this a better experience for you and she said all she wanted was a private bathroom that she could throw up then and not have everybody around her and you're like Ok that's it's such a powerful thing to go when you think you're going to get bright colors and games and things and it really wasn't about that at all so it's just it's really important to talk to them.

[00:38:40]
I give me of this these are all the community partners for. The International Children's advisory network that on on the president but but this shows you is that we really work all across the industry in pediatric health care we work with the yesterday we work with the with north.

[00:38:58]
All sorts of international groups as well and so we really are involved all around the world n.p.t. Africa health care and again we think that's good because you get everybody working together for the same goal and that really makes a difference. So here I'm going to talk just super brazenly about a couple of projects that we have just so you can kind of get a high level idea of some of the things that we've done we have worked in the policy areas so looking at what is it cost the system if you diagnose does diagnose a child with autism at 5 years old versus 18 months can we convince lawmakers that you're going to save so much money over the life of that child it is well worth it to require testing at 18 months as opposed to when those kids figure it out when they go off to school.

[00:39:45]
Pietro kids in the state of Georgia so how does that work for you if you are somebody with asthma and you don't live near a center who specializes in asthma What does that look like and how bad is that and so they figured out that the outcomes for these kids who are 20 miles or more away from a specialist in asthma care that their outcomes are much worse and so what this was used for is to help Children's Healthcare of Atlanta look at the state of Georgia and figure out where they want to place the next urgent cares or specialist in the state to help combat this problem this is a defining cancer care for kids by kids it started out as a something on paper where the kids feel scared where they want to have different things happen and now it's turned into an i Pad app that is being used at the s. like cancer center and I get a lot of this is about you know how can we ease the anxiety that some of these kids feel throughout the process where where's that really kick in and where can we help improve and by the way the bathrooms are being private bathrooms are being installed in the new children's hospital when they build it in the section so that these kids can have that.

[00:40:54]
Data analytics so let's see what is this one out yeah this is pretty cool so one of the things that they're working on is for kids with severe with autism severe behavior problems they have a lot of repetitive motions and they can be very violent but a lot of times it's a very it's repeatable you can see the the motion that they have and.

[00:41:16]
So what they're doing is they're going to use sensors more on the children so that they can pick up whether or not the treatment is working it. At the at the home when they go home after training at the facility and so it's a way to monitor and just kind of get on is on an understanding of truly what is it like when the child is homes when they're there.

[00:41:39]
M.r.i. system again you think bright colors maybe fish on the wall or whatever but they figured out that is. It's a control what like color does that's what made them happy. Because it was something that they had control over and I know that that's a big thing when someone who said especially something chronic You really don't have a whole lot of.

[00:41:59]
Control over a lot of things in your life so any little bit that you can give them a chance to help this is another one and I think this is my last one this is an app that we made for the Marcus Autism Center and started well as I had Autism Center project here but it's a it's an app we built for them a guy there works on something called the feeding program so kids who do not eat it's not that they have eating disorders it's that they have feeding disorders they're not going to take food by mouth for a variety of reasons usually it relates to some sort of bad health care experience when they're an infant and there's a program that somebody at Marcus figured out how to do this is a 6 week program it's very time intensive multiple visits per week to the specialist here in Atlanta and this is for kids all over the country and it's around $60000.00 a child Well what we did is we took this protocol out of its head and off of the stack of papers that literally had suits wired all over it and we.

[00:42:56]
Put it in show now so very obvious right like that super easy it makes a huge difference that you know analyze the data right away but the other thing that it does is it allows other clinicians to use a protocol to help get these kids off the wait list faster he thinks about 90 percent of kids can use this app to start eating again and then he can really focus on the most severe 10 percent that really need his expertise but then the other thing that happened through this is he realized that he could give this to parents after he diagnose them and explain to them what they need to do and they could take this home and now they can do this at home so they don't have to travel to Atlanta anymore parents not taking off or kids not getting out of school.

[00:43:38]
Overall I and plus about half these kids have autism and so taking them out of their routine is not necessarily fabulous for these kids so. So we really changed the way that he thought about getting care he had never considered this possible and we made it possible really just by creating a digital version of what he was already doing.

[00:43:58]
So symptom management is one more is an app for kids with color rectal issues and I know this one isn't this one is for products and just measuring the color off of the color 2nd a year and it so they can take better records of what their protein levels look like in the urine We have no other one for color rectal management just to help kids kind of if they take better care of themselves and they're more aware of what's going on and they're to have less accidents so this is classy classy essential on infection terrible for hospitals it's what hospitals kind of get sued for all the time when they happen infections happen in the hospital because the way you connect all the i.v. lines and they you know those will tell at sea my clean your hands with.

[00:44:43]
You know that or skin a little square package that's what they currently use at the hospital on joints of these things that actually have it's like a cup with something sticking out of it and that's what this other tube fits into we can imagine that the towel doesn't reach down in there were actually all of the problem is and so we have a guy in material science on campus and he came up with a specially a sugar cube of material that soaked in alcohol and you can see that it's just ridiculous how much better it is than even what's the standard in health care right now last one mission sorry so this is the infiltration also a bad problem for the hospital lots of multimillion dollar lawsuits if this happens and it gets bad because they wind up.

[00:45:30]
If you're having Nike and you're getting a drug and let's say that they break and so that liquid or it's missed completely the liquid goes into your skin area you can see the horrible things that it causes it can completely eat away at all the muscle the nerves everything and causes children permanent damage it happens in adults too so we have a group in mechanical engineering who's making a system that's going to go in the dressing around the the site to using flexible at Tronics to detect when this happens they asked they said if we could do this within 30 minutes it would be wonderful and we think going to do it about 2 so we're super excited about that one and that's it so I don't know if anyone has any questions that feel free to ask them and also reach out any time I'm happy to talk about what we have going on or ideas you have or whatever it might be thank you.

[00:46:46]
Yeah yes I mean again another one of the things about us is we do kind of we do work on real world problems so we like to have real world since that can actually be who I think that that is a little bit differently everybody that we work with really appreciates that people been working on things that now they have a real problem that they kept so for a very different talk now.

[00:47:22]
I want to welcome my colleague Dr Yasmin May and he's going to talk about bios because that's a bit you know and I didn't you come here I'm to yeah the priest is talking for your privacy on the amount not talking so I look here I want to emphasize that it's not someone who you know sort of is sort of part opportunity at it you'll get 3 weekly comes out in sessions if you're so kind of a lack of today doing your research your son and.

[00:48:33]
So. Originally And so for those students or Paul stalks and will you have any studies of questions and it will come. And my sessions and also for fact useful. Sort of like church so you know Pastor why don't we have it down the. Passages comes out of sessions or you know general support would it be kind of also some questions around from design experiments or some bus sized calculations or power analyses and longitude and.

[00:49:09]
Syria. Back in there more than a minute a machine learning mansard and so on and you know pass it on we are still walking around to do it and the kind of lack of paid posts and and the some kind of look at a review comma and sign that we see how to address it or some.

[00:49:29]
Kind of lack of their ground approach they need us some support. From the user support but we will be are still very happy to provide it and the civil discourse and sometimes so I can give you one concrete example. There's a one plus. Haven't yet had an Emory Yemi and get a grant proposal and they get the kind that will ask the revisions and the review economists say Ok how do you justify They they somebody size and and you know proposal and.

[00:50:06]
Research Yes they always do that's the whole much money I have a to but you know talk Also you cannot have adequate so we recommend recommended them so you need to back that call analyses and a sample size of cargo regimes so that's a general kind of law firm or soft support and we will we will provide.

[00:50:29]
So. You said he is a supported by the so called today that you speak if you need all of that research to design and we quoted the core of that you're saying I'm not sure. If you mean it was C.T.'s they are not just the t.s.a. use a condom.

[00:50:54]
Is a big night to go round from. Many And from. Emory and make sure you have and you. And more house and a for part of my job you need to provide upper hand if not always use of the co-director for the forward and I to do it yet and.

[00:51:17]
Walk around and they say have a program manager. And try to feel you and then. Maybe next year Brownie will come back and maybe. Support and support and so forth program and have a kind of like the 2 kind of leaders and so we are no fun at your tax and on one side.

[00:51:46]
Of caution from. You said if you read. Here say I did I did you had prevented. From Emory is a bit direct or direct. Deposited at Emory about that he states he is a b. And so for the also of the research just new lease rule and he specially in your research and particular graduate students you have any set of.

[00:52:18]
Questions and feel free to let us know we will be. With you. And I also want to briefly mention my size Elise kind of comes out in. And the surveys that we've provided all I have provided you know a lot of separate and ours a want in my own personal research and focus on the high level on a particular one project and in general my research or what walk under the sun.

[00:52:49]
Related analyses and or quick student actions and for that methodology research yesterday and the s. and s. after Korea war in that the directions and and also the and are still walking around some lack of Ballston and. Careless research and by using my experiments. In cancer research. Seattle and over there I learned more like the infection these east and dynamics and recently ours a walk around.

[00:53:27]
Which is that we will talk about it next and we also work on some in your applications and they specially the. Quality control and the since and then I mentioned theories and I try to keep a massive kind of like a simple and probably should have been about my last slice and so I just a walk around a one submissiveness that you can have under to you Kate what kind of research we have and both collaborative research and most of all that you raise and and at least one of the you talk of out of them are you probably.

[00:54:07]
Know Rhea Connery's the 2nd the most commonly reported the disease in the United States and the increased. Rate over the passages and in particular you can see at least. Some. So you can see the they are not even a group of. Rule populations and this is a ratings most survey able there are so I'm not a group.

[00:54:46]
Or some other group so you can you can see that you see the rate of increase dramatically and the so so d.c. city scene. How did I find the recommended a treatment option and 2006 and I don't now these days I have only one treatment the left and it due to the market dropping resistance and the United States actually has a better shape of that European ice in the in the Germany so now there's a case he says there are no treatment of a loved one and so due to their crockery resists and so soon the c.d.c. is kind of like that they.

[00:55:25]
There are so you have a date hard to c.d.c. and in particular they have this national no to d.c. surveillance system and since 44 hours of. Colleagues at the c.d.c. they have a say and they have. Had a c.d.c. Hardwell they were not to be able to get to the individual personalized they do it at their privacy so a lot of our research and we'll talk with the high level we the United States get and the data from c.d.c. Web site is very kind of.

[00:56:07]
Interesting show you how to get it and also you d.v.d.-a day how we try to have a collaborator. Professor Dr Khurana and from there for 800 the kinds of resources and kind of a large medical doctors and a provide a software. And I just let you know a small part of so if you're in this part yet.

[00:56:31]
Once you know we want to talk with them there. Are certain guidelines to say probably a whole would have you come Dr the driving reason attest to this 1st a trendy 5 men at each at each of the use of clinics and then using that data to make a young prince of out of a lot of populations and the sometimes.

[00:56:58]
There are people who you know room will immediately see that there is somebody in bottles just because of them employ relations a higher risk you cannot adjust the sample in or you. That will demand populations so kind of low but I also talked with them why are you doing this why any kind of particular reason this.

[00:57:19]
Discovery of them Arsenal particular reasons they just simple is a day policy it was a use established that you were in a 2nd the water war and people just using it for ever and for their for their for their laziness of maybe for a convenience and solace kind of part of the act.

[00:57:38]
And. Sat for a project ongoing projects and why is that. Fair to the students and to walk to look at the feet when they. Look at the funding they're trying to Lisa's Servatius temporal chorus of a 50. United States a different state using their hands. Out of their methodology table and table and we are so right with the students on a walk on the banded of problems and they say a basic Lisa somebody how do you see Charlie and the sample in so that you can test a group whole populations these are rated well you're creating or there is this integrated Well you know.

[00:58:24]
So I was. Just letting you know so we have a go back there so we have to be able to do in there so that is it comes audio all that most of all your development and hopefully you will use it a resource that you'll get. If you.

[00:58:58]
Are just you have anybody. We want you for the other ones that we want put it. Out there all right well now last but not least is Dr who is also a colleague. Who writes Good afternoon everyone my name is. Now I am so shit Professor add to school 5 spotty here at Georgia Tech so my main research area is in statistics and sure any answer today I would like to share with you one particular project I've done in this domain and this is a collaboration ways.

[01:00:10]
Are your students my students she who was so close and the engineer and our colleague Joe so clearly n.p.r. and so. So in this project we want to study this is more a focus on less knowledge in developing decision offices to those who are actually core interests. Ok so here's the description of the problem so the motivation is to look at literature honest as we're in transplant and so a lot of the small existing feature is actually built based on adults data.

[01:00:48]
And so many people look at it care to 6 of adults pediatric patients organ transplant is very different so fortunately for a look at it so well we raised this in playing pediatric patients has much longer so vial much better so well to organ transplant and you can look further you actually find out actually the decision variables which decide what is the best much were organ transplant and recipients it's different it's going to be different so that makes a lot of sense because kids to Children's are different in size in their body has to be fully developed so the decision variables should be different.

[01:01:23]
So the goal in this project is you want to look at the data and develop new survival analysis monos to help doctors to match and receive. Donors to help these kids or the transplants and why do they need to just decide the small is very important because you are the tools that doctors why do you decide you have to decide quickly and typically in Oregon once the Speakon swimmable it's only available for probably 48 hours of what he's learned from by the way I should mention this is a collaboration with c.d.c..

[01:01:58]
Brian telling us you know typically it's only available 48 hours so Dr given the minimal time you have to make it quick decision. These 2 is combined with their knowledge so having a precise so well prediction out and seeing based on this characteristic this matching of the c.p.s. and donor How likely is it going to be successful how long it's going to be successful so it's very important how they're precise model and that aspect is typically it's impossible to find a perfect match between the or getting and because organ is a very scarce scares you know resource so if there's no perfect match then and maybe you can tolerate a little bit of mismatch in some characteristic that's not super important but this can provide more in transplant much sooner for the kids and that's also very important so the goal of their search is to build a model and also identify which factors are most important in making its.

[01:02:57]
Ok so here is you know summarizing what I just said what is our goal is we want to develop the so so while will not let this model based on data to predict today how successful haul the case is going to survive after organ transplant so if I give you the characteristic of this patient and all the pressures of the donor and by matching to cash mistakes how likely is it going to be successful how many years he's going to supply it after this this or precision and identify the most important factors not only you have mana but in terms of all that Dr King uses knowledge to make their own judgment as well and so this is our contribution so as you can see is you know focus you want to take us but I think it's probably it's like being you know I think you coincides with the earlier presentation.

[01:03:46]
In pretty nicely about p.t.s.d. Chris if this is one particular aspect of pediatric research so that the small so this new methods a new model to how they can decisions and here is a. Summary how to be approaching this problem. So the data source we have is from the you know United Network for Organ Sharing and so it's very comprehensive dataset contains data actually ranging from 9087 to 2 salt in 14 and if you look into the data there are 19000 cases of different cases and you particularly will look at so they keep me translating this data set actually contains besides key but also other organ sharing here it's you know it's 1st is that we look at one particular organ transplant.

[01:04:35]
And so so the way we did this is we built the survival analysis model and we used to approach one he is a classical sit to school approach based on the so-called coax model coax proportional hazards model and the other is more of a machine learning modesty's machine learning is very popular now and he believes this is going to be some times more robust in capturing all the area and you know data so we tried both approach and comparison.

[01:05:04]
And so in terms of performance matching we consider to commonly adopted mention the c. in deaths and writers who are in comparing this method and so wanted us moto compare between school and mature when you approach and also compared East new methods with. Earlier development that message based on it only and seemed didn't really have improved in my view needs model and going yes is indeed it's worse to develop these models and we also select the variables.

[01:05:35]
Identify which factor is most important so you see it's very interesting that the factors forge citing for for pediatric patients is going to be different from the adult and so the data is going to be not homogeneous because so how do you decide it's a con also kids how young is a kid is it going to be 2 year old 3 year old maybe 7 years or so how do we decide so we decided to use change when detection which is one of my research area to decide the cutoff and so the kind of means that ablation build a dedicated model for disco tickle age group and maybe the cutoff also besides age are some other there so how do they decide.

[01:06:17]
Ok song is going to give you 2 slices peak into the mess analogy here and so the 1st one is to face graphs based variable selection so this is basically coupled their selection meaning selecting the most important factors together ways the survival analysis and so this is our militia incident and this part actually has to have the some such as Comesa biology no way sure why didn't consider us regularize 0 based the method is because we have both 14800 some features and these features includes the genetic features and the medical tests and gender and a lot of formation and these variables are typically correlated this graph structure will help us to incorporate this car Lieschen structure between the variables and also help us incorporate current knowledge medical knowledge into making the decision and choosing based factors important Ok so audition is just one is out we have on variable selection and so this is one of the plugs and this is comparing basically the variable so action coming out of it suggests a whole machine learning approaches compared with the base the only model so these are the variables selected from those models so you know one a couple of takeaways from this is actually identifying d. they are different factors for example we find out for pediatric patients.

[01:07:43]
I think one thing should be the donor age is very important so I guess it makes sense because you have to consider the difference in age in the case where says who is you know I don't know if you know this organ to the kids and the others important yes the weight of the patient so I think it also makes sense because you know the bottom stage of the case depending on how have the rows they are.

[01:08:07]
The actual the actual age they find porn in as well and this is going to be the grouping part so for example we find out. How do we do. How to cut out to build different models so they decide to build 3 models in the end and that kind of age are going to be recipients ages $1.00 and $1.00 so so this basically means that.

[01:08:30]
Kids the low year 1.5 year old is going to be one model and then between $1.00 actually tonight euro is going to be anonymous so also makes sense because for babies and for slightly older kids and you should probably have different models and so this is the final flight I have is the comparison of this approach for the existing approach and indeed there's you message which is tailored to data and selecting variables according to pediatric characteristics much better prediction as to why a race compared to exist so this is your research stage so far and we're hoping to actually making this model will go to a practitioner and doctors and you know patients when in the decisions about this and we hope this can help case you know patients and be more active making their disease thank you so much for attention.

[01:10:01]
So yes yes so that you know that's the bottom line is because the bottom line is why do we need. The characteristic data and distribution of the data is different for very young kids versus older kids versus adopts the one what is going to be the precise decision nominee is it going to be one year old 2 year old so how do you decide so we used to decide these cut offs.