Events Add an event Speakers Talks Collections
 
Duration 20:14
16+
Play
Video

Productionising Machine Learning and Shiny in Healthcare Settings: a Case Study (Chris Beeley)

Chris Beeley
Senior Analyst at Nottinghamshire Healthcare NHS Foundation Trust
  • Video
  • Table of contents
  • Video
R/Medicine 2020
August 29, 2020, Online, USA
R/Medicine 2020
Request Q&A
R/Medicine 2020
From the conference
R/Medicine 2020
Request Q&A
Video
Productionising Machine Learning and Shiny in Healthcare Settings: a Case Study (Chris Beeley)
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Add to favorites
416
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speaker

Chris Beeley
Senior Analyst at Nottinghamshire Healthcare NHS Foundation Trust

I have been working with data in applied healthcare settings in Nottinghamshire since 2002. I am interested in statistical and quantitative methods particularly those that are germane for evaluation and research in healthcare settings such as mixed effects and other regression models, and psychometrics, dabbling in other things such as complex survey techniques and cluster analysis as appropriate. My other great interest is statistical and interactive graphics and data and programming techniques to improve the reporting of routinely collected data within the NHS. I led on the data methodologies for Nottinghamshire Healthcare's patient feedback website (http://feedback.nottinghamshirehealthcare.nhs.uk/) and I am a keen user of literate programming, R, LaTeX, HTML, Shiny interactive graphics (http://www.rstudio.com/shiny/) and anything else I can get my hands on that allows clinicians, managers, and the public to better understand, access, and explore routinely collected data.

View the profile

About the talk

This video is part of the R/Medicine 2020 Virtual Conference.

Share

Welcome to this next set of sessions Krispy, Lee is going to be talking about machine learning and shiny and how this works in production or in predicting in at the clinic. Okay, thanks. So I just signed my timer. Yes, so I'm Chris Bailey, I work in the NHS, in the, the UK has talkin to you about using machine learning with shiny, in in, in production processing. To just give a little bit of background. I think a lot of people to hood of the NHS but maybe people that always has

Solange is single-payer healthcare system in the world and I work in one of those little things. So as a consequence, I think we don't wear that we don't listen to super slick. Operation the people. Sometimes it seemed that way off. So I will get to the NHS using off about 10 years. I've been using all the shiny for 7. I started using it with patient experienced. I took with your smiles. Love and life. About and Johnny are going to be using all and shiny with Clinical Director for about 3 days. So we have fragmented day seems we

have three different teams. Welcome it takes up so I'm in the Des Moines team. Did ex want to see him on the whole entire and shiny. We use a pair of pants and having that was starting to see that your engineering team for a spreadsheet and we've analyzed using all the things. You didn't take, the meantime are using Microsoft sequel server. And so the overall effect, this is that everything is all the fragment it so that the dangerous is fragmented. So we have insights and variables and information that they changed

everything. I may have information that we don't have on for the users, what? Also this fragmentation. So we have multiple dashboards running on multiple Platforms, in multiple locations. Instead of the the result for I use has basically Extra work and confusion having up some things in different places. So, I've been trying to. I am, I want to talk quite recently from, I'll see that Comforts Arena, delivery, via the link on the side of you say that I want a thing, she says not talk cuz he talked about that

being one. DeSantis 410 engine is, I don't think that's probably roughly 2 in my contacts. I haven't done the exact and what he's talking about. Talk is hargeisa, science teacher smooth, so we can meet up, we can be interested, we can test things, we can hitch a ride and then we need to move on and do something else which is really good. But the problem that I have faced is how do we share the work that we, how do we share the way we don't, we don't all you since I want to just talk about you and you really. So if I don't started about three years ago, when we started using the That point was we

run on kissing. Someone say this in the chance that she asked me did do that exact thing. So we hooked into the house and around select oil from like about 7. And then we use Lowe's about her five thousand lines, literally five thousand lines to sort and filter, and the computers useful inside. And then try to communicate that back to their children in the folder 5000 and the problem with that, then it's a waste of these supposed to have a son. That's what I want to do

this shit. Now, I would say with exclusive rights, slap bang in the middle of phase, two of them, and what I'm talking about today. And so, with the green shoots of face free face to face wondering if it was still price. Very large dice at 8 which is not a complication, the sensible thing to do told, but once it has been trying to move some of the dates processing back to single, so we might do some compensation of some exploration or somebody is what we're trying to do is feed them back and push them back to back to back to tables that you miss

the days when I asked her to resend. The first one is because he's a lot easier in a big sequel environment than it is on the small bass licks. It was a small series of smaller but more reason, why I thought so important is because we want to have one truth or at least once at the trees. So I can talk to say the fragmentation so we have different definitions of things and you know it's it's quite confusing when looking at things from different ways, not so helpful. So they surprised that I want to talk. Today, we just we pretty much we picked it up

with this project, right on this project is deliberately simple because I'm kind of working on the coach are, right. And I didn't want to work on the couch too complicated right now, because once the clinic, who's going to come in, who is the area of the seven? Five won't be telling people by sneeze. If you give me a thousand appointments, I can say you should bring these 50 people. If you bring those sixty people 20, those people would too, she hadn't broken and if you can buy that was just guessing, right? It would be 7/2, people would

come so it's no. I think you could do with more information in different. Text me where it's a very large organization. That was weird seeing lots of different types of patience. And I think, if you choose them over within those particular within each of those contracts, I think you're funny. We were best buddies. Why would say good enough? This is so I know the flock that I've been espousing bracelet is this? I did it again because we were small team. You know what up Google? We caught me. So just making fractional improvements we want to just make a change, makes me useful building out

and you know it. Yes, sir, Michelle we can do lots of things and all, but if nobody ever sees them and nobody can understand them if we start doing something else and then don't say what we've been trying to do is we're trying to build simple stuff. I'm going to build it and communicate in a carbon frame work week on how to say there are many many clinical environment for patients. We can hope to do all the trading information ourselves and we don't need to. Could be done quite easily bought a bunch

of pick up more technically, competent for the test against Russian could be done by us some of the early. What can be interested rapidly along with the shiny? Can I see who? I don't know. It says I'm a quote by to pay United friend. Went fishing applications emphasize. If I see a particular way of filling out applications, I think it's very useful. I don't be using it for that long but I started to build more and more stuff like that. So emphasizes particular ways of doing things of emphasizes use emojis. But it also emphasizes package like functionality, like testing and the

idea of it allows you to be a fairly agnostic at the way you going to deport. So I think the reason why we stressed about it was because they wanted to have a code base of a shiny coat, basically lots of different ways without rewriting some of the underlying structures. So what I asked you to connect to a crime package or whatever it is I'm surprised me what were what we're trying to do which is to minimize the friction of these communication flow. This is where we're running the dates of life in the day to my house instead of saving sing these girls need to understand, can

you please do a big joint and you know get a hold for us which is more work for them. We can just go and get it all ourselves, and we could use Google in China eat to build something emotional function. So we enforce the separation basically, so they're supposed to date a liar to the application today to buy Max uses the couch, which is useful for managing legal connections. Michigan, I don't have time to talk about it. So useful thing you want to do that or the application

to find outside of her contacts. So you couldn't contain itself, it should be defined and stuff to values. Sit in the static in garments and then take his as Imports arguments and the effects of doing all of those different things is to produce that could be reasonable and communication in pieces. That's the idea of a Different Strokes for different things to avoid. This is the opposite of the approach, while you run everything under the sun, and it's very hard to breathe.

Everything in my department, it's early days for us, but we have, I been so stuff. Can we plan to him? Cuz I was more tired because I think it would be more confusing than anything else. Funny on Twitter @chris baby. The simplest way to find a mate. I promise I'll get back to you when there is when the risk of chocolates. Absolutely. Our intention. I'm right, so just us a little bit more about those individual stages. So the model training, as I mentioned to Stone by some stray off the path, which is very useful because it allows you to use that, use the power of the sequel service

instead of doing the filtering, the sourcing of the arranging and all kind of stuff instead of bringing everything into a little older and then you bring it up to you what you need and Knotts Berry, that's very liberating, the office on the street, which is about a mile off. Trey itself is also built in a Sousa, March of the way and precise has to use the pipelines. And again, the idea is that they can be so you can have a particular invitation statacorp to get algorithm but you can then swapping out with a different one. Another the rest of

the chamber will work out stuff, that's profit than the feeling sucks. So what we've built in the shiny side facing this or two applications. I'm talking about, they may be more, but I think I need some more trouble. Still, can I say she's first to my first idea? Was the idea was that people be kept coming in the morning, turn on the computer and I can meet you at Hazel. The people are coming in today and he is. So I'm going to rain in such person cuz I can say that maybe we

can make a decision soon. Percentage is it doesn't say who you should. I do have a Coke. So I can see the other view, interesting would be for people who are interested in a hole, Clara people looking at the focus of it so they might be interested in the April, right? If do not attend, they might be interested in the percentage of people that we've called the plenty of those that we told him still didn't come with any people that we didn't call didn't come out. Things just to get an idea about how it's going to be single. So

weather is changing. You know what, how well it was working, whether it's making our practice faster in a walk tonight. Tortoise. I said go and it's really useful in this in this respect because it allows for the day till I was totally separate from the rest of the application. So that's a separate building for the day to end, its suffering, multiple for each of the views, and that's very liberating on the shiny side, because it means that you couldn't, you can swap them out and reuse it. So it may be that you write something over here that could be used somewhere else or sick, or vice

versa. And again, thinking about the business aspect of things does it say? This is not really gotten, this is good practice and I think I'm just helped you working this way. So did you define all of the Apple server example, the table of the clinic view which shows all the patients and that would be to find the way they work. Sweet. I'm outside of the of the context, and that allows you to to get good testing, and it was very nice about it. It's not as a good practice for shiny, but it means that those components of reusable, so they were usable by

yourself. If you make it down there, or you want to make it 8 p.m. Whatever is it's not fine. It's it's static. It's it's rays. Will have the box. Have you got a lot? What do we do with it? Also the easiest thing to do, as I mentioned before, he got his fairy, the oldest with over say, very naturally goes into, the are not, something will be able to do anything without all this stuff, even if we've done that quite. So I can't because all the way we could celebrate or posting our garment comes now, more when we talkin to the rest of the

organisation said, the first thing we can do is we can take the bottle which we've got to say, we've jest the right test so we can decide what we trying to hold this tight to the old adage was pulled from using crazy, that's going to get inside. This is how we did it. You can do it. I'm assuming again outside because the Shawnee application is itself modular, it's very easy them for society, example of which come from the de celaya, which is itself too much alike. And as of right to find such about what's with your this allows us

to take my individual piece, that's why I want to do. I want to go to an individual power to buy and Beyond develop, take the individual case and say this is how this works. This T's the inputs and outputs and you can just do it. And then simply with the performance of you, that might go to a different different team on. Again, you can just pull it off the shelf inside stop. This is the opposite of the 2009 China because that you've got to take him to soccer games out what I should. You should know that for something else. Does that supposed to mean? That's the

kind of things that we want to do, was trying to avoid to make it easier to talk about. So how was this helps us to what? Becca choose the whole forward to being trying to do stuff like that. So we can just still kind of holding them, we're still kind of in charge of them, says no idea. But that's something that I can stop to understand that I mentioned, then they can then can bring them by the bottle in the dates when I asked to sign the bottom and just talk to you

pretty much from that. Run sequel, tables. And then they can, then ran from the shine application. So we can say to these people that these comments about if we make these changes, and here is his on his, this bad news. Turn summary. What I've come to talk tell you about today just to sum up. Basically, it's a, I watch the torch was ran into Winchell, we are small team, are we ready to punch of a white? And we need in order to do that, we need to be able to build new stuff. I need to test it, make sure it works. And

then we need to just get out and just move on with you to hand it over to people, pretty good at your scaling things in a big architectures, not kind of thing that we don't have those skills, and we don't have that, my power left for that myself. I didn't know what to do that. We need to be used Frameworks that help us to help, help us to communicate with the text came in second place overall benefit. We need to use framework that helps to reason about all the commercials on, for those who work in China, who maybe haven't tried modules over that you have. That's the first thing you

notice is, when you start using this as a psycho, I'm just very encouraged you to be free, you up and started to help you. Think about your work and I end up in a month. So I know I can be very liberating when you don't have to worry about how things are set up and having all these unpleasant interactions with application and and that's what about sex right now. Well, thank you very much, Chris. So we have a question here. Is recent attendance in, emergency a predictor of no-show for follow-up

appointment. Oh, that's a very excellent question but I'm afraid I'm not the person to answer that question because I don't work and not kind of environment. So I will. Can we deliver a mental health services on way to live Community Health Services? So I can tell me off the top, my head, the thing. That's the thing that has predicted that you'd expect. That right there is three of these things are predicted to. The first one is just how many times we see you on the second one is how many times you haven't come before, which is probably something at all. So I guess but also something

else that we were expecting to find which wasn't bad but it's worth noting. This app is also deprivation. So we find that the people in the place where I went, just have quite lost it deprives individuals within this area. And we find those people act like it's an issue because it means so much to me that those who most need of my services at least wants you to come and see them. Can you briefly say anything about the conversion to power? Bi Do you have any code available on them

going just yet? Either side plays out, I wanted to bring more already but it's it's sort of a long time together with string and I'm so pleased. Alright, well thank you for that talk.

Cackle comments for the website

Buy this talk

Access to the talk “Productionising Machine Learning and Shiny in Healthcare Settings: a Case Study (Chris Beeley)”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free

Ticket

Get access to all videos “R/Medicine 2020”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “Medicine, Health and MedTech”?

You might be interested in videos from this event

August 18 - 20, 2020
Online
6
40
bud, compliance, covid-19, hospital pharmacies, pharmaceutical compounding, preparation, science, stability testing

Similar talks

Cass Wilkinson Saldaña
Data Instructional Specialist II at Children's Hospital of Philadelphia
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Brennan Beal
Quantitative Scientist at Flatiron Health
+ 1 speaker
Beth Devine
Professor at University of Washington
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Vincent Major
Assistant Professor at NYU Grossman School of Medicine
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free

Buy this video

Video

Access to the talk “Productionising Machine Learning and Shiny in Healthcare Settings: a Case Study (Chris Beeley)”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
816 conferences
32658 speakers
12329 hours of content