A Healthcare AI Suite Built for More Time to Care

Hi to everyone, glad to be here. Today we are not presenting just a product, we're not making a demo, we are presenting the reasoning behind a suite of four products that are, let's say, complementary to each other.

Alessia was saying before, health care is impacting a lot of society, so the butterfly effect could be of course very huge, just deep dive more in there, yeah the whole presentation

is more time to care because we think that physicians need more time to care, like care in actual way, think about health care of people.

An Ontology-Driven Approach to Healthcare AI

So on this path I'm an analyst in analyst research so I'm researching also new ways to impact in the health care journey and in AI innovation manager in P4B in which we are developing new ways to match AI and society needs in terms of health care matching.

Let's say just the point of difference without saying everything that is very obvious and clear for an AI understanding person. Our point of difference is the ontology. We are ontology driven. It means that the reasoning behind the AI is very focused in this case on the healthcare.

We have other sectors so of course in other sectors it is focused on its own dominion but in this case each artificial intelligence we are training or building it's closely developed and thought to be an expert dominion on that matter so this is actually what we think it's our difference in terms of building AI applicatives so here we have it and of course without our target hospital

Addressing Emergency Department Needs

local authorities regions ministries we are working order you know to scale the healthcare industries and also you know average costs and profits of of the structures let's say that we made a very nice market sound negativity so in this market sound negativity we actually got the main needs.

The main needs, they are in different fields but today we're bringing something about ED, emergency direction, in Italian it is called pronto soccorso, it's that specific division in which, you know, a lot of conflicts, a lot of urgency, a lot of in a hurry.

It's filled with this problem. It's normal, let's say, for our healthcare sector, for our public healthcare, but it could be addressed better with the AI, with the capacity of the AI.

We are showing four ways to do that. There are four modules of the same product that it's a product very dedicated and focused on the ED directory.

Four AI Modules for the Emergency Department

The first one is for Scribe, just take a step behind, I didn't say that our ontology and our company is called Primordi, the ontology called Primordi it is related with the primordial human being that it was in our vision in this case before the ai so each title of each module of the product that you are going to

see it is a synonymous with primo with the first so in this case we have the synonymous in english first then you will see synonymous in italian that is translated first in italian it's already primo

PrimoFirst: Explainable Prescription Support

first right it's an ai assistant the check prescription let's say that prescription in In the ED, they are made very fast, very quickly, because you need to do that. Not because physicians don't care to focus on that, but because they need to be in a hurry. They need to be fast because they are in a hurry in an ID situation. So in this case, AI is just doing a very simple but good work.

it is looking for sources because as we said before we are thinking about AI as a dominio expert so he knows perfectly what he's doing because we gave it all the information to do

a job at the best then you have another concept in here in this dashboard that is explainable AI AI should explain you how she got, or he got, or it got to that result. You don't want to trust AI just because you have to trust it. You want to trust AI because you are okay with its reasoning. You think she taught in a good way.

So let's say that just the physician is checking all the items from the person that is going to the ED directory and then is prescribing something, an exam, it could be a treatment and in this case the eye is crossing all the data it has and saying yeah I think this prescription is okay or maybe I think this is not". You could try also to check this other way to treat this problem you're trying to solve.

And here you have three sources linked in which you can check if the eye is right or not.

PrimoCare: Keeping Families Informed

The other one is PrimoCare. You know that

in the ED when you bring someone to the ED on me for example if I'm sick and I have to go to ED, when I say ED I say every time prontos of course for for Italian speakers in that case then you want to know how the person is the ED she's or he is so you cannot bother anytime the physicians people in the hospital they are there to care of people to take care of the health of the people not to inform us about how people is doing so in this case the shift is

just let physician being physician and you can talk with the chatbot that is connected to the hospital database is gathering from the diary clinic information that can develop deliver to the person who is asking we are removing the person in this case the physician from an activity from a task that he doesn't want to do that and we are just providing a 24 7 service very accurate because the only

PrimoTriage: Supporting Clinical Prioritization

source of those information they are coming from the hospital database another one is primo triage in which we are doing more or less the same thing of the slide before but on the triage do you know

what is triage yeah okay so let's say that sometimes it is very normal we are not charging physician in any time in this case we are trying to help them it's not easy to do any time to sign a code to someone in few minutes just on some few cues so in this case we are trying to help that person with some

information let's say that you are charging what the result of the first analysis you are just charging what a couple of things that the person who sick told you so you are putting all the information you have and this proposing your code you of course as before you can check if do you think the reason

is right because of course what one our mindset is human in the loop we are saying that the AI has to calculate but not to decide you are the one who decide the physician is the one which evaluates the reasoning that made very fast the AI because this is the one point that the AI is better than the physician and then is deciding plus one is everything else the workflow so we

PrimoFlow: Improving Shift Handover and Operations

are presenting also primo flow because it's about operations do you know that for example in the ED when one when there is a shift when a physician let's say at 8 p .m. is leaving the hospital and it's changing with another one they They have to pass them information with, yes, yes, a brief, a phone call, WhatsApp, you leave a note, but sometimes you just forgot something because you, of course, you want to leave or you are exhausted or whatever.

Maybe the one that is taking on the shift has not the time because maybe he's coming to the ED with an urgency. So in this case, the database is linked to the assistant and you can just ask the assistant. And you can understand, for example, who has been checked, who has not.

Or, for example, you can apply or not some suggestion. Because we have data analysis, you know, about which is the workflow, which are the peak times and all of that.

A Unified Goal: More Time to Care

I just understand that for products in 12 minutes it's complicated this is because sorry I just wait this is why we have this right that is just trying to sum up everything prescribed for prescription very easy fast going and quick prescription checking on your prescription suggestion you can decide

PrimoPair chatbot in order to ask how my cousin, how my uncle is doing.

Very easy, you remove all the tasks not clinic that physician nowadays they are doing and you leave them just what it matters to them and they are the best to do that. Care for ourselves.

PrimoTriage in which you check if the code assigned is right or not in a very quick and fast way. You don't get wrong, you don't get other complication because you maybe got wrong and that's it.

And then Firm of Law in which operations like night shift or other kind of issue they are addressed just better with AI.

Everything for a concept saying there is more time to care. That's it. Thanks.