Hero AI Seeks to Transform Mindsets in the Lagging Healthcare Sector

Episode 30 Hero AI Seeks to Transform Mindsets in the Lagging Healthcare Sector

The resistance to change is emblematic in a sector that has idealized practices and the legacy systems that have been working for decades. One founder seeks to disrupt this.

AI adoption continues to lag in the healthcare sector. The complexity and sensitivity of patient data, how they’re governed and communicated within systems and between institutions requires significant time and investment. The last decade saw health information data slowly becoming digitized however strict privacy regulations, security requirements, and liability concerns–have all made AI challenging to fully integrate into the industry.

The integration of artificial intelligence (AI) in healthcare holds immense potential for transforming patient care and outcomes. Navigating the ethical landscape of healthcare data usage is a crucial consideration.

I met with Devin Singh, the founder and CEO of Hero AI, an early-stage health tech company, to delve into the challenges surrounding healthcare data usage, the need for advocacy as he pushes for significant change within a highly regulated sector.

Transcript

Hessie Jones

Hi everyone. I’m Hessie Jones and welcome back to Tech Uncensored. We are here at collision 2023 and here I am speaking today to Devin Singh who is the founder and CEO of Hero AI. I welcome Devin.

 

Devin Singh

That’s right. Oh, it’s a pleasure to be here. And what an incredible conference this is. Being able to meet folks like you and to talk about all things tech, AI, ethics, responsibility, super interested in this conversation we are going to have.

 

Hessie Jones

OK, awesome. So, let’s talk about, first of all, hero AI tell us the barebones. What is it about? What are you attempting to solve?

 

Devin Singh

Well, you know what I’m going to do is I’ll tell you a story that is actually the catalyst point that caused us to create Hero AI and it’s a bit of a Side Story, to be honest. So, I’m also an emergency doctor at SickKids Hospital. Many years ago, when I was a junior doctor, we had a young patient die, a preventable death, partially because they were waiting too long and it was really difficult, right? I remember doing CPR. The team came together, provided outstanding care. But many aspects of this case, whether it was presenting too late to hospital, prolonged wait times, different resources contributed to this terrible outcome. And I just remember being so angry that this could happen. And it drove me to think through, well, how do we really think in a dramatically different way on solving meaningful problems in healthcare? And so that’s what led me to go from doctor to computer scientists in AI. So, I did a Masters in Computer Science and artificial intelligence, and we started to realize that we needed to find a way to empower hospitals. To do more with less and we needed to find a way to empower hospitals to liberate their electronic health record data so that they could solve problems like this in a really agile way. So that’s what here what it does. So, we essentially, we take real time streaming health record data, we feed them into decision logic or machine learning algorithms and then we output to highly customizable front end user faces which is a web app, a patient app, a provider app and figuring out how do you do that in a way where it literally takes hours to customize and deploy rather than weeks to months, is the innovation that here I sought for.  So, what that means is that we can actually be in a meeting, discover that there’s a problem, like a real clinical problem in a hospital and rapidly spin up solutions and get to testing and deployment at a speed that’s really never been seen before in Healthcare.

Hessie Jones

OK, so you said something interesting there liberate patient data, which is scary, especially for if you talk to the Privacy Commissioner, you talk to anybody. Even in the healthcare space, because as you know, patient data is considered one of the most sensitive information out there. So how do you do something like that within the guise of compliance without putting risk to the information especially? From an app perspective, that probably will have to go through an equivalent of like a Health Canada or an FDA approval.

 

Devin Singh

It’s a really great question because we must remember that at the core of healthcare and providing outstanding care to a patient is the trust that you have between a health system and a patient or it’s that trust that you have between a physician and that person you’re caring for. So that trust is critical. And so one of the things that we’ve done is we’ve done obsessive stakeholder engagement to interview not only just parents and families. It’s like it’s but the kids themselves, right? Like, how do you feel about the way your data might be used? And so there is that point of understanding that the user really well and trying to gather what do they want to see happen to their data and what are their concerns about their data. But then there’s also that point around like legislation, privacy law. You know, we’re working with a set of laws that really wasn’t built, to think about big data and so how do you then take these opinions that are really important from those patients? How do you harmonize that with? What are the boundaries of the law? And then how do you add layers of cybersecurity? How do you add layers of patient safety and be really obsessed about delivering a product that checks all those boxes? That was really difficult. Like it took us years to go through a really authentic engagement with both lawmakers with Health Canada and with our patients and and even like privacy commissioners and experts to figure out how do we harmonize those things together in a way that’s really genuine and authentic. And that’s what. Is there a single clear path right now in Canada to go from creating an AI innovation like what we’ve done to then getting it deployed safely and in a way that respects privacy? There isn’t, because the laws weren’t written for AI specifically, but that’s part of the advocacy that our group does as well. So I have my AI researcher hat at Sickkids as well, and part of the research I do is around regulatory reform, privacy reform. How do we think through creating streamlined pathways that can ensure we can deliver AI solutions to patients, but do it in a way that is obsessed with privacy? Obsessed with equity, right? Obsessed with fairness to make sure that these solutions are actually closing health gaps rather than widening them.

 

Hessie Jones

So, we’ll unpack that a little bit because from your, my perspective when I think of systemic racism, it’s included within the hospital system, there’s only limited amounts of data that we have that can treat patients based on a limited set of circumstances. And so, what you’re advocating for is aggregation of data from almost anywhere within the healthcare system that allows you to surface some insights about a certain disease and to be able to have it be inputted from all from all areas, from, all from, from, genders, from different ethnicities. So, you can understand the impacts and also potential solutions that that actually impact some of those different groups.

 

 

Devin Singh

And so well, what we do is rather than thinking of such big picture because that could be scary, even hearing you describe that I’m a little bit scared about privacy and technology. And I think that patients and the public get scared. About these big sweeping ideas that we’re trying to absorb all the data and we’re trying to build these, these all-knowing AI solutions, that’s not necessarily it at all. So, what we’re really thinking about is what might be a very specific clinical problem that an emergency department might be struggling with. So, let’s say wait times for getting an abdominal ultrasound, right? Let’s say you might have appendicitis. For you to get that surgery, you’re going to arrive at the emergency department. You’re going to get that ultrasound done, and then that’s going to give the diagnosis. To get that surgical intervention you really need, right? So, we want that time to be really short. So we will discover that here’s a problem in healthcare, and then we will figure out what is the specific data that we need around that problem to build an AI model that might then be able to diagnose that you likely have an appendicitis, and then while you’re waiting those four hours, why don’t we just order the ultrasound, right? Why don’t we get the test done while you’re waiting? That’s research that we do. As kids right now, as we’re thinking through, how can we use AI and automated ordering potentially in a really precise way to advance care moving faster. But then you raised some really interesting. We could deploy that, and we’ve got really great perspective trial data that shows that this works really, really well, right. But here’s a really interesting point. How do we know that that model is going to perform equally depending on your language preference or depending on your gender or depending on your race? From my perspective as an emergency clinician, as an AI researcher at SickKids and as an innovator in the industry space. We must stand tall and proud and say to the public, here’s a solution that we’re deploying. Yes, you’re going to get that ultrasound way faster now. But it’s going to be fair regardless of what you look like, regardless of how old you are, what your gender is. And let’s say that there are areas where the performance slightly differs. We have to be transparent about that. So how do you do that? You need that data we can build the models off of the clinical data like the triage notes, the symptoms that people have, and we can build high performing models. But if I don’t know in that data set what your ethnicity is, I could never do the audit on the model to be able to say that. Yes, it truly is equitable. And so that’s a really key point of advocacy here is that if we just say ignore collecting that data, let’s just not even look at it. Let’s assume care is equal for everyone. Well, your initial premise was that care isn’t necessarily equal for everyone. We need that data to demonstrate where those deficiencies might lie for certain groups of people. And we also need that data to then really challenge the model and demonstrate that it is fair, or it isn’t fair, right? That transparency is needed. So that’s the thing that I advocate for deeply is around how do you create a regulatory landscape that promotes doing that, but then you can’t regulate something if there’s no data to then expect the innovator to be able to have access to do it right. And so it’s trying to merge these worlds together. And really this comes from like the patient perspective, we have to do this for our patients.

 

Hessie Jones

It just seems like an uphill battle like I know the way that the Canadian data landscape works. I know about health information custodians, who will die protecting that data. It’s not supposed to be shared. I mean that’s under the law. So how are you dealing with HIPAA Compliance and how are you trying to make them think differently from the perspective of patient outcomes? Are they receptive to rewriting or augmenting laws to be able to do that?

 

Devin Singh

Well, you know, it’s so powerful about healthcare is that when you drill down to that initial story that I told you about. Right. That’s really touching. Like this isn’t about circumventing laws and trying to figure out how do we just get something to market really fast this is. Think about trying to genuinely improve the way we provide care to patients and literally save lives like it’s written on their shirt here, transform, care, save lives. So, when you go into these meetings with privacy experts, regulatory experts, and you really galvanized support around like, this is why we were doing this. Yes, it’s an uphill battle. And it’s going to be an upper bound. I start off with these meetings. This is going to be challenging to work through, but remind ourselves why we’re doing it, right. We’re genuinely trying to improve the lives of Canadians in the healthcare system, right? A system that’s in crisis and overrun, and so that as a starting point, you’d be surprised, is incredibly powerful. Everything we do, it’s the kids, everything we do at hero AI all falls under the banner of being HIPAA compliant. It has to be right or else you’re breaking the law like it has to be. But you’d be surprised how people can help you navigate that complicated space when they’re genuinely motivated about trying to save a kid’s life because that’s really what this is all for, right? I never thought I would become a computer scientist by any means. Nor did I thought I would be at Collision trying to launch a new health tech company. But the real mode of is that case when that case happened, a fire ignited in me, and I said can’t be how healthcare in Canada is right now. It just can’t. And so, I am running up that hill really fast and we’re making the change to see these things happen.

 

Hessie Jones

OK, so what is your goal this year and what kind of hurdles have you run into and or have you kind of exceeded some of those hurdles?

 

Devin Singh

So all of those uphill battles that you were talking about around privacy, exploring like, what are the regulatory boundaries access to the data? Thankfully, we’ve run up that hill and we were now sort of at the top, being able to take a breath because we’ve got really great foundations in place. Like I said, we’ve done that patient stakeholder engagement to really, genuinely understand what their concerns are so that we respect that and now this year is all about deployment and impact. So, we’ve got models, for example, where if you’re a high-risk patient, I’m coming into sick, it’s the emergency department, whether it be a transplant. Or a cardiac issue. Our model will see if you’re waiting too long and flag that and send an alert to an app to make sure. You don’t wait too long, right? Something that someone said to me just earlier today was, you know, wait times are inconvenient for most people, but that’s often all it is. But for some people, wait, wait times is life and death. So, our algorithm is really detecting those people who really can’t wait and elevating them so that they get seen faster, right? We’re deploying that and the impact that we’re going to see from that is just incredible, right. We’ve also got a use case that we’re looking to deploy. It’s in beta testing right now around automating the advocacy for people with mental health issues. So now when a kiddo in a mental health crisis comes to our emergency department very soon. What will happen is, and if they’re in a real crisis, an automated alert will get sent to the psychiatry team. And that’s just the way technology saying, hey, there’s someone who needs your help. They haven’t yet seen the doctor, but why don’t you come down and begin providing that care? Think of how powerful that is with automation. Children with autism come into our emergency department. It can be really challenging to wait in a highly stimulating environment like the ED . We automate an alert to a child life specialist so that they can come and support that family during the process. Learning all through process automation all through that template of streaming health record data, decision logic, Rai algorithm and then output to a front end user interface right that that template allows us to just unlock the creativity and a lot of it you’ll see is like we’re just advocating for patients like using AI to advocate for patients’ need at massive scale, but that’s what’s so rewarding about all this.

 

Hessie Jones

That’s amazing. So I just want to tell you a really. Quick story before we wrap up my son. He was diagnosed with tonsillitis back in 2014. He was a little bit old to actually get the surgery because it would take much longer for him to recover, so we went through three specialists who all did not agree, one said make him get the surgery, the other one said. He’ll get over it. We took a third shot at a doctor at sick kids. And he said, yeah, just give him amoxicillin. It should be fine. So, seven years later, my son still has breathing problems. His tonsils are covering his airwaves. We finally gave him the surgery and it’s not only about the wait time. It’s about trying to make sure that the diagnosis is consistent, but also the remediation has to be consistent because if he had that, you know 7-8 years ago, it would have stopped a lot of the headaches that we. Experience today and his recovery time would been faster and he’d be able to breathe. And that’s the big, big part of it so.

 

Devin Singh

And and I know that those symptoms are so bothersome to kiddos, adults. I mean, imagine you as an adult, let alone a little kiddo. Right, so bothersome. But here’s where I’m hopeful in that exact story. Is that probably for 9 out of the 10 kids. Maybe it would have gotten better, and maybe that’s the experience that these clinicians encounter. But is there something unique about your son in this case where you’re right, maybe that decision should have been to do surgery faster. This is where this concept of precision medicine starts to come into play, right? How can we leverage AI? How can we leverage data and a precision medicine type framework to, you know what, actually, you should probably get the surgery earlier because it seems like according to this prediction, you might not respond in the way that everyone else typically does, right? That’s the promise of AI and healthcare right now, it’s definitely a major focus for the hospital. People at the kids and for a lot of these big research institutes around the world is pivoting from this one-size-fits-all type of medicine to a precision medicine framework is going to be incredibly revolutionary. And I’m so proud that Hero AI is trying to, like, bring that to the bedside and is contributing to that advancement in care.

 

Hessie Jones

Yeah, I I’m really happy that you’re here and I’m happy you’re contributing at a time when we have an aging population and our hospitals are probably going to be overfilled with those types of patients, so. Thank you for coming up today and I wish you luck, OK.

 

Devin Singh

Thank you so much for the time. Appreciate it.

Host Information
Hessie Jones

Hessie Jones is an Author, Strategist, Investor and Data Privacy Practitioner, advocating for human-centred AI, education and the ethical distribution of AI in this era of transformation. 

She currently serves as the Innovations Manager at Altitude Accelerator. She provides the necessary support for Altitude Accelerator’s programs including Incubator and Investor Readiness. She will be the liaison among key stakeholders to provide operational support and ultimately drive founder success. 

LinkedIn

You can also listen to this podcast on Transistor.

Please subscribe to our weekly LinkedIn Live newsletters.