S3E11: Provider Led Efficiency Gains (ft. Yair Saperstein, Avo, Mount Sinai Health System)

Yair (Avo)   1:20
Yeah, so the story actually starts when I was in residency and at the time I really wanted to be a whole person. You know, for me as a practicing doc, I was pegged as you are a doc and this is your whole life, you know? But I really wanted to be able to have dinner with my family. I wanted to be able to go for a walk.
Walks. I'm a whole person. And so I wanted something that would give me efficiency to be a whole person. Simultaneously, though, I'm a doc, right? I wanted to be a good clinician. I wanted to be able to deliver value to my patients. And the best way to do that was with quality. And how can I have efficiency with quality? Those kinds of things didn't.
There were either the quality systems, like best practice alerts, but there were so many that I I got overwhelmed, right? It's like with all these alerts. And then on the other side is like the efficiency plays, but the efficiency plays were like, yeah, we'll do an 80% good enough job. It's like, no, I'm trying to really do as well as I can.

Jordan Cooper   2:01
Yeah.
OK.

Yair (Avo)   2:19
And so I really wanted the ability to have those two together as an individual clinician, as an individual physician, you know, Fast forward a little bit and I became chief resident. I was involved on the quality side and you know, I started to think more from a system perspective. I actually was an informatics liaison to be able to help.
Install the new Epic system as we were transitioning where I was chief resident over to to our Epic system and so really got more of an overview on the system side. I had also just completed my MPH in hospital policy and management, right? So more of like this.
Broader holistic looking down umbrella perspective on from a system side more than as an individual physician to complement that with you know what do we need as a system and what would be important from a system level where it was really much more about OK.
Make sure that our physicians, our clinicians are happy, obviously, right. So if they're burning out, we got to solve that. But we also need from our system side, the financial ROI to be able to keep us afloat, right, keep us going up and up. And so thinking creatively about these different ways of putting those two together.
Which actually is a larger macro perspective of what my micro perspective as an individual clinician was. And from that came out Evo and Evo as a tool, right? Evo as a platform is this melding of the two like quality and efficiency and the way that we do it in a word.
Is that we have a back end clinical AI apps builder and so this builder, we kind of call it an operating system for building clinical AI apps is able to solve these different problems. You want to solve physician burnout. You have an app for that. You want to be able to solve your throughput and capacity.
We'll talk about that. We have an app for that. And over this time, over the last bunch of years, right, as I've been interacting with various other health systems through our progress as a company, right, with my company hat on, really have heard from the leading health systems and the issues that are.
Being faced, you know, through all of these systems and heard some of the resonance of what people are facing and what the systems are facing, where one of the major things that I've been hearing about is capacity. How do we solve our capacity problems? How do we improve throughput? How do we allow for more patients to get in by getting more patients out? And so thinking about the.
Aspect of how to solve that from a broader perspective and the clinical AI apps that can help with that is something that's definitely been on my mind for the last for the last bunch of time.

Jordan Cooper   5:04
So here you covered a lot of different issues right now. We're going to drill into one of them, but let me just recapitulate for our listeners. We talked about quality of care, efficiency and delivery of care, physician satisfaction and burnout and capacity and throughput. I'd like to start with quality of care. I understand that there are society guidelines and best.
Practices put out by the Association of whatever American, the AP. There are many different associations, pulmonary, cardiac, and they have best practices and you have content aggregators to summarize the guidelines and then integrate them into the EHR to.
Tee them up for clinicians to provide more standard care. If that's not accurate, please just clarify how are you addressing quality of care?

Yair (Avo)   5:53
Yeah, so similar dual perspective, right? So everyone and their friends are all using AI scribes, and if you're not using an AI scribe, well, get on the bandwagon already. The thing is that AI scribes have difficulty in integrating quality.
Into the overall workflow because it's really contingent on whatever I'm saying. So there's no easy way to be able to guide me into a proper direction. The best place to be able to guide is in the chart summarization piece, right? Before I see a patient, I normally do chart review and when I go through my chart review, I'll take a look at the.
Notes and labs and meds, whatever my role is, right? I'll do this as an outpatient doc. I'll do this as a hospitalist. I'll do this as a consultant to be able to review the church, whatever it is. And then for me, I need to synthesize that and think about, OK, what is the proper thing that I need to do regardless of where that's coming from? Is it a society guideline?

Jordan Cooper   6:33
OK.
OK.

Yair (Avo)   6:52
And is it coming from a guideline aggregator? Is it coming from utilization management? And if I should admit or discharge the patient, right, wherever that decision point is, I'm getting it from a reliable source, which is going to guide me. And sometimes that source is something that I've learned beforehand and then I'm just incorporating whatever I need to do.
In order to actually have decision support, right, the way that people think about it is alerts in the EHR based on things that are going on in the EHR. But if there was some way that I can have it in the context of my chart summarization so that whenever I'm opening up my chart summarizer or what we call chart assist, then I'm able to get the decision support together with that chart assist.
Here's the summary of your patient. Here's our recommendations of what you might want to do, and here's a pre-drafted note that actually incorporates both of them together. Well, all of a sudden then I have the ability to really drive forward in my path that I'm doing right, because I'm getting more efficiency as a clinician, but with the guidance of decision support and.
The workflow that I prefer right? Because I want to do this, it saves me time. And on the Avo side, we show that we can actually reduce node bloat by doing it this way in addition to improving quality in addition to saving time, right? So like this triple aim of having quality that is true quality for.
The patient reflected in documentation and then time efficiency gains for clinicians. But regardless of the specific answer, it's thinking creatively about how to actually put the effective content of what you want to do into a workflow that people will respond to. And people used to think, well, if you have a hard stop, if you have an alert, that's gonna force.
to go in whatever direction you want, that'll work. And they realize that people are just writing QWERTY as an answer to whatever it is, like, hey, this is a hard stop, you must respond. QWERTY, enter. And then people are ignoring those hard stops because there's too many. And instead people want an easy pass. We're in New York, so it probably won't say Sunshine Pass like in the South.
Right that people want an easy pass to be able to actually go down and once you're on that easy pass route, then you can get whatever your recommendations are or advertisements right in the car example. But the right path that you wanna go to, maybe not the best corollary with advertisements, right? But the proper path of where you wanna go cuz you're already on that easy path.
Yes.

Jordan Cooper   9:11
So you're looking to reduce provider variability, improve throughput by making more valuable and less frequent BPAS, and then reducing friction between case managers and clinicians and trying to help providers better adhere to best practices. Is that right?

Yair (Avo)   9:31
Yeah, so so once one specific example right for throughput and this is really where right at the the best value can be found so.

Jordan Cooper   9:42
M.
Yeah.
OK.
OK.
I.

Yair (Avo)   10:02
Basically, case managers would use these guidelines to help determine if the patient, let's say, discharged, right? So an admitted patient, are they eligible for discharge or not? If they match the criteria that are in the care guideline criteria, right, that's OK, yeah, they match each of these things.
They're stable, they're able to walk, they they're they don't have abnormal labs and each of the criteria specific to their disease states, to their conditions, and if they match well, then they are eligible for discharge, right? And those guidelines that arbitrator is trusted by health systems and by insurance companies.
Now, the way that happens for me as a hospitalist, right? So let's go back in time a little bit is I would show up at an interdisciplinary rounds and the case manager would say, is the patient ready for discharge? And I would say, I don't know, the patient's still sick. And then it's OK, well, should I get the discharge paperwork ready? I'll let you know.
The patient.
Better. And then there's this whole back and forth because they're taking a look at the guidelines where the patient's actually matching as they're cross checking everything, which takes them a long time if they actually go through it. For me, I'm like, hey, I would just thought that the patient's sick and I'll let you know. And then this afternoon, oh, everything's actually resolved. Yeah, we can discharge. Unfortunately, it's already Friday afternoon and the.
You know, shelter's not going to take the packet and we got to wait until Monday. That's a longer length of stay and it's, you know, that's repetitive across the different patients across their disease states. Over here, what we do is we partnered with MCG to be able to take their content into the overall workflow so that when I open up my chart summarization, right, my chart assist.
In the morning, when I'm pre-rounding on my patients as a hospitalist or as a case manager, it will automatically summarize the patient's information compared to the MCG content, tell you if the patient's eligible for discharge. If the patient is eligible for discharge, pre-draft you a discharge summary. It means that the hospitalist and the.
Case manager on the same page. That's how to get patients out appropriately because you suddenly reduce that variation and the back and forth between these different parties and then allow for the patients to be discharged appropriately, reducing length of stay, improving throughput and thereby improving capacity. So it's a specific use case.
To solve that ROI with the same idea, right? It's the same structural framework. It's the same scaffolding. You're doing patient data extraction, checking against a knowledge source and having the AI take the patient data against the knowledge source to see where it matches and then producing outputs.
or orders as appropriate with a specific eye towards utilization management of the example of the MCG guidelines.

Jordan Cooper   12:48
And when I'd like to, I think some of our listeners might be interested in learning about some of the clinical outcomes of implementing these interventions. Would I be correct to say that you've seen improved HCAP scores, better guideline medication adherence, savings at one site in the millions of dollars, reduced admissions delays and.
Improved time saving and is it not accurate?

Yair (Avo)   13:11
That is, yeah. And we've seen it across these different ROI trackers. So in addition to all the things that you mentioned, I had mentioned we saw reduced note bloat in addition by eliminating copy forward, which is good for quality, significant time reduction in documentation.
So, right, it's across different areas where you'll see these improvements, obviously cost savings, better medication adherence, right, as you mentioned, decreased admission delays, right. So across the spectrum where we have shown significant ROI with these tools.

Jordan Cooper   13:45
I.
I.

Yair (Avo)   13:51
And then the nice thing about it is when people first started implementing AI in health systems, it was a like, OK, we have to get this in because everybody else has it. You know, we have to get this in because it's something which seems to have a significant impact. But now we're at a point of, OK, we want to get these tools in because we want.

Jordan Cooper   14:06
OK.

Yair (Avo)   14:11
Significant financial ROI. In addition, where are we going to see this ROI that the system will pay for itself over the first year?

Jordan Cooper   14:17
So, Yair, I'd like to pivot. We we only have about a few more minutes left in this episode, and I want to pivot to the idea of being an entrepreneurial provider. So health systems across the United States have incubators where they try to.
Invest in startup ideas and generate new revenue streams for the health system by holding an equity stake in these startups. Also, health systems have providers like you who see problems in the health system as they're delivering care that they have ideas for how to solve. So I'd like to ask you to speak about.
Both those topics separately. One is if a health system executive is listening to this episode right now, how would you advise them to respond to providers in their own system who have identified problems and identify would like to experiment with possible solutions to improve the problems facing them every day?
Start there.

Yair (Avo)   15:18
Yeah. So when I was in residency, I was in SUNY downstate. And one of my favorite things about it was that the leadership, the executives were so supportive of my ideas to be able to guide me to the right people who can help to the right resources.
I was able to work with the CMIO at Kings County Hospital doing electives and then later as an informatics liaison to be able to help install Epic. And there was so much that I learned along the way because of the support that I got. Fast forward and we were actually, once I started Avo, the first company in Elemental Labs.

Jordan Cooper   15:57
I.

Yair (Avo)   15:58
In Mount Sinai. So when that started as an incubator, we were the first company to go through and we were learning together, right, how to be able to support these ideas that are coming through. Not that I was a doc at Mount Sinai and being supported from inside that we actually applied externally.

Jordan Cooper   16:17
I.

Yair (Avo)   16:17
To be able to join Elementa and I had to be a path that was very explicit about like how we would chart this out because I was per diem at Mount Sinai, right, even at the time. And so we had to figure out that path that we would navigate.

Jordan Cooper   16:32
Mhm.

Yair (Avo)   16:37
Exactly what I could do and couldn't do within that context of the path. So I think a couple things to call out. Number one is being supportive of the physicians and of the clinicians that have these ideas will allow them to feel empowered to make the real changes that we're seeing. So support your clinicians, support your physicians #2 is.

Jordan Cooper   16:48
I.
I.
I.

Yair (Avo)   16:57
Make sure to handle the conflict of interest needs. Well, you know, that was something which was a little bit difficult to navigate, but especially they were super helpful for me to, you know, figure out, OK, what was OK for me? What should be handled by others within my company? What was I allowed to do or not? You know, could we join this or not?
At different times to just to make sure that everything was was clear and I took their guidance and was super appreciative for that guidance to make sure that everything was in line. Third is you know, thinking from a system perspective.
There is money opportunity, right? It's like if companies are going to do well and you have equity stake in these companies, well, that's great. Then you're able to do well from continuing to do what you're doing, just inculcating this idea of innovation. And it's not just that you are innovative, which itself has a lot of value, but also the financial returns that are possible. So with those.
I would say definitely from my side within the vendor perspective, it was amazing to be able to have these experiences and to go through the elemental labs in Mount Sinai and in general to be supported by the health systems that we were working with.
Both from my time as a in residency and then and then beyond and then from a system level perspective, like there's a lot of gain that can happen there. So it's not just a service, it's you know, you can do good and do well at the same time.

Jordan Cooper   18:26
Right. So it sounds like you health system may benefit from supporting providers who'd like to start their own solutions both by having equity and successful start-ups and generating cash flow that way and also introducing solutions for providers at their health system to problems that exist. So kind of improving provider retention.
And all the different issues we've covered today from quality to efficiency to to improve profitability. We are approaching the end of this podcast episode. So final question, Yair, I'd like to ask if looking back, is there if you could have one wish of something you could change or looking forward if there's one thing you'd like to have?
You know, you think health system leadership could help you with or help providers if they're looking to improve throughput and improve outcomes and improve efficiency. You know, what's kind of the one thing that you would either like differently or like in the future or one kind of take away for our exec leadership audience?
Today.

Yair (Avo)   19:27
Yeah. Biggest take away is kind of like your last point, which is listen to your physicians and clinicians on the ground to be able to synthesize the things that are important and the way that things are actually working.

Jordan Cooper   19:35
OK.

Yair (Avo)   19:43
Which you don't always get to experience unless you're actually in the trenches. And then with that you're going to be able to understand, OK, what are the real problems and you know what's going on and with the capacity and throughput examples, right? It's like if the.

Jordan Cooper   20:12
I.

Yair (Avo)   20:12
10 rounds and it's like, OK, cool. Like would it help to get something which would be able to be like an arbitrator that has the guidelines that can help be present there? And then it's like you have something to disagree about as opposed to just disagreeing like that can help solve that problem as one example, right? But the key is listening.

Jordan Cooper   20:19
Mhm.

Yair (Avo)   20:32
Listening to your clinicians, that's, uh, that's the secret.

Jordan Cooper stopped transcription