A Quest for Better Patient Outcomes Through AI and Compassion with Narinder Singh, the co-founder and CEO of Look Deep Health

[00:00:02] Hey everybody, welcome back to the Outcomes Rocket, Salmar Kest here and today I have a

[00:00:10] privilege of hosting a really awesome guest. His name is Narender Singh. He's the co-founder

[00:00:16] and CEO of Look Deep Health. Prior he was co-founder and president of Aperio, a pioneer

[00:00:23] in cloud computing and first cloud investment of Salesforce.com and Sequoia Capital. The

[00:00:29] company grew over 1200 employees and was acquired in 2016 by a Y-Pro. Previous to Narender worked

[00:00:36] in the office of the CEO at SAP and ran product development at Web Methods. He's just an incredible

[00:00:43] person, 40 under 40 on CNN, the Daily Show. Just countless attention that he's gotten.

[00:00:49] So we're privileged to have time with him today. Narender, thanks for joining us.

[00:00:53] Thanks, all that was before kids. So now I'm a slower paced person.

[00:00:59] Love it, man. It's a pleasure to have you here on the podcast. And before we dive into Look Deep Health and the work

[00:01:07] that you and the team do there, tell us a little bit about you and what inspires your work in healthcare.

[00:01:11] You covered the professional background as we were ending and about to sell the company. I just was more

[00:01:17] and more enamored with two things. One, we had bought a million person developer and algorithm

[00:01:22] data science community called TopCoder and I became obsessed with AI. So that was like just

[00:01:27] inkling at me at a time when AI was not commercially viable thing to sell the enterprises. The second

[00:01:32] was I've been very involved in philanthropy for many years and this burning desire to bring my passions

[00:01:39] after work into my daytime experience. So I decided to get into healthcare. I'm my wife's a professor,

[00:01:44] we moved back to California. I did a master's in translational medicine with a bunch of 20-somethings

[00:01:48] at UCSF in Berkeley. And that was a wonderful experience for your older listeners. It was

[00:01:53] like Rodney Dangerfield and back to school, really transformative. But I got my real education

[00:01:58] in healthcare after that. My mom had interstitial lung disease. She went into the hospital for a

[00:02:02] biopsy which apparently went well but then she developed ARDS and crash while she was in the

[00:02:07] hospital. She immediately had to be portrayed. They said she can't come out. She's got

[00:02:12] to get a lung transplant. So she had to qualify for a lung transplant while in this

[00:02:16] precarious situation. Then eventually made it on the list, had to get ECMO to extend her

[00:02:21] time, had a successful bilateral lung transplant, multiple infections. So she was inpatient at

[00:02:27] UCSF for 12 weeks and then we had another six weeks where we were in the city taking care

[00:02:32] of her own health facility. So I spent a thousand hours in the ICU on the floor of

[00:02:35] the hospital watching how every element of healthcare worked from the patient perspective

[00:02:40] and frankly amazed because for my mom to make it out, two grandkids five years later

[00:02:45] is because a thousand things had to go right. But you notice the things that were everything

[00:02:51] was critical. And for example, my favorite nurse was this woman named Meg. If she's listening,

[00:02:55] I hope I get a chance to thank her in hunger. She was such an inspiration through but not

[00:03:00] because she was caring and compassionate. Almost all the nurses worked. The thing

[00:03:03] that Meg was very good at is when she didn't have another patient, she was very good at spotting

[00:03:08] problems but before they became an issue. So I started thinking about why can't Meg be

[00:03:12] my mom all the time? And then why can't Meg be with everyone's mom all the time?

[00:03:16] And we're not calling the AI Meg, don't get me wrong. But we started thinking about AI and

[00:03:21] computer vision that could have watched the patient at every instance because on the average

[00:03:24] shift a nurse is in a patient's room an hour or two out of 12, a doctor's there a few minutes a day.

[00:03:30] Most of the time there's not attention on the patient and every clinical provider will

[00:03:35] say there's no substitute for seeing feeling interacting with the patient. And so that

[00:03:39] really inspired me for healthcare and really what the company sought to do is use artificial

[00:03:45] intelligence and computer vision to help watch patients, to help those guardians,

[00:03:49] those caregivers provide the best care, care possible even when they weren't in the room,

[00:03:53] would you love to one? Yeah that's a great story. And what a blessing that your mom made it,

[00:03:59] Narender. Like many people don't and especially going through ECMO and the transplant and man

[00:04:06] what a blessing that she's still with us. So I'm just so grateful to hear that for you and

[00:04:11] sounds like you really pulled out a lot of gems and diamonds from that observation,

[00:04:17] that thousand hours of observation. Talk to us about Look Deep. How is the business adding value

[00:04:23] to the healthcare ecosystem? Yeah so as I start it's interesting because we've learned a lot.

[00:04:28] You start with this kind of noble aspiration of my mom needs attention and your mom needs

[00:04:33] attention and everyone's mom needs attention. Then you get into the reality of the business of

[00:04:37] healthcare, right? Our focus we were like we're going to build artificial intelligence so we did

[00:04:41] research partnerships with Duke and UCSF to build that technology and then we discovered wait a

[00:04:45] second the cost of deployment is so high because video in the hospital costs $10 or $15,000

[00:04:50] a room. So we're like okay I guess we have to integrate with hardware and build our own

[00:04:55] hardware simply so we can give it away for free so we can lower the cost of what's there.

[00:05:01] And then we went through and said wait there's like five or six different solutions here

[00:05:05] for EICU, for virtual sitting that doesn't make any sense. We have to build a software platform

[00:05:10] that covers inpatient scenarios generically. So Look Deep helps hospitals be present at every

[00:05:15] moment for every patient and we do that through software that's flexible, hardware that's included

[00:05:21] without capital expense and then AI that helps nudge the attention of caregivers

[00:05:26] towards the patients that need them. And it's really important because there's so much

[00:05:30] hype around AI. Hospitals and healthcare are too critical to rely on AI alone today,

[00:05:35] just like full self-driving cars been almost there for a decade but driver assist has been

[00:05:40] something that's saving lives for a decade so that's our models using AI to nudge the attention

[00:05:44] of caregivers towards the patients that need them. Zooming back for a second the fundamental

[00:05:49] problem for hospitals for the next decade is going to be this imbalance between patients are

[00:05:54] getting more acute they're getting older the number of people over 65 is going to double in

[00:05:58] the next 50 years. The less acute patients are getting diverted to outpatient-centric care

[00:06:03] repair times. The average staffing level is flat or declining so if you've got flat or declining

[00:06:09] staffing more acute patients and financial constraints some things got to give and so our

[00:06:15] innovation is we're not providing better care for more money we're trying to reduce cost and

[00:06:20] increase the scope of care at the same time it's the ultimate quadruple aim of healthcare

[00:06:25] and so that's our focus on the inpatient side is really trying to commoditize cost of

[00:06:30] core inpatient telemedicine and use AI to rise the level of care and support these nurses and

[00:06:36] doctors who are caring for our loved ones. Well that's really great and it sounds like rather

[00:06:41] than have something that's capital heavy you guys are integrating to existing OEM monitors

[00:06:47] different vendors providing the hardware at no cost but providing the intelligence that rests

[00:06:53] on that hardware as part of what you guys do to enable more more scalable care what would you say

[00:06:59] makes what you do different or better than what's out there is it a lot of us you already talked

[00:07:04] about or is there more? I'd say yes it's what I talked about but the real difference is

[00:07:08] everybody talks and so we're in the hype cycle of AI and as I look out there there's so many

[00:07:14] people that have what I call is drawing boxes on pictures and saying we have AI too so one of

[00:07:19] the things we've tried to do is approach this with seriousness I mentioned our research partners

[00:07:24] we've you know presented at the Society of Hospitals Medicine the American Thoracic Society

[00:07:28] Duke Grand Rounds so we do that to expose ourselves to talk about what we've done where there's

[00:07:35] potential where there's limitations and so from that perspective on the AI side we have

[00:07:40] we believe more training data than anybody in in the industry and we're using that and

[00:07:44] sharing the results of that broadly and publicly in a way that we feel aligns with what we do for

[00:07:52] clinical interventions not the tech hand-waving where we try to do marketing through so I think the AI

[00:07:57] the reality of our AI capability is definitely the most significant differentiator and then yes

[00:08:02] reducing the cost of hardware providing our owns that hospitals don't have to I can't oversell

[00:08:07] how important that is because the first thing we get to is that if you try to apply AI in every

[00:08:11] 10th room imagine having an EHR at every 10th room it doesn't matter what the system does you just can't

[00:08:17] make it part of workflow so that fundamental step of saying no capital expense you can deploy

[00:08:22] gladly and count on the technology being there is really a minor but fundamental transformation

[00:08:28] that pairs with the AI to allow hospitals to see strategic change from this not just another

[00:08:34] niche innovation that gets applied we're trying to be a platform for every patient in every

[00:08:38] room in the hospital that's fascinating and now you know the the front door to healthcare is not

[00:08:44] just one is there any applications of what you do here in the home yeah so part two I think that

[00:08:52] makes this unique is focus and the reason we're so focused is because of the AI capabilities

[00:08:56] taking software and moving from the hospital to a assisted living facility is not very hard

[00:09:01] taking the AI models and having them translate is very difficult and so I would say this our

[00:09:06] focus is on acute care so that means hospitals, LTACs, some SNFs and potentially hospital at home

[00:09:14] because there it's a room setting it's trying to replicate the hospitals like capability

[00:09:18] in that room setting there we're very strong outside of that I think a lot of our technology

[00:09:23] applies but our differentiator is not so we're pretty focused on the acute care paradigm from

[00:09:27] home to hospital that's fantastic thank you for sharing that and folks it's something to

[00:09:33] think about and I want to just pause here for a second or internet to just point out like when

[00:09:37] you've got a business and you're wanting to establish an edge it's important that you're

[00:09:43] clear and vocal about what you don't do because that is oftentimes more important I was at Metronic

[00:09:48] before and when Omar Ishtrak was the CEO man he just did this masterfully he would sit in front

[00:09:54] of a crowd and he'd be like this is what we do and this is what we don't do and let me extend

[00:10:00] the list so I think you're doing a fantastic job of keeping us focused here and on what is

[00:10:05] you guys operate acute it's specific to that particular area and no questions asked there

[00:10:11] so how have what you guys do really has it improved outcomes can you give us some stories

[00:10:17] there yeah absolutely so think about like our spectrum if we get into acute care as we think

[00:10:22] about things that relate to safety like virtual sitting like a lot of your listeners may not

[00:10:27] even be aware that in the average hospital there's three to five percent of patients that are

[00:10:32] watched with what's called a sitter that's not a nurse not an assistant just somebody who sits in

[00:10:36] the room and says so please stay in bed right now you're in a hospital trying to prevent falls

[00:10:41] and other kind of extreme activities there's been a market for virtual sitting for a decade where

[00:10:45] people will do that same thing over video what we're doing is expanding that dramatically

[00:10:49] because the AI is nudging the person so they can watch for more people and so we've seen

[00:10:54] incredible outcomes on our virtual sitting capabilities and what we're really targeting is

[00:11:00] not just great outcomes on five percent of patients how do we expand that to 15 20 30 50

[00:11:06] percent of patients so we're seeing some of our hospitals deploy 20 30 percent of their

[00:11:10] census through this this approach where's in the past they've only done five percent

[00:11:15] and our goal is to get that to 100 percent so Sanctius wave one and now we're engaging with

[00:11:20] several systems on virtual nursing so taking the same way that we nudge a virtual sitter to say

[00:11:25] this person might be getting out of bed will nudge a virtual nurse to say hey this patient is not

[00:11:30] moving and it's just static in bed and nobody's seen them for a couple hours should we talk to

[00:11:35] them about your pressure bundle so nudging the nurse's attention so they can a virtual nurse

[00:11:39] can support their bedside nurses are watching more patients and and that becomes the second

[00:11:44] wave and then finally transfers of care virtual medicine moving people from the ICU to the

[00:11:49] floor of the hospital this is where just now starting to get into is seeing that can we provide

[00:11:54] overlaps so that you can move people at the appropriate time but still give them additional

[00:11:59] attention so that you can have ICU like attention as they're transitioning down it's somewhat odd

[00:12:05] and if you've ever been in a hospital you go from the ICU or you have one nurse for two

[00:12:09] patients to the floor we have one nurse for six and you feel like you've been abandoned

[00:12:14] as a patient and so it always struck me that there's no smooth transition between those two

[00:12:19] and some of the things that we're working on the virtual side is providing that and so

[00:12:22] I would say virtual sitting virtual nursing virtual medicine and in that order is where we've

[00:12:27] had we were turning those things into hospital wide initiatives versus niche telemedicine solutions

[00:12:33] that's awesome not thank you thank you for that and then I do want to take a moment to

[00:12:37] dig into the nudge and really help the listeners understand better and myself too

[00:12:42] like I'm curious so when you say a nudge a nudge could be a lot of things a nudge could be a prompt

[00:12:47] on a mar screen it could be a vocera message it could be a page so talk to us a little bit

[00:12:53] about what that looks like to make it more real for everybody listening wonderful question if you

[00:12:58] have it like Richard Taylor who won the Nobel Prize has a wonderful book called nudge that I

[00:13:03] would highly recommend if he was the father of behavioral economics and nudge is like one of

[00:13:08] the paramount places on how you can with small changes drive larger changes and getting people

[00:13:14] to do the right thing for themselves so in this context here's a couple things that are important

[00:13:18] first is you got to respect the clinical work you can't try to go in and say I'm going to change

[00:13:22] how all nurses do their job and hey I'm going to train you on AI this magical thing that you're

[00:13:27] already a little scared of because you saw the Watson stuff never go anywhere etc so you've

[00:13:31] got to respect the human change management aspect so what we look at is human plus AI as I

[00:13:36] mentioned so in all of our scenarios take the virtual nursing there's a virtual nurse and today

[00:13:42] even before look deep people are using virtual nurses to do things like admissions and discharges

[00:13:47] or try to provide support remotely so for the bedside nurse a virtual nurse is a good thing

[00:13:52] it's a one of your people one of your compatriots who's helping and keeping an eye on things so

[00:13:57] the first thing we said was that's the person we should nudge because then the interface to

[00:14:01] the clinician on the ground the bedside team stays the same it's still one of your colleagues

[00:14:06] and then for that virtual nurse what we're doing is saying let's say we're watching several units

[00:14:10] 100 patients we'll say hey here's the patients that we think are right now at high risk and in

[00:14:17] this case an example I gave a pressure high risk there's usually a pretty defined protocol

[00:14:21] he was returned people over a couple hours and we can tell you then we're nudging you based

[00:14:25] on something you understand it's not AI magic the people who are in bed not moving and

[00:14:31] have not been visited we're going to we're going to escalate them to the top and so when the virtual

[00:14:35] nurse will see that they'll say I already understand what that is oh those are the people

[00:14:38] that are not moving in there so we're falling into things they understand not saying there's

[00:14:43] some black box algorithm we're trying to lean towards transparency of approach so we get more

[00:14:49] adoption and now that virtual nurse can use our prompt of these are high to look at the EHR

[00:14:56] for what meds they're on or things that might have changed they can look at our data that

[00:14:59] shows day over day movement patterns they can use our instant video to say Mr Smith looks like

[00:15:04] you're a little lethargic this morning are you feeling the effects of the drug change we made

[00:15:09] and then they can interface with their bedside team or order the pressure bundle in their EHR

[00:15:13] they can bring their whole context to make the decision so all we did is take 100 patients

[00:15:19] and said which of the first 10 you should look to around this so we're simply getting them

[00:15:23] to use their expertise more efficiently we're not trying to take over the decision making

[00:15:28] certainly not now and that makes it so the AI can be deployed right now not in 10 years

[00:15:34] that's awesome thank you for diving into that a little bit deeper and at the end of the day folks

[00:15:39] it's about exception based management what are the exceptions out of all of the vitals

[00:15:45] and information that you're getting and those exceptions the ones that are troubling

[00:15:51] that need attention that's where we're focused here with Narender and what his team are doing

[00:15:54] so love that I think about setbacks as opportunities personally as an entrepreneur

[00:16:00] so let's talk about one of those for you what's an experience you've had as a setback with the

[00:16:05] company and a key learning that's come out of it so we started the company a little over four

[00:16:10] years ago which is before COVID and so certainly COVID was on one end a massive setback but it

[00:16:16] did make people more accepting a video in the hospital the real setback I've alluded to

[00:16:21] already which is just understanding that the reality of it doesn't matter you have to have a

[00:16:28] financial argument and that financial argument plus better care that's great better care long

[00:16:34] term financial argument just doesn't work and that drove us to expand our solution much

[00:16:39] wider we thought we were an AI only company and really were a solution we have to be a solution

[00:16:45] we had what does the customer need to succeed what does the hospital need to succeed so our

[00:16:49] setback was having to move from just AI to say we do care about hardware we can't defer that problem

[00:16:55] we do care about software we do care about integration we have to care about plenable

[00:17:00] workflow on the floor like those are all of our problems we cannot be an AI only company in healthcare

[00:17:06] and we learned that lesson very early and I think it's been pretty profound for us and it

[00:17:10] really affects every aspect of how we interact even our business model is we charge per patient day

[00:17:16] not installed not if you use it you pay us if you don't you don't pay us and the reason we do that

[00:17:22] is to align our incentives with the hospitals so that if they put it on the shelf in a shelf

[00:17:26] where we suffer so it's our responsibility to make them successful with the solution

[00:17:33] no matter how small or big that is how close or not it relates to other features that we provide

[00:17:37] we're in this together with them and so that to me has been a profound release and

[00:17:42] uplifting it says there is no excuses it's always our problem to make them be successful with

[00:17:48] this new technology because their plate is full we need to make it so obvious that they want to go

[00:17:53] forward with this and then they want to expand the usage that's great it seems like a de-risk

[00:17:59] approach for anybody wanting to test out the platform and certainly a really smart way to

[00:18:05] approach the the market and or so kudos to you and your team over there of thought leaders

[00:18:11] and again like the thing that you mentioned and I'll double down on right now is this idea of being

[00:18:16] a platform company not a pipeline company the platform company like Narendra takes a view of

[00:18:23] the entire market need and helps address it so love that you we've been talking about this

[00:18:28] on several of our past podcasts and you literally hit the sweet spot with it so thank you

[00:18:32] Narendra what are you most excited about so I think the things that I'm excited about

[00:18:37] is I've never had an opportunity in my career where I feel this alignment between mission

[00:18:44] and opportunity and so a lot of and maybe this is the naivety of only being in healthcare for four

[00:18:49] or five years right I haven't gotten beaten down I know it's hard I know sales cycles are long

[00:18:54] I hear no a lot I hear all the reasons people can't change but it feels like

[00:19:00] right now and I'm sure many a down driver said this it feels like something's got to give

[00:19:05] what's happening is there's so much attention outside of the hospital but I laid out all the

[00:19:10] facts of why hospital care has to change in the next decade and covid exposed many of those

[00:19:15] so I think there's a willingness to change that has not been there when the healthcare

[00:19:21] hospital system revenue was up into the right so there's a willingness to change there's

[00:19:26] all sorts of barriers to that change but that willingness to change I think is a once in a

[00:19:31] generation opportunity I don't think that's existed I think there's been a lot of times where

[00:19:36] healthcare 10 years later looks like healthcare 10 years ago and I think everybody is pretty

[00:19:40] pretty clear that change is going to come they may argue how fast it comes but when we talk to

[00:19:45] people who are thought leaders in the industry they're like this is gonna happen it's just a

[00:19:49] question of when I feel very blessed that we get an opportunity to have that in our backs

[00:19:53] as we get to try to help hospitals provide better care for patients yeah no I think that's a

[00:19:58] great call and now is the time and look there's the shift that the shifts that are happening in

[00:20:03] healthcare the new entrance into healthcare people are making moves and it's a 4 trillion

[00:20:10] 4.3 trillion dollar pie in the U.S. alone annually yeah so there's definitely incentive to make

[00:20:18] change happen and so of that 4.3 trillion about a trillion and a half is in hospitals

[00:20:23] and our argument is that the number of startups and innovation new innovations in that section

[00:20:28] is 1 to 100 less than hospitals so we might be crazy and going out for the highest hill

[00:20:33] but we think there's such opportunity for innovation there because in the past

[00:20:37] people have focused so much on the outpatient side that I think there's an appetite for hospitals

[00:20:42] who want to change with new solutions hey so give me that stat again it's one in a hundred

[00:20:48] give me that one again so out of the 4.3 trillion about 25 to 35 percent of that spend is in hospitals

[00:20:54] okay yep if we go and look at the funding of healthcare investors the last n years in startups

[00:21:04] and series seed and series especially that has been 50 to 101 outpatient and even when it's

[00:21:12] inpatient it's like revenue cycle management barely ever the care pieces the care innovation

[00:21:17] has only been the metronics of the world or medical devices a lot of things we've seen

[00:21:22] have really neglected those hospitals because they've been like oh hospitals that's

[00:21:26] different and I think that truism is true until it's not and I think we're seeing with COVID

[00:21:31] that it's not like they were seeing the opportunity for change to happen

[00:21:34] and just like a lot of exponential changes they feel like they drip along and then

[00:21:39] they're everywhere and I think we've seen that with AI and the problem with AI in hospitals

[00:21:43] how do you bring it in pragmatically and I think we've at least taken a good

[00:21:47] trip crack at solving that so I think we're going to see this kind of change

[00:21:51] in hospitals look like it's nowhere and then everywhere all at once

[00:21:55] love that thanks for double clicking on that one for me and render definitely wanted a

[00:21:59] dig deep into that stat and yeah look I'm a fan of going the opposite way

[00:22:05] it oftentimes has helped me in my career and in my moves so I think it's a great call out and

[00:22:11] an interesting fact so definitely one will call out in the show notes and folks by the way the

[00:22:15] show notes have everything you need links to things we've talked about information about

[00:22:20] look deep health information about Narender so make sure you check that out and find ways to

[00:22:25] engage with him and the team there we're here at the end Narender so I love if you could

[00:22:30] just close us out with a closing thought and where you want people to get in touch with you and your

[00:22:34] team my closing thought is this it's my favorite quote of all time it's by this guy you might have

[00:22:40] heard of Albert Einstein as simple as possible but no simpler and I think that covers so much

[00:22:47] of what we're looking at is that we don't need magical treatments that have this incredible

[00:22:53] complexity but we also cannot oversimplify the problems and the reality of tensions we have

[00:22:58] in the healthcare system especially in hospitals so to me that's our guiding principle it should be

[00:23:03] for hospitals as simple as possible but no simpler we'd love to engage with you as you mentioned

[00:23:08] our website is lookdeep.health my personal twitter is singns lookdeep.health is on twitter we're happy

[00:23:15] to engage folks and I think you said something in the middle of the conversation we're all about

[00:23:19] prove it you shouldn't believe what I said you should force us to prove it our business

[00:23:23] model or entire approach is on that so we're happy to step up to the challenge

[00:23:27] of comparing against current standards of care in any of these areas because if we're not better

[00:23:31] people shouldn't use us and we think that's a driving force for this industry of change and

[00:23:36] we'd like to lean into that so I hope your listeners will hold us to account on that.

[00:23:41] Love that Narendra thanks for that bold close there and for anybody wanting to take a challenge

[00:23:47] to really try out the technology that Narendra and his team are doing it seems pretty de-risked

[00:23:52] to me so take a shot these podcasts we have them so that you could actually get after it

[00:23:58] and make a difference with the stuff that we do the rocket is about 10x not 10% so if you're looking

[00:24:04] for that type of growth link up with Narendra such a pleasure to have you on and looking forward

[00:24:08] to staying in touch my friend I love the engagement thank you so much.