1 00:00:00,460 --> 00:00:01,540 - This is Gracelyn Keller 2 00:00:01,540 --> 00:00:02,585 with the Becker's Healthcare Podcast, 3 00:00:02,845 --> 00:00:06,745 and we are live at the 11th annual CEO and CFO Roundtable. 4 00:00:07,245 --> 00:00:09,105 I'm joined currently by Drew Goldstein 5 00:00:09,245 --> 00:00:11,585 and Jeremy David, who are both the co-head 6 00:00:11,585 --> 00:00:13,065 of healthcare at Palantir. 7 00:00:13,405 --> 00:00:15,625 So Drew and Jeremy, thanks so much for joining me today. 8 00:00:15,765 --> 00:00:18,585 Um, and could you just quickly each introduce yourself? 9 00:00:20,165 --> 00:00:21,505 - Hey, I'm Drew Goldstein. 10 00:00:21,645 --> 00:00:23,665 Uh, I'm one of the co-heads of healthcare 11 00:00:23,765 --> 00:00:26,345 for Palantir Technologies, and I'm here with Jeremy. 12 00:00:27,305 --> 00:00:28,685 - Hey, uh, thanks for having us. 13 00:00:29,005 --> 00:00:31,165 I have the same, uh, introduction as Drew. 14 00:00:31,875 --> 00:00:34,125 - Wonderful. Well, drew, if you wanna just take it away 15 00:00:34,125 --> 00:00:35,725 and explain a little bit more about 16 00:00:36,005 --> 00:00:37,245 Palantir as we get started. 17 00:00:37,955 --> 00:00:39,605 - Yeah, happy to. So, you know, 18 00:00:39,605 --> 00:00:41,645 Palantir works in a number of different industries. 19 00:00:42,145 --> 00:00:45,565 Um, Jeremy and I lead Palantir's Healthcare Business, um, 20 00:00:45,585 --> 00:00:47,085 you know, across each of these industries, 21 00:00:47,175 --> 00:00:49,685 Palantir's Focus is providing software 22 00:00:49,685 --> 00:00:52,085 that creates an interface between the data 23 00:00:52,105 --> 00:00:54,845 and information that sits across various systems throughout 24 00:00:54,845 --> 00:00:58,045 an organization and the frontline operators that need 25 00:00:58,045 --> 00:00:59,205 that information to be able 26 00:00:59,205 --> 00:01:02,365 to make the real day-to-day decisions to run the business. 27 00:01:02,715 --> 00:01:05,445 What we're doing in healthcare is taking that same approach 28 00:01:05,785 --> 00:01:08,845 and deploying it against core hospital operational 29 00:01:09,125 --> 00:01:11,205 challenges with health systems across the country. 30 00:01:11,545 --> 00:01:14,125 Uh, we went from working with just about 1% 31 00:01:14,265 --> 00:01:17,565 of the US healthcare system at the beginning of this year, 32 00:01:17,865 --> 00:01:20,885 so now working with 16% of the US healthcare system. 33 00:01:21,305 --> 00:01:22,605 Uh, and we work with, you know, 34 00:01:22,605 --> 00:01:24,165 several health systems across the country. 35 00:01:24,465 --> 00:01:27,445 Uh, we have a team of now about 75 engineers 36 00:01:27,445 --> 00:01:30,125 that are fully dedicated, um, to our work directly 37 00:01:30,125 --> 00:01:31,205 with our health system partners. 38 00:01:32,345 --> 00:01:34,355 - Wonderful. Thanks for that explanation. 39 00:01:34,655 --> 00:01:37,155 Um, and Jeremy, I would love to start off 40 00:01:37,155 --> 00:01:39,395 with our first question, uh, throwing this to you, 41 00:01:39,665 --> 00:01:41,115 what are you most excited about right now? 42 00:01:42,265 --> 00:01:44,315 - Yeah, I, I think it's funny, like, uh, 43 00:01:44,365 --> 00:01:48,035 paler has a relatively unique way of approaching with, uh, 44 00:01:48,095 --> 00:01:49,235 of partnering with health systems. 45 00:01:49,775 --> 00:01:50,875 Um, and, uh, 46 00:01:50,875 --> 00:01:51,955 the way it works is we have these 47 00:01:51,955 --> 00:01:53,555 forward deployed engineers. 48 00:01:53,985 --> 00:01:56,755 They, uh, meet with the actual operators 49 00:01:56,755 --> 00:01:58,275 and leaders at these health systems 50 00:01:58,775 --> 00:02:01,075 and respond to what the hardest problem 51 00:02:01,255 --> 00:02:03,875 or what the biggest fire at that health system currently is. 52 00:02:04,255 --> 00:02:06,275 So, like the cheeky way to answer the question is like, 53 00:02:06,275 --> 00:02:08,355 we're most excited by whatever the folks we're interfacing 54 00:02:08,385 --> 00:02:11,035 with, uh, whatever they're most, most excited about. 55 00:02:12,185 --> 00:02:14,405 And I think over the past, uh, year 56 00:02:14,405 --> 00:02:17,165 or two, there's been a couple of verticals that have been, 57 00:02:17,385 --> 00:02:19,085 um, uh, most exciting to us. 58 00:02:19,245 --> 00:02:22,485 I think the first I'll mention is like, uh, we've, uh, 59 00:02:22,655 --> 00:02:24,125 partnering with HCA 60 00:02:24,125 --> 00:02:26,565 and a, a couple other folks, uh, we've really come 61 00:02:26,585 --> 00:02:30,445 to a different way to approach, uh, nurse scheduling. 62 00:02:30,785 --> 00:02:32,405 And so if you go to any, you know, 63 00:02:32,405 --> 00:02:33,845 health system leadership right now 64 00:02:33,985 --> 00:02:36,965 and ask what the hardest problem is, most often you'll get, 65 00:02:37,105 --> 00:02:40,205 uh, as part of the, you know, top three at least, uh, 66 00:02:40,345 --> 00:02:41,365 is the nursing shortage 67 00:02:41,465 --> 00:02:43,805 and not having enough staff to, to staff those beds. 68 00:02:44,265 --> 00:02:47,285 And I think if you go back, uh, a couple of years to our, 69 00:02:47,425 --> 00:02:50,365 our, our first partnerships with the Cleveland Clinic and, 70 00:02:50,365 --> 00:02:52,685 and, and Tampa General, uh, like out 71 00:02:52,685 --> 00:02:54,205 of one year we would be hearing 72 00:02:54,315 --> 00:02:57,005 that there's this like massive nursing shortage in the 73 00:02:57,005 --> 00:02:58,565 country, reading the first page 74 00:02:58,565 --> 00:02:59,820 of the Wall Street Journal saying, 75 00:02:59,820 --> 00:03:01,765 we don't have enough nurses to take care of the patients, 76 00:03:01,825 --> 00:03:02,965 uh, we have in this country. 77 00:03:03,225 --> 00:03:05,285 And then out of the other, uh, year we'd like watch 78 00:03:05,345 --> 00:03:07,685 how this was actually done on the front lines. 79 00:03:08,345 --> 00:03:12,645 And it was like a lot of, uh, frankly brilliant work, but, 80 00:03:12,645 --> 00:03:17,125 but manual work done by, uh, nurse managers, uh, CNOs, 81 00:03:17,245 --> 00:03:20,405 staffing office folks, um, uh, manually and, and, 82 00:03:20,425 --> 00:03:23,365 and with like virtually no data, no software. 83 00:03:23,585 --> 00:03:25,045 And I think there's a lot of reasons for that 84 00:03:25,045 --> 00:03:26,045 that we don't have to go into. 85 00:03:26,265 --> 00:03:28,085 But over the past couple of years, we've sort of, 86 00:03:28,085 --> 00:03:31,165 from the bedside out actually built systems 87 00:03:31,225 --> 00:03:34,165 and algorithms to support those folks bringing data from the 88 00:03:34,365 --> 00:03:37,085 EMR, from other clinical systems, as well as, uh, 89 00:03:37,325 --> 00:03:40,285 building new, uh, ways to, uh, new staffing systems 90 00:03:40,665 --> 00:03:43,765 to bring the information together to, to best align, uh, 91 00:03:43,765 --> 00:03:45,325 patient demand with nursing supply. 92 00:03:45,545 --> 00:03:48,045 Uh, the, it's a sort of long explanation on 93 00:03:48,045 --> 00:03:49,965 how it actually works, but the, the broad strokes 94 00:03:49,965 --> 00:03:51,925 of it is it's not self-scheduling anymore. 95 00:03:52,385 --> 00:03:54,605 It, it's a workflow by which nurses 96 00:03:54,865 --> 00:03:55,965 can enter their preferences. 97 00:03:56,385 --> 00:03:59,805 And there's an AI optimizer that then, uh, matches between 98 00:04:00,525 --> 00:04:03,245 forecasted patient demand that determines, uh, uh, 99 00:04:03,425 --> 00:04:06,405 the required nursing workload, uh, and the staff 100 00:04:06,465 --> 00:04:08,605 and preferences of those staff that have available. 101 00:04:08,825 --> 00:04:10,685 And it generates the optimal nursing schedule. 102 00:04:10,965 --> 00:04:12,285 I think there's like a ton of benefits 103 00:04:12,285 --> 00:04:14,245 to this from a CFO's perspective, 104 00:04:14,665 --> 00:04:18,005 but importantly, like, uh, to get a nurse manager or, or, 105 00:04:18,025 --> 00:04:20,845 or nursing leadership to adopt the thing, it actually saves, 106 00:04:20,945 --> 00:04:24,605 uh, between 15 and 20 hours of the, the, uh, uh, 107 00:04:24,785 --> 00:04:27,845 the scheduling time that goes back into the, uh, the, 108 00:04:27,845 --> 00:04:29,085 the pocket of that nurse manager. 109 00:04:29,545 --> 00:04:33,405 Um, and, uh, it, it's really like a, a, a paradigm shift in 110 00:04:33,405 --> 00:04:36,245 how nursing, uh, nurse scheduling is done today. 111 00:04:36,465 --> 00:04:38,085 Uh, and I think that that's incredibly exciting. 112 00:04:38,085 --> 00:04:39,405 There's, there's various other things 113 00:04:39,405 --> 00:04:42,725 that we've built on top of the, like data asset. 114 00:04:42,725 --> 00:04:45,725 We've began to build at each health system on, uh, patients 115 00:04:45,725 --> 00:04:47,565 and nurses like a couple are using, uh, 116 00:04:47,565 --> 00:04:51,245 large language models on top of Palantir's a IP platform 117 00:04:51,305 --> 00:04:54,285 to do better and more automated revenue cycle management. 118 00:04:54,585 --> 00:04:57,205 Um, but ultimately we'll continue to react to the, 119 00:04:57,345 --> 00:04:59,165 the problems health system leaders bring to us. 120 00:04:59,625 --> 00:05:01,405 Um, but those are two examples. 121 00:05:03,345 --> 00:05:06,205 - And Drew, next question is for you, what are some 122 00:05:06,205 --> 00:05:08,565 of the issues that are taking up most of your time? 123 00:05:09,115 --> 00:05:10,685 - Yeah. We're finding that the issues 124 00:05:10,685 --> 00:05:12,685 that we're focusing on today are the issues 125 00:05:12,685 --> 00:05:13,845 that are most top of mind 126 00:05:13,865 --> 00:05:15,445 for health systems across the country. 127 00:05:15,865 --> 00:05:17,125 And that's not an accident. 128 00:05:17,125 --> 00:05:20,285 That's the result of that four deployed engineering model 129 00:05:20,315 --> 00:05:21,765 that Jeremy described before. 130 00:05:22,145 --> 00:05:23,885 And some of the examples of what we're seeing 131 00:05:23,885 --> 00:05:26,125 with our partners today at Cleveland Clinic, 132 00:05:26,135 --> 00:05:27,245 we've increased the rate 133 00:05:27,245 --> 00:05:31,205 of accepting patients via their transfer center by over 10%. 134 00:05:31,585 --> 00:05:33,765 At Tampa General, we've decreased the length of stay 135 00:05:33,765 --> 00:05:34,925 for their septic patients, 136 00:05:35,155 --> 00:05:38,165 reduced the hold times in their PACU by over 30%, 137 00:05:38,505 --> 00:05:41,285 and onboarded 15 analysts from their analytics team 138 00:05:41,545 --> 00:05:42,845 to run their end-to-end data 139 00:05:42,845 --> 00:05:45,085 and analytics program directly in our platform. 140 00:05:45,505 --> 00:05:47,285 And at HCA with the staffing 141 00:05:47,285 --> 00:05:49,365 and scheduling solution that Jeremy described, 142 00:05:49,575 --> 00:05:51,205 we've onboarded nine hospitals 143 00:05:51,225 --> 00:05:53,405 and thousands of nurses this year so far, 144 00:05:53,785 --> 00:05:55,525 and we've reduced the time on scheduling 145 00:05:55,525 --> 00:05:59,210 by over 90% while capturing nearly 99% adherence 146 00:05:59,305 --> 00:06:01,125 to their staff member preferences. 147 00:06:01,585 --> 00:06:03,045 Um, we've been providing speed 148 00:06:03,045 --> 00:06:05,685 to value in these areas in a way that's been key both 149 00:06:05,685 --> 00:06:08,525 to the adoption that we've seen from our platform at these 150 00:06:08,525 --> 00:06:11,445 systems, um, but also to providing increased access to care 151 00:06:11,445 --> 00:06:14,005 for patients, uh, across each of our partners. 152 00:06:16,725 --> 00:06:18,525 - Absolutely. And Jeremy, I'm gonna throw it back to you 153 00:06:18,525 --> 00:06:19,805 for our final question here. 154 00:06:20,025 --> 00:06:22,605 Uh, what do you think the most effective healthcare leaders 155 00:06:22,705 --> 00:06:23,805 are going to need to be 156 00:06:23,805 --> 00:06:25,325 successful in the next two to three years? 157 00:06:26,355 --> 00:06:30,125 - Yeah, I think we're obviously like experiencing a boom 158 00:06:30,505 --> 00:06:33,325 of AI over the past six months to a year. 159 00:06:33,785 --> 00:06:36,405 And I think like asking ourselves the question like, 160 00:06:37,065 --> 00:06:40,405 how can we get, um, AI actually to the front lines 161 00:06:40,625 --> 00:06:41,685 of healthcare today? 162 00:06:41,835 --> 00:06:44,005 Like whether it's at the point of care, uh, 163 00:06:44,035 --> 00:06:46,445 from a physician interacting with a patient or, 164 00:06:46,745 --> 00:06:49,645 or to a more back office operational workflow, 165 00:06:49,835 --> 00:06:52,845 like thinking about the point of care, if you suppose in, 166 00:06:52,905 --> 00:06:53,965 you know, 15, 10, 167 00:06:54,275 --> 00:06:58,325 however many years, uh, physicians will have access 168 00:06:58,425 --> 00:07:01,845 to these more advanced AI algorithms to help make sense 169 00:07:01,845 --> 00:07:03,645 of all this information about the patient 170 00:07:03,705 --> 00:07:05,925 and similar patients, uh, to, to, 171 00:07:05,925 --> 00:07:07,085 to make clinical decisions. 172 00:07:07,475 --> 00:07:10,485 Like I think you have to think about, uh, where's the data 173 00:07:10,835 --> 00:07:14,085 that these AI algorithms are going to be using and, 174 00:07:14,105 --> 00:07:17,005 and what software systems are going to be used at the point 175 00:07:17,005 --> 00:07:19,165 of care, uh, to enable those physicians? 176 00:07:19,465 --> 00:07:20,565 And, and, and if you look at 177 00:07:20,565 --> 00:07:23,005 what software is actually in place at the point 178 00:07:23,005 --> 00:07:26,005 of care today, it, it's really only one system. 179 00:07:26,275 --> 00:07:27,365 It's the EMR, 180 00:07:27,585 --> 00:07:29,485 and it's the sort of same EMR system 181 00:07:29,485 --> 00:07:32,365 that's been at the point of care over the past, 182 00:07:32,465 --> 00:07:33,525 you know, however many years. 183 00:07:34,525 --> 00:07:37,305 And I think, uh, this will happen very soon. 184 00:07:37,305 --> 00:07:39,425 Like we've had recent conversations with a lot 185 00:07:39,425 --> 00:07:41,825 of healthcare leaders that are really excited about the 186 00:07:41,825 --> 00:07:44,625 concept of taking the EMR 187 00:07:44,625 --> 00:07:47,625 and re-imagining it to be more of a, of a platform 188 00:07:48,005 --> 00:07:50,985 by which folks can harness the data that exists within, 189 00:07:51,165 --> 00:07:53,025 within, uh, these health systems 190 00:07:53,285 --> 00:07:55,665 and actually apply a algorithms to the point of care. 191 00:07:55,925 --> 00:07:59,025 And, and, and it requires an open system, a system by which, 192 00:07:59,325 --> 00:08:02,665 you know, the 22-year-old Stanford new grad can actually 193 00:08:02,665 --> 00:08:03,905 like, build upon, uh, 194 00:08:03,905 --> 00:08:06,185 and get excited about these workflows to, to, to, 195 00:08:06,185 --> 00:08:07,425 to start the next generation 196 00:08:07,425 --> 00:08:09,185 of these companies and innovations. 197 00:08:09,365 --> 00:08:12,425 The chief digital officers, uh, of all of these companies 198 00:08:12,425 --> 00:08:15,385 that have been moving from technology to healthcare, 199 00:08:15,775 --> 00:08:17,705 like currently, they're effectively blocked 200 00:08:17,765 --> 00:08:20,465 by the software monopoly that exists at the point of care. 201 00:08:20,845 --> 00:08:22,185 And I think what we'll see in 202 00:08:22,185 --> 00:08:23,505 what health system leaders should be pushing 203 00:08:23,565 --> 00:08:28,425 for is something that's more open, something where more, uh, 204 00:08:28,575 --> 00:08:29,625 like smart folks 205 00:08:29,725 --> 00:08:34,585 and in innovative folks can, can get access to, uh, 206 00:08:35,045 --> 00:08:38,305 uh, uh, a reasoning about how we can deploy, uh, 207 00:08:38,305 --> 00:08:39,865 deploy AI at the point of care. 208 00:08:39,885 --> 00:08:42,025 So I, I think we're gonna see a lot more restructuring 209 00:08:42,165 --> 00:08:45,265 and openness of, uh, EMR systems, 210 00:08:45,565 --> 00:08:47,865 and I think they'll start to look a little bit less like 211 00:08:47,865 --> 00:08:51,305 medical records and more like care delivery platforms 212 00:08:51,565 --> 00:08:53,465 by which folks can, can build upon and, 213 00:08:53,465 --> 00:08:55,205 and get AI into these workflows. 214 00:08:55,545 --> 00:08:58,805 Um, and I think that's gonna have tremendous impact on, uh, 215 00:08:58,805 --> 00:09:01,565 quality and the number of mistakes that that are made. 216 00:09:01,665 --> 00:09:03,645 But it all starts with making sure we have the right 217 00:09:03,925 --> 00:09:05,725 software at the point of care to harness these, 218 00:09:05,725 --> 00:09:06,725 these new algorithms. 219 00:09:07,325 --> 00:09:09,965 - A hundred percent. Well, Jeremy Drew, thanks so much 220 00:09:09,965 --> 00:09:11,885 for joining me today on the Becker's Healthcare Podcast. 221 00:09:12,095 --> 00:09:14,165 Again, we're live at the 11th annual 222 00:09:14,405 --> 00:09:16,365 CEO and CFO round table. Thanks so much. 223 00:09:16,665 --> 00:09:18,445 - Thanks. Thank you. .