1 00:00:00,560 --> 00:00:02,720 This is Laura Dirdo with the Becker's Health 2 00:00:02,720 --> 00:00:04,960 Care podcast. I'm thrilled today to be joined 3 00:00:04,960 --> 00:00:07,839 by doctor Philip Payne, inaugural chief health AI 4 00:00:07,839 --> 00:00:10,660 officer at Washington Medicine and BJC Health System, 5 00:00:10,719 --> 00:00:12,719 as well as Deborah O'Dell, chief data and 6 00:00:12,719 --> 00:00:15,785 analytics officer at BJC Health System. Doctor Payne, 7 00:00:15,785 --> 00:00:16,984 Deborah, it's a pleasure to have you on 8 00:00:16,984 --> 00:00:18,045 the podcast today. 9 00:00:19,064 --> 00:00:20,605 Great. Thanks for having us. 10 00:00:21,224 --> 00:00:23,704 I'm excited for our conversation today because, you 11 00:00:23,704 --> 00:00:25,544 know, we're really gonna zero in on a 12 00:00:25,544 --> 00:00:27,945 collaboration or partnership between you both as well 13 00:00:27,945 --> 00:00:31,149 as the Washington University and BJC Health System, 14 00:00:31,289 --> 00:00:33,950 which created a new center for health AI. 15 00:00:34,329 --> 00:00:37,049 How do these strengths of each organization complement 16 00:00:37,049 --> 00:00:39,469 each other in advancing AI in health care, 17 00:00:39,530 --> 00:00:42,030 and how will this collaboration evolve over time? 18 00:00:42,250 --> 00:00:43,865 I would love to hear from your perspectives 19 00:00:43,865 --> 00:00:47,065 what's really important in the development process and 20 00:00:47,065 --> 00:00:49,065 and how you plan to continue to grow, 21 00:00:49,385 --> 00:00:50,045 the partnership. 22 00:00:51,065 --> 00:00:52,265 So maybe I'll start. 23 00:00:52,984 --> 00:00:55,065 You know, I think there's this unique opportunity 24 00:00:55,065 --> 00:00:57,164 that exists at the intersection of both 25 00:00:57,630 --> 00:01:00,189 the Washington University School of Medicine and then 26 00:01:00,189 --> 00:01:01,489 our partners at BJC. 27 00:01:02,270 --> 00:01:03,950 And when you think about it, really what 28 00:01:03,950 --> 00:01:06,189 we're trying to do is advance the design 29 00:01:06,189 --> 00:01:07,890 and adoption and use of AI 30 00:01:08,269 --> 00:01:11,150 in a way that really drives excellence in 31 00:01:11,150 --> 00:01:13,584 both operations and ultimately care delivery. 32 00:01:14,045 --> 00:01:15,805 And when you think about the combination of 33 00:01:15,805 --> 00:01:17,725 our 2 organizations, we have a wealth of 34 00:01:17,725 --> 00:01:20,784 expertise in that medical school from a biomedical 35 00:01:20,924 --> 00:01:24,540 research and computational science and AI perspective. And 36 00:01:24,540 --> 00:01:26,299 that really provides us with this sort of 37 00:01:26,299 --> 00:01:28,240 academic foundation for our work, 38 00:01:28,780 --> 00:01:31,579 and that leads to activities like cutting edge 39 00:01:31,579 --> 00:01:34,319 algorithm design or training new models or validating 40 00:01:34,459 --> 00:01:34,959 systems. 41 00:01:35,980 --> 00:01:37,760 And then at the same time, BJC 42 00:01:38,140 --> 00:01:39,359 Health System is 43 00:01:42,194 --> 00:01:43,119 and then at the same time, BJC Health 44 00:01:43,119 --> 00:01:43,807 System is a leading not for profit delivery 45 00:01:43,807 --> 00:01:44,791 system and really provides the ideal real world 46 00:01:44,791 --> 00:01:46,435 environment in which we can implement and scale 47 00:01:46,435 --> 00:01:47,655 these AI solutions. 48 00:01:48,114 --> 00:01:49,715 If you think about it, it's really sort 49 00:01:49,715 --> 00:01:51,974 of, as I just mentioned, a living laboratory 50 00:01:52,034 --> 00:01:53,635 in which we can think about how these 51 00:01:53,635 --> 00:01:55,840 technologies scale, how they have impact, 52 00:01:56,299 --> 00:01:59,420 and really how they provide value in real 53 00:01:59,420 --> 00:02:00,719 world complex workflows. 54 00:02:01,179 --> 00:02:02,780 And that means the innovations that we'll be 55 00:02:02,780 --> 00:02:04,700 developing won't just work in the laboratory, but 56 00:02:04,700 --> 00:02:06,859 they'll really have a direct and practical impact 57 00:02:06,859 --> 00:02:09,205 on daily patient care. And this new center 58 00:02:09,205 --> 00:02:11,284 is really all about bridging that gap and 59 00:02:11,284 --> 00:02:13,944 connecting sort of those 2 sets of capabilities 60 00:02:14,004 --> 00:02:15,144 so we can really, 61 00:02:15,925 --> 00:02:16,905 navigate that 62 00:02:17,444 --> 00:02:20,485 journey from identifying opportunities where AI can provide 63 00:02:20,485 --> 00:02:23,284 value to deploying technologies, to evaluating them and 64 00:02:23,284 --> 00:02:24,240 then scaling them. 65 00:02:24,960 --> 00:02:26,480 And I also think this is emblematic of 66 00:02:26,480 --> 00:02:29,360 just the collaborative nature of the relationship between 67 00:02:29,360 --> 00:02:30,260 our 2 organizations, 68 00:02:31,120 --> 00:02:33,520 especially as we work, even more closely together 69 00:02:33,520 --> 00:02:37,460 to drive, care delivery for a substantial population 70 00:02:37,599 --> 00:02:38,099 across 71 00:02:38,400 --> 00:02:40,394 not only Missouri, but a number of surrounding 72 00:02:40,394 --> 00:02:40,894 states. 73 00:02:42,394 --> 00:02:44,555 That's amazing to hear. You know? It really, 74 00:02:45,034 --> 00:02:46,174 very much, it penalizes 75 00:02:46,474 --> 00:02:47,775 how these collaborations 76 00:02:48,155 --> 00:02:50,555 between academic side and health system side can 77 00:02:50,555 --> 00:02:52,394 really make a big difference. Deborah, is there 78 00:02:52,394 --> 00:02:53,854 anything you wanted to add here? 79 00:02:55,040 --> 00:02:56,960 I just echo what Philip said. I think 80 00:02:56,960 --> 00:02:57,599 this is, 81 00:02:58,159 --> 00:03:00,639 for us, the opportunity for us to bring 82 00:03:00,639 --> 00:03:03,360 together both the benefits of the what the 83 00:03:03,360 --> 00:03:05,060 university brings in terms 84 00:03:05,360 --> 00:03:05,919 of their, 85 00:03:06,480 --> 00:03:09,300 advanced research and what they bring from biomedical 86 00:03:09,895 --> 00:03:12,935 research and AI development. But BJC brings the 87 00:03:12,935 --> 00:03:15,254 place where we can operationalize that and scale 88 00:03:15,254 --> 00:03:17,974 that. So it's a really effective partnership that 89 00:03:17,974 --> 00:03:20,794 we have developed and, looking forward to continuing 90 00:03:20,854 --> 00:03:23,034 to advance AI capabilities through that. 91 00:03:24,150 --> 00:03:25,750 That makes a lot of sense. Now I 92 00:03:25,750 --> 00:03:28,310 know AI has the potential to significantly enhance 93 00:03:28,310 --> 00:03:29,209 patient care. 94 00:03:29,829 --> 00:03:31,909 Deborah, starting with you, can you share some 95 00:03:31,909 --> 00:03:34,229 of the specific ways that technology is being 96 00:03:34,229 --> 00:03:36,949 applied right now to improve patient outcomes and 97 00:03:36,949 --> 00:03:38,329 overall patient experience? 98 00:03:40,014 --> 00:03:40,495 So, 99 00:03:40,974 --> 00:03:43,455 you know, AI is already transforming health care 100 00:03:43,455 --> 00:03:46,754 operations and clinical outcomes both, through our university, 101 00:03:47,534 --> 00:03:49,935 Washington University School of Medicine and BJC Health 102 00:03:49,935 --> 00:03:53,215 Care. So we have, for example, predictive algorithms 103 00:03:53,215 --> 00:03:55,120 are optimizing patient flow, 104 00:03:55,900 --> 00:03:57,840 looking at real time analysis, 105 00:03:58,139 --> 00:04:00,159 historical data to help us identify 106 00:04:00,939 --> 00:04:04,000 patient trajectories and optimize transitions in care. 107 00:04:04,620 --> 00:04:07,275 AI tools are are enhancing early detection and 108 00:04:07,754 --> 00:04:11,674 diagnosis through medical imaging data and identifying conditions 109 00:04:11,674 --> 00:04:13,055 like cancer or cardiovascular 110 00:04:13,435 --> 00:04:13,935 degree, 111 00:04:14,474 --> 00:04:17,454 disease. So we are, really helping 112 00:04:17,914 --> 00:04:21,134 to support clinicians in making timely and precise 113 00:04:21,354 --> 00:04:21,854 decisions. 114 00:04:22,800 --> 00:04:24,500 AI is also being used 115 00:04:24,879 --> 00:04:28,000 to personalize treatment plans by analyzing electronic health 116 00:04:28,000 --> 00:04:29,699 records and genomic data, 117 00:04:30,480 --> 00:04:31,540 to really recommend 118 00:04:32,000 --> 00:04:34,720 therapies that are tailored and particular in fields 119 00:04:34,720 --> 00:04:35,460 like oncology. 120 00:04:36,154 --> 00:04:38,014 So predictive analytics are driving, 121 00:04:39,275 --> 00:04:42,395 population health, identifying high risk patients. So there's 122 00:04:42,395 --> 00:04:44,415 so many opportunities where we're seeing, 123 00:04:44,955 --> 00:04:47,375 the demonstration of how AI can, 124 00:04:47,835 --> 00:04:50,819 enhance both the efficacy and the the operations 125 00:04:50,879 --> 00:04:53,040 so that we can provide superior quality of 126 00:04:53,040 --> 00:04:53,540 care. 127 00:04:54,639 --> 00:04:55,939 Philip, what would you add? 128 00:04:56,400 --> 00:04:57,920 No. I mean, I think that was an 129 00:04:57,920 --> 00:04:59,759 excellent summary. I mean, I think a lot 130 00:04:59,759 --> 00:05:01,935 of this is anchored on, I think, 131 00:05:02,415 --> 00:05:04,995 a driving philosophy here, which is that 132 00:05:05,694 --> 00:05:08,675 our goal is to use AI to complement 133 00:05:08,735 --> 00:05:10,735 the unique strengths of our providers and our 134 00:05:10,735 --> 00:05:13,214 clinical workforce so that it's really this combination 135 00:05:13,214 --> 00:05:15,794 of humans and computers working together 136 00:05:16,735 --> 00:05:18,920 that is superior to either working alone. And 137 00:05:18,920 --> 00:05:20,360 I think that's borne out in a lot 138 00:05:20,360 --> 00:05:22,220 of the examples that Deborah gave, 139 00:05:22,680 --> 00:05:24,279 which really come down to how do we 140 00:05:24,279 --> 00:05:27,160 sort through and understand these massive quantities of 141 00:05:27,160 --> 00:05:28,920 data and evidence and knowledge we have and 142 00:05:28,920 --> 00:05:29,914 make better, 143 00:05:30,254 --> 00:05:30,754 safer, 144 00:05:32,375 --> 00:05:35,254 care related decisions for our patients and improve 145 00:05:35,254 --> 00:05:36,935 the the value of that care along the 146 00:05:36,935 --> 00:05:37,435 way. 147 00:05:38,454 --> 00:05:40,055 I love that. It it seems like such 148 00:05:40,055 --> 00:05:43,134 a meaningful way to apply technology into, the 149 00:05:43,134 --> 00:05:44,470 health system and health system and health care 150 00:05:44,470 --> 00:05:44,970 organizations 151 00:05:45,349 --> 00:05:47,750 for patients as well as, the team members? 152 00:05:47,750 --> 00:05:49,909 And and speaking of your team, how do 153 00:05:49,909 --> 00:05:52,629 you envision AI tools improving health care employee 154 00:05:52,629 --> 00:05:53,129 experience? 155 00:05:53,430 --> 00:05:55,669 Are there any early signs of success or 156 00:05:55,669 --> 00:05:57,669 feedback that you've received from those on the 157 00:05:57,669 --> 00:05:58,409 front lines? 158 00:05:59,665 --> 00:06:01,425 Yeah. I mean, I think you're sort of 159 00:06:01,425 --> 00:06:03,264 touching on probably one of the most important 160 00:06:03,264 --> 00:06:03,764 areas 161 00:06:04,064 --> 00:06:05,845 that we have been looking at, 162 00:06:06,625 --> 00:06:08,944 which is how do we remove sort of 163 00:06:08,944 --> 00:06:12,004 these high friction, low value tasks, 164 00:06:13,104 --> 00:06:13,604 that 165 00:06:14,100 --> 00:06:16,819 are taking our providers away from focusing on 166 00:06:16,819 --> 00:06:17,560 our patients? 167 00:06:18,180 --> 00:06:19,620 And in a similar way, how do we 168 00:06:19,620 --> 00:06:21,300 get sort of the computer out of the 169 00:06:21,300 --> 00:06:24,120 middle of those interactions, again, between our 170 00:06:24,660 --> 00:06:27,375 providers and patients so that we restore a 171 00:06:27,375 --> 00:06:29,535 certain degree of humanism to how we practice 172 00:06:29,535 --> 00:06:30,035 medicine. 173 00:06:30,654 --> 00:06:34,415 So, of course, like many large delivery systems, 174 00:06:34,415 --> 00:06:35,074 we've been 175 00:06:35,615 --> 00:06:36,115 leveraging 176 00:06:36,654 --> 00:06:40,175 ambient AI technology to help automate documentation during 177 00:06:40,175 --> 00:06:42,220 patient visits and we've had some very promising 178 00:06:42,759 --> 00:06:45,080 results not only in primary care settings but 179 00:06:45,080 --> 00:06:47,319 now in procedural disciplines and we're looking at 180 00:06:47,319 --> 00:06:48,060 other domains. 181 00:06:48,680 --> 00:06:51,000 And this has been described by some of 182 00:06:51,000 --> 00:06:52,300 our providers as transformative. 183 00:06:53,564 --> 00:06:55,725 We have individuals who used to spend several 184 00:06:55,725 --> 00:06:58,785 hours a day after normal business hours documenting 185 00:06:59,004 --> 00:07:00,764 that are no longer doing that because they're 186 00:07:00,764 --> 00:07:02,525 able to be more efficient and timely with 187 00:07:02,525 --> 00:07:03,185 their documentation 188 00:07:03,725 --> 00:07:05,004 during the course of their, 189 00:07:05,725 --> 00:07:06,705 clinical work. 190 00:07:07,459 --> 00:07:08,279 We're also 191 00:07:08,660 --> 00:07:10,420 using those same types of technologies to look 192 00:07:10,420 --> 00:07:11,540 at how can we do a better job 193 00:07:11,540 --> 00:07:14,180 of summarizing the content of electronic health records 194 00:07:14,180 --> 00:07:15,779 so that we can get the right information 195 00:07:15,779 --> 00:07:18,120 to our providers more quickly, whether they're 196 00:07:18,580 --> 00:07:20,120 getting ready to admit a patient 197 00:07:20,580 --> 00:07:22,819 or they are seeing a new patient for 198 00:07:22,819 --> 00:07:23,854 the first time, 199 00:07:24,235 --> 00:07:26,254 or any number of other use cases where 200 00:07:26,954 --> 00:07:28,894 prior workflows would involve them spending 201 00:07:29,354 --> 00:07:31,914 potentially up to several hours trying to prepare 202 00:07:31,914 --> 00:07:34,074 for these types of clinical encounters. And now 203 00:07:34,074 --> 00:07:36,634 we've turned that summarization activity into sort of 204 00:07:36,634 --> 00:07:39,329 a matter of minutes and mouse clicks. So 205 00:07:39,329 --> 00:07:41,269 for us, it's all about how do we, 206 00:07:41,410 --> 00:07:43,970 again, restore that humanism in medicine and really 207 00:07:43,970 --> 00:07:46,129 let our providers focus on our patients by 208 00:07:46,129 --> 00:07:48,610 taking those other tasks out of the mix 209 00:07:48,610 --> 00:07:51,089 by automating them. So we're really excited and 210 00:07:51,089 --> 00:07:53,009 I think our providers are really excited about 211 00:07:53,009 --> 00:07:53,509 this. 212 00:07:55,165 --> 00:07:57,165 That's amazing to hear. And, certainly, you know, 213 00:07:57,165 --> 00:07:58,045 what what a great, 214 00:07:58,605 --> 00:08:00,764 an interesting way to think about technology and 215 00:08:00,764 --> 00:08:03,564 using that to then give the providers back 216 00:08:03,564 --> 00:08:05,824 time to be more human with their patients 217 00:08:06,205 --> 00:08:08,305 and have that really human to human connection. 218 00:08:08,659 --> 00:08:10,099 Deborah, is there anything else you wanted to 219 00:08:10,099 --> 00:08:12,039 speak to when you think about the employee 220 00:08:12,180 --> 00:08:14,819 experience and those that you're working with, to 221 00:08:14,819 --> 00:08:17,060 really truly elevate the way things are at 222 00:08:17,060 --> 00:08:18,360 BJC Health Care? 223 00:08:19,300 --> 00:08:20,979 The only other thing that I'll add is 224 00:08:20,979 --> 00:08:23,219 that, when we think about what the new 225 00:08:23,219 --> 00:08:25,754 center can can do and how we can 226 00:08:25,754 --> 00:08:29,435 leverage technology, it's it's also our employees and 227 00:08:29,435 --> 00:08:29,935 staff 228 00:08:30,235 --> 00:08:32,875 that might not provide direct patient care. So 229 00:08:32,875 --> 00:08:35,034 where can we look at opportunities to help 230 00:08:35,034 --> 00:08:35,115 our, 231 00:08:36,620 --> 00:08:38,860 our other capabilities, whether they be in our 232 00:08:38,860 --> 00:08:42,079 finance team or supply chain to help with 233 00:08:42,220 --> 00:08:44,699 tools that make day to day life more 234 00:08:44,699 --> 00:08:46,000 efficient, more effective? 235 00:08:46,779 --> 00:08:49,575 So I think the center is is here 236 00:08:49,714 --> 00:08:52,754 primarily to help that clinical that clinician and 237 00:08:52,754 --> 00:08:55,095 that patient experience, but we do see opportunities 238 00:08:55,154 --> 00:08:56,754 that can help us as a health system 239 00:08:56,754 --> 00:08:57,654 to also be 240 00:08:57,955 --> 00:08:59,654 more effective, which really does, 241 00:08:59,955 --> 00:09:02,434 ultimately have a positive impact on the patients 242 00:09:02,434 --> 00:09:03,095 as well. 243 00:09:04,159 --> 00:09:06,159 Absolutely. That's such a great point and and 244 00:09:06,159 --> 00:09:07,059 such an important, 245 00:09:07,600 --> 00:09:09,120 thing to remember when you're looking at the 246 00:09:09,120 --> 00:09:10,879 system as a whole. Now I know the 247 00:09:10,879 --> 00:09:13,199 Center For Health AI will offer opportunities for 248 00:09:13,199 --> 00:09:15,600 medical students and residents to train on AI 249 00:09:15,600 --> 00:09:18,035 driven care. How do you see this influencing 250 00:09:18,035 --> 00:09:20,035 the next generation of health care leaders, and 251 00:09:20,035 --> 00:09:21,795 what skills will they need in order to 252 00:09:21,795 --> 00:09:22,295 succeed, 253 00:09:23,554 --> 00:09:26,355 in an AI enhanced medical landscape? What are 254 00:09:26,355 --> 00:09:27,975 your thoughts on this, doctor Payne? 255 00:09:28,759 --> 00:09:30,200 Yeah. I mean, I think that's a really 256 00:09:30,200 --> 00:09:32,299 important question because the fundamental 257 00:09:32,759 --> 00:09:35,159 environment in which medicine is being practiced is 258 00:09:35,159 --> 00:09:37,339 changing as a function of AI. And so 259 00:09:37,399 --> 00:09:39,559 the question is, really, how do we prepare 260 00:09:39,559 --> 00:09:41,100 our future practitioners, 261 00:09:41,480 --> 00:09:43,754 researchers, and leaders to thrive in that 262 00:09:48,715 --> 00:09:50,794 fortuitous timing for us that we've just gone 263 00:09:50,794 --> 00:09:51,695 through a 264 00:09:52,075 --> 00:09:54,394 refresh of our medical curriculum here at Washington 265 00:09:54,394 --> 00:09:54,894 University. 266 00:09:55,274 --> 00:09:56,715 And as part of that, we thought about 267 00:09:56,715 --> 00:09:58,894 what are really the core areas of expertise 268 00:09:58,955 --> 00:10:01,040 that the physician leaders of the future need 269 00:10:01,040 --> 00:10:02,980 to have. And one of them was clearly, 270 00:10:03,759 --> 00:10:04,259 informatics, 271 00:10:04,639 --> 00:10:06,180 AI and data science. 272 00:10:06,639 --> 00:10:09,139 And that means we're now integrating that into 273 00:10:09,440 --> 00:10:12,320 our medical curriculum. We're teaching our medical students 274 00:10:12,320 --> 00:10:12,945 to understand 275 00:10:13,825 --> 00:10:15,845 how the use of technology impacts 276 00:10:16,225 --> 00:10:18,225 their decision making and how they need to 277 00:10:18,225 --> 00:10:20,304 sort of consider that whether that be for 278 00:10:20,304 --> 00:10:23,504 diagnostic purposes or therapy planning or any number 279 00:10:23,504 --> 00:10:24,725 of other scenarios. 280 00:10:25,940 --> 00:10:28,500 We're also using AI powered simulation tools now 281 00:10:28,500 --> 00:10:31,620 to create really realistic scenarios so students can 282 00:10:31,620 --> 00:10:33,620 practice clinical decision making in sort of a 283 00:10:33,620 --> 00:10:35,379 safe space before they end up in the 284 00:10:35,379 --> 00:10:38,419 clinical environment. And that provides instant feedback and 285 00:10:38,419 --> 00:10:40,634 insights into their performance that are highly tailored. 286 00:10:41,035 --> 00:10:42,975 So much like we talk about precision medicine 287 00:10:43,115 --> 00:10:45,535 for patients, this is really precision curricula. 288 00:10:46,634 --> 00:10:49,115 And then I think much like it has 289 00:10:49,115 --> 00:10:51,514 historically been the case that we teach sort 290 00:10:51,514 --> 00:10:52,174 of future, 291 00:10:52,554 --> 00:10:53,799 again, practitioners 292 00:10:54,339 --> 00:10:56,679 and researchers and leaders to critically evaluate, 293 00:10:57,620 --> 00:10:59,939 the literature. We're now teaching them how to 294 00:10:59,939 --> 00:11:02,120 critically evaluate the output of, 295 00:11:02,899 --> 00:11:04,600 AI tools like predictive algorithms 296 00:11:04,980 --> 00:11:06,914 so that they understand sort of both the 297 00:11:06,914 --> 00:11:08,855 strengths and weaknesses of these types 298 00:11:10,115 --> 00:11:12,514 of predictive tools and how they can and 299 00:11:12,514 --> 00:11:15,715 should influence clinical decision making in equal measure 300 00:11:15,715 --> 00:11:16,215 where 301 00:11:16,595 --> 00:11:18,774 sort of appropriate caution may be necessary. 302 00:11:20,590 --> 00:11:22,269 And the last area that I would emphasize 303 00:11:22,269 --> 00:11:22,769 is 304 00:11:23,149 --> 00:11:25,809 we're also giving our students the opportunity to 305 00:11:26,029 --> 00:11:29,470 really analyze extremely large datasets and gain insights 306 00:11:29,470 --> 00:11:29,970 around 307 00:11:30,750 --> 00:11:33,009 phenomena that might be of interest around 308 00:11:33,444 --> 00:11:33,944 policy 309 00:11:34,404 --> 00:11:36,985 or population health or other sort of 310 00:11:38,084 --> 00:11:40,804 interventions that happen beyond the individual patient in 311 00:11:40,804 --> 00:11:43,284 the exam room or hospital room. And we're 312 00:11:43,284 --> 00:11:45,845 doing that by really democratizing how people access 313 00:11:45,845 --> 00:11:48,404 data and analyze data, again, using AI tools 314 00:11:48,404 --> 00:11:50,440 that sort of guide the trainee through that 315 00:11:50,440 --> 00:11:50,940 process. 316 00:11:51,399 --> 00:11:53,340 So really, when you put that all together, 317 00:11:53,960 --> 00:11:54,860 we're not only 318 00:11:55,559 --> 00:11:57,879 improving the quality and sort of the completeness 319 00:11:57,879 --> 00:12:00,120 of the education we provide, but we're also 320 00:12:00,120 --> 00:12:01,879 ensuring that those students are ready to go 321 00:12:01,879 --> 00:12:03,580 out and practice in a world where 322 00:12:04,200 --> 00:12:06,914 AI is going to be ubiquitous and, you 323 00:12:06,914 --> 00:12:07,975 know, really a 324 00:12:08,274 --> 00:12:10,054 sort of, in some ways, 325 00:12:10,595 --> 00:12:13,554 sort of key dimension of the modern practice 326 00:12:13,554 --> 00:12:14,134 of medicine. 327 00:12:15,634 --> 00:12:17,955 Absolutely. Well, that's fascinating to hear, you know, 328 00:12:17,955 --> 00:12:19,919 and certainly, as you think about those next 329 00:12:20,079 --> 00:12:21,459 generations of medical students, 330 00:12:21,919 --> 00:12:23,360 what a a great tool to have in 331 00:12:23,360 --> 00:12:25,519 their back pocket with AI and understanding its 332 00:12:25,519 --> 00:12:27,059 limitations, but also opportunities. 333 00:12:27,600 --> 00:12:29,039 Deborah, is there anything you wanted to add 334 00:12:29,039 --> 00:12:29,700 in there? 335 00:12:30,799 --> 00:12:33,120 I'll just say I think, BJC wants to 336 00:12:33,120 --> 00:12:34,339 continue to be, 337 00:12:35,105 --> 00:12:37,424 a world class institution that brings world class 338 00:12:37,424 --> 00:12:39,684 providers here, and so we need to stay, 339 00:12:40,384 --> 00:12:42,324 at the forefront of this technology 340 00:12:42,625 --> 00:12:43,125 that, 341 00:12:44,464 --> 00:12:47,105 our providers are learning to use in school, 342 00:12:47,105 --> 00:12:49,184 and we want them to have access to 343 00:12:49,184 --> 00:12:49,684 that, 344 00:12:50,139 --> 00:12:52,399 at a scale of a system our size 345 00:12:52,700 --> 00:12:55,279 and breadth. So, this is why the partnership 346 00:12:55,339 --> 00:12:58,220 between BJC and WashU is so important that 347 00:12:58,220 --> 00:12:59,659 we can continue to do this in a 348 00:12:59,659 --> 00:13:02,299 very unique way, to the medical school and 349 00:13:02,299 --> 00:13:04,159 then the size and scale of the BJC 350 00:13:04,220 --> 00:13:04,720 system. 351 00:13:05,834 --> 00:13:07,514 That's amazing to hear, you know, and really 352 00:13:07,514 --> 00:13:09,194 great that on the health system level, you 353 00:13:09,194 --> 00:13:10,254 can have that foresight 354 00:13:10,634 --> 00:13:12,794 and and support what the medical school is 355 00:13:12,794 --> 00:13:14,714 doing. Now I have a few questions for 356 00:13:14,714 --> 00:13:16,394 you, doctor Payne, and then we'll follow you 357 00:13:16,394 --> 00:13:18,660 back to you, Deborah. But, doctor Payne, one 358 00:13:18,660 --> 00:13:20,980 example that you've shared, is using AI to 359 00:13:20,980 --> 00:13:21,960 predict excessive 360 00:13:22,340 --> 00:13:25,379 blood loss during surgery, which helps improve resource 361 00:13:25,379 --> 00:13:28,259 management. Can you expand how AI is improving 362 00:13:28,259 --> 00:13:32,200 surgical procedures, diagnostic accuracy, and precision medicine overall? 363 00:13:33,245 --> 00:13:34,445 Yeah. I mean, it's a it's a great 364 00:13:34,445 --> 00:13:36,764 question and it's a project I'm particularly excited 365 00:13:36,764 --> 00:13:37,264 about. 366 00:13:37,804 --> 00:13:40,605 There's actually a research team here that spans 367 00:13:40,605 --> 00:13:43,485 not only our Institute For Informatics and our, 368 00:13:43,725 --> 00:13:44,865 Department of Anesthesiology 369 00:13:45,245 --> 00:13:47,024 and Department of Surgery, but also, 370 00:13:47,610 --> 00:13:48,909 key partners at BJC. 371 00:13:49,529 --> 00:13:52,029 And that group has really sort of demonstrated 372 00:13:52,330 --> 00:13:54,809 in a very compelling way how you can 373 00:13:54,809 --> 00:13:56,909 use machine learning models to predict 374 00:13:57,210 --> 00:13:57,710 blood 375 00:13:58,409 --> 00:14:01,514 loss during surgical procedures. And if you're familiar 376 00:14:01,514 --> 00:14:03,355 with this domain, sort of the current state 377 00:14:03,355 --> 00:14:05,134 of the art is to use sort of 378 00:14:05,274 --> 00:14:06,735 historical and largely, 379 00:14:07,194 --> 00:14:09,694 I would say out of date data concerning 380 00:14:09,914 --> 00:14:12,074 predictive blood loss as a function of how 381 00:14:12,074 --> 00:14:13,590 the average patient may 382 00:14:14,389 --> 00:14:15,050 sort of 383 00:14:15,430 --> 00:14:17,990 experience blood loss during surgery. And the result 384 00:14:17,990 --> 00:14:20,149 was a lot of blood products being used 385 00:14:20,149 --> 00:14:20,970 in the OR 386 00:14:22,149 --> 00:14:24,389 that were necessary but an equal measure of 387 00:14:24,389 --> 00:14:26,070 blood products being sent to the OR that 388 00:14:26,070 --> 00:14:28,550 were never used and therefore wasted. And so 389 00:14:28,550 --> 00:14:29,684 this team actually used 390 00:14:30,065 --> 00:14:32,625 over 4,000,000 surgical records from the American College 391 00:14:32,625 --> 00:14:36,084 of Surgeons National Surgical Quality Improvement Program database 392 00:14:36,544 --> 00:14:39,504 and actually built a highly accurate model that 393 00:14:39,504 --> 00:14:41,284 recommended type and screen orders, 394 00:14:42,065 --> 00:14:45,379 for patients during the sort of perioperative management 395 00:14:46,240 --> 00:14:47,540 sort of period. 396 00:14:47,920 --> 00:14:48,500 And actually, 397 00:14:49,200 --> 00:14:51,680 this resulted in, if you sort of compare 398 00:14:51,680 --> 00:14:54,720 the traditional model versus current model, a difference 399 00:14:54,720 --> 00:14:55,540 of 96% 400 00:14:55,920 --> 00:14:57,220 sensitivity in anticipating 401 00:14:58,720 --> 00:15:01,634 perioperative blood loss using the machine learning model 402 00:15:01,634 --> 00:15:03,815 as compared to sort of around 50% 403 00:15:04,674 --> 00:15:05,174 sort 404 00:15:05,554 --> 00:15:07,875 of sensitivity in a baseline model. And that 405 00:15:07,875 --> 00:15:10,995 results in real savings of blood products. So, 406 00:15:10,995 --> 00:15:12,419 in a lot of ways, this is about 407 00:15:12,419 --> 00:15:13,320 how we can use personalized 408 00:15:13,779 --> 00:15:16,500 data driven interventions in order to enhance that 409 00:15:16,500 --> 00:15:19,799 perioperative care period and at the same time, 410 00:15:20,259 --> 00:15:22,019 really be much more judicious in how we 411 00:15:22,019 --> 00:15:22,519 manage 412 00:15:22,820 --> 00:15:25,539 a very finite resource that's critical to surgical 413 00:15:25,539 --> 00:15:27,294 safety. So, that's I think, a a great 414 00:15:27,294 --> 00:15:29,315 example of how all these pieces come together. 415 00:15:31,134 --> 00:15:33,054 That it really colors the work that you're 416 00:15:33,054 --> 00:15:33,454 doing, 417 00:15:34,014 --> 00:15:36,735 well and and spotlights the outcomes and impact 418 00:15:36,735 --> 00:15:38,334 it can have not only on patients but 419 00:15:38,334 --> 00:15:41,070 the organization too. You mentioned AI can help 420 00:15:41,070 --> 00:15:44,110 identify the most effective treatments for individual patients 421 00:15:44,110 --> 00:15:47,309 as well. How is this process being applied 422 00:15:47,309 --> 00:15:49,250 in practice at WashU and BJC, 423 00:15:49,710 --> 00:15:51,809 especially for the complex or rare diseases? 424 00:15:52,934 --> 00:15:53,735 Yeah. So, 425 00:15:55,415 --> 00:15:56,934 I think one of the things that we 426 00:15:56,934 --> 00:15:59,115 know is that especially in rare disease, 427 00:15:59,415 --> 00:16:01,975 the availability of therapeutic options is very limited 428 00:16:01,975 --> 00:16:05,095 because they are frankly not as appealing to 429 00:16:05,095 --> 00:16:05,595 pharmaceutical 430 00:16:05,975 --> 00:16:08,769 or bio technology companies. And so we often 431 00:16:08,769 --> 00:16:10,549 are in a position where we have to 432 00:16:10,690 --> 00:16:12,629 create tailored therapeutic 433 00:16:13,330 --> 00:16:14,870 approaches for those individuals. 434 00:16:15,490 --> 00:16:17,029 Great examples, we have a research 435 00:16:17,490 --> 00:16:20,309 program here working with our neurofibromatosis 436 00:16:21,090 --> 00:16:23,144 type 1 center or NF1 center. And this 437 00:16:23,144 --> 00:16:23,644 is, 438 00:16:23,945 --> 00:16:26,745 the most common rare genetic diseases in children. 439 00:16:26,745 --> 00:16:28,684 It's relatively easily diagnosed. 440 00:16:29,384 --> 00:16:31,705 But then there's a multitude of phenotypes that 441 00:16:31,705 --> 00:16:33,784 patients will present with. Some patients will be 442 00:16:33,784 --> 00:16:36,584 fine and will not experience any meaningful symptoms. 443 00:16:36,584 --> 00:16:37,565 Some will develop 444 00:16:37,940 --> 00:16:40,839 optic pathway gliomas or plexiform neurofibromas, 445 00:16:41,139 --> 00:16:43,459 which are painful nerve sheet tumors. Some will 446 00:16:43,459 --> 00:16:46,339 develop ADHD. Some will develop scoliosis. I could 447 00:16:46,339 --> 00:16:48,419 keep on going down a lengthy list. And 448 00:16:48,419 --> 00:16:50,019 the problem is we're not very good at 449 00:16:50,019 --> 00:16:50,519 anticipating 450 00:16:50,899 --> 00:16:53,079 which of those phenotypes patients will 451 00:16:53,495 --> 00:16:55,254 exhibit it. So we tend to manage those 452 00:16:55,254 --> 00:16:58,394 patients extremely aggressively trying to find early signs 453 00:16:58,855 --> 00:17:01,495 of those various sort of deleterious outcomes, but 454 00:17:01,495 --> 00:17:03,274 we're not always good at finding them. 455 00:17:03,975 --> 00:17:06,294 And we've actually been able to, again, using 456 00:17:06,294 --> 00:17:09,000 artificial intelligence methods and a combination of imaging 457 00:17:09,000 --> 00:17:12,279 data, data from the EHR and then various 458 00:17:12,279 --> 00:17:12,779 neurobehavioral 459 00:17:13,079 --> 00:17:15,500 phenotyping instruments applied to patients directly, 460 00:17:16,440 --> 00:17:18,700 we're able to generate highly accurate 461 00:17:19,079 --> 00:17:21,339 predictive models that can be used to forecast 462 00:17:21,805 --> 00:17:23,565 sort of the disease course for children with 463 00:17:23,565 --> 00:17:25,565 NF1. And so what that means now is 464 00:17:25,565 --> 00:17:28,285 instead of every child having to have a 465 00:17:28,285 --> 00:17:30,865 large amount of imaging or other potentially challenging 466 00:17:31,005 --> 00:17:31,505 diagnostic 467 00:17:32,285 --> 00:17:35,085 procedures, we can focus on those individuals who 468 00:17:35,085 --> 00:17:37,420 are most at risk and often detect the 469 00:17:37,420 --> 00:17:40,240 outcomes earlier, which leads to better therapeutic outcomes. 470 00:17:41,019 --> 00:17:43,180 And I often say to people, sometimes precision 471 00:17:43,180 --> 00:17:45,580 medicine is about finding the right treatment, but 472 00:17:45,580 --> 00:17:48,700 sometimes precision medicine is just about demystifying diseases. 473 00:17:48,700 --> 00:17:50,460 So we know what to expect and we 474 00:17:50,460 --> 00:17:52,140 can be smarter about how we manage those 475 00:17:52,140 --> 00:17:53,494 symptoms. And I think this is a great 476 00:17:53,494 --> 00:17:56,454 example where we're getting smarter using AI so 477 00:17:56,454 --> 00:17:58,454 that we can be very precise in how 478 00:17:58,454 --> 00:17:59,194 we manage 479 00:17:59,494 --> 00:18:01,974 the clinical course of kids with a very 480 00:18:01,974 --> 00:18:03,355 challenging genetic disease. 481 00:18:04,214 --> 00:18:05,460 And we need to do more of that. 482 00:18:05,700 --> 00:18:07,380 There's no doubt that we have a wealth 483 00:18:07,380 --> 00:18:09,539 of data to build these models. It's just 484 00:18:09,539 --> 00:18:11,220 that we need to start doing the hard 485 00:18:11,220 --> 00:18:12,919 work of bringing those data together, 486 00:18:13,460 --> 00:18:16,200 working with our clinical colleagues and frankly running 487 00:18:16,740 --> 00:18:19,140 large studies to validate that the models provide 488 00:18:19,140 --> 00:18:21,325 clinical value, which is another big part of 489 00:18:21,325 --> 00:18:23,265 the project I just described to you. 490 00:18:24,365 --> 00:18:26,445 Absolutely. Thank you so much for that example. 491 00:18:26,445 --> 00:18:28,684 And real quick, what steps are you taking 492 00:18:28,684 --> 00:18:30,765 to ensure that those AI technologies are used 493 00:18:30,765 --> 00:18:32,785 responsibly and safely in patient care? 494 00:18:33,609 --> 00:18:34,109 Yeah. 495 00:18:34,569 --> 00:18:37,210 That is a huge and really important question. 496 00:18:37,210 --> 00:18:37,710 Right? 497 00:18:38,409 --> 00:18:39,769 How do we make sure that when we 498 00:18:39,769 --> 00:18:41,710 deploy ai, it's safe, it's efficacious 499 00:18:42,009 --> 00:18:43,789 that we are considering the ethical 500 00:18:44,089 --> 00:18:46,585 and legal and social implications of these emerging 501 00:18:46,585 --> 00:18:47,484 technologies. And 502 00:18:47,785 --> 00:18:49,785 I would start by saying we're committed across 503 00:18:49,785 --> 00:18:52,184 Washington BJC to making sure that everything we 504 00:18:52,184 --> 00:18:54,125 do through this new Center For Health AI 505 00:18:54,585 --> 00:18:55,085 emphasizes 506 00:18:55,944 --> 00:18:58,105 delivering these tools in a secure, scalable and 507 00:18:58,105 --> 00:19:00,029 reliable manner, but also in a way that 508 00:19:00,029 --> 00:19:02,130 prioritizes patient safety and privacy 509 00:19:02,670 --> 00:19:03,490 and confidentiality. 510 00:19:04,190 --> 00:19:05,549 And that means we have to really sort 511 00:19:05,549 --> 00:19:06,990 of use what is often referred to as 512 00:19:06,990 --> 00:19:09,150 sort of a privacy by design approach, which 513 00:19:09,150 --> 00:19:12,430 means that data governance and the application of 514 00:19:12,430 --> 00:19:16,065 appropriate regulatory standards as well as more technical 515 00:19:16,065 --> 00:19:16,565 controls, 516 00:19:16,944 --> 00:19:18,565 and then engagement of 517 00:19:18,865 --> 00:19:21,265 key stakeholder groups to define early and often 518 00:19:21,265 --> 00:19:21,924 the ethical 519 00:19:22,464 --> 00:19:24,865 frameworks in which these tools exist, all of 520 00:19:24,865 --> 00:19:26,784 that has to happen at the beginning of 521 00:19:26,784 --> 00:19:28,784 projects, not at the end of projects. And 522 00:19:28,784 --> 00:19:30,704 I think one of our challenges often is 523 00:19:30,704 --> 00:19:32,960 that we treat these, again, ethical, legal, and 524 00:19:32,960 --> 00:19:34,419 social implications of technologies 525 00:19:35,039 --> 00:19:35,700 as something 526 00:19:36,319 --> 00:19:38,500 that is assessed after the fact. 527 00:19:38,879 --> 00:19:41,039 But the likelihood of success is greatly diminished 528 00:19:41,039 --> 00:19:43,039 when we do that. And I think the 529 00:19:43,039 --> 00:19:45,434 cornerstone of all of this is transparency, right? 530 00:19:45,434 --> 00:19:47,694 Making sure that we're empowering our patients, 531 00:19:47,994 --> 00:19:50,474 our providers and all the other stakeholders involved 532 00:19:50,474 --> 00:19:52,394 in our delivery system so they know when 533 00:19:52,394 --> 00:19:54,255 and how we're using AI. 534 00:19:55,595 --> 00:19:56,494 And I 535 00:19:57,230 --> 00:19:59,970 think ultimately patient safety and these ethical considerations, 536 00:20:00,190 --> 00:20:02,269 they're gonna be central to every decision we 537 00:20:02,269 --> 00:20:03,650 make about AI deployment. 538 00:20:04,269 --> 00:20:05,950 And that means that the teams that we 539 00:20:05,950 --> 00:20:07,650 put together will have to be interdisciplinary 540 00:20:08,029 --> 00:20:10,465 and include not only technologists and clinicians 541 00:20:10,924 --> 00:20:14,144 but ethicists and patient advocates and community representatives. 542 00:20:14,525 --> 00:20:16,144 We're gonna have to make sure that we 543 00:20:16,365 --> 00:20:18,384 focus on and commit to really rigorous 544 00:20:18,765 --> 00:20:20,545 validation of these tools that includes 545 00:20:20,924 --> 00:20:23,484 an assessment of sort of the ethical and 546 00:20:23,484 --> 00:20:25,789 social impact of their use. And then we 547 00:20:25,789 --> 00:20:28,109 have to create these continuous monitoring and feedback 548 00:20:28,109 --> 00:20:31,309 loops so that we understand unintended consequences or 549 00:20:31,309 --> 00:20:31,809 outcomes 550 00:20:32,190 --> 00:20:34,430 of the use of AI rather than treating 551 00:20:34,430 --> 00:20:36,589 the process as somewhat unidirectional where we declare 552 00:20:36,589 --> 00:20:39,154 success when the tool works, deploy it in 553 00:20:39,154 --> 00:20:40,914 the healthcare delivery system, and then move on 554 00:20:40,914 --> 00:20:41,815 to the next project. 555 00:20:42,755 --> 00:20:43,815 And I would say, 556 00:20:44,674 --> 00:20:47,075 much as people have talked about the need 557 00:20:47,075 --> 00:20:49,875 to be vigilant about adverse outcomes with new 558 00:20:49,875 --> 00:20:52,269 therapeutics as they come to market. I think 559 00:20:52,269 --> 00:20:54,190 the same will be very much the case 560 00:20:54,190 --> 00:20:54,930 with AI. 561 00:20:55,549 --> 00:20:57,789 And I'll just wrap up by saying, in 562 00:20:57,789 --> 00:20:59,490 my mind, it is our responsibility 563 00:20:59,869 --> 00:21:01,410 at Washington Medicine and BJC 564 00:21:01,789 --> 00:21:02,769 to set the benchmark 565 00:21:03,375 --> 00:21:05,695 for responsible, safe, and ethical AI innovation in 566 00:21:05,695 --> 00:21:08,095 health care. So I don't believe we're satisfied 567 00:21:08,095 --> 00:21:10,894 just with, adopting best practices from others, but 568 00:21:10,894 --> 00:21:12,974 we really wanna set the standard for what 569 00:21:12,974 --> 00:21:14,355 those best practices are. 570 00:21:15,349 --> 00:21:17,349 That's amazing to hear. A big job, but 571 00:21:17,349 --> 00:21:19,289 one certainly that you're set up well-to-do 572 00:21:20,230 --> 00:21:20,730 responsibly. 573 00:21:21,269 --> 00:21:23,430 Now, Deborah, from your perspective, one of the 574 00:21:23,430 --> 00:21:26,309 primary goals of the center is to improve 575 00:21:26,309 --> 00:21:29,865 operational efficiency, including streamlining scheduling and reducing administrative 576 00:21:29,924 --> 00:21:31,765 burdens. Can you walk us through some of 577 00:21:31,765 --> 00:21:34,404 the AI tools currently being piloted and how 578 00:21:34,404 --> 00:21:36,404 they're making a tangible difference for health care 579 00:21:36,404 --> 00:21:37,545 providers and patients? 580 00:21:38,565 --> 00:21:39,545 Yeah. We're actually, 581 00:21:39,924 --> 00:21:43,259 really looking at what our current vendor partners 582 00:21:43,480 --> 00:21:45,799 offer embedded in tools that we might already 583 00:21:45,799 --> 00:21:47,240 have. And I think this is very common 584 00:21:47,240 --> 00:21:47,740 across, 585 00:21:48,599 --> 00:21:50,599 the health care environment today that we're in 586 00:21:50,599 --> 00:21:52,619 any environment, actually, where they're seeing, 587 00:21:53,559 --> 00:21:54,059 existing 588 00:21:54,680 --> 00:21:57,315 software and tools that we're using and they're 589 00:21:57,315 --> 00:21:59,894 adding in components. So whether that be our 590 00:22:00,115 --> 00:22:02,694 EHR vendor or our ERP vendor, 591 00:22:03,234 --> 00:22:05,494 we are we are looking at those opportunities 592 00:22:05,634 --> 00:22:07,554 to see how we can implement them. I 593 00:22:07,554 --> 00:22:08,054 think 594 00:22:08,355 --> 00:22:09,634 as Philip and I, 595 00:22:09,954 --> 00:22:11,654 are standing up this new center, 596 00:22:12,000 --> 00:22:14,019 we really talk about how we can 597 00:22:14,320 --> 00:22:16,960 help move these projects along, how we can 598 00:22:16,960 --> 00:22:17,460 unstick, 599 00:22:18,240 --> 00:22:21,220 where things get get stuck with new technology 600 00:22:21,440 --> 00:22:21,940 and, 601 00:22:22,400 --> 00:22:25,515 implementation and change of processes. So we're looking 602 00:22:25,515 --> 00:22:27,275 at really how we can leverage what we 603 00:22:27,275 --> 00:22:28,974 already have today as well as, 604 00:22:29,595 --> 00:22:32,154 the experience and knowledge of all the examples 605 00:22:32,154 --> 00:22:34,414 that Philip just gave that come from our 606 00:22:34,555 --> 00:22:36,815 university research and provider side. 607 00:22:37,355 --> 00:22:39,434 So we have many different pilots going on 608 00:22:39,434 --> 00:22:41,990 in those spaces where it might feel, 609 00:22:42,450 --> 00:22:44,849 like, we don't see it every day, because 610 00:22:44,849 --> 00:22:45,349 we're 611 00:22:46,369 --> 00:22:46,869 processing 612 00:22:47,570 --> 00:22:50,789 payment or we're ordering supplies or we're working, 613 00:22:51,329 --> 00:22:52,230 through vendor 614 00:22:52,529 --> 00:22:55,009 agree legal agreement. But those are places where 615 00:22:55,009 --> 00:22:55,509 AI 616 00:22:55,904 --> 00:22:57,904 tools can really help us be more effective 617 00:22:57,904 --> 00:22:59,845 and efficient and move things through faster. 618 00:23:01,105 --> 00:23:03,424 That's great to hear. You mentioned that AI 619 00:23:03,424 --> 00:23:05,744 could help predict demand for equipment and staff 620 00:23:05,744 --> 00:23:06,244 resources. 621 00:23:06,545 --> 00:23:08,225 What are some of the real world examples 622 00:23:08,225 --> 00:23:10,590 of how AI is improving logistics and resource 623 00:23:10,590 --> 00:23:13,070 management in health systems, especially in large and 624 00:23:13,070 --> 00:23:14,930 complex organizations like BJC? 625 00:23:16,430 --> 00:23:18,350 I think today in health care, a lot 626 00:23:18,350 --> 00:23:20,670 of health systems are really struggling with similar 627 00:23:20,670 --> 00:23:21,809 things around capacity, 628 00:23:22,734 --> 00:23:26,034 and throughput. And so understanding how we expect 629 00:23:27,214 --> 00:23:29,134 people to come through, the needs and equipment 630 00:23:29,134 --> 00:23:30,674 to come through, whether it's shortages 631 00:23:31,214 --> 00:23:32,914 that we have in certain supplies, 632 00:23:33,375 --> 00:23:35,990 and being able to project and predict when 633 00:23:35,990 --> 00:23:37,509 those things are going to happen and what 634 00:23:37,509 --> 00:23:39,130 the impact is going to be. 635 00:23:39,750 --> 00:23:40,569 So I think 636 00:23:40,950 --> 00:23:43,109 most of the country is experiencing some of 637 00:23:43,109 --> 00:23:45,349 those at any given time, whether that be, 638 00:23:45,589 --> 00:23:48,149 we have backups in our emergency department or 639 00:23:48,149 --> 00:23:48,809 we have, 640 00:23:49,109 --> 00:23:50,865 IV shortages due to 641 00:23:51,265 --> 00:23:51,765 unforeseen, 642 00:23:52,785 --> 00:23:54,005 world situations, 643 00:23:54,704 --> 00:23:55,505 we need to, 644 00:23:55,904 --> 00:23:58,484 we need to be move quickly at helping 645 00:23:59,105 --> 00:24:01,904 improve how we can answer those problems and 646 00:24:01,904 --> 00:24:05,190 address those resource needs. And so, those are 647 00:24:05,190 --> 00:24:06,630 things that we do on a daily basis, 648 00:24:06,630 --> 00:24:08,549 and we're hoping the new center can really 649 00:24:08,549 --> 00:24:10,390 help us to figure out how to continue 650 00:24:10,390 --> 00:24:12,329 to drive those sorts of decision making. 651 00:24:13,990 --> 00:24:15,750 I love that. What a great use case 652 00:24:15,750 --> 00:24:17,990 and and really strong example. Before we wrap 653 00:24:17,990 --> 00:24:19,565 up here, I have one more quick question. 654 00:24:19,704 --> 00:24:21,944 How do you plan to scale successful AI 655 00:24:21,944 --> 00:24:24,924 tools and solutions from pilot projects to widespread 656 00:24:24,984 --> 00:24:27,384 adoption across the health system? What challenges do 657 00:24:27,384 --> 00:24:29,544 you anticipate in this process, and how can 658 00:24:29,544 --> 00:24:31,085 the center help overcome them? 659 00:24:32,140 --> 00:24:34,859 I think, the the challenges are gonna be 660 00:24:34,859 --> 00:24:37,500 challenges with any new technology, new change, new 661 00:24:37,500 --> 00:24:38,000 processes, 662 00:24:39,019 --> 00:24:41,759 learning what works, what doesn't, trying some things. 663 00:24:42,539 --> 00:24:44,240 Process change is always, 664 00:24:44,975 --> 00:24:47,055 getting some people will be on board. Some 665 00:24:47,055 --> 00:24:48,595 people will need more, 666 00:24:49,934 --> 00:24:52,174 support and coaching around how this can be 667 00:24:52,174 --> 00:24:53,955 different and how this can change their, 668 00:24:54,494 --> 00:24:57,394 their workflow and their day. So I think, 669 00:24:58,820 --> 00:25:00,900 luckily, I think Philip and I are privileged 670 00:25:00,900 --> 00:25:01,960 to work in an organization 671 00:25:02,339 --> 00:25:02,839 that, 672 00:25:03,220 --> 00:25:06,180 really wants to learn and grow, and we 673 00:25:06,180 --> 00:25:08,099 are often just trying to keep up with 674 00:25:08,099 --> 00:25:10,099 the needs and demand around how we can 675 00:25:10,099 --> 00:25:12,740 do things more efficiently, more effectively, and provide 676 00:25:12,740 --> 00:25:13,559 better care. 677 00:25:13,974 --> 00:25:16,534 So I think the center will we will, 678 00:25:17,095 --> 00:25:20,134 we're really here to help our leaders and 679 00:25:20,134 --> 00:25:20,954 our providers 680 00:25:21,494 --> 00:25:22,934 do the best that they can for our 681 00:25:22,934 --> 00:25:25,115 patients. So however that, 682 00:25:25,909 --> 00:25:27,909 however that manifests itself in what we can 683 00:25:27,909 --> 00:25:30,230 do every day to support them, I think, 684 00:25:30,630 --> 00:25:33,029 I'm excited about it. I'm excited to continue 685 00:25:33,029 --> 00:25:34,869 to work with Philip in in how we 686 00:25:34,869 --> 00:25:37,829 can scale this and really support what our 687 00:25:37,829 --> 00:25:39,769 providers do every day for our community. 688 00:25:41,164 --> 00:25:43,484 That's amazing to hear. Doctor Payne, Deborah, thank 689 00:25:43,484 --> 00:25:44,605 you so much for joining us on the 690 00:25:44,605 --> 00:25:46,765 podcast today. This has been such a full 691 00:25:46,765 --> 00:25:48,444 and rich conversation, and I look forward to 692 00:25:48,444 --> 00:25:50,525 connecting with you again soon and catching up 693 00:25:50,525 --> 00:25:51,884 on some of the great things that you 694 00:25:51,884 --> 00:25:53,505 continue to do with the center. 695 00:25:54,045 --> 00:25:55,424 Well, thank you for the opportunity.