1 00:00:00,915 --> 00:00:02,826 This is Laura D with the Becker Health 2 00:00:02,826 --> 00:00:03,542 care podcast. 3 00:00:03,940 --> 00:00:05,533 I'm thrilled today to be joined by Doctor. 4 00:00:05,692 --> 00:00:08,814 Y T mitsubishi, Chief Medical Informatics officer an 5 00:00:08,814 --> 00:00:11,451 interim medical director for Ovation and Director of 6 00:00:11,451 --> 00:00:13,928 clinical research informatics at the Metro health system. 7 00:00:14,408 --> 00:00:15,766 Doctor. T she she's a pleasure to have 8 00:00:15,766 --> 00:00:16,725 you on the podcast today. 9 00:00:17,618 --> 00:00:19,768 Thank you for having me. I'm really looking 10 00:00:19,768 --> 00:00:21,758 forward to our discussion and diving a little 11 00:00:21,758 --> 00:00:23,429 bit deeper to what you're doing at both 12 00:00:23,429 --> 00:00:24,942 ovation as well as the Metro health system, 13 00:00:25,101 --> 00:00:27,195 but before we dive in, I was wondering 14 00:00:27,195 --> 00:00:28,155 if you could tell us a little bit 15 00:00:28,155 --> 00:00:29,355 more about yourself in your background? 16 00:00:30,315 --> 00:00:32,954 Sure. So I am a pulmonary and critical 17 00:00:32,954 --> 00:00:34,727 care physician and position in 18 00:00:35,761 --> 00:00:37,988 Who's been doing a lot more of the 19 00:00:37,988 --> 00:00:39,102 latter in my career. 20 00:00:40,056 --> 00:00:42,225 I've been at Mental Health now for almost 21 00:00:42,225 --> 00:00:42,784 9 years. 22 00:00:43,424 --> 00:00:43,924 And 23 00:00:44,545 --> 00:00:45,045 increasingly 24 00:00:46,144 --> 00:00:46,644 doing 25 00:00:47,344 --> 00:00:50,144 things in the data science and clinical decision 26 00:00:50,144 --> 00:00:52,630 support space as well as more recently in 27 00:00:52,630 --> 00:00:54,243 the virtual care space 28 00:00:54,618 --> 00:00:56,765 through this ovation venture, which is a metro 29 00:00:56,765 --> 00:00:57,083 health and 30 00:00:58,991 --> 00:01:00,684 Collaboration. So multi health care system 31 00:01:02,031 --> 00:01:04,742 virtual care platform, which is built frankly on 32 00:01:04,742 --> 00:01:05,242 interoperability 33 00:01:05,619 --> 00:01:08,729 and system integration as a differentiating factor. 34 00:01:12,406 --> 00:01:14,474 That's really cool to hear. And and definitely 35 00:01:14,474 --> 00:01:17,018 exciting to see how the technology can really 36 00:01:17,018 --> 00:01:18,609 inform and influence the things you're doing at 37 00:01:18,609 --> 00:01:20,770 the health system level as well. And and 38 00:01:20,770 --> 00:01:22,994 given your perspectives in both fronts, what are 39 00:01:22,994 --> 00:01:24,424 some of the biggest issues that you're following 40 00:01:24,424 --> 00:01:25,536 in healthcare today? 41 00:01:27,046 --> 00:01:28,261 Sure. So I think 42 00:01:28,634 --> 00:01:30,040 artificial intelligence is 43 00:01:30,475 --> 00:01:31,664 obviously on everybody's in mind. 44 00:01:33,170 --> 00:01:34,304 When I think about 45 00:01:35,468 --> 00:01:37,237 healthcare systems through the 46 00:01:37,624 --> 00:01:40,113 clinical informatics lens, a lot of it 47 00:01:40,488 --> 00:01:43,432 least from our perspective right, sits on this 48 00:01:43,432 --> 00:01:44,489 cushion of data 49 00:01:44,879 --> 00:01:47,024 and how we can turn data, 50 00:01:47,660 --> 00:01:50,544 data input into meaningful outputs to help drive 51 00:01:50,918 --> 00:01:51,395 patient care. 52 00:01:52,284 --> 00:01:53,712 And so as you think about kind of 53 00:01:53,712 --> 00:01:55,854 the evolution of where we've been and what 54 00:01:55,854 --> 00:01:57,203 we've been able to achieve in the health 55 00:01:57,203 --> 00:01:58,100 informatics space 56 00:01:58,631 --> 00:01:59,131 around 57 00:01:59,679 --> 00:02:01,914 really getting everybody on an electronic health records, 58 00:02:02,153 --> 00:02:02,653 standardizing 59 00:02:03,031 --> 00:02:04,867 data to the degree that we've been able 60 00:02:04,867 --> 00:02:07,261 to do so and increasingly more looking at 61 00:02:07,261 --> 00:02:07,761 interoperability 62 00:02:08,472 --> 00:02:10,244 I think that's created an incredible 63 00:02:10,698 --> 00:02:13,185 really charged atmosphere or 64 00:02:14,115 --> 00:02:16,285 bigger and better things that we can do 65 00:02:16,341 --> 00:02:17,238 with this data 66 00:02:17,548 --> 00:02:20,492 including obviously implementing Ai intervention. 67 00:02:21,208 --> 00:02:21,947 And so 68 00:02:22,322 --> 00:02:23,379 I think that that 69 00:02:24,152 --> 00:02:26,005 really becomes the key driver 70 00:02:26,634 --> 00:02:28,708 of where we go next? We can talk 71 00:02:28,708 --> 00:02:31,021 a lot about some of those aspects. So 72 00:02:31,100 --> 00:02:32,478 I think about things like 73 00:02:33,174 --> 00:02:35,668 how does data standards and interoperability 74 00:02:37,097 --> 00:02:37,597 drive 75 00:02:38,692 --> 00:02:40,388 evolution in the Ai 76 00:02:40,845 --> 00:02:42,041 clinical and support space, 77 00:02:42,759 --> 00:02:44,433 and that comes down to things like being 78 00:02:44,433 --> 00:02:46,943 able to build and scale and fed 79 00:02:47,400 --> 00:02:48,778 learning and models 80 00:02:49,235 --> 00:02:51,251 in a way that really is meaningful 81 00:02:51,868 --> 00:02:52,906 and externally valid. 82 00:02:53,479 --> 00:02:55,077 And also how do we bring those insights 83 00:02:55,077 --> 00:02:56,376 and models 84 00:02:56,755 --> 00:02:59,312 and knowledge back into our healthcare system. And 85 00:02:59,312 --> 00:03:00,850 so talking about things like 86 00:03:01,481 --> 00:03:03,469 Ai assurance labs and how do we scale 87 00:03:03,469 --> 00:03:06,434 those? Like how do we know that models 88 00:03:07,365 --> 00:03:10,228 approved or sanctioned by other organizations or entities? 89 00:03:10,959 --> 00:03:12,471 Really come back into our system and make 90 00:03:12,471 --> 00:03:14,143 a difference. So a lot of what I 91 00:03:14,143 --> 00:03:15,200 focused this on is 92 00:03:15,973 --> 00:03:17,269 thinking about Ai 93 00:03:17,645 --> 00:03:18,145 implementation 94 00:03:18,679 --> 00:03:19,896 at the local level 95 00:03:20,843 --> 00:03:23,567 and us kind of reached this conclusion that 96 00:03:23,941 --> 00:03:27,278 really Ai is a implementation, frankly is a 97 00:03:27,278 --> 00:03:27,778 local 98 00:03:28,152 --> 00:03:28,652 concept 99 00:03:29,359 --> 00:03:30,976 and the value of a model 100 00:03:31,592 --> 00:03:32,810 is only really 101 00:03:34,304 --> 00:03:34,804 made 102 00:03:35,580 --> 00:03:37,893 obvious when it's implemented and valid at the 103 00:03:37,893 --> 00:03:38,452 local setting. 104 00:03:39,264 --> 00:03:41,175 And so I like to think about how 105 00:03:41,175 --> 00:03:41,994 all of these 106 00:03:42,370 --> 00:03:43,985 tools in these ecosystems 107 00:03:44,600 --> 00:03:45,578 evolve to 108 00:03:46,352 --> 00:03:47,547 support that mission. 109 00:03:48,199 --> 00:03:49,390 And I get we can get into some 110 00:03:49,390 --> 00:03:51,534 of the details, but I think that's both 111 00:03:51,534 --> 00:03:53,757 the opportunity and I think the excitement in 112 00:03:53,757 --> 00:03:54,234 this space. 113 00:03:55,266 --> 00:03:56,957 I think about the challenges it's 114 00:03:57,663 --> 00:03:58,638 really the hype 115 00:03:59,010 --> 00:03:59,510 and 116 00:04:00,040 --> 00:04:02,656 making sure as an informatics as a physician 117 00:04:02,656 --> 00:04:04,265 executive, you're not making the wrong 118 00:04:05,367 --> 00:04:06,423 decisions and choices 119 00:04:06,956 --> 00:04:09,261 when it comes to which solutions to implement, 120 00:04:09,420 --> 00:04:11,248 right? So the obvious 1 being saying away 121 00:04:11,248 --> 00:04:13,894 from point solutions and making good solid 122 00:04:14,683 --> 00:04:18,054 decisions that enhance your digital ecosystem and infrastructure 123 00:04:18,511 --> 00:04:19,410 and that can 124 00:04:19,788 --> 00:04:21,542 allow you to build on top of that 125 00:04:21,542 --> 00:04:22,282 and develop 126 00:04:22,659 --> 00:04:24,984 the Ai driven analytics that are going to 127 00:04:24,984 --> 00:04:27,208 propel your organization and healthcare forward. 128 00:04:28,638 --> 00:04:30,488 So very high level generic 129 00:04:30,862 --> 00:04:31,974 take on the Ai space. 130 00:04:33,105 --> 00:04:34,625 But I'm happy to get into the details 131 00:04:34,625 --> 00:04:35,345 if you're interested. 132 00:04:36,625 --> 00:04:38,865 Yeah. Absolutely. And I I really appreciate your 133 00:04:38,865 --> 00:04:41,113 analysis here because you know, I I know 134 00:04:41,113 --> 00:04:43,102 some money leaders and organizations are trying to 135 00:04:43,102 --> 00:04:45,011 figure out the best way to move forward 136 00:04:45,011 --> 00:04:47,814 with Ai and technology and the different types 137 00:04:47,814 --> 00:04:50,130 of investments that they're trying to make. Now 138 00:04:50,130 --> 00:04:52,868 when you talk about artificial intelligence especially for 139 00:04:53,167 --> 00:04:55,576 organizations that, you know, may have started to 140 00:04:55,576 --> 00:04:57,170 work through some of that and and figure 141 00:04:57,170 --> 00:04:57,807 out what 142 00:04:58,524 --> 00:05:01,074 their their mission is and how Ai can 143 00:05:01,074 --> 00:05:03,066 support their mission. How do you then grow 144 00:05:03,066 --> 00:05:05,392 into more sophisticated space? What you need from 145 00:05:05,392 --> 00:05:07,884 a team level as well as a technology 146 00:05:08,021 --> 00:05:10,093 level that, will really make a difference and 147 00:05:10,093 --> 00:05:11,607 propel you quickly into the future. 148 00:05:12,403 --> 00:05:12,642 Yeah. 149 00:05:13,678 --> 00:05:14,178 So 150 00:05:14,809 --> 00:05:16,185 I'll say that a lot of 151 00:05:16,721 --> 00:05:20,545 the infrastructure that you need to embrace this 152 00:05:20,545 --> 00:05:21,364 evolving technology. 153 00:05:22,153 --> 00:05:23,584 Now at the end of the day, it's 154 00:05:23,584 --> 00:05:26,072 a tool. Obviously, it's a different tool 155 00:05:26,526 --> 00:05:27,026 with 156 00:05:27,560 --> 00:05:28,855 unique implications 157 00:05:29,389 --> 00:05:31,081 and considerations from 158 00:05:31,869 --> 00:05:34,444 several perspectives, right? So like I said, this 159 00:05:34,503 --> 00:05:36,737 notion of truly validating that a model is 160 00:05:36,737 --> 00:05:38,493 going to work for your patient's place in 161 00:05:38,493 --> 00:05:41,560 time that it's going to be fair and 162 00:05:41,618 --> 00:05:45,068 unbiased and even beyond that actually being able 163 00:05:45,206 --> 00:05:47,837 to implement it in a way that actually 164 00:05:47,837 --> 00:05:48,610 does not 165 00:05:49,046 --> 00:05:52,064 exacerbate disparities, which is especially important for a 166 00:05:52,064 --> 00:05:54,128 place like metro health where we're safety healthcare 167 00:05:54,128 --> 00:05:56,908 care systems that we see patients that are 168 00:05:56,908 --> 00:05:57,646 at risk 169 00:05:58,434 --> 00:06:00,133 minorities and soc economic 170 00:06:00,512 --> 00:06:03,410 deprivation being a common theme here. So 171 00:06:03,948 --> 00:06:05,168 when you think about 172 00:06:05,707 --> 00:06:07,545 what is the skills set required? 173 00:06:08,118 --> 00:06:09,786 Right? To bring that technology in. 174 00:06:10,738 --> 00:06:12,112 I think the foundation 175 00:06:12,724 --> 00:06:15,662 is fairly similar to really everything we do 176 00:06:15,662 --> 00:06:16,694 in the informatics space. 177 00:06:17,423 --> 00:06:18,693 Those understanding systems, 178 00:06:19,724 --> 00:06:22,421 understanding that any technology you bring in enters 179 00:06:22,421 --> 00:06:22,921 a 180 00:06:23,531 --> 00:06:24,586 person process 181 00:06:25,926 --> 00:06:28,701 a complex person process associated with technical space, 182 00:06:29,336 --> 00:06:30,549 right that you need to 183 00:06:31,239 --> 00:06:32,770 appreciate and understand 184 00:06:34,108 --> 00:06:34,608 and 185 00:06:34,980 --> 00:06:37,278 the implications too of change, 186 00:06:38,229 --> 00:06:38,467 innovation, 187 00:06:39,894 --> 00:06:40,607 implementation science, 188 00:06:41,098 --> 00:06:43,398 quality improvement. So all of these things that 189 00:06:43,398 --> 00:06:45,619 we we've been doing, right, from the outset, 190 00:06:46,174 --> 00:06:48,078 implementing all these technologies in a thoughtful and 191 00:06:48,078 --> 00:06:48,713 meaningful way. 192 00:06:49,363 --> 00:06:50,556 Do you have to take all of that 193 00:06:50,556 --> 00:06:52,089 solid base that you developed 194 00:06:52,544 --> 00:06:54,395 and add this new tooling to it 195 00:06:54,770 --> 00:06:56,622 and the stretch that you have to make 196 00:06:57,076 --> 00:06:58,349 when you're getting to that point. 197 00:06:59,240 --> 00:07:02,060 Is you need somebody who understands enough about 198 00:07:02,279 --> 00:07:03,100 these models 199 00:07:03,639 --> 00:07:04,600 and the implication, 200 00:07:05,240 --> 00:07:06,680 right to bring it into the space. 201 00:07:07,333 --> 00:07:08,230 Because the converse 202 00:07:08,684 --> 00:07:10,749 unfortunately is it's not true, right? So you 203 00:07:10,749 --> 00:07:13,711 can have amazing Ai and amazing technology, 204 00:07:14,403 --> 00:07:16,548 but shoving it into a healthcare system. 205 00:07:17,119 --> 00:07:19,511 Is not always the best way, right? That's 206 00:07:19,511 --> 00:07:21,345 just a clear path of failure. 207 00:07:22,063 --> 00:07:22,802 And so 208 00:07:23,259 --> 00:07:25,911 I think that the model that I 209 00:07:26,382 --> 00:07:27,837 I think speaks to this 210 00:07:28,291 --> 00:07:31,017 is this concept of a like quan 211 00:07:31,472 --> 00:07:32,926 and that somebody who 212 00:07:33,698 --> 00:07:35,471 understands enough about the technology 213 00:07:36,977 --> 00:07:40,167 implications and potential, right, to bring it into 214 00:07:40,167 --> 00:07:41,066 their environment 215 00:07:41,523 --> 00:07:43,437 and then bring it into their environment into 216 00:07:43,437 --> 00:07:45,612 this complex healthcare system 217 00:07:46,242 --> 00:07:47,695 in the way that it actually 218 00:07:48,386 --> 00:07:51,483 feeds and enables people to do the things 219 00:07:51,483 --> 00:07:53,548 that they're setting out to do, which obviously 220 00:07:53,548 --> 00:07:55,318 in this space primarily is around 221 00:07:55,628 --> 00:07:56,581 providing excellent care. 222 00:07:58,249 --> 00:08:00,633 Understand that the implications understands the risk and 223 00:08:00,633 --> 00:08:01,983 biases that they have to address. 224 00:08:02,713 --> 00:08:04,643 And so you obviously need some 225 00:08:06,288 --> 00:08:09,069 that informatics type individual coming out from healthcare. 226 00:08:09,798 --> 00:08:11,860 With enzyme understanding into the data science and 227 00:08:11,860 --> 00:08:12,970 the artificial intelligence, 228 00:08:13,842 --> 00:08:15,665 also obviously paired with somebody who can do 229 00:08:15,665 --> 00:08:16,299 to data science. 230 00:08:17,824 --> 00:08:19,665 I think we're slowly moving away from the 231 00:08:19,665 --> 00:08:20,564 need to have 232 00:08:21,105 --> 00:08:22,564 hardcore data scientists. 233 00:08:23,985 --> 00:08:25,764 I'm also optimistic that 234 00:08:26,398 --> 00:08:27,774 as the tools evolve 235 00:08:28,389 --> 00:08:30,938 that these vendors that are producing these solutions, 236 00:08:32,293 --> 00:08:32,793 package 237 00:08:33,184 --> 00:08:36,612 their solutions with ideally open source tools that 238 00:08:36,612 --> 00:08:37,431 have been 239 00:08:37,888 --> 00:08:39,881 really kind of taken through their paces, 240 00:08:40,519 --> 00:08:41,019 standardized 241 00:08:42,048 --> 00:08:42,445 vetted, 242 00:08:44,034 --> 00:08:46,654 allowing these organizations then to kind of flex 243 00:08:46,654 --> 00:08:50,014 up with whatever resources they have to reach 244 00:08:50,070 --> 00:08:53,191 into that space. And bring that solution in. 245 00:08:53,829 --> 00:08:55,825 And so that way you can start to 246 00:08:55,825 --> 00:08:57,362 think about how you're democrat 247 00:08:57,820 --> 00:09:00,787 the Ai to the smaller healthcare organization. We 248 00:09:00,787 --> 00:09:02,694 don't have the money and the resources and 249 00:09:02,694 --> 00:09:03,352 the personnel 250 00:09:04,521 --> 00:09:07,246 to individually develop their own models 251 00:09:08,032 --> 00:09:10,413 and their own Ai expertise. Now those places 252 00:09:10,413 --> 00:09:11,469 exist, obviously 253 00:09:12,000 --> 00:09:12,738 academic centers 254 00:09:14,540 --> 00:09:16,047 obviously, I could name some organizations, 255 00:09:17,175 --> 00:09:19,582 we all know which ones they are. It's 256 00:09:19,798 --> 00:09:21,864 the same organizations who are leading a lot 257 00:09:21,864 --> 00:09:24,169 of The charge and this color is the 258 00:09:24,169 --> 00:09:26,171 rest of the discussion that I I kind 259 00:09:26,171 --> 00:09:26,966 of brought in earlier. 260 00:09:28,080 --> 00:09:28,739 I think 261 00:09:29,590 --> 00:09:31,339 this notion of Ai assurance, 262 00:09:32,055 --> 00:09:34,540 right, that an organization can tell this model 263 00:09:34,540 --> 00:09:36,480 looks good because we tested in our data 264 00:09:37,500 --> 00:09:41,100 as someone who's independently implemented and validated Ai 265 00:09:41,100 --> 00:09:42,060 solutions at the point of care. 266 00:09:43,033 --> 00:09:46,452 Models are messy inherently biased and they don't 267 00:09:46,452 --> 00:09:49,234 always work well for your patient population. And 268 00:09:49,234 --> 00:09:51,062 this has been true throughout time. It doesn't 269 00:09:51,062 --> 00:09:52,749 not be a sophisticated model to do that, 270 00:09:52,989 --> 00:09:54,528 even models at pretty risk 271 00:09:54,907 --> 00:09:56,985 like heart disease for instance. When we bring 272 00:09:56,985 --> 00:09:58,764 them into a system like Metro health 273 00:09:59,462 --> 00:10:02,020 and more diverse and our risk coa population, 274 00:10:02,674 --> 00:10:05,390 we can see that the model actually tends 275 00:10:05,390 --> 00:10:08,346 to under predict risk of cardiac, disease and 276 00:10:08,346 --> 00:10:09,246 poor outcomes 277 00:10:09,720 --> 00:10:11,399 which case, we actually could be doing our 278 00:10:11,399 --> 00:10:12,519 patients on disservice, 279 00:10:13,320 --> 00:10:15,240 a model like that is. So this notion 280 00:10:15,240 --> 00:10:17,660 of rules and models being biased 281 00:10:17,974 --> 00:10:19,172 just the world we live in and it's 282 00:10:19,172 --> 00:10:19,731 not new. 283 00:10:20,369 --> 00:10:21,269 And I think 284 00:10:21,806 --> 00:10:24,122 when you take these models out from systems 285 00:10:24,122 --> 00:10:26,677 that they've been developed, we bring them into 286 00:10:26,677 --> 00:10:27,177 systems 287 00:10:27,888 --> 00:10:29,104 like metro health or 288 00:10:29,558 --> 00:10:30,058 F 289 00:10:31,784 --> 00:10:33,533 or other similar systems. 290 00:10:34,010 --> 00:10:36,475 There's always that fair and concern that you're 291 00:10:36,475 --> 00:10:38,953 actually going to be doing and disadvantage. And 292 00:10:38,953 --> 00:10:39,985 so some of the work that we do 293 00:10:39,985 --> 00:10:40,461 for instance, 294 00:10:42,128 --> 00:10:42,866 we've done 295 00:10:43,954 --> 00:10:44,692 implementation of 296 00:10:45,636 --> 00:10:48,176 potentially challenging, predictive models like we predicted the 297 00:10:48,176 --> 00:10:49,446 risk that a patient when to show up 298 00:10:49,446 --> 00:10:50,081 to a clinic. 299 00:10:50,954 --> 00:10:53,756 And what that brings with it is obviously 300 00:10:54,304 --> 00:10:56,614 more often than not more likely, you're actually 301 00:10:56,614 --> 00:10:58,207 going to be targeting. Whatever intervention you build 302 00:10:58,207 --> 00:10:59,583 on that, you're going to be targeting 303 00:11:00,119 --> 00:11:02,190 patients that are at risk, right? So usually 304 00:11:02,190 --> 00:11:02,690 minorities. 305 00:11:03,639 --> 00:11:06,284 And actually did the implementation through a study 306 00:11:06,658 --> 00:11:08,248 where we decided we're not going to over 307 00:11:08,248 --> 00:11:10,257 book because we think that that actually worsen 308 00:11:10,313 --> 00:11:12,240 disparity in this but we're actually going to 309 00:11:12,240 --> 00:11:14,392 provide more resources for outreach to bring patients 310 00:11:14,392 --> 00:11:16,783 that risk not showing up into the system. 311 00:11:17,420 --> 00:11:19,750 And then doing so actually reduced d show 312 00:11:19,750 --> 00:11:21,990 rate in our Black patients by 15 percent. 313 00:11:22,629 --> 00:11:23,830 And so we kind of hold that up 314 00:11:23,830 --> 00:11:25,929 as an example of how an 315 00:11:26,389 --> 00:11:26,889 organization 316 00:11:27,444 --> 00:11:28,424 can responsibly 317 00:11:28,882 --> 00:11:30,101 validate implement 318 00:11:31,040 --> 00:11:34,076 a model that could otherwise worsen disparities in 319 00:11:34,076 --> 00:11:36,399 a way where not only do all new 320 00:11:36,399 --> 00:11:37,852 tied with all boats, but 321 00:11:38,464 --> 00:11:41,109 really you were able to generate a benefit 322 00:11:41,562 --> 00:11:43,174 to your patient population 323 00:11:43,484 --> 00:11:46,105 in a very equitable way. We're actually the 324 00:11:46,105 --> 00:11:48,750 outside effect help build that were less fortunate. 325 00:11:49,601 --> 00:11:50,419 So that's 326 00:11:50,728 --> 00:11:52,716 kind of the work that I do that's 327 00:11:52,716 --> 00:11:54,250 where I find myself 328 00:11:54,624 --> 00:11:55,919 spending a lot of time thinking 329 00:11:56,374 --> 00:11:57,964 about these problems and thinking about how to 330 00:11:57,964 --> 00:11:58,623 bring these 331 00:11:59,093 --> 00:12:01,245 solutions in. And I think the opportunity like 332 00:12:01,404 --> 00:12:02,223 I said earlier 333 00:12:02,998 --> 00:12:04,911 is the tooling needs a lot of work, 334 00:12:05,150 --> 00:12:06,607 but I'm excited that 335 00:12:06,918 --> 00:12:10,513 with better standards, data standardization and interoperability, 336 00:12:10,886 --> 00:12:12,259 getting people on the same 337 00:12:12,632 --> 00:12:15,508 kind of playing field where I can share 338 00:12:15,508 --> 00:12:17,042 my model and my experience 339 00:12:17,576 --> 00:12:18,314 and data 340 00:12:19,167 --> 00:12:20,917 in a way where I can validate your 341 00:12:20,917 --> 00:12:21,289 model 342 00:12:21,727 --> 00:12:24,137 right? On my data for instance, Having that 343 00:12:24,194 --> 00:12:25,172 interchange happen 344 00:12:25,945 --> 00:12:27,321 in a way where I get it's 345 00:12:27,855 --> 00:12:28,253 protected, 346 00:12:28,969 --> 00:12:30,402 no particular information is passed, 347 00:12:30,974 --> 00:12:33,839 I think that's incredibly powerful. So I see 348 00:12:33,839 --> 00:12:35,612 this as a continuum of 349 00:12:36,147 --> 00:12:38,398 everything that we've been doing in informatics 350 00:12:38,869 --> 00:12:41,500 making that world just so much much more 351 00:12:41,500 --> 00:12:43,196 potentially powerful and empowering 352 00:12:43,813 --> 00:12:45,726 some of the organizations with western resources is 353 00:12:45,726 --> 00:12:47,855 to join the Ai revolution that 354 00:12:48,292 --> 00:12:50,357 otherwise, may just be relegated to, you know, 355 00:12:50,755 --> 00:12:52,662 the the large academic centers. 356 00:12:54,012 --> 00:12:55,919 That's fascinating to hear. Thank you so much 357 00:12:55,919 --> 00:12:57,921 for going through that with us dig you 358 00:12:57,921 --> 00:13:00,304 1 bit deeper, especially on Ai. Now what 359 00:13:00,304 --> 00:13:01,814 are you most excited about for the future? 360 00:13:01,894 --> 00:13:02,847 And what makes you nervous? 361 00:13:05,565 --> 00:13:06,065 So 362 00:13:06,440 --> 00:13:07,656 I'm excited about 363 00:13:08,349 --> 00:13:09,405 the opportunity 364 00:13:09,780 --> 00:13:10,496 that we have, 365 00:13:11,132 --> 00:13:11,870 I think 366 00:13:12,579 --> 00:13:14,753 maybe even sufficiently excited about 367 00:13:15,448 --> 00:13:17,122 the dying down of the Ai hike. 368 00:13:17,999 --> 00:13:19,296 So the more pragmatic 369 00:13:20,071 --> 00:13:23,024 folks rate that weather the storms of new 370 00:13:23,024 --> 00:13:23,524 technologies 371 00:13:23,897 --> 00:13:26,039 and the hype cycles, right? We'll tell you 372 00:13:26,039 --> 00:13:28,182 that we really just want to know what 373 00:13:28,182 --> 00:13:30,428 works, right and what works for our patients 374 00:13:30,428 --> 00:13:32,346 and what enable us to provide the care 375 00:13:32,346 --> 00:13:33,065 that we'd like. 376 00:13:33,865 --> 00:13:34,365 So 377 00:13:34,824 --> 00:13:37,002 I think that the current energy 378 00:13:37,461 --> 00:13:38,280 is around 379 00:13:39,073 --> 00:13:42,412 trying to validate an Fda regulate and even 380 00:13:42,412 --> 00:13:42,912 create 381 00:13:43,445 --> 00:13:43,945 assurance 382 00:13:44,478 --> 00:13:46,727 bodies to basically sanction 383 00:13:47,102 --> 00:13:48,269 certain models as 384 00:13:48,866 --> 00:13:49,366 usable 385 00:13:49,740 --> 00:13:50,296 or safe. 386 00:13:51,409 --> 00:13:53,737 I would say that really we should 387 00:13:54,668 --> 00:13:55,803 I worry a 388 00:13:56,195 --> 00:13:57,975 about the amount of money that we're spending 389 00:13:58,195 --> 00:13:59,014 as the healthcare 390 00:13:59,315 --> 00:14:00,754 organizations following that high cycle. 391 00:14:01,315 --> 00:14:04,123 So I am looking forward to that dying 392 00:14:04,123 --> 00:14:06,188 down, right? And for us to really to 393 00:14:06,188 --> 00:14:08,752 see these the real solutions come out also 394 00:14:09,205 --> 00:14:09,705 to 395 00:14:10,555 --> 00:14:12,326 likely actually lead to more 396 00:14:13,990 --> 00:14:16,779 or should I should say surface the more 397 00:14:16,779 --> 00:14:19,409 obvious integrated solutions rather than the point solutions 398 00:14:19,409 --> 00:14:19,909 that 399 00:14:20,286 --> 00:14:22,694 are more hype than actual value. 400 00:14:24,129 --> 00:14:25,645 So that's the concern right. We spend more 401 00:14:25,645 --> 00:14:28,197 money there that concerns me. But the opportunity 402 00:14:28,197 --> 00:14:30,111 again, is once the field starts to kind 403 00:14:30,111 --> 00:14:30,611 of 404 00:14:32,363 --> 00:14:34,621 or I should say once the weaker options 405 00:14:35,080 --> 00:14:37,717 shake out, I think the path forward for 406 00:14:37,717 --> 00:14:40,055 driving true value from this 407 00:14:40,369 --> 00:14:41,588 Ai clinical 408 00:14:41,966 --> 00:14:44,941 revolution, I think it becomes more obvious. And 409 00:14:45,159 --> 00:14:47,554 that's the kind of work that many us 410 00:14:47,554 --> 00:14:49,151 in this area are excited about. 411 00:14:49,965 --> 00:14:50,465 Really 412 00:14:50,845 --> 00:14:53,404 getting beyond the hype cycle goal and actually 413 00:14:53,404 --> 00:14:54,785 starting to integrate 414 00:14:55,485 --> 00:14:55,985 these 415 00:14:57,004 --> 00:14:59,345 solutions into our process and the workflow 416 00:14:59,659 --> 00:15:00,537 in a meaningful way. 417 00:15:01,494 --> 00:15:03,807 The other parallel to that for me in 418 00:15:03,807 --> 00:15:05,663 the virtual care space. So 419 00:15:06,040 --> 00:15:07,976 mentioned that we're building this cross 420 00:15:08,991 --> 00:15:09,491 organization 421 00:15:09,963 --> 00:15:10,463 so 422 00:15:11,080 --> 00:15:13,414 Metro health and Have partnered together 423 00:15:14,191 --> 00:15:15,408 to build a 424 00:15:15,945 --> 00:15:18,896 common virtual care platform, which is meant to 425 00:15:18,896 --> 00:15:19,396 be 426 00:15:20,027 --> 00:15:22,892 inter and integrated with healthcare care systems that 427 00:15:22,892 --> 00:15:24,347 really want to 428 00:15:24,882 --> 00:15:26,553 extend the care that they do. 429 00:15:27,365 --> 00:15:29,625 And the differentiation for us there is 430 00:15:30,485 --> 00:15:32,325 being able to provide the additional service in 431 00:15:32,325 --> 00:15:34,325 a way where really from the patient perspective 432 00:15:34,325 --> 00:15:35,845 it's a seamless which part of their journey. 433 00:15:36,499 --> 00:15:38,735 That when you have reasons to come back 434 00:15:38,735 --> 00:15:41,209 into and organization's brick and mortar services, 435 00:15:42,567 --> 00:15:43,306 that happens 436 00:15:43,684 --> 00:15:45,621 seamlessly. And so that also 437 00:15:46,253 --> 00:15:48,101 common theme here being built on 438 00:15:49,503 --> 00:15:50,637 great data standards 439 00:15:51,246 --> 00:15:51,746 and 440 00:15:52,118 --> 00:15:52,618 interoperability. 441 00:15:53,719 --> 00:15:54,913 And so these are the kinds of things 442 00:15:54,913 --> 00:15:55,971 that you can unlock 443 00:15:56,506 --> 00:15:58,576 when we all start speaking that same language. 444 00:15:58,815 --> 00:16:01,681 And that for me is very exciting. And 445 00:16:01,681 --> 00:16:03,610 so to kind of bring those 2 examples 446 00:16:03,610 --> 00:16:04,328 for circle, 447 00:16:05,125 --> 00:16:07,598 we need to be continuing to spend more 448 00:16:07,598 --> 00:16:08,098 money 449 00:16:08,555 --> 00:16:09,852 on the right things 450 00:16:10,310 --> 00:16:11,767 and the right things are 451 00:16:12,319 --> 00:16:14,256 continuing to integrate our systems 452 00:16:14,714 --> 00:16:16,869 to create an opportunity across the board for 453 00:16:16,869 --> 00:16:17,848 a lot of these 454 00:16:18,466 --> 00:16:21,041 initiatives and collaborations to really 455 00:16:22,232 --> 00:16:24,783 to break through and really deliver on that 456 00:16:24,783 --> 00:16:25,102 promise, 457 00:16:25,819 --> 00:16:27,276 rather than spending money 458 00:16:27,733 --> 00:16:28,233 on 459 00:16:28,690 --> 00:16:29,827 shiny objects 460 00:16:30,616 --> 00:16:32,442 The other thing in terms of kind of 461 00:16:32,442 --> 00:16:35,697 spending money is new regulation, so that concerns 462 00:16:35,697 --> 00:16:37,705 be coming around the corner. So 463 00:16:38,254 --> 00:16:40,808 how is the Fda planning to regulate a 464 00:16:40,808 --> 00:16:41,707 lot of this technology? 465 00:16:42,563 --> 00:16:45,038 And what is that do to the innovation 466 00:16:45,038 --> 00:16:48,485 space, right? So does that increase the barrier 467 00:16:48,485 --> 00:16:51,120 of entry? Does it take away the ability 468 00:16:51,120 --> 00:16:51,859 to democrat 469 00:16:52,637 --> 00:16:53,775 the both 470 00:16:55,351 --> 00:16:55,851 and 471 00:16:56,964 --> 00:16:58,503 of of some of these Ai 472 00:16:59,042 --> 00:16:59,361 initiatives. 473 00:17:01,679 --> 00:17:03,677 It's fascinating in here. You know, and really, 474 00:17:04,316 --> 00:17:05,890 cool to to think through 475 00:17:06,248 --> 00:17:08,801 all the possibilities that are ahead. Before we 476 00:17:08,801 --> 00:17:10,476 wrap up here, I'm wondering, could you talk 477 00:17:10,476 --> 00:17:12,789 about what the most effective healthcare leaders will 478 00:17:12,789 --> 00:17:14,464 need in order to be successful over the 479 00:17:14,464 --> 00:17:16,553 next 2 to 3 years, especially plans as 480 00:17:16,633 --> 00:17:19,267 Ai continues to evolve and in grow this 481 00:17:19,267 --> 00:17:21,422 space and technology becomes an even bigger part 482 00:17:21,422 --> 00:17:23,758 of health delivery and in systems 483 00:17:24,216 --> 00:17:24,376 operations. 484 00:17:26,869 --> 00:17:27,849 Yes. So 485 00:17:28,630 --> 00:17:29,849 I'll recap that 486 00:17:30,630 --> 00:17:31,769 to create change 487 00:17:32,470 --> 00:17:33,849 in a healthcare environment 488 00:17:34,963 --> 00:17:35,939 I think we have 489 00:17:36,391 --> 00:17:36,891 already 490 00:17:37,422 --> 00:17:40,198 a slew of amazing models to build off. 491 00:17:41,150 --> 00:17:41,839 And so 492 00:17:42,757 --> 00:17:45,788 modeling around this notion of learning health systems, 493 00:17:46,028 --> 00:17:46,927 right, building 494 00:17:48,262 --> 00:17:49,035 organizations and 495 00:17:49,474 --> 00:17:52,595 that attempt to be agile and learn from 496 00:17:52,595 --> 00:17:55,154 the data they generate to create this loop 497 00:17:55,154 --> 00:17:56,755 of continuous quality improvement. 498 00:17:57,474 --> 00:17:58,295 How do you 499 00:17:59,009 --> 00:18:01,643 how do you change, right? And how do 500 00:18:01,643 --> 00:18:04,118 you manage change And how do you implement 501 00:18:04,118 --> 00:18:06,853 things that you know should be implemented and 502 00:18:07,071 --> 00:18:07,811 how do you 503 00:18:08,204 --> 00:18:11,236 monitor the effects from a quality improvement perspective. 504 00:18:11,874 --> 00:18:13,332 I think these are the many 505 00:18:14,188 --> 00:18:14,688 very 506 00:18:15,066 --> 00:18:15,566 specific 507 00:18:16,198 --> 00:18:19,569 areas that somebody who's in clinical informatics 508 00:18:20,186 --> 00:18:22,180 is a clinician leader even needs to think 509 00:18:22,180 --> 00:18:23,318 about right from an 510 00:18:23,695 --> 00:18:24,812 organizational perspective. 511 00:18:25,862 --> 00:18:27,370 If you want to add Ai to that, 512 00:18:28,323 --> 00:18:30,307 I'll recap this notion of the Light Quan, 513 00:18:30,466 --> 00:18:33,585 which is start understanding enough about the technology 514 00:18:34,609 --> 00:18:38,286 to understand the implication, the limitations and the 515 00:18:38,286 --> 00:18:38,786 opportunities 516 00:18:39,325 --> 00:18:41,023 that allow to bring it back 517 00:18:41,642 --> 00:18:42,441 into the system. 518 00:18:43,001 --> 00:18:43,820 And so 519 00:18:45,570 --> 00:18:46,687 what I guess I'm trying to say in 520 00:18:46,687 --> 00:18:48,681 the nutshell is, we have these really great 521 00:18:48,681 --> 00:18:51,414 tools, right in the informatics universe that 522 00:18:51,792 --> 00:18:53,727 we leverage to do what we do 523 00:18:54,358 --> 00:18:56,687 And so building on that, I think 524 00:18:57,539 --> 00:19:00,425 where an informatics leader specifically 525 00:19:01,356 --> 00:19:03,662 used to sit, right, 10 to 20 years 526 00:19:03,662 --> 00:19:06,474 ago they were the physician champion of the 527 00:19:06,474 --> 00:19:07,194 electronic health record, 528 00:19:07,835 --> 00:19:11,134 right but that digital ecosystem now has expanded 529 00:19:11,609 --> 00:19:13,301 And so now they need to be thinking 530 00:19:13,755 --> 00:19:15,050 much more broadly 531 00:19:15,822 --> 00:19:16,322 about 532 00:19:16,696 --> 00:19:19,025 where we meet patients, how we meet patients 533 00:19:19,815 --> 00:19:22,214 and that's exciting, right, because now we're getting 534 00:19:22,214 --> 00:19:23,974 into the act I think the point of 535 00:19:23,974 --> 00:19:24,474 contact 536 00:19:25,174 --> 00:19:26,934 and the doors if you will have expanded 537 00:19:26,934 --> 00:19:28,055 in terms of opportunity. 538 00:19:28,789 --> 00:19:30,402 And so they need to be thinking about 539 00:19:31,015 --> 00:19:33,162 digital care. They need to be thinking about 540 00:19:33,162 --> 00:19:34,831 patients as consumers with choice. 541 00:19:35,547 --> 00:19:37,399 And beyond that, they need to be thinking 542 00:19:37,773 --> 00:19:38,273 less 543 00:19:38,584 --> 00:19:41,234 tactically, like how do we survive the day 544 00:19:41,450 --> 00:19:42,508 and more strategically, 545 00:19:43,202 --> 00:19:46,149 why should we acquire and implement this platform 546 00:19:46,149 --> 00:19:46,808 or technology? 547 00:19:47,439 --> 00:19:49,437 How can we carry that forward? Great? How 548 00:19:49,437 --> 00:19:50,636 can we how can we turn this into 549 00:19:50,636 --> 00:19:52,494 a successful return our investment 550 00:19:53,113 --> 00:19:55,511 for our organization? And so I really do 551 00:19:55,511 --> 00:19:56,570 just feel like this 552 00:19:56,884 --> 00:19:59,278 a more initially limited role of a clinical 553 00:19:59,278 --> 00:20:00,336 and informatics leader 554 00:20:01,112 --> 00:20:02,947 has just kind of blown up. And I 555 00:20:02,947 --> 00:20:03,846 think you've seen 556 00:20:04,395 --> 00:20:05,450 pieces of that right 557 00:20:05,980 --> 00:20:08,700 specializations now and there's your Chief Digital Officer 558 00:20:08,992 --> 00:20:09,492 and 559 00:20:10,102 --> 00:20:11,315 your Chief Ai 560 00:20:11,624 --> 00:20:14,036 sir and it's really just shown the need 561 00:20:14,252 --> 00:20:17,222 for people to start exploring some of the 562 00:20:17,359 --> 00:20:19,988 expanding tentacles of the health technology states. 563 00:20:21,117 --> 00:20:23,205 I guess like conclusion really, there's 564 00:20:23,736 --> 00:20:26,197 a lot that we already know about how 565 00:20:26,197 --> 00:20:29,634 to create effective change and lead within healthcare 566 00:20:29,634 --> 00:20:31,315 and from within healthcare, which is I think 567 00:20:31,315 --> 00:20:32,434 where most of the gist are going to 568 00:20:32,434 --> 00:20:32,755 happen. 569 00:20:33,394 --> 00:20:35,154 It's just a matter of where you think 570 00:20:35,154 --> 00:20:37,394 you need to sit in terms of driving 571 00:20:37,394 --> 00:20:39,880 that change. So Ai is not for everybody. 572 00:20:40,039 --> 00:20:41,733 It's going to be ubiquitous, but it's 573 00:20:42,108 --> 00:20:44,359 definitely going to come to a point where 574 00:20:45,132 --> 00:20:47,219 there's going some people don't understand enough about 575 00:20:47,219 --> 00:20:48,653 a drive forward and the rest of us 576 00:20:48,653 --> 00:20:50,348 are just basically going to have to 577 00:20:51,283 --> 00:20:52,319 so that's just going to be a part 578 00:20:52,319 --> 00:20:52,717 of life. 579 00:20:53,847 --> 00:20:55,439 Doctor Terry. Thank you so much for joining 580 00:20:55,439 --> 00:20:56,871 us on the podcast today. This has been 581 00:20:56,871 --> 00:20:58,701 a fascinating discussion, and I look forward to 582 00:20:58,701 --> 00:20:59,655 connecting with you again soon. 583 00:21:01,023 --> 00:21:02,852 Yeah. Thank you for having. It's it's been 584 00:21:02,852 --> 00:21:03,091 fun.