1 00:00:05,040 --> 00:00:07,720 Hello everyone, and welcome to the Quorum podcast. 2 00:00:07,800 --> 00:00:11,440 This is where academic medicine meets remote, austere and 3 00:00:11,440 --> 00:00:17,520 resource limited areas. Welcome back to the podcast. 4 00:00:17,520 --> 00:00:20,960 This is April Kelly. This week we're with Doctor 5 00:00:20,960 --> 00:00:23,640 Michael Clapper from South Africa. 6 00:00:23,760 --> 00:00:28,520 Michael's been on faculty with the college for well, I decade 7 00:00:28,520 --> 00:00:30,880 or so. He first came to Malta and ages 8 00:00:30,880 --> 00:00:34,080 ago, but and he just earned his doctorate. 9 00:00:34,080 --> 00:00:37,960 He got his DHS through core one of our first graduates. 10 00:00:38,080 --> 00:00:40,800 Michael, welcome to the podcast. Thank you a brick. 11 00:00:40,800 --> 00:00:42,760 It's always a pleasure to do these with you. 12 00:00:43,200 --> 00:00:46,120 So tell us a bit about yourself and what's keeping you busy. 13 00:00:46,120 --> 00:00:50,680 So at the moment I just successfully defended my 14 00:00:50,720 --> 00:00:56,520 dissertation 2 weeks ago. So up until then, the DHS was 15 00:00:56,560 --> 00:00:58,600 pretty much keeping me the busiest. 16 00:00:59,840 --> 00:01:05,880 And alongside my other day job, which is working in AI, just 17 00:01:05,880 --> 00:01:10,440 helping organisations navigate that that ecosystem. 18 00:01:10,800 --> 00:01:14,200 But it's also really working, helping the college set up 19 00:01:14,200 --> 00:01:17,280 various programmes, getting systems to work and just, yeah, 20 00:01:17,280 --> 00:01:20,720 just generally trying to bridge the gap between the world of 21 00:01:20,720 --> 00:01:24,080 software and the world of Austria and remote medicine. 22 00:01:24,120 --> 00:01:28,400 And now that the the DHS is finally behind me, I'd like to 23 00:01:28,400 --> 00:01:31,800 say it's less work, but I'm quickly realising it's even more 24 00:01:31,800 --> 00:01:34,520 work ahead of me. But we'll, we'll get to your 25 00:01:34,560 --> 00:01:38,040 doctoral thesis on AI. But let's let's start way back. 26 00:01:38,040 --> 00:01:41,080 So what? That was your first connection 27 00:01:41,080 --> 00:01:44,560 with the college. Well, interestingly, it all 28 00:01:44,560 --> 00:01:49,840 started with an e-mail I received from you, probably way 29 00:01:49,840 --> 00:01:55,120 back hazard a guess, E 13 E 14. You at the time were visiting 30 00:01:55,120 --> 00:01:59,480 Johannesburg to set up clinical placements, I think with ER 24 31 00:01:59,840 --> 00:02:05,200 and you and I connected through the WMS form register and we 32 00:02:05,200 --> 00:02:07,200 thought, hey, I'm in Johannesburg, you're in 33 00:02:07,200 --> 00:02:09,280 Johannesburg, let's just have a pint together. 34 00:02:09,280 --> 00:02:13,360 And from there I think I was sort of very taken in the bar. 35 00:02:13,360 --> 00:02:16,800 Your vision for remotes and Austria Mits and in general, I 36 00:02:16,800 --> 00:02:20,680 mean, from the that's when the the roller coaster ride started 37 00:02:20,680 --> 00:02:23,880 working with you getting those clinical placements up, you 38 00:02:23,880 --> 00:02:26,400 know, helping look after the medics when they came here. 39 00:02:26,760 --> 00:02:31,280 At the time I was also doing my national registry paramedic 40 00:02:31,320 --> 00:02:35,960 through Percom and at the time you had just receive permission. 41 00:02:35,960 --> 00:02:39,520 I think you were the first organisation to be able to offer 42 00:02:39,520 --> 00:02:45,120 the psychometric evaluation outside of the continental USI. 43 00:02:45,120 --> 00:02:49,400 Then came across to Malta for a couple of weeks, I think it was 44 00:02:49,400 --> 00:02:54,960 the AEMT evaluation. So I completed that with you and 45 00:02:54,960 --> 00:02:57,480 I think there were a few other students at the time. 46 00:02:57,880 --> 00:03:02,720 Then after that came back for a month just to sit in on your 47 00:03:03,600 --> 00:03:06,520 industry paramedic course. So just helped out there and got 48 00:03:06,520 --> 00:03:12,520 a bit more familiar with how the college operates and teachers 49 00:03:12,520 --> 00:03:16,160 and what it was all about. Then carried on pretty much with 50 00:03:16,160 --> 00:03:19,640 my own mission, which was getting my Fellowship of the 51 00:03:19,640 --> 00:03:22,280 Academy, Waters meds and finishing up my national 52 00:03:22,280 --> 00:03:25,960 registry paramedic, and also working or quite heavily 53 00:03:25,960 --> 00:03:29,640 involved in modern search and rescue in South Africa. 54 00:03:29,760 --> 00:03:32,920 And really at the time just bolstering my buildings 55 00:03:32,920 --> 00:03:36,720 education wherever I could. Then one of the more ambitious 56 00:03:36,720 --> 00:03:40,440 projects that worked on with Korum was setting up the field 57 00:03:40,440 --> 00:03:42,040 guide or the digital field guide. 58 00:03:42,280 --> 00:03:45,920 And that was just in such a way that it would really work in 59 00:03:45,920 --> 00:03:48,680 these very limited connectivity environments. 60 00:03:48,800 --> 00:03:51,760 And I'm pleased to say that project has been going for quite 61 00:03:51,760 --> 00:03:55,680 some time. Then as the college grew, just 62 00:03:55,680 --> 00:03:59,360 was always happy to sort of help out, try to get a few programmes 63 00:03:59,360 --> 00:04:02,880 running in South Africa. Yeah, I was just generally, I 64 00:04:02,880 --> 00:04:06,000 would say along for the ride with Koram as it grew to what it 65 00:04:06,000 --> 00:04:08,600 is today. So you have been part of the 66 00:04:08,600 --> 00:04:14,040 family for over a decade now. Yes, if I had to count exactly, 67 00:04:14,040 --> 00:04:16,600 I'd probably say 1314 years. Right. 68 00:04:16,600 --> 00:04:20,720 And I think that we were talking before I even left Ireland with 69 00:04:20,720 --> 00:04:22,880 Remote medicine Ireland moving down to Malta. 70 00:04:23,000 --> 00:04:25,240 I think we had shouted before that. 71 00:04:25,240 --> 00:04:29,560 And then we first met in Johannesburg when I flew down in 72 00:04:29,560 --> 00:04:33,160 2014, I believe it was. And we set up clinical 73 00:04:33,160 --> 00:04:36,880 placements for the R-24. They had asked us for a critical 74 00:04:36,880 --> 00:04:40,440 care class which we we ran a class for them. 75 00:04:40,440 --> 00:04:42,240 I don't remember if you were on that one or not. 76 00:04:43,280 --> 00:04:45,080 Yeah. Do you remember I think that was 77 00:04:45,080 --> 00:04:49,600 the first ever critical K paramedic proctored exam in 78 00:04:49,600 --> 00:04:51,440 South Africa? That's right. 79 00:04:51,920 --> 00:04:54,520 That one I just sat in just helped. 80 00:04:54,520 --> 00:04:57,600 Just a bit of the facilitation of that. 81 00:04:57,800 --> 00:05:00,760 But I think that was only the first trip where you and I met 82 00:05:00,760 --> 00:05:03,320 is getting that all set up. Yeah, back then we're doing 83 00:05:03,320 --> 00:05:06,440 quite a few trips. We we ran a full paramedic 84 00:05:06,440 --> 00:05:11,040 programme out of East London there in the southern coast of 85 00:05:11,200 --> 00:05:13,480 of South Africa. Yeah, that's right. 86 00:05:13,480 --> 00:05:17,520 I do remember doing a couple of trips with you and also a few on 87 00:05:17,520 --> 00:05:20,120 my own. That was a very underserved area 88 00:05:20,120 --> 00:05:23,800 in terms of EMS. And I do remember working on quite a few 89 00:05:23,840 --> 00:05:27,960 interesting calls while working with the Koram students. 90 00:05:28,200 --> 00:05:29,520 Yeah. So we had the clinical 91 00:05:29,520 --> 00:05:33,160 placements going down there, then setting up some of the 92 00:05:33,200 --> 00:05:38,080 clinical placements in Johannesburg at the TMH in the 93 00:05:38,080 --> 00:05:40,280 hospital there. And yeah, also try to, I'm 94 00:05:40,280 --> 00:05:42,320 trying to think of the names of these, some of the other 95 00:05:42,320 --> 00:05:44,760 courses, they escape me at the moment. 96 00:05:45,000 --> 00:05:48,520 But yes, a lot of time spent setting up clinical placements, 97 00:05:48,520 --> 00:05:52,280 going around and just doing as much networking as possible in 98 00:05:52,280 --> 00:05:53,280 South Africa. Right. 99 00:05:53,360 --> 00:05:55,640 I'm trying to, I'm trying to remember what are the courses 100 00:05:55,640 --> 00:05:59,080 where we just ran the paramedic and then the critical care 101 00:05:59,080 --> 00:06:01,400 paramedic in South Africa, if I remember right. 102 00:06:01,480 --> 00:06:02,680 Yeah. And then the clinical 103 00:06:02,680 --> 00:06:05,720 placements, I mean right back in the beginning with your remote 104 00:06:05,720 --> 00:06:09,160 Nets in Ireland, I did run one or two wilderness first aid 105 00:06:09,160 --> 00:06:10,840 courses, but not too many of those. 106 00:06:11,680 --> 00:06:15,520 So Michael, you also have your your farm, you're a fellow of 107 00:06:15,520 --> 00:06:18,400 Academy Wilderness Medicine through the Wilderness Medical 108 00:06:18,400 --> 00:06:21,840 Society, and you've had it now for over 5 years. 109 00:06:22,080 --> 00:06:24,040 Have you thought about getting your M? 110 00:06:24,040 --> 00:06:27,560 Farm Heavens once I have the the form and if you have the 111 00:06:27,560 --> 00:06:29,760 calendar, I think I've had it for over 10 years. 112 00:06:30,600 --> 00:06:33,880 To my belief, I think it was the first person in Africa to get 113 00:06:33,880 --> 00:06:36,320 the award. I do stand to be corrected on 114 00:06:36,320 --> 00:06:39,600 that. The inform was not a high 115 00:06:39,600 --> 00:06:42,080 priority for me. Obviously I had to focus on the 116 00:06:42,080 --> 00:06:44,400 things like my national registry paramedic. 117 00:06:44,520 --> 00:06:46,280 Then we went on to the wilderness paramedic 118 00:06:46,280 --> 00:06:49,360 certification and then on to the DHS. 119 00:06:49,560 --> 00:06:53,480 But now that I had that behind me, it's being the eternal 120 00:06:53,480 --> 00:06:55,000 students. That's something that's 121 00:06:55,000 --> 00:06:59,200 definitely on my radar just to really could be anything that 122 00:06:59,200 --> 00:07:02,720 just helps with my by not just my own knowledge and experience 123 00:07:02,720 --> 00:07:05,480 of our about an osteo medicine, but that anything that just 124 00:07:05,480 --> 00:07:09,680 helps me help my students. So how do you see the form 125 00:07:09,840 --> 00:07:14,480 helping your students? So the form was one of my main 126 00:07:14,480 --> 00:07:18,160 foundational wilderness myths and qualifications. 127 00:07:18,680 --> 00:07:22,600 The one thing I like about the form is just the diversity of 128 00:07:22,600 --> 00:07:27,160 the topics that you have or the, or the diversity of topics in or 129 00:07:27,160 --> 00:07:29,720 particularly in, in modern medicine or wildness medicine. 130 00:07:29,960 --> 00:07:34,280 That's one thing I do my best to try and install in my students 131 00:07:34,600 --> 00:07:38,440 is when you're out in whatever deployment that you're on, you 132 00:07:38,440 --> 00:07:41,760 just need this massive breadth of knowledge and understanding. 133 00:07:42,000 --> 00:07:45,360 You don't necessarily need the exact answer, but you do need to 134 00:07:45,360 --> 00:07:50,560 know that things exists, way to get help, who to speak to and so 135 00:07:50,560 --> 00:07:52,240 on. And that's just something that I 136 00:07:52,240 --> 00:07:54,160 found. The form really helped me 137 00:07:54,160 --> 00:07:56,400 initially is getting that breadth of knowledge. 138 00:07:56,480 --> 00:07:59,920 I found that the form, the form was difficult for me. 139 00:07:59,920 --> 00:08:03,920 I'll be honest, I started in what, 2007 I think it was, and 140 00:08:03,920 --> 00:08:06,040 it took me almost seven years to get it. 141 00:08:06,080 --> 00:08:09,600 No, it took me 5 years. But what I missed was a lot of 142 00:08:09,600 --> 00:08:13,240 the the paediatric stuff are the things that we just didn't get 143 00:08:13,240 --> 00:08:18,680 in the army and I wound up with 160 points, of which only 100 of 144 00:08:18,680 --> 00:08:22,840 the points counted to get me past the finish line and earn my 145 00:08:22,840 --> 00:08:25,000 form. Was that similar with yourself? 146 00:08:25,360 --> 00:08:26,520 Yes. And I've actually had a few 147 00:08:26,520 --> 00:08:28,400 colleagues who've expressed the same. 148 00:08:28,400 --> 00:08:31,880 And I think that is by design. The form is one of those 149 00:08:31,880 --> 00:08:36,039 qualifications which to my understanding, the DMS doesn't 150 00:08:36,039 --> 00:08:40,200 want form candidates to be able to RIP the proverbial bandit off 151 00:08:40,240 --> 00:08:42,640 and get the qualification. I think it is designed to. 152 00:08:42,640 --> 00:08:47,040 I don't mean to use the word force in a negative way, but to 153 00:08:47,080 --> 00:08:50,840 encourage candidates to get that breadth of knowledge. 154 00:08:51,400 --> 00:08:55,200 And oftentimes I did find you'd get these elusive credits that 155 00:08:55,200 --> 00:08:58,600 were very hard to get to. But it was actually the tipping 156 00:08:58,600 --> 00:09:01,640 scale for me on the form was when I went and actually 157 00:09:01,640 --> 00:09:03,440 participated in one of the conferences. 158 00:09:03,440 --> 00:09:07,760 And there that's just a treasure trove of training information, 159 00:09:07,760 --> 00:09:10,680 both practical and theoretical. And I mean, I walked away from 160 00:09:10,680 --> 00:09:14,720 that conference and not just with a couple of the the elusive 161 00:09:14,720 --> 00:09:19,760 form credits in hand, but just a much more robust level of 162 00:09:19,760 --> 00:09:24,240 experience level training and this understanding of of the 163 00:09:24,240 --> 00:09:26,400 wilderness myths and environments. 164 00:09:26,400 --> 00:09:28,200 It is something that I, I push for. 165 00:09:28,200 --> 00:09:32,160 So our, our college is one of the largest providers of farm 166 00:09:32,160 --> 00:09:36,720 credits outside the US and we have four of our courses that 167 00:09:36,720 --> 00:09:40,560 give farm credits and we have our student interest group 168 00:09:40,560 --> 00:09:44,160 because every month we, we hold the ES SIG. 169 00:09:44,160 --> 00:09:46,040 So people who get foam credit there. 170 00:09:46,400 --> 00:09:49,240 Well, because of COVID, one, one of the few things that was 171 00:09:49,240 --> 00:09:54,600 beneficial from that fiasco was WMS has allowed virtual 172 00:09:54,600 --> 00:09:58,880 attendance for, that's one of the requirements for your farm 173 00:09:58,880 --> 00:10:01,760 for your fellowship is you must attend one of the conferences. 174 00:10:01,760 --> 00:10:05,960 And I, I totally understand that, but it was just unbearably 175 00:10:06,200 --> 00:10:10,120 painful to pay 3 grand to fly somewhere in the US. 176 00:10:10,400 --> 00:10:13,600 And they always hold it in resorts that are 200 a night. 177 00:10:13,680 --> 00:10:16,920 And, and it's just that the costs were prohibited and now 178 00:10:16,920 --> 00:10:19,920 you can attend virtually. And they've, they've not dropped 179 00:10:19,920 --> 00:10:21,880 that. They've kept the virtual option 180 00:10:22,000 --> 00:10:25,760 even now that COVID is over. So there is no reason why 181 00:10:25,760 --> 00:10:29,840 someone living here and Europe or in Africa or Asia to not get 182 00:10:29,840 --> 00:10:31,640 their fellowship. Absolutely. 183 00:10:31,640 --> 00:10:35,640 It's living in South Africa where essentially you would take 184 00:10:35,640 --> 00:10:37,880 a dollar or euro and multiply it by 20. 185 00:10:38,160 --> 00:10:42,040 That's one of the common complaints about getting the 186 00:10:42,040 --> 00:10:44,280 form. It just doesn't equate into the 187 00:10:44,280 --> 00:10:47,320 local currency. There is a huge demand in South 188 00:10:47,320 --> 00:10:50,840 Africa among physicians who would like to get the form as 189 00:10:50,840 --> 00:10:52,640 well as the diploma Martin Medicine. 190 00:10:52,720 --> 00:10:56,480 And I think now that we have the MUM and the virtual conferences 191 00:10:56,480 --> 00:10:59,640 available, I think the uptake is increasing because as you 192 00:10:59,640 --> 00:11:03,640 mentioned, travel to anywhere from South Africa that holds his 193 00:11:03,640 --> 00:11:06,400 conferences. We'll easily set you back a few 194 00:11:06,400 --> 00:11:08,600 grand. So, Michael, let's talk about 195 00:11:08,680 --> 00:11:11,760 AI. So you have been studying AI 196 00:11:11,760 --> 00:11:15,320 long before you started the DHS. Now you have a doctorate and AI 197 00:11:15,320 --> 00:11:18,960 in medicine, and you have been helpful with the college getting 198 00:11:18,960 --> 00:11:20,840 in the digital field you created. 199 00:11:21,000 --> 00:11:24,960 You invented a field guide app just for the college, and you've 200 00:11:24,960 --> 00:11:29,840 now updated it again. But tell us about your thesis 201 00:11:29,840 --> 00:11:33,720 about AI in medicine. It's interestingly, when I 202 00:11:33,720 --> 00:11:37,600 started the The Doctor about four years ago, I think at the 203 00:11:37,600 --> 00:11:43,760 beginning of 2022, initially I was not planning on doing AI. 204 00:11:44,200 --> 00:11:45,840 And so interesting. I remember having a 205 00:11:45,840 --> 00:11:49,000 conversation, I think with yourself and one of the founders 206 00:11:49,000 --> 00:11:53,240 of the programme. The advice you gave to me is 207 00:11:53,360 --> 00:11:57,960 whatever you think your final project is today. 208 00:11:57,960 --> 00:11:59,600 So this is four years in the past. 209 00:11:59,760 --> 00:12:04,560 It will not be the case when you actually end up defending your 210 00:12:04,560 --> 00:12:07,040 dissertation. And that was very true when I 211 00:12:07,040 --> 00:12:09,760 started out. I was just really going to do a 212 00:12:09,760 --> 00:12:12,800 bit of work around how the journals operates, what the data 213 00:12:12,800 --> 00:12:14,320 inside all of those journals mean. 214 00:12:14,920 --> 00:12:17,880 But it was only when we had that AI boom. 215 00:12:18,040 --> 00:12:23,160 So the introduction of ChatGPT and basically what most people 216 00:12:23,160 --> 00:12:27,520 think of as AI is I started thinking is what is this 217 00:12:27,520 --> 00:12:30,280 incredible invention called? A large language model? 218 00:12:30,480 --> 00:12:32,480 What impact is that going to have? 219 00:12:32,680 --> 00:12:36,480 I'm sure anyone who's listening, they have probably tried either 220 00:12:36,480 --> 00:12:39,560 a ChatGPT, Gemini, Claude and so on. 221 00:12:39,680 --> 00:12:42,520 They've asked it medical questions, possibly even 222 00:12:42,520 --> 00:12:45,280 research questions and the answers have come back are 223 00:12:45,280 --> 00:12:48,120 always very plausible and very convincing. 224 00:12:48,280 --> 00:12:50,640 That's oftentimes, and I'm sure you've experienced this 225 00:12:50,640 --> 00:12:54,680 yourself, a brick is the answer is completely or seemingly 226 00:12:54,680 --> 00:12:56,960 fabricated or just hallucinating. 227 00:12:58,200 --> 00:13:05,120 So what I did for or started out with my dissertation and what I 228 00:13:05,120 --> 00:13:08,640 started out with was I received a lot of help from the 229 00:13:08,880 --> 00:13:11,760 supervisors and the the head of the programme and, and sort of 230 00:13:11,760 --> 00:13:18,040 forming what my dissertation was going to be is a let me back up 231 00:13:18,040 --> 00:13:20,160 one second. So when, whenever a medical 232 00:13:20,160 --> 00:13:23,680 person is doing research, typically you will go to Google 233 00:13:23,680 --> 00:13:28,640 Scholar, you'll go to pub Med or whichever journal or repository 234 00:13:28,640 --> 00:13:31,440 that you you'd like to use. You will go in and you'll type 235 00:13:31,440 --> 00:13:34,120 in the keywords. Say, for example, you want to 236 00:13:34,120 --> 00:13:37,400 search for the topic of middle area or any other tropical 237 00:13:37,400 --> 00:13:39,440 disease or anything you want in the medical sphere. 238 00:13:40,000 --> 00:13:44,440 Pub meds for example will return you specific results based on 239 00:13:44,440 --> 00:13:46,800 that search. So my thinking for the 240 00:13:46,800 --> 00:13:53,440 dissertation was these databases are all built on technology that 241 00:13:53,440 --> 00:13:57,720 is 1020, even more sort of decades older. 242 00:13:58,200 --> 00:14:02,800 What is it going to look like if AI gets involved in the search 243 00:14:02,800 --> 00:14:08,120 method? So my research question was 244 00:14:08,120 --> 00:14:15,200 quite simple is I wanted to see how the current way of searching 245 00:14:15,280 --> 00:14:20,000 would compare against the call it the new AI way of searching 246 00:14:20,000 --> 00:14:23,600 these databases. And what does that mean in terms 247 00:14:23,640 --> 00:14:27,040 of complying to the evidence based medicine standards? 248 00:14:27,200 --> 00:14:30,320 I'm talking things like Prisma, things like the the Nas 249 00:14:30,320 --> 00:14:33,920 framework, which chats about or framework for adopting new 250 00:14:33,920 --> 00:14:36,520 technology. So it wasn't just a technical 251 00:14:36,520 --> 00:14:39,720 question, it was a sociotechnical question as what 252 00:14:39,720 --> 00:14:42,320 is going to be the overall impact A on performance? 253 00:14:42,440 --> 00:14:45,440 Will the AI search do better or worse than their current search 254 00:14:45,440 --> 00:14:47,440 methods? And what does it mean in terms 255 00:14:47,440 --> 00:14:51,520 of adoption and governance and what impact will that have on 256 00:14:51,520 --> 00:14:54,600 healthcare? So when I started off with this 257 00:14:54,600 --> 00:14:59,200 and the first time I sat with my supervisors, Dr Von Reckling 258 00:14:59,200 --> 00:15:03,680 Housing and Doctor Zuba, they immediately cut the scope of my 259 00:15:03,680 --> 00:15:07,200 dissertation. I was planning on doing an AI 260 00:15:07,200 --> 00:15:12,040 search of hundreds of thousands, if not millions of documents, 261 00:15:12,560 --> 00:15:17,760 and being a lone researcher, we actually set that set and cut 262 00:15:17,760 --> 00:15:24,120 that corpus down to just 439 documents, but specifically in 263 00:15:24,120 --> 00:15:28,200 the ACLS domain. And the reason they did that was 264 00:15:28,200 --> 00:15:30,400 very simple. So if I got into trouble or 265 00:15:30,400 --> 00:15:35,440 wasn't sure about what a lexical search was returning or what a 266 00:15:35,440 --> 00:15:38,680 semantic search was returning, I could very easily sit with those 267 00:15:38,680 --> 00:15:41,200 400 odd documents and work it out myself. 268 00:15:41,200 --> 00:15:44,720 Or A, it's a very small amount for one person to work through 269 00:15:44,840 --> 00:15:49,200 and B, understood the material. And that would basically help me 270 00:15:49,200 --> 00:15:52,600 sort of pull in the reins and just help me understand my own 271 00:15:52,600 --> 00:15:56,400 research. So what we ended up finding or 272 00:15:56,400 --> 00:16:01,680 what I ended up finding was although the AI search 273 00:16:01,680 --> 00:16:06,080 understood my query a lot better, increased the recall or 274 00:16:06,080 --> 00:16:08,640 the amount of articles that are relevant to my query. 275 00:16:08,800 --> 00:16:13,920 The AI search brought back a very different set of documents 276 00:16:13,920 --> 00:16:18,480 than the lexical search. And that is a problem because 277 00:16:18,480 --> 00:16:21,280 one of the things that are surfaced in that research is 278 00:16:21,280 --> 00:16:25,640 with a fixed corpus of information is you could run the 279 00:16:25,640 --> 00:16:29,120 exact queries or two researchers, they could ask the 280 00:16:29,120 --> 00:16:32,120 exact same question and depending on which search 281 00:16:32,120 --> 00:16:35,320 mechanism they used, they get back very different results. 282 00:16:35,320 --> 00:16:39,400 Often in research, we talk about the the smoking gun article, and 283 00:16:39,400 --> 00:16:42,440 that is that elusive article where they could potentially 284 00:16:42,440 --> 00:16:46,120 either derail or significantly shift the outcomes of your 285 00:16:46,120 --> 00:16:48,880 research. And if you have this mismatch 286 00:16:48,880 --> 00:16:52,640 between your traditional search methods, your lexical search and 287 00:16:52,640 --> 00:16:56,600 your AI search, that is going to create a real problem if you're 288 00:16:56,600 --> 00:16:59,040 going to miss very relevant articles. 289 00:16:59,480 --> 00:17:03,920 So my research basically did quite a deep introduction into 290 00:17:03,920 --> 00:17:07,040 how these AI or LLM search mechanisms worked. 291 00:17:07,160 --> 00:17:11,480 We took a a bit of a tour of how the current lexical systems 292 00:17:11,480 --> 00:17:13,640 were. Too often people will take the 293 00:17:13,640 --> 00:17:16,160 underlying mechanisms for pubmitting Google Scholar for 294 00:17:16,160 --> 00:17:17,520 granted. So we went into that. 295 00:17:17,800 --> 00:17:21,839 They did a comparison did did quite deep into the statistical 296 00:17:21,839 --> 00:17:25,000 side of things, but did a very clear comparison of the results 297 00:17:25,000 --> 00:17:27,760 of the AI search. The traditional search wrote up 298 00:17:27,760 --> 00:17:29,880 what that all meant. And the bottom line or the 299 00:17:29,880 --> 00:17:34,080 conclusion of either research is that it's not about the 300 00:17:34,120 --> 00:17:38,400 technology, it's that it's about the the evidence itself. 301 00:17:38,440 --> 00:17:41,600 And that search, whether it's traditional search, IR search, 302 00:17:41,600 --> 00:17:45,200 is an epistemic gatekeeper. So whenever you conduct 303 00:17:45,200 --> 00:17:47,800 research, you have to pay real attention to the search 304 00:17:47,800 --> 00:17:50,640 mechanism you're using, understand how it's going to 305 00:17:50,640 --> 00:17:52,920 bring up the results relative to your research. 306 00:17:52,920 --> 00:17:58,000 And that itself needs to be discussed and disclosed in 307 00:17:58,000 --> 00:18:00,280 whichever research paper you're working on. 308 00:18:03,200 --> 00:18:08,160 So what is your comment to a academic researcher who just 309 00:18:08,160 --> 00:18:11,840 refuses to accept anything coming out of generative AI? 310 00:18:11,840 --> 00:18:17,080 So interestingly, I would agree with that is if you're using a 311 00:18:17,080 --> 00:18:21,240 general purpose generative AI like check GPT or Gemini is 312 00:18:21,240 --> 00:18:23,520 fantastic. Of those products are is I would 313 00:18:23,520 --> 00:18:25,840 agree with them. Don't use those for your 314 00:18:25,840 --> 00:18:30,800 research, however. They are the AI search 315 00:18:30,800 --> 00:18:34,520 mechanisms that are very powerful in helping surface 316 00:18:34,520 --> 00:18:37,440 information from very large bodies of data. 317 00:18:37,440 --> 00:18:39,720 The one is conducting the research. 318 00:18:39,720 --> 00:18:42,240 You really have to be a master of the tools that you're using, 319 00:18:42,240 --> 00:18:45,120 or at least understand where you should use something and where 320 00:18:45,120 --> 00:18:47,920 you shouldn't. And I have found that when 321 00:18:47,920 --> 00:18:51,880 discussing with AI, it's always more useful to discuss 322 00:18:52,520 --> 00:18:55,760 specifically which part of AI you're using because the term AI 323 00:18:55,760 --> 00:18:58,720 encompasses probably at the moment, hundreds of thousands, 324 00:18:58,720 --> 00:19:00,160 if not millions of different tools. 325 00:19:00,280 --> 00:19:03,880 So it helps be very specific. At this moment, I still 326 00:19:03,880 --> 00:19:07,840 discourage the use of generative AI for research, but I do 327 00:19:07,840 --> 00:19:12,200 encourage the other AI search mechanisms for just helping 328 00:19:12,720 --> 00:19:16,360 trawl through these millions of documents, part of the results 329 00:19:16,360 --> 00:19:19,760 of my dissertation. So looking forward to the future 330 00:19:20,120 --> 00:19:25,960 is that hybrid use of the traditional way and the AI way 331 00:19:26,080 --> 00:19:29,560 are very effective in surfacing additional results and just 332 00:19:29,560 --> 00:19:31,560 helping you understand what you have. 333 00:19:32,000 --> 00:19:37,960 So throughout my DHS I've gone through to almost 14,000 334 00:19:37,960 --> 00:19:41,520 documents, mostly at the abstract level, but without some 335 00:19:41,520 --> 00:19:44,400 of these more sophisticated AR tools it would have been very 336 00:19:44,400 --> 00:19:47,520 difficult to get any sort of understanding out of that sheer 337 00:19:47,520 --> 00:19:50,880 volume of data. So what is your suggestion since 338 00:19:50,920 --> 00:19:55,160 you've done all the work to the the medic, the nurse, the doc 339 00:19:55,160 --> 00:19:58,920 out there who is using AI? What platform do you suggest? 340 00:19:59,320 --> 00:20:01,520 I've been looking at perplexity as an option. 341 00:20:01,600 --> 00:20:05,080 Obviously the big ones like ChatGPT, everything that it 342 00:20:05,080 --> 00:20:08,560 comes up with is bollocks. That's my assumption, but what 343 00:20:08,560 --> 00:20:11,240 about some of these? That are specifically aimed at 344 00:20:11,240 --> 00:20:14,560 academia. I had to chocolate that comments 345 00:20:14,640 --> 00:20:17,200 a brick. So I heard another comment from 346 00:20:17,200 --> 00:20:20,520 a conference I was watching where the speaker said if you 347 00:20:20,520 --> 00:20:24,920 want ChatGPT on LLM to give you an answer, you have to tell it 348 00:20:24,920 --> 00:20:26,960 what the answer is. And that's obviously I think 349 00:20:27,040 --> 00:20:29,440 call out some of your frustrations with the tool. 350 00:20:30,120 --> 00:20:33,840 What I would suggest on the research side is again, do your 351 00:20:33,840 --> 00:20:35,840 research. There are more and more 352 00:20:35,840 --> 00:20:39,800 medically rated or research or medical research rated tools 353 00:20:39,800 --> 00:20:41,760 coming out. I'm sure that the, the bigger 354 00:20:41,760 --> 00:20:44,920 platforms, your pub meds, your Google scholars, many of the 355 00:20:44,920 --> 00:20:49,400 journals are starting to augment the AI or augment the existing 356 00:20:49,400 --> 00:20:52,440 searches with AI. When you are embarking on your 357 00:20:52,440 --> 00:20:58,680 research is take the tools that you're used to go and have a 358 00:20:58,680 --> 00:21:02,760 look to see have they augmented that with any sort of Arkansas 359 00:21:02,760 --> 00:21:07,840 tooling and then try and understand how do I leverage 360 00:21:07,840 --> 00:21:11,400 that to help my research. The big take away is if you 361 00:21:11,400 --> 00:21:14,520 don't fully understand the consequences of using Arkansas 362 00:21:14,520 --> 00:21:18,400 in your research, you'll probably sort of best to either 363 00:21:18,400 --> 00:21:21,360 reach out to someone who does understand how to use the tools 364 00:21:21,920 --> 00:21:25,560 or just proceed slowly on a very sandboxed environment. 365 00:21:25,800 --> 00:21:28,640 But definitely don't go in pedal to the middle. 366 00:21:28,760 --> 00:21:33,240 Start bringing AI in different flavours into your research. 367 00:21:33,320 --> 00:21:36,400 You all get to start getting things like semantic drift, 368 00:21:36,400 --> 00:21:40,360 model drift and stuff which can really mess up your results. 369 00:21:40,360 --> 00:21:43,160 I. Am I'm a poor viewer for a few 370 00:21:43,160 --> 00:21:46,000 academic journals. And the first thing that I 371 00:21:46,000 --> 00:21:50,200 always do is I look up all of their their journals that 372 00:21:50,200 --> 00:21:53,040 they've cited and look for DOI numbers. 373 00:21:53,080 --> 00:21:57,600 And that is a very quick way to see if one of our students or 374 00:21:57,600 --> 00:22:01,040 artist, whoever submitted the journal or article created 375 00:22:01,040 --> 00:22:04,240 bollocks using AI. And, and I have at least six 376 00:22:04,240 --> 00:22:08,680 Times Now have immediately kicked back a paper because 377 00:22:08,680 --> 00:22:12,560 their DOI numbers were fake or the research that they're 378 00:22:12,560 --> 00:22:15,280 quoting is fake. And that's before I even start 379 00:22:15,520 --> 00:22:18,040 looking line by line at what they road. 380 00:22:18,120 --> 00:22:23,440 Yeah, that's a real problem, not just in research is part of our 381 00:22:23,440 --> 00:22:25,560 work. I do a lot of medical education 382 00:22:25,560 --> 00:22:29,600 developments in the US and the exact problem you describe is 383 00:22:29,600 --> 00:22:32,840 rearing its head in that environment where the issue is. 384 00:22:32,840 --> 00:22:36,280 If you use something like chat TPT to generate training 385 00:22:36,280 --> 00:22:40,120 material, the issue is that particular system is designed 386 00:22:40,200 --> 00:22:44,960 for generalised day-to-day use, the layperson, so not 387 00:22:44,960 --> 00:22:46,840 specifically medical professionals. 388 00:22:47,000 --> 00:22:50,600 The one thing with AI is the output depends on how it's been 389 00:22:50,600 --> 00:22:52,880 trained. So if you are going to use it to 390 00:22:52,880 --> 00:22:56,880 develop very specialised medical material like we do at Coram in 391 00:22:56,880 --> 00:23:00,440 our environments, you do really need to, if you are going to use 392 00:23:00,440 --> 00:23:04,880 it, use a system that supports that particular medical medical 393 00:23:04,880 --> 00:23:06,960 knowledge. There are systems out there that 394 00:23:06,960 --> 00:23:10,920 will only generate specific or accurate citations, But even 395 00:23:10,920 --> 00:23:14,760 then, the researcher or the author of that paper, I must go 396 00:23:14,760 --> 00:23:19,320 off, not just take the AI, this word for it and go off read the 397 00:23:19,320 --> 00:23:23,520 DRI, make sure it actually correctly supports your your 398 00:23:23,520 --> 00:23:26,560 claim in your paper. But I think you and I take the 399 00:23:26,560 --> 00:23:29,080 similar approach. My approach is a just read the 400 00:23:29,080 --> 00:23:32,160 abstract, then go and check the citations. 401 00:23:32,160 --> 00:23:37,080 I like to make sure that we're checking a citation is, is it 402 00:23:37,080 --> 00:23:40,400 just a random citation put there to support effect and you will 403 00:23:40,400 --> 00:23:44,160 have a 200 page citation article and all you're doing is 404 00:23:44,160 --> 00:23:47,080 surfacing 1 fact. And that's where I find AI can 405 00:23:47,080 --> 00:23:50,520 actually help. If you have a citation again to 406 00:23:50,520 --> 00:23:54,480 our 200 page article to take that article, put it through an 407 00:23:54,480 --> 00:24:00,280 LLM and just help me understand how does the spec relate to the 408 00:24:00,280 --> 00:24:03,280 material that's talking about? Because often it's not clear if 409 00:24:03,280 --> 00:24:07,280 the article is not cited correctly that this is something 410 00:24:07,280 --> 00:24:10,200 we are working on at the college, slowly but surely. 411 00:24:10,200 --> 00:24:13,920 In fact, there's one of my tasks is actually developping a 412 00:24:13,920 --> 00:24:19,520 training module to guide our students on how to use AI safely 413 00:24:19,520 --> 00:24:22,160 and correctly to produce research material. 414 00:24:22,160 --> 00:24:26,760 And it's, it's a pretty complicated field, but ChatGPT, 415 00:24:26,840 --> 00:24:29,920 it's a blessing and a curse because it's so easy to use. 416 00:24:29,920 --> 00:24:32,480 And those results are very plausible. 417 00:24:32,640 --> 00:24:35,960 But if you don't do the work to go and check every little word 418 00:24:35,960 --> 00:24:39,080 that comes out of it, you're going to end up in academic 419 00:24:39,160 --> 00:24:42,320 trouble, unfortunately. It's interesting to see how 420 00:24:42,320 --> 00:24:46,600 quick that AOI is taking off. So 12 months ago I couldn't 421 00:24:46,600 --> 00:24:51,360 spell ChatGPT and now I'm using it every day and, and making 422 00:24:51,360 --> 00:24:55,240 sure that everything that I see come out of it is is verified on 423 00:24:55,240 --> 00:24:58,360 on some way. So from 12 months ago, having 424 00:24:58,360 --> 00:25:03,040 very limited AI capability to the point where now you can type 425 00:25:03,040 --> 00:25:07,600 in create a PhD for me in this and then it would be a 75 page 426 00:25:07,760 --> 00:25:10,000 dissertation. What's going to happen in 12 427 00:25:10,000 --> 00:25:12,880 months? Yes, definitely on a exponential 428 00:25:12,920 --> 00:25:16,240 trend ever. There is a fair amount of 429 00:25:16,240 --> 00:25:21,800 research coming out that LLMS are not going to keep growing 430 00:25:21,800 --> 00:25:24,640 exponentially. It's just physically impossible. 431 00:25:24,640 --> 00:25:26,600 In terms of the the processing power. 432 00:25:26,800 --> 00:25:30,600 The one thing that is evolving very quickly is the tooling 433 00:25:30,600 --> 00:25:34,640 around the LLMS. And specifically there's a term 434 00:25:34,640 --> 00:25:37,360 called RAG or retrieval of meta generation. 435 00:25:37,560 --> 00:25:41,000 Say for example, you wanted to write that 75 page dissertation 436 00:25:41,000 --> 00:25:44,760 on a tropical disease. That's a malaria for example. 437 00:25:45,560 --> 00:25:49,640 It's simply a case of you provide the LLM with as much 438 00:25:49,640 --> 00:25:54,920 valid material as possible and just use the tool to generate 439 00:25:58,160 --> 00:26:01,120 material based on that and you sort of read it for accuracy. 440 00:26:01,120 --> 00:26:03,320 And I found this during my own dissertation chat. 441 00:26:03,320 --> 00:26:06,680 TPT and Gemma were a great help. Taking the material I had 442 00:26:06,680 --> 00:26:09,600 created and checked helped me improve the flow and 443 00:26:09,600 --> 00:26:11,680 occasionally would throw something else ahead. 444 00:26:11,680 --> 00:26:14,160 Did you consider this? But then it's a case of do not 445 00:26:14,160 --> 00:26:16,760 take his word for it. Take that suggestion, go and 446 00:26:16,760 --> 00:26:20,600 research it yourself and if it's valid, bring it back into your 447 00:26:20,600 --> 00:26:22,080 whatever paper you're working on. 448 00:26:22,280 --> 00:26:26,160 But if you take all your work, bundle it up, throw it over the 449 00:26:26,160 --> 00:26:33,240 fence into AI's backyard is you are going to get back something 450 00:26:33,240 --> 00:26:35,720 that is just not of an acceptable standard and 451 00:26:35,720 --> 00:26:38,680 potentially even dangerous. Good advice. 452 00:26:39,760 --> 00:26:44,040 So, Doctor Clapper, what advice do you have for the new medic, 453 00:26:44,040 --> 00:26:47,320 the new nurse, the new doc who is just starting out on the 454 00:26:47,320 --> 00:26:51,840 career in austere medicine? So my advice for that is 455 00:26:51,920 --> 00:26:55,040 whatever environments you're going to, whether that's the, 456 00:26:55,200 --> 00:26:59,160 the mountains, the deserts or the jungle is before you pick up 457 00:26:59,160 --> 00:27:02,880 the the books that understand all the diseases, the complex 458 00:27:02,880 --> 00:27:05,800 medical stuff that we have to deal with that environment is 459 00:27:05,800 --> 00:27:10,960 make sure that you are personally or physically fit and 460 00:27:10,960 --> 00:27:13,400 capable of being in that environments. 461 00:27:13,520 --> 00:27:17,040 Whether it's the physical or the the mental rigour that you're 462 00:27:17,040 --> 00:27:20,600 going into is an issue that I've encountered time and time again. 463 00:27:20,600 --> 00:27:23,560 Whether it's just a lack of physical fitness is if you get 464 00:27:23,560 --> 00:27:26,960 deployed into one of these environments, those deployments, 465 00:27:26,960 --> 00:27:29,440 first of all, are very expensive to get you into that 466 00:27:29,440 --> 00:27:31,880 environment. A huge amount of money has been 467 00:27:31,880 --> 00:27:36,600 spent to transport you the set up, whatever living environment 468 00:27:36,600 --> 00:27:38,800 you need, whatever clinics and so on. 469 00:27:39,360 --> 00:27:42,840 If you arrive there and decide physically you can't handle it, 470 00:27:42,840 --> 00:27:47,320 this is not for you and so on, that can become a problem. 471 00:27:47,320 --> 00:27:49,840 And I do understand there's an acclimatisation to any 472 00:27:49,840 --> 00:27:52,840 deployment that you might do. But my advice, and this is the 473 00:27:52,840 --> 00:27:57,160 advice I gave to my your rescuers this morning is before 474 00:27:57,160 --> 00:28:00,640 you go into the difficult rescue stuff in the wilderness, get 475 00:28:00,640 --> 00:28:04,400 yourself out there, take a backpack, go and spend a few 476 00:28:04,400 --> 00:28:08,280 nights ago and pivot in one of the the safer sort of wilderness 477 00:28:08,280 --> 00:28:11,480 environments. Go hiking, go climbing, be a 478 00:28:11,480 --> 00:28:15,400 Mountaineer, be a hiker, go on a lightweight expedition. 479 00:28:15,480 --> 00:28:18,720 So don't just build up your medical knowledge, just build up 480 00:28:18,720 --> 00:28:22,800 your personal experience in those environments. 481 00:28:22,880 --> 00:28:26,560 And without that, you do risk potentially either injuring 482 00:28:26,560 --> 00:28:30,240 yourself or becoming a liability to the mission that you're 483 00:28:30,240 --> 00:28:33,080 trying to so nobly help. That's a challenge. 484 00:28:33,200 --> 00:28:37,360 From deployment to deployment, you're going to different areas, 485 00:28:37,360 --> 00:28:40,920 aren't you? Yeah, it's as much as you can 486 00:28:40,920 --> 00:28:47,440 try prepare for everything. It's there's always something 487 00:28:47,440 --> 00:28:50,200 new, whether it's just the severity of the patients you 488 00:28:50,200 --> 00:28:52,680 might deal with, if you're not used to altitude, if you're not 489 00:28:52,680 --> 00:28:55,160 used to the humidity. So whatever environment you're 490 00:28:55,160 --> 00:29:00,000 going into is just to make sure that you understand what the 491 00:29:00,000 --> 00:29:03,120 environment is going to be like. And again, do whatever you can 492 00:29:03,120 --> 00:29:07,480 to prepare yourself physically for operating in that 493 00:29:07,480 --> 00:29:09,760 environment. And before we started recording, 494 00:29:09,760 --> 00:29:12,920 you mentioned something about your local Air Force. 495 00:29:12,960 --> 00:29:16,360 What was that about? So this morning I actually was 496 00:29:16,360 --> 00:29:19,600 asked just to go and help the local mouse and search and 497 00:29:19,600 --> 00:29:23,720 rescue and some of the other teams was to just help them with 498 00:29:23,720 --> 00:29:28,640 a dry run for helicopter rescue. And I dispense the same advice 499 00:29:28,640 --> 00:29:31,480 to them. Is that before you decide to get 500 00:29:31,480 --> 00:29:37,200 on the helicopter and one of the biggest risks to the aircraft or 501 00:29:37,200 --> 00:29:41,200 is the rescuer themselves if they don't have all the kits 502 00:29:41,200 --> 00:29:44,040 squared away, if they aren't very familiar with all the 503 00:29:44,040 --> 00:29:47,560 equipment that they are using there, It's a Caribbean safety 504 00:29:47,560 --> 00:29:48,880 lanyard. If they're not comfortable with 505 00:29:48,880 --> 00:29:51,400 their helmets, they're not comfortable with the PPE, the 506 00:29:51,400 --> 00:29:54,880 harness doesn't fit right. They're not quite sure how to 507 00:29:54,880 --> 00:29:56,720 dress for whether to lay a property. 508 00:29:56,880 --> 00:30:00,280 The second you step into the aviation environmental onto a 509 00:30:00,280 --> 00:30:04,520 helicopter, all of those seemingly small nitpicky issues 510 00:30:04,600 --> 00:30:07,520 immediately get compounded. And I demonstrated the several 511 00:30:07,520 --> 00:30:12,240 times where we had over 20 rescuers and the capacity of the 512 00:30:12,280 --> 00:30:15,760 aircraft to 16. If I'm trying to get 16 rescuers 513 00:30:15,760 --> 00:30:19,880 out the aircraft, but if there's even a 32nd delay where someone 514 00:30:19,880 --> 00:30:22,880 is tumbling a Caribbean can't get the air pro on need to 515 00:30:22,880 --> 00:30:25,880 adjust the goggles, that's times 20. 516 00:30:26,040 --> 00:30:29,200 You're looking at now potentially a 10/20 even a 517 00:30:29,200 --> 00:30:33,520 longer minute delay in getting everybody safely deployed and at 518 00:30:33,520 --> 00:30:36,160 the circles in back to my original point is just to make 519 00:30:36,160 --> 00:30:38,440 sure that you are very comfortable in that environment 520 00:30:38,440 --> 00:30:40,560 and that you can deal with the the task load. 521 00:30:40,680 --> 00:30:43,800 And the way you do that is take care of the small stuff first 522 00:30:43,800 --> 00:30:46,480 yourself. And I even gave the advice is 523 00:30:46,880 --> 00:30:50,520 right next to the TV remote in your lounge, put a carabiner to 524 00:30:50,520 --> 00:30:53,280 sit there, get it to a subconscious level, click it 525 00:30:53,280 --> 00:30:56,360 open, you know, and get a feel for exactly how it works. 526 00:30:56,560 --> 00:30:59,680 So when it comes time to use any of that equipment and anger, it 527 00:30:59,680 --> 00:31:02,760 is basically instinctual and second nature. 528 00:31:02,800 --> 00:31:07,160 It's the same advice I give to my students with A10 mil syringe 529 00:31:07,600 --> 00:31:09,280 in class. They should just have it in 530 00:31:09,280 --> 00:31:12,440 their hand and learn how to manipulate it with right hand, 531 00:31:12,440 --> 00:31:16,440 left hand, a whole bunch of different, different ways. 532 00:31:16,440 --> 00:31:18,000 It's it's the hand knowledge, isn't it? 533 00:31:18,040 --> 00:31:21,160 Yeah. It's a muscle memory. 534 00:31:21,160 --> 00:31:26,280 The when I first started out in in Martin Rescue took the same 535 00:31:26,280 --> 00:31:30,120 advice I'm giving now. I joined the Master Club of 536 00:31:30,120 --> 00:31:33,240 South Africa and every single weekend, every single 537 00:31:33,240 --> 00:31:37,760 opportunity I got, I was out there hiking, mining, technical 538 00:31:37,760 --> 00:31:41,720 Rd equipment and that pays huge dividends to this day in the 539 00:31:41,720 --> 00:31:43,760 rescue environment. That's good advice. 540 00:31:43,800 --> 00:31:48,400 So Michael, thank you for being on our podcast again and thank 541 00:31:48,400 --> 00:31:53,160 you for the 10 years that you've given to the college and you've 542 00:31:53,160 --> 00:31:56,280 gotten a doctorate out of it. I understand is reciprocal, but 543 00:31:56,480 --> 00:31:59,200 you are a valued member of our faculty. 544 00:31:59,200 --> 00:32:00,760 Thank you. Thank you every and I look 545 00:32:00,760 --> 00:32:06,000 forward to the next 10 years. Has been a presentation from the 546 00:32:06,000 --> 00:32:09,680 College of Remote and Offshore Medicine if you would like to 547 00:32:09,680 --> 00:32:13,040 earn CPD credit. For this podcast, you can join. 548 00:32:13,320 --> 00:32:16,520 The Council of members being a member of the college, gives you 549 00:32:16,520 --> 00:32:20,800 free CPD credit, free access for a virtual field guide and 550 00:32:20,800 --> 00:32:22,840 discounts on our e-learning courses. 551 00:32:23,240 --> 00:32:26,600 You can join the team. On our college website at 552 00:32:26,600 --> 00:32:29,160 Quorum. Dot Edu dot Mt.