WEBVTT 00:00.850 --> 00:02.850 Uh I'm gonna sit here because we're 00:02.850 --> 00:05.079 gonna basically be presenting um so 00:05.090 --> 00:07.630 great segue from the previous presenter , 00:07.639 --> 00:10.449 uh the , the Statistic . So what is , 00:10.460 --> 00:12.627 what , who are we , where we have ? Uh 00:12.627 --> 00:14.238 The Chief Digital Artificial 00:14.238 --> 00:16.293 Intelligence Office was developed in 00:16.293 --> 00:18.460 2022 . It previously was uh used to be 00:18.460 --> 00:20.293 part of what they call the joint 00:20.293 --> 00:21.904 artificial intelligence . Um 00:21.904 --> 00:23.849 Simplistically , our mission is to 00:23.849 --> 00:25.849 maximize data sharing uh to provide 00:25.895 --> 00:28.117 operability across the services and the 00:28.225 --> 00:30.385 cravat demands . Uh We do that and we 00:30.395 --> 00:33.785 recently sponsored the uh 2024 excuse 00:33.875 --> 00:37.076 me , 2020 3d MD uh Data Analytics and A 00:37.085 --> 00:39.116 I strategy um which you guys can 00:39.125 --> 00:41.125 actually go and look up and it will 00:41.125 --> 00:43.765 really start to paint a picture of what 00:43.776 --> 00:46.805 we're going to present here today . Um 00:46.905 --> 00:49.346 So who am I ? I am rich . Got to keep 00:49.546 --> 00:51.922 it . I lead the health portfolio . I'm 00:51.931 --> 00:54.098 the Chief of Health Data and Analytics 00:54.098 --> 00:55.875 at the Chief Digital Artificial 00:55.875 --> 00:58.032 Intelligence Agency . And then I have 00:58.041 --> 01:00.881 uh Doctor Charles Bosch , he is my lead 01:00.891 --> 01:02.724 data scientist . And today we're 01:02.724 --> 01:04.558 actually going to walk through a 01:04.558 --> 01:06.669 particular product uh line that we've 01:06.669 --> 01:08.724 been working for quite some time and 01:08.724 --> 01:10.947 show how you can actually get return on 01:10.947 --> 01:13.141 investment related to readiness . Um 01:13.152 --> 01:15.263 You know , we've worked , everybody's 01:15.263 --> 01:17.540 chasing metrics um for what , whatever 01:17.550 --> 01:19.717 reason or purpose , whether that's all 01:19.717 --> 01:21.717 the way up to the depth se . And so 01:21.717 --> 01:24.050 even the depth se dep is looking at how , 01:24.050 --> 01:26.161 how do we evaluate our force , how do 01:26.161 --> 01:28.272 we evaluate our readiness ? Um How do 01:28.272 --> 01:30.383 we look holistically across on what's 01:30.383 --> 01:32.940 our return on investment ? Um So that's 01:32.949 --> 01:35.116 what we're gonna be discussing today , 01:35.139 --> 01:37.529 our overall objectives . Um without 01:37.540 --> 01:39.596 reading the slides , I'll simply put 01:39.596 --> 01:41.651 this today . We're here to introduce 01:41.651 --> 01:43.818 who CD A O is , who we're partnering , 01:43.818 --> 01:46.239 what we're working on . And then we're 01:46.250 --> 01:48.209 gonna go through a particular uh 01:48.220 --> 01:50.300 machine learning model that showed a 01:50.309 --> 01:52.420 return on investment that we did with 01:52.420 --> 01:54.531 the Armed Forces , Wellness Centers . 01:54.531 --> 01:56.642 And then the last piece is the aspect 01:56.642 --> 01:59.250 of privacy . Um So when we talk about 01:59.260 --> 02:00.982 big big data exchange and data 02:00.982 --> 02:02.982 interoperability , it's key that we 02:02.982 --> 02:05.093 always uh look at it from the privacy 02:05.093 --> 02:07.316 components . Um Whether that's wearable 02:07.316 --> 02:09.371 devices , whether that's the medical 02:09.371 --> 02:11.371 information , but how do we collect 02:11.371 --> 02:13.593 that big data um to be able to drive uh 02:13.593 --> 02:17.119 decision making and outputs . So again , 02:17.130 --> 02:19.399 a little bit about two CD A O is and 02:19.410 --> 02:21.740 what our mission is . Uh CD OS mission 02:21.750 --> 02:24.470 is the to accelerate dod adoption of 02:24.479 --> 02:26.257 data analytics and A I from the 02:26.257 --> 02:28.479 boardroom to the battlefield . Uh Hence 02:28.479 --> 02:30.479 the title today is how can we drive 02:30.479 --> 02:32.610 readiness and let the let from the 02:32.619 --> 02:34.675 boardroom to the battlefield . Um If 02:34.675 --> 02:36.897 you look on the right hand side of this 02:36.897 --> 02:39.063 image right here , you can see that it 02:39.063 --> 02:41.175 starts with quality data , then it is 02:41.175 --> 02:43.119 driven by insightful analytics and 02:43.119 --> 02:45.286 metrics . And that actually how we get 02:45.286 --> 02:47.619 to respond to A I . Um everybody's like , 02:47.619 --> 02:47.509 oh , it's a I it's fancy , it's cool , 02:47.520 --> 02:49.576 right ? Um So , and , and there is , 02:49.576 --> 02:51.798 there's a lot of great capability about 02:51.798 --> 02:54.020 A I but it starts with foundation , the 02:54.020 --> 02:56.076 quality data , the collection of the 02:56.076 --> 02:58.242 data um where we actually store that . 02:58.242 --> 03:00.131 How do we put that into a data la 03:00.131 --> 03:02.298 concept or what we call API management 03:02.298 --> 03:04.464 um to be able to get access to that to 03:04.464 --> 03:06.759 actually drive the insightful analytics 03:06.770 --> 03:09.240 and um uh the psychoanalytic and 03:09.250 --> 03:13.020 metrics who is the health portfolio . 03:13.320 --> 03:15.320 So the health portfolio actually uh 03:15.320 --> 03:17.431 sits underneath what we call business 03:17.431 --> 03:19.376 analytics . And the purpose of the 03:19.376 --> 03:22.179 health portfolio is um simplistically 03:22.190 --> 03:25.330 is we focus on public health metaphor 03:25.899 --> 03:28.010 of human performance . It's where all 03:28.010 --> 03:30.010 three of those bubbles on the right 03:30.010 --> 03:32.169 hand side um come together and are 03:32.179 --> 03:34.346 currently our structure on the block . 03:34.346 --> 03:36.568 You'll see that we have uh we have what 03:36.568 --> 03:38.623 they call customer focussed , lot of 03:38.623 --> 03:40.790 efforts and then we have core funded , 03:40.790 --> 03:42.846 lot of efforts . So our core funding 03:42.846 --> 03:42.796 line of efforts is um again , 03:42.805 --> 03:44.749 maximizing data sharing to support 03:44.749 --> 03:46.815 combatant transit services . Um Our 03:46.826 --> 03:49.065 customer focused line of effort is we 03:49.076 --> 03:51.416 are working with Total Force Fitness um 03:51.425 --> 03:53.314 to help establish the Total Force 03:53.314 --> 03:55.425 fitness framework . And that includes 03:55.425 --> 03:57.203 the analytical piece and the uh 03:57.203 --> 03:59.425 respective data sources that are needed 03:59.425 --> 04:01.647 to be able to drop measures and metrics 04:01.647 --> 04:03.258 for total force fitness . Um 04:03.258 --> 04:05.092 Additionally , we have , we have 04:05.092 --> 04:07.203 another use case , we're working with 04:07.203 --> 04:09.258 us Metcom . And that's really how we 04:09.258 --> 04:11.036 increase the efficiency and the 04:11.036 --> 04:13.258 throughput of bringing in recruits into 04:13.258 --> 04:15.258 the Meps locations and getting them 04:15.258 --> 04:17.425 through the Meps locations at a faster 04:17.425 --> 04:19.536 speed uh given the quantity of data . 04:21.040 --> 04:23.262 And then the last piece uh a little bit 04:23.262 --> 04:25.373 more about the the health portfolio . 04:25.373 --> 04:27.373 And this again ties to the strategy 04:27.373 --> 04:29.484 that I talked about earlier . Um It's 04:29.484 --> 04:31.707 around our service offerings and how we 04:31.707 --> 04:33.707 um plan to achieve enter enterprise 04:33.707 --> 04:35.873 wide data access and availability . Um 04:35.873 --> 04:37.818 Some of you guys may have heard of 04:37.818 --> 04:39.707 Ivanna , but Ivanna is one of our 04:39.707 --> 04:41.707 platforms that we work on and it is 04:41.707 --> 04:43.596 considered the Dod Federated data 04:43.596 --> 04:46.274 catalog . Um So our initiative there is 04:46.285 --> 04:48.605 to bring in all the respective data 04:48.614 --> 04:50.684 sources , at least at minimum from a 04:50.695 --> 04:52.639 catalog perspective . If it can be 04:52.639 --> 04:54.934 connected to an API we would be able to 04:54.945 --> 04:57.375 post that and show that . Um But to be 04:57.385 --> 04:59.825 able to go and find what data do I need , 04:59.834 --> 05:02.056 what data elements do I need ? How do I 05:02.056 --> 05:04.375 tie readiness to healthcare to physical 05:04.385 --> 05:06.554 to all the other um different domains 05:06.565 --> 05:08.676 whether that's under H two F or under 05:08.709 --> 05:10.542 to force fitness . So we do that 05:10.542 --> 05:12.765 through data discovery , data catalog , 05:12.765 --> 05:15.098 and data access , uh and data analytics . 05:15.098 --> 05:17.320 And the last thing I would say is um CD 05:17.320 --> 05:19.320 O is a platform agnostic . So we're 05:19.320 --> 05:21.431 working across uh with all services . 05:21.431 --> 05:23.542 Um Some of you guys made from an army 05:23.542 --> 05:25.542 perspective , their big platform is 05:25.542 --> 05:27.653 called Vantage . Um So we have direct 05:27.653 --> 05:29.765 connections working with Vantage . Uh 05:29.765 --> 05:31.709 Navy operates on a platform called 05:31.709 --> 05:33.931 Jupiter . It sits on top of the and the 05:33.931 --> 05:36.042 Air Force has one quality vision . Um 05:36.042 --> 05:37.820 So again , feeding back off the 05:37.820 --> 05:39.987 previous conversation you had . Um the 05:39.987 --> 05:42.153 services in the grad fans are thinking 05:42.153 --> 05:42.024 through this of how we integrate and 05:42.035 --> 05:44.202 how we connect these together . And so 05:44.202 --> 05:46.313 we provide , we help provide the tech 05:46.313 --> 05:48.479 stack . Uh whether that's on a payment 05:48.479 --> 05:50.424 data is a capability or another um 05:50.424 --> 05:52.368 platform to be able to provide the 05:52.368 --> 05:54.591 tools and capabilities um to be able to 05:54.591 --> 05:57.540 integrate data uh over . So for today's 05:57.549 --> 05:59.660 purpose , though , we're gonna really 05:59.660 --> 06:01.438 start diving into the readiness 06:01.438 --> 06:03.382 prediction model . Um So with that 06:03.382 --> 06:05.382 being said , I want to turn over to 06:05.382 --> 06:07.493 doctor Trump that's been leading this 06:07.493 --> 06:09.493 effort for quite some time um on my 06:09.493 --> 06:11.660 team and to talk through the readiness 06:11.660 --> 06:13.827 and the TFF framework and specifically 06:13.827 --> 06:15.549 the outcomes that we've seen . 06:17.989 --> 06:20.510 So the original task was identify the 06:20.519 --> 06:22.470 measures and metrics for the total 06:22.480 --> 06:24.591 course fitness domains that should be 06:24.591 --> 06:27.359 recommended in policy and processes for 06:27.399 --> 06:29.399 all the programs including physical 06:29.399 --> 06:32.239 fitness . So at the time , we asked , 06:32.250 --> 06:34.417 ok , there's two ways to do this . You 06:34.417 --> 06:36.639 need to look at the commonality between 06:36.639 --> 06:38.861 the measures of metrics or you can look 06:38.861 --> 06:40.694 at your definition of individual 06:40.694 --> 06:42.861 readiness and we can see what physical 06:42.861 --> 06:45.529 fitness and other metrics predict load 06:45.619 --> 06:47.649 on individual readiness . 06:49.440 --> 06:51.529 And they said we don't have a 06:51.540 --> 06:54.000 definition of , right . I was like , 06:54.010 --> 06:56.010 huh , we're the office of personnel 06:56.010 --> 06:58.232 Raiders . And I was new to the military 06:58.232 --> 07:00.232 and I was like , Moneyball came out 07:00.232 --> 07:02.177 like 15 years ago . You don't have 07:02.177 --> 07:01.640 money ball in the military . You don't 07:01.649 --> 07:04.130 have s metrics on soldiers yet . And 07:04.140 --> 07:06.251 they're like , well , and I read Drug 07:06.251 --> 07:08.473 club , one tactical readiness , nuclear 07:08.473 --> 07:10.584 readiness , no individual readiness . 07:10.929 --> 07:12.707 And we listened to a lot of the 07:12.707 --> 07:14.762 stakeholders , we went to SOCOM , we 07:14.762 --> 07:16.985 went to different research environments 07:16.985 --> 07:18.985 like army , a group and they talked 07:18.985 --> 07:21.040 about how the closest thing they had 07:21.040 --> 07:22.818 for a definition , a measurable 07:22.818 --> 07:22.809 definition . And then you can predict 07:22.820 --> 07:25.149 so you can do back and stuff was 07:25.160 --> 07:27.104 individual medical readiness . But 07:27.104 --> 07:29.271 that's what not what they meant . That 07:29.271 --> 07:29.059 was the checklist of , can you get out 07:29.070 --> 07:31.126 the door ? Have you done your will ? 07:31.126 --> 07:33.292 Have you done your vaccines ? Really ? 07:33.292 --> 07:35.348 What they meant was , can you deploy 07:35.348 --> 07:37.403 and can you complete your deployment 07:37.403 --> 07:39.514 without coming back early for disease 07:39.514 --> 07:42.730 found in ? So we wrote that down and 07:42.739 --> 07:45.220 that is how the chairman's draft 07:45.250 --> 07:47.194 definition of individual ratings , 07:47.194 --> 07:49.306 individual ratings too . So then what 07:49.306 --> 07:51.417 we decided to do is ok , let's see if 07:51.417 --> 07:53.528 we can measure that per service . And 07:53.528 --> 07:55.639 then let's see if we can predict that 07:55.639 --> 07:57.806 per service . But before we try to fix 07:57.806 --> 07:59.972 something , we can't predict something 07:59.972 --> 08:02.083 that's random . So we decided to look 08:02.083 --> 08:04.250 at a bunch of deployments from 2011 to 08:04.250 --> 08:06.339 2017 and say when you look at the 08:06.350 --> 08:08.929 deployment and you look at the timeline 08:08.940 --> 08:10.996 of the deployment , the disease , no 08:10.996 --> 08:13.051 battle injury happened randomly like 08:13.051 --> 08:15.218 uniformly throughout the deployment or 08:15.218 --> 08:17.329 did it look like a resiliency issue ? 08:17.329 --> 08:19.329 Like the fallout rate of a marathon 08:19.329 --> 08:21.273 that has just been predictable and 08:21.273 --> 08:25.109 preventable ? So to the 08:25.119 --> 08:27.540 left shows the disease , non Matal 08:27.549 --> 08:30.089 injury timing for appointment to the US 08:30.100 --> 08:33.969 army in 2011 and 2017 . So we decided 08:33.979 --> 08:37.020 to try to predict this . So we found a 08:37.030 --> 08:39.252 research environment that already had a 08:39.252 --> 08:41.141 lot of the data you need across a 08:41.141 --> 08:42.929 medical , dental , physical um 08:42.940 --> 08:45.051 deployment history and the rest . And 08:45.051 --> 08:48.030 we said , let's reuse that platform 08:48.039 --> 08:49.817 because they already had a good 08:49.817 --> 08:51.983 research facilitation , IRB . They had 08:51.983 --> 08:54.206 to do that , they put it all together . 08:54.206 --> 08:56.428 So Google tried to commit suicide which 08:56.428 --> 08:58.483 is top of 3500 but disease not bowel 08:58.483 --> 09:01.760 injury is 7% . You can use A I models 09:01.770 --> 09:04.890 to predict a 7% outcome . So we did 09:05.539 --> 09:07.539 and predicted it really well at the 09:07.539 --> 09:10.169 unit level really well . We're not at 09:10.179 --> 09:12.346 the individual level , when you add up 09:12.346 --> 09:14.512 all the individual level predictions , 09:14.512 --> 09:16.623 you can see at the unit level , which 09:16.623 --> 09:18.735 ones we're gonna have 2% force loss , 09:18.735 --> 09:20.679 4% force loss and 6% force loss in 09:20.679 --> 09:22.790 theater . Based on what we knew about 09:22.790 --> 09:25.012 the deployment , what we knew about the 09:25.012 --> 09:27.179 service numbers before they deployed , 09:27.179 --> 09:29.346 that's pretty valuable to commanders , 09:29.346 --> 09:31.123 to trainers , to all the rest , 09:31.123 --> 09:33.179 especially if you can start breaking 09:33.179 --> 09:35.401 down to why or what measures of me . So 09:35.401 --> 09:37.789 we bucket it down . Also the types of 09:37.799 --> 09:41.299 metrics . Type one , pure predictors is 09:41.309 --> 09:43.531 what I mean by pure predictors . You're 09:43.531 --> 09:45.587 not gonna change policy , you're not 09:45.587 --> 09:45.530 gonna do anything but still predict 09:45.539 --> 09:47.559 your outcome . That's type one top 09:47.570 --> 09:49.559 right patch of 09:52.330 --> 09:55.750 this . It's hard to see on the map . So 09:56.049 --> 09:59.359 this is your failure rate , this is 09:59.369 --> 10:02.260 your variable . So we wanted to see 10:02.270 --> 10:04.880 things that moved , ok ? And the first 10:04.890 --> 10:08.440 one was the , we saw something 10:08.450 --> 10:11.479 that 0 to 6 of this thing , what would 10:11.489 --> 10:13.750 happen . But once you have 6789 of 10:13.760 --> 10:15.593 these things , your failure went 10:15.593 --> 10:18.539 through the roof . That was was the 10:18.549 --> 10:20.660 number of central nervous system drug 10:20.660 --> 10:22.605 prescriptions within the 12 months 10:22.605 --> 10:26.590 before the 0.06 or 56789 Hill . 10:26.849 --> 10:30.150 It goes up by now , not need it between 10:30.159 --> 10:32.215 the doctor and patient and tell them 10:32.215 --> 10:34.048 they can't have a prescription . 10:34.048 --> 10:33.880 They're getting all these prescriptions 10:33.890 --> 10:36.057 for a reason . But if you're forced to 10:36.057 --> 10:38.057 plan and you know , you have a unit 10:38.057 --> 10:39.890 that's hopped up on 6789 central 10:39.890 --> 10:41.946 nervous system and drugs and they're 10:41.946 --> 10:43.946 gonna have more disease , no injury 10:43.946 --> 10:46.168 than the other unit . You might want to 10:46.168 --> 10:46.039 overstep that unit . So it's relevant 10:46.049 --> 10:48.271 from a protection perspective , even if 10:48.271 --> 10:50.327 you're not going to change policy or 10:50.327 --> 10:50.280 intervene in any way within the service 10:50.289 --> 10:52.345 number or the soldier . Second one , 10:52.710 --> 10:55.080 this is policy relevant barriers . We 10:55.090 --> 10:57.349 looked at the to call ratio that was 10:57.359 --> 10:59.470 monotonic in the way that you think . 10:59.830 --> 11:01.719 But the more interesting ones was 11:01.719 --> 11:03.608 months since last deployment , we 11:03.608 --> 11:05.774 actually saw a really interesting help 11:05.774 --> 11:08.669 from zero to here was around 24 months 11:08.679 --> 11:11.400 every extra month to dwell . After your 11:11.409 --> 11:13.631 last deployment , the failure rate went 11:13.631 --> 11:15.742 down and then there was an area where 11:15.742 --> 11:18.489 the failure was flattened from 24 to 36 11:18.500 --> 11:20.611 months and then afterwards started to 11:20.611 --> 11:22.778 go back up . Now this is a correlation 11:22.778 --> 11:24.833 one . So it's not causation . So you 11:24.833 --> 11:27.056 can't say , well , this is people being 11:27.056 --> 11:29.167 rusty or this is selection , but it's 11:29.167 --> 11:28.859 still interesting information that can 11:28.869 --> 11:31.760 help in form one flow policy . So that 11:31.770 --> 11:33.937 went up to the chairman , Secretary of 11:33.937 --> 11:36.159 Defense . And they said now I want this 11:36.159 --> 11:38.270 thing to inform all sorts of policy , 11:38.270 --> 11:40.492 not just TFF policy reading this policy 11:40.492 --> 11:40.090 and we want to move it to an 11:40.099 --> 11:43.049 operational system . Last one is where 11:43.059 --> 11:45.349 I went for your community . We found we 11:45.359 --> 11:47.303 thought looked like a magic pill . 11:47.303 --> 11:50.760 Where did you take the pill ? Then we 11:50.770 --> 11:52.770 rate from disease , not that injury 11:52.770 --> 11:56.710 went down . The more dose response in 11:56.719 --> 11:58.719 terms of the decrease in disease on 11:58.719 --> 12:00.886 bowel got . And so when we looked up , 12:00.886 --> 12:02.941 what is this magic pill ? And it was 12:02.941 --> 12:04.886 how the army Wellness Centers were 12:04.886 --> 12:06.997 keeping track of their appointments . 12:06.997 --> 12:09.219 They send a prescription for someone to 12:09.219 --> 12:11.441 go , but they actually went , it showed 12:11.441 --> 12:13.663 up in the the prescription records . So 12:13.663 --> 12:15.663 we can see the number of army . Now 12:15.663 --> 12:15.575 their armed forces on the center visits 12:15.655 --> 12:18.025 and that they have a dose response to 12:18.034 --> 12:20.294 increase readiness is to find that it 12:20.304 --> 12:22.895 decreases as like failure or force loss 12:22.905 --> 12:24.905 from people . So it's a pretty hard 12:24.905 --> 12:27.109 connection to readiness . And so that 12:27.119 --> 12:29.909 also then became the highlight . And I 12:29.919 --> 12:32.086 think one of the reasons um well , our 12:32.086 --> 12:34.840 arm armed forces was army will centers . 12:34.849 --> 12:37.090 Now our borders law centers are a rich 12:37.099 --> 12:39.377 evidence base to do what they're doing . 12:39.377 --> 12:41.488 But we helped uncover a direct tie of 12:41.488 --> 12:43.432 their evidence based practice to a 12:43.432 --> 12:45.655 reading this outcome . And this type of 12:45.655 --> 12:48.150 man can be used for a couple of things 12:49.210 --> 12:52.080 can help evaluate the good programs out 12:52.119 --> 12:54.929 there . If you have a good program , be 12:54.989 --> 12:58.330 like track your participation , give it 12:58.340 --> 13:00.229 to someone like CD L and they can 13:00.229 --> 13:03.969 directly tie your actions to reminisce 13:03.979 --> 13:06.146 if you have one because there's models 13:06.146 --> 13:09.150 for that . The second one is this could 13:09.159 --> 13:11.270 be given to commanders . But then you 13:11.270 --> 13:13.492 have a spreadsheet that says here's the 13:13.492 --> 13:15.492 forecasted readiness of your troops 13:15.492 --> 13:17.603 based on historic data for the rest . 13:17.603 --> 13:19.659 But we're gonna get the privacy in a 13:19.659 --> 13:22.559 second . I understand . Now , we good . 13:24.140 --> 13:27.409 Uh , we ask a question like , what is 13:27.500 --> 13:30.299 it that is a matter which domain 13:30.309 --> 13:32.420 they're , they're interacting with or 13:32.420 --> 13:34.642 is it just they're interacting with the 13:34.642 --> 13:36.865 wallet ? And does it matter why they're 13:36.865 --> 13:38.587 going , like , uh it's kind of 13:38.587 --> 13:41.380 important that the model also control 13:41.390 --> 13:43.580 for their physical fitness , their 13:43.590 --> 13:45.590 height , their weight , their other 13:45.590 --> 13:47.701 factors . Because if you only look at 13:47.701 --> 13:49.923 the people who went to arm and Wellness 13:49.923 --> 13:52.090 Center , their failure rate was higher 13:52.090 --> 13:54.257 than the , than the , than the average 13:54.257 --> 13:56.257 soldier . So the acts , the way the 13:56.257 --> 13:56.059 numbers I like to use is if the , if 13:56.070 --> 13:58.460 the failure rate on average was 4% the 13:58.469 --> 14:00.469 people who were being prescribed an 14:00.469 --> 14:02.302 army Wellness Center visit had a 14:02.302 --> 14:04.358 failure rate around 8% . But if they 14:04.358 --> 14:06.358 went , it dropped to 6% and because 14:06.358 --> 14:08.636 we're controlling for the factors that , 14:08.636 --> 14:10.636 that then would have sent them to a 14:10.636 --> 14:10.250 Wellness Center in the first place , 14:10.260 --> 14:12.260 their height , their weight , other 14:12.260 --> 14:14.482 risk factors , that's why we're picking 14:14.482 --> 14:16.427 up the positive effect of actually 14:16.427 --> 14:18.649 going to the Wellness Center as opposed 14:18.649 --> 14:18.549 to being identified with someone at 14:18.559 --> 14:20.559 risk who would be told to go in the 14:20.559 --> 14:22.726 first place . And that is why bringing 14:22.726 --> 14:24.615 together all of the data from the 14:24.615 --> 14:26.670 physical fitness domain , from the , 14:26.670 --> 14:28.726 from from everything is important so 14:28.726 --> 14:30.559 that you can control those other 14:30.559 --> 14:32.781 factors to the extent possible that the 14:32.781 --> 14:34.948 data and the mathematical models allow 14:34.948 --> 14:34.830 and , and to answer your question 14:34.840 --> 14:37.609 further related to uh does it matter 14:37.619 --> 14:41.599 which domain ? Um this when , when this 14:41.609 --> 14:43.720 was done ? Uh a couple of years ago , 14:43.880 --> 14:46.539 um we , we were not looking at 14:46.739 --> 14:49.330 specifically domains , we were again 14:49.340 --> 14:51.510 just looking at individual readiness 14:51.619 --> 14:53.730 and what can we predict on individual 14:53.730 --> 14:55.786 readiness ? Uh Fast forward , we are 14:55.786 --> 14:57.841 now specifically starting to look at 14:57.841 --> 15:00.210 each domain . Um The focus right now is 15:00.219 --> 15:02.369 really the framework . What what data 15:02.380 --> 15:04.547 do we need to be able to collect under 15:04.547 --> 15:06.602 each of the domains and what are the 15:06.602 --> 15:08.658 respective measures and metrics that 15:08.658 --> 15:10.658 drive individual readiness um under 15:10.658 --> 15:12.380 each of those domains ? Um And 15:12.380 --> 15:14.602 specifically , you know , uniquely here 15:14.602 --> 15:16.989 is Total Force fitness framework has 15:17.000 --> 15:19.222 eight domains . I know that HTF has the 15:19.222 --> 15:21.278 five . So we're looking at the eight 15:21.278 --> 15:23.222 specific domains under total force 15:23.222 --> 15:25.389 answer how long until that , until you 15:25.389 --> 15:27.278 have like information this year ? 15:30.969 --> 15:33.309 Yeah , great segue . Let me , let me 15:33.320 --> 15:36.289 jump into privacy and uh no , thanks . 15:37.460 --> 15:41.429 Um All right . So 15:42.719 --> 15:43.719 that one again , 15:46.669 --> 15:48.891 in the research environment of a AG and 15:48.891 --> 15:51.058 the research protocols , we got access 15:51.058 --> 15:52.725 to health services . Half the 15:52.725 --> 15:54.979 predictive power of whether or not 15:54.989 --> 15:57.045 service members can complete what it 15:57.045 --> 15:59.211 comes from the health records and half 15:59.211 --> 16:01.539 come from other records . You cannot do 16:01.549 --> 16:03.549 it without the medical record , you 16:03.549 --> 16:05.438 can't do it with just the medical 16:05.438 --> 16:07.659 record and the outcome variable is 16:07.669 --> 16:09.725 normally trained on si pr outcomes . 16:09.725 --> 16:11.780 When were you supposed to , when did 16:11.780 --> 16:14.113 you come back ? Did you come back early ? 16:14.113 --> 16:16.225 And , and the more precise definition 16:16.225 --> 16:18.169 of that SI R ? So you need to have 16:18.169 --> 16:20.690 access to those data elements from 16:20.700 --> 16:22.756 electronic health records that we've 16:22.756 --> 16:25.049 demonstrated that it readiness in order 16:25.059 --> 16:27.119 to run these models . And there's , 16:27.130 --> 16:29.352 there's exemptions for that . There's , 16:29.352 --> 16:31.297 there's 12 exemptions in HP A that 16:31.297 --> 16:33.297 allow for things like public health 16:33.297 --> 16:35.297 that allow for a central government 16:35.297 --> 16:37.408 function including military command . 16:37.408 --> 16:39.741 And so it's feasible that you could say , 16:39.741 --> 16:41.963 hey , we need these exact data elements 16:41.963 --> 16:41.320 so we can run a readiness prediction 16:41.330 --> 16:43.552 model and then run it with the physical 16:43.552 --> 16:45.441 data , with the other data that's 16:45.441 --> 16:47.774 required . But there's been some snacks . 16:48.450 --> 16:50.672 So that , that's what we're gonna , I'm 16:50.672 --> 16:52.894 just gonna hit on privacy real quick to 16:52.894 --> 16:56.460 given the time . So privacy is I I 16:56.469 --> 16:58.691 introduced this at the beginning . It's 16:58.691 --> 17:00.747 a component that we have to consider 17:00.747 --> 17:02.636 and think of . Um in the research 17:02.636 --> 17:04.858 environment , it's relatively easy . We 17:04.858 --> 17:07.080 can go through the research exemption . 17:07.080 --> 17:09.302 We will be able to get access to this . 17:09.302 --> 17:11.469 Um Matter of fact , DH a privacy civil 17:11.469 --> 17:13.358 liberties office um calls herself 17:13.358 --> 17:15.709 because they have over 500 data sharing 17:15.719 --> 17:18.180 agreements , but 90% of their data 17:18.189 --> 17:20.133 sharing agreements is for research 17:20.133 --> 17:22.300 purposes that tells us we have the gap 17:22.300 --> 17:23.800 of the ability to actually 17:23.810 --> 17:26.250 operationalize this . But we have these 17:26.260 --> 17:29.069 gaps because we don't fully understand 17:29.079 --> 17:31.839 how to use the authority given to us , 17:31.849 --> 17:34.319 whether that authority is again through 17:34.329 --> 17:36.619 us as CD A O or the authority that's 17:36.630 --> 17:38.564 actually given entitled to and uh 17:38.574 --> 17:40.944 uniform service code to the actual 17:40.984 --> 17:43.285 combat commands and the services . So 17:43.295 --> 17:45.351 here's some of the , the interesting 17:45.351 --> 17:47.573 authoritative uh documents . Um Again , 17:47.573 --> 17:49.684 I'm gonna hit on a couple of things . 17:49.684 --> 17:51.851 Um The first one I wanna hit on is the 17:51.851 --> 17:54.204 one on the left hand side . And it was 17:54.214 --> 17:56.325 uh the memo that came out from the de 17:56.325 --> 17:58.474 se de called Creating Data Advantage 17:58.484 --> 18:00.834 memo came out on Cinco de Mayo of May 18:00.844 --> 18:04.364 2021 in May 2021 . We were in the 18:04.375 --> 18:08.130 middle of the pandemic . The DEVS de 18:08.219 --> 18:10.339 released a memo to say creating data 18:10.349 --> 18:13.500 Advantage memo that all dod data should 18:13.510 --> 18:16.540 be to the maximum extent possible share . 18:18.390 --> 18:20.729 I say this because what's interesting 18:20.739 --> 18:23.449 is I've been told by legal and other 18:23.459 --> 18:25.819 people that she didn't mean health day . 18:26.329 --> 18:28.273 We were in the middle of a opinion 18:28.349 --> 18:30.293 surely when she said all day , she 18:30.293 --> 18:33.300 meant all day . Fast forward . Here we 18:33.310 --> 18:35.532 are to answer the gentleman's questions 18:35.532 --> 18:38.130 earlier . How where is our blockers ? 18:38.140 --> 18:40.689 Our blockers is trying to get the the 18:40.699 --> 18:42.810 prediction power of the health record 18:42.810 --> 18:44.810 and combining that in the readiness 18:44.810 --> 18:47.088 data , combining that with people data , 18:47.088 --> 18:49.310 combining that with wearable data . And 18:49.310 --> 18:51.366 then actually getting , once we push 18:51.366 --> 18:53.532 that readiness . We need to be able to 18:53.532 --> 18:53.369 have it at the right classification 18:53.380 --> 18:55.699 level because now we're showing a 18:55.709 --> 18:58.109 classification based on a vulnerable 18:58.130 --> 19:00.680 potential vulnerability to a particular 19:00.689 --> 19:04.180 location um that does not exist on the 19:04.189 --> 19:06.640 health side . They have a nipper 19:06.650 --> 19:08.594 environment . We got to be able to 19:08.594 --> 19:10.650 bring that from nipper , use a cross 19:10.650 --> 19:12.761 domain solution and get into a sipper 19:12.761 --> 19:16.479 capability . So here 19:16.489 --> 19:20.260 is the , using Charles analogy , here's 19:20.270 --> 19:24.239 the magic pill . There's 12 exemptions 19:24.369 --> 19:26.770 in under I A law . The first one is 19:26.780 --> 19:30.280 required by law and the one that 19:30.290 --> 19:32.234 applies to you'll , you'll hear it 19:32.234 --> 19:34.068 under HIPA L it's called central 19:34.068 --> 19:37.260 government functions , but in dod uh M 19:37.270 --> 19:40.160 60 25 18 , which is actually our HP A 19:40.170 --> 19:42.469 manual , it says military command 19:42.479 --> 19:45.819 authority , military command 19:45.829 --> 19:48.939 authority allows for the services and 19:48.949 --> 19:50.739 the combatant commands of any 19:50.750 --> 19:53.459 individual wearing a uniform underneath 19:53.469 --> 19:55.640 their command to be able to get access 19:55.650 --> 19:57.594 to the respective data they need . 19:57.729 --> 19:59.840 Historically , in the nineties , this 19:59.840 --> 20:01.729 was done predominantly if I was a 20:01.729 --> 20:03.951 commander of a unit and I needed to ask 20:03.951 --> 20:06.007 something about a readiness , then I 20:06.007 --> 20:08.062 had that right to be able to talk to 20:08.062 --> 20:07.750 that provider and get that information 20:08.079 --> 20:10.357 um with the exception of mental health . 20:10.357 --> 20:13.079 And so some other areas . Um but now 20:13.089 --> 20:15.979 we're in a world of big data and so we 20:15.989 --> 20:18.100 should be able to look at this from a 20:18.100 --> 20:20.322 unit perspective . So this is the magic 20:20.322 --> 20:22.545 piece we are continuing to advocate for 20:22.545 --> 20:24.767 this um be able to talk to the services 20:24.767 --> 20:26.656 in the command to start trying to 20:26.656 --> 20:28.878 exercise this . It allows us to be able 20:28.878 --> 20:30.822 to bring that data start doing the 20:30.822 --> 20:32.878 prediction power , combine that with 20:32.878 --> 20:35.100 the physical and the other domains . Um 20:35.100 --> 20:37.156 So with that being said , um this is 20:37.156 --> 20:37.010 our closure . So open it up to any 20:37.020 --> 20:40.910 questions . Yes sir . 20:41.410 --> 20:44.060 Hey , so uh as , as you guys go through 20:44.069 --> 20:46.800 this , you're in sync . How long do you 20:46.810 --> 20:48.977 reasonably think it'll take before the 20:48.977 --> 20:51.479 06 level commands ? And the H two F 20:51.489 --> 20:54.079 performance teams are able to nest 20:54.089 --> 20:56.660 their mops and their modes within this 20:56.670 --> 20:59.219 larger total force fitness metric 20:59.229 --> 21:02.579 system . Realistically . Um My opinion , 21:02.589 --> 21:04.756 I think we're looking at , I , I think 21:04.756 --> 21:07.033 we're still looking at core foundation . 21:07.920 --> 21:11.329 OK ? And a necessary precondition . 21:12.099 --> 21:15.310 So this was tasked the CEO to make 21:15.319 --> 21:17.810 available throughout the enterprise . 21:18.579 --> 21:20.801 But in order to get the data , you need 21:20.801 --> 21:22.968 to serve military mandatory and the CD 21:22.968 --> 21:25.489 O is not a military mandatory . So CEO 21:25.500 --> 21:27.722 S in this position to make it available 21:27.722 --> 21:30.079 and make those with military command 21:30.089 --> 21:33.660 authority aware of a thing that can 21:33.670 --> 21:36.520 predict readiness if they ask for it 21:37.060 --> 21:39.349 and assert their military command 21:39.359 --> 21:41.470 authority . And then there's two ways 21:41.470 --> 21:43.470 that you can use this capability in 21:43.470 --> 21:45.581 short order . The best one is program 21:45.581 --> 21:47.839 of data , put it in a research market 21:47.849 --> 21:49.849 and you guys are all army . So army 21:49.849 --> 21:52.071 analytics group is fine and use them to 21:52.071 --> 21:54.293 rerun the right explanation model . But 21:54.293 --> 21:56.516 add the data from your programs into it 21:56.516 --> 21:58.738 to get a return on the basement on each 21:58.738 --> 21:58.180 of your programs . That's probably the 21:58.189 --> 22:00.189 most tactical parochial one for you 22:00.189 --> 22:02.300 guys here . But at the same time , if 22:02.300 --> 22:04.522 you're doing that , because you want to 22:04.522 --> 22:06.522 man , if you also have the remit of 22:06.522 --> 22:08.745 informing your command on the readiness 22:08.745 --> 22:10.856 of your service members , they can go 22:10.856 --> 22:13.022 up their chain and ask for this . They 22:13.022 --> 22:15.189 can ask this 12 or they can ask for it 22:15.189 --> 22:17.411 on . But it has to be the services that 22:17.411 --> 22:19.133 ask for this capability that's 22:19.133 --> 22:22.479 available with CD A . What's the lowest 22:22.489 --> 22:24.711 authority level that can request that ? 22:25.060 --> 22:27.599 There's a disagreement between DH A and 22:27.609 --> 22:29.665 HHS on the answer to that question . 22:31.130 --> 22:34.989 What is the , there is 22:35.890 --> 22:38.520 PC LT is assisting in bringing 22:38.530 --> 22:40.586 resolution of that question . Yeah , 22:40.586 --> 22:42.586 our , our , we recently . So yeah , 22:42.586 --> 22:44.530 we're working with privacy , civil 22:44.530 --> 22:47.599 liberties , transportation to help uh 22:47.609 --> 22:49.831 and health affairs . Uh General Counsel 22:49.831 --> 22:51.900 is also addressing but really it's 22:52.280 --> 22:54.447 right now , the only thing we're going 22:54.447 --> 22:56.558 off of is the , the requests that are 22:56.558 --> 22:59.359 in place today . So of course , volume 22:59.369 --> 23:02.959 speaks works . The D DJ PC L 23:02.969 --> 23:05.530 OS is comfortable with the individual 23:05.540 --> 23:07.989 commander , low level commander 23:08.000 --> 23:10.750 requesting data on their direct troops . 23:10.760 --> 23:13.150 They're very comfortable with that . Um 23:13.930 --> 23:16.152 They are less comfortable on ball , but 23:16.152 --> 23:19.989 there has been uh questions about that 23:20.000 --> 23:22.167 that have been circulated on what that 23:22.167 --> 23:24.333 means . And we've seen more success in 23:24.333 --> 23:26.380 lower level units and commands at a 23:26.390 --> 23:29.219 small um uh 23:29.229 --> 23:31.459 population . Then we have a large 23:31.469 --> 23:33.469 population . That's , it gets scary 23:33.469 --> 23:35.636 when you get into larger populations . 23:35.640 --> 23:37.418 We have time for maybe one more 23:37.418 --> 23:39.696 question . All right . Yes , sir , sir . 23:39.696 --> 23:41.751 So when you're identifying low-level 23:41.751 --> 23:43.862 populations , are you talking for gay 23:43.862 --> 23:45.973 and below ? Or battalion below ? Just 23:45.973 --> 23:48.010 battalion battalion below ? Ok . 23:53.150 --> 23:53.790 Thanks . Thank you .