Meeting Title: Brainforge <> Chang <> HPHF Date: 2024-05-15 Meeting participants: Unknown
WEBVTT
00:18:03.000 ⇒ 00:18:13.000 Response you Who are to identify a person who is sitting next to you? So that is what our problem in 2,018 to 20.
00:18:13.000 ⇒ 00:18:23.000 Then myself and AR and then, teamed up. To solve this problem. To identify the individual in less than a 3 min time.
00:18:23.000 ⇒ 00:18:35.000 That is what we started. That does scale of 80 million. So, 80 million people across the state around 30,000 users for 30,000 users.
00:18:35.000 ⇒ 00:18:43.000 It has to come in less than 3 min, but we achieved a license 100, milliseconds.
00:18:43.000 ⇒ 00:18:44.000 Cool.
00:18:44.000 ⇒ 00:18:50.000 5 search and mainly it is offline. And we can able to handle the calls with the team of 10.
00:18:50.000 ⇒ 00:18:57.000 Think of 80 million population. Handled with 10 people at the state. And that is how the architecture was for us.
00:18:57.000 ⇒ 00:19:07.000 It’s all with the concerns from governments like this. This is how it.
00:19:07.000 ⇒ 00:19:19.000 It was started like that. So with the summing above all this the experience we thought, why can’t we, Why can’t we able to do the same?
00:19:19.000 ⇒ 00:19:27.000 Whatever we learn from one of the state because we want the service delivery to go combined to it. Then the option will be higher.
00:19:27.000 ⇒ 00:19:36.000 This is our evolution like, we are adapting, we are iterating, we are learning from the field and we are comprehensively packaging.
00:19:36.000 ⇒ 00:19:37.000 I’m not only that digital solution, but also the option. So we want to give a service delivery.
00:19:37.000 ⇒ 00:19:47.000 We want to give a trainee. Yeah, capacity building, monitoring and evaluation as a comprehensive package.
00:19:47.000 ⇒ 00:19:51.000 Doctor, are, are you sharing a, are you sharing a deck?
00:19:51.000 ⇒ 00:19:52.000 Yeah, I’m pulling up the slide with them. Yeah.
00:19:52.000 ⇒ 00:20:01.000 Okay. Okay, yes. So I thought, I thought you were reading from it. I was like, you can share on the video.
00:20:01.000 ⇒ 00:20:02.000 Yeah.
00:20:02.000 ⇒ 00:20:03.000 No, it’s not really, it’s what we are living on. So, the timeline.
00:20:03.000 ⇒ 00:20:11.000 Okay. Same like you’re reading in but it’s okay. No, for
00:20:11.000 ⇒ 00:20:12.000 Yeah. So.
00:20:12.000 ⇒ 00:20:13.000 Yeah. Hmm.
00:20:13.000 ⇒ 00:20:15.000 Basically, he’s multitasking. He was talking and he was searching for the Okay.
00:20:15.000 ⇒ 00:20:16.000 Yeah.
00:20:16.000 ⇒ 00:20:21.000 Okay. No, I hope it might be easier for you.
00:20:21.000 ⇒ 00:20:28.000 See then we able to summarize it from 3 domains basically. What we are operating is Like you are seeing 3 people right?
00:20:28.000 ⇒ 00:20:38.000 Our London is primarily on the tank part, but he is also exposed to public health and what we do in the field.
00:20:38.000 ⇒ 00:20:55.000 And how we engage for business development with the government entities enterprise as well as with private sector hospitals. On the other side, is the one who is formulating how this formalization of engagement and what is the return of investment with the partners.
00:20:55.000 ⇒ 00:21:04.000 So he is working on that field and myself working on governance and healthcare mainly at the population scale are option.
00:21:04.000 ⇒ 00:21:15.000 So given this 3 we have a farm matrix of combination of research. Same time implementation for research as well as population scale implementation for impact.
00:21:15.000 ⇒ 00:21:31.000 So this is how, we are placed. But 2 in a sharp. The whole project is the empowering citizen by giving the benefits to the citizen.
00:21:31.000 ⇒ 00:21:38.000 So it may seem easy if you can refer the health system of Europe and US. Like the fewer data will reach the government after 6 months.
00:21:38.000 ⇒ 00:21:50.000 After passing through insurance Sometimes it may go to HMOs, sometime it will go to the, epidemiologist unit in the state.
00:21:50.000 ⇒ 00:21:57.000 I worked closely with previously at one closely with CDC Atlanta in 18 2,016 to 19 time frame.
00:21:57.000 ⇒ 00:22:09.000 So be able to compare with what is the advantage of India with the, US and UK. So the hard one is here is we can able to make our own SOPs.
00:22:09.000 ⇒ 00:22:20.000 We can able to make our own protocols. Dear privacy and cybersecurity is much compromised till 2021.
00:22:20.000 ⇒ 00:22:30.000 So now it is given that is a data. Privacy policy is in place. And there are much more policies which are helping us in forming the data.
00:22:30.000 ⇒ 00:22:36.000 At the end of the day, what we are trying to establish is the data judiciary. So you can understand the data for you, right?
00:22:36.000 ⇒ 00:22:44.000 That is a independent entity which can able to. Collect the record and matt govern the access of the records So that is what we are establishing in the name of population governance and research.
00:22:44.000 ⇒ 00:22:58.000 So what the proposal shared with you is talking about how we are establishing the model site, how we are engaging the partners.
00:22:58.000 ⇒ 00:23:05.000 At the end of the day, what we are bringing is best practices from the model site. So that will help us in.
00:23:05.000 ⇒ 00:23:16.000 Fine-tuning most of the diagnostics algorithm. Or a cluster generating algorithm. Or some of the best practices that can help in.
00:23:16.000 ⇒ 00:23:22.000 Populations can service deliver. I think. Yeah, it’s tough. The slides are also loaded.
00:23:22.000 ⇒ 00:23:32.000 So I just show you the slide. So whatever, it’s, I described it will be in one or 2 slides, right?
00:23:32.000 ⇒ 00:23:42.000 That’s give me a second. Hey, they want to add anything?
00:23:42.000 ⇒ 00:23:43.000 Oh.
00:23:43.000 ⇒ 00:23:44.000 Hello, no, no, just please carry on.
00:23:44.000 ⇒ 00:23:50.000 So, see what we did is in, so while doctor gave a overall, from an experience point of view.
00:23:50.000 ⇒ 00:23:59.000 He, had the system running in one of the key states in India. For close to 6 months live, right?
00:23:59.000 ⇒ 00:24:16.000 So does a pilot 6 months live. We were able to cover 25 million and service delivery several times to the right.
00:24:16.000 ⇒ 00:24:22.000 This was unique in Indigenous 20 with 25 million people are covered. With, service, a little bit to them more than once.
00:24:22.000 ⇒ 00:24:43.000 Okay, so the, does success. In the adoption is 2 things right one. People who are educated to a level of 10 standard something like a not preschool but middle school are more than middle school, right?
00:24:43.000 ⇒ 00:25:07.000 So they were able to use the app and able to do service delivery at the most. Without much fiction because the solution that we developed and the training both the solution as well as the training went and there right And also the solution is same as a value at by those individuals who are using it.
00:25:07.000 ⇒ 00:25:12.000 Not as a deterrent, right? So they were using it and they’ll say, okay, my name is beginning.
00:25:12.000 ⇒ 00:25:25.000 Second, the citizens also were able to see the difference in. And they are started asking and respect the people who is coming to the rows.
00:25:25.000 ⇒ 00:25:35.000 So one of the things that we used to tell is you can pay a person, say, for ₹5,000, which is nothing but save.
00:25:35.000 ⇒ 00:25:45.000 15 pounds a month, right? And you ask them to walk around and provide services to citizens.
00:25:45.000 ⇒ 00:25:52.000 They don’t only do it for money, they do it for the respect they gets in the community.
00:25:52.000 ⇒ 00:26:05.000 So those are something which we learned over the process of, the pilot. So currently we are, we are focusing on Southeast Asia.
00:26:05.000 ⇒ 00:26:17.000 Man, we, we are, Actively, to close a model site. In 2 island nations and 2 states small states in India.
00:26:17.000 ⇒ 00:26:31.000 So that is where we are currently at. So the discussions are in. Advanced stages. So we want we want them to come get that converted by June, July.
00:26:31.000 ⇒ 00:26:39.000 And then the model side will start us, give us the start. Because Kamil Nadu is, is, they should not think the Tamil Nadu’s will be close, right?
00:26:39.000 ⇒ 00:26:48.000 We want to. Thank you to more than one site and make it a lines. And from there also learnings will come for us both from.
00:26:48.000 ⇒ 00:27:02.000 System adoption point of view. As well as the take point of you on the expedition, So that is what we are currently working on, channel.
00:27:02.000 ⇒ 00:27:09.000 Thank you. Yeah, thanks to both of you guys. That’s great. And can you just clarify a few terms, movies?
00:27:09.000 ⇒ 00:27:17.000 I think for you guys, it might be. Assumed. But it’s it’s just really important for I guess.
00:27:17.000 ⇒ 00:27:28.000 Someone who like me to clearly is a bit more educating on this front. What do you what exactly do you mean by service delivery in the context in this context that you’ve just mentioned?
00:27:28.000 ⇒ 00:27:36.000 So, so is delivery can be as simple as. Going, Mr. Sika, there are some, 4 people in the family, right?
00:27:36.000 ⇒ 00:27:41.000 6 people, in, in, in Indian family 6 people. So you have father mother.
00:27:41.000 ⇒ 00:27:52.000 And then the children and probably their grandmother. So you for them to get a health care delivery. So they have to go to India with PC probably around Connecticut as a 5.
00:27:52.000 ⇒ 00:27:53.000 Yes.
00:27:53.000 ⇒ 00:28:00.000 This is a primary health center, right? So they have to do a routine routine checkup of their blood sugar, VP.
00:28:00.000 ⇒ 00:28:01.000 Yep.
00:28:01.000 ⇒ 00:28:10.000 And also get their medication which is given by the given by the state for free because state is ultimately the biggest provider of medical services in India.
00:28:10.000 ⇒ 00:28:11.000 Cool.
00:28:11.000 ⇒ 00:28:14.000 Think about those services being delivered at your doorstep. So those are the services that you do that was built.
00:28:14.000 ⇒ 00:28:26.000 So they did. Routine follow-up on their chronic, chronic, this, non-communal, NPC, right?
00:28:26.000 ⇒ 00:28:36.000 Non-cmical diseases, diabetic hypertension. CQD even in some cases cancers where they do a reprative follow.
00:28:36.000 ⇒ 00:28:46.000 On the treatment portal, whether the particular person is taking the, whether the diabetes never has come down after taking medicine.
00:28:46.000 ⇒ 00:28:51.000 If it is not come down, they will refer you to a facility, a doctor, again.
00:28:51.000 ⇒ 00:28:55.000 So instead of I have to go every month or I leave it, I don’t, I don’t want to go.
00:28:55.000 ⇒ 00:28:56.000 Yep.
00:28:56.000 ⇒ 00:29:03.000 This, this team went and picked that service delivery under the. That’s what the, the, When we go as video
00:29:03.000 ⇒ 00:29:05.000 60 TC, yes, yes.
00:29:05.000 ⇒ 00:29:10.000 Let me go. So that is the pilot that we did, right? That’s only medical services.
00:29:10.000 ⇒ 00:29:19.000 But when we speak about PGA, the, the, can be much more broader. It can be asked simple as getting you an ID card.
00:29:19.000 ⇒ 00:29:20.000 Yep.
00:29:20.000 ⇒ 00:29:31.000 National ID card at your doorstep. All the way up to major services are subsidy reaching a particular subsidy or a are, incentive reaching you, your bank account directly.
00:29:31.000 ⇒ 00:29:46.000 So all of these are services can be different departments. Have the government. So, easier as a platform, enable the government.
00:29:46.000 ⇒ 00:29:51.000 Yep.
00:29:51.000 ⇒ 00:29:52.000 Yes.
00:29:52.000 ⇒ 00:30:00.000 To reach the citizens without much friction. And also it covered the citizens with the departments. Easily and also track it as a proper proper system.
00:30:00.000 ⇒ 00:30:19.000 That’s great. Thank you, The new Dr. V are you able to give us an example or me an example of a digital solution you’ve implemented in the context of and now because one of the from having spoken to others, other collaborators based in India.
00:30:19.000 ⇒ 00:30:32.000 It seems to me that there is no legal differentiation or regulatory differentiation between a digital health tool. And a medical device.
00:30:32.000 ⇒ 00:30:44.000 That is, even if the health tour, whatever you built. Helps to diagnose patients. That is not under the same legal nor regulatory framework.
00:30:44.000 ⇒ 00:30:54.000 As so I’ve been on par with a physician. Having the responsibility of making that diagnostic call.
00:30:54.000 ⇒ 00:31:02.000 Because it can influence the diagnosis of a physician. As you can imagine, over in Europe at least, that’s been deemed to be far higher risk.
00:31:02.000 ⇒ 00:31:09.000 Before it’s approved. Either by the FTA or in the US or via equivalent body here in the UK or in Europe.
00:31:09.000 ⇒ 00:31:16.000 Just some different. And as far as I know, there is. I’d love some clarification of that.
00:31:16.000 ⇒ 00:31:23.000 If you share some insights into an example of a digital solution you’ve implemented.
00:31:23.000 ⇒ 00:31:36.000 Yeah, it’s. See what we have implemented is a solution for stitching the health record and building a longitudinal.
00:31:36.000 ⇒ 00:31:37.000 Okay.
00:31:37.000 ⇒ 00:32:03.000 Record for individuals and families. So, like the query is about can we able to give a solution that can diagnose is what Like the is trying to do right we are trying to pass through a medical exam we are trying to pass through an algorithm for diabetes we are trying to I have agreement with the physician diagnosis of diabetic food with the A.
00:32:03.000 ⇒ 00:32:13.000 So, I in that context, the Indian system is more with the physical examination or a diagnosis by a doctor.
00:32:13.000 ⇒ 00:32:14.000 Yes.
00:32:14.000 ⇒ 00:32:23.000 So these 2 come much ahead, like we can able to use the distance solution for Screening. That means.
00:32:23.000 ⇒ 00:32:38.000 We have 7 levels of diagnosis in, in our system. So what it will start is if you are a individual you are giving a symptomatic report like a participant report that can be augmented with the digital solution.
00:32:38.000 ⇒ 00:32:49.000 So that part is exist in practice. Where we will combat the parameters not the algorithms. So the algorithms will end up in.
00:32:49.000 ⇒ 00:32:57.000 Suspected case, right? We are not giving even the provisional diagnosis with or any of the digital solution.
00:32:57.000 ⇒ 00:33:03.000 It has to go with the lab diagnosis. It has to go with the physical examination by your medical pee.
00:33:03.000 ⇒ 00:33:11.000 Let me, registered medical practitioner in India. So this is what the law is about.
00:33:11.000 ⇒ 00:33:19.000 What we can able to do is we can able to have medical device the solution can become medical device.
00:33:19.000 ⇒ 00:33:35.000 The moment it is linked with diagnosis part of it like if you want to have a for anemia the less, he more broken than we want to have a That will go as a medical device alternative for a lab test.
00:33:35.000 ⇒ 00:33:36.000 Yeah.
00:33:36.000 ⇒ 00:33:45.000 That did not come as a tool for diagnosing an. So this is what the difference majorly how the.
00:33:45.000 ⇒ 00:33:46.000 Yep.
00:33:46.000 ⇒ 00:33:55.000 Legal entities function. In India. So we will try to equate with either lab or diagnostic methodology, but we are not comparing with the physicians diagnosis.
00:33:55.000 ⇒ 00:33:56.000 Understood. Okay.
00:33:56.000 ⇒ 00:34:03.000 In chat, this is what happens. Including the technical reality or diagnosing the x-ray with the conditions.
00:34:03.000 ⇒ 00:34:17.000 It’s more in the higher level like we have meaningful symbolic outputs. We are not having a definitive diagnosis.
00:34:17.000 ⇒ 00:34:18.000 Yep.
00:34:18.000 ⇒ 00:34:32.000 The probability of agreement is not that good. So this is the practice in India like If you want more into regulations like rules Yes, we can able to, dive it, but, that require, longer session.
00:34:32.000 ⇒ 00:34:41.000 Like it’s majorly. That digital devices are equated with. Diagnostics and at the very basic endromic levels.
00:34:41.000 ⇒ 00:34:42.000 Not at the diagnost itself.
00:34:42.000 ⇒ 00:34:44.000 Yeah.
00:34:44.000 ⇒ 00:34:46.000 I see.
00:34:46.000 ⇒ 00:35:06.000 So in our system we use, we have 15 diseases to suspect at the doorstep. Almost the similar number of diagnosis is made at the primary health center level But what we help is to document things and to create a longitudinal meaningful records.
00:35:06.000 ⇒ 00:35:15.000 That can be aggregated at the state level.
00:35:15.000 ⇒ 00:35:28.000 Which follows the privacy policy which is not mandated by. Bye, Youified health interface promoted by the federal government.
00:35:28.000 ⇒ 00:35:29.000 I see.
00:35:29.000 ⇒ 00:35:34.000 Yeah. But we follow the concert procedure. We follow the.
00:35:34.000 ⇒ 00:35:38.000 And differential privacy policy for the individuals. So need to.
00:35:38.000 ⇒ 00:35:41.000 And somebody’s longitudinal. Sorry, very good, please go on.
00:35:41.000 ⇒ 00:35:48.000 To know basis people who are interacting will know more about the user.
00:35:48.000 ⇒ 00:35:49.000 I see. So with these longeritudinal records that you said you’ve already implemented this solution to an extent, is that right?
00:35:49.000 ⇒ 00:36:03.000 You’re saying that you’ve been helping to develop these. Long as usual records for patients.
00:36:03.000 ⇒ 00:36:12.000 Documentation, aggregation at the state level. Where are these? So that makes that sort of implies to me that you’ve already worked out.
00:36:12.000 ⇒ 00:36:20.000 In terms of data governance and you know speaking of fiduciaries. Where the custodian ship.
00:36:20.000 ⇒ 00:36:28.000 Lies with respect to the data, where the data is stored. In what format, whether it’s interoperable from state to state.
00:36:28.000 ⇒ 00:36:45.000 These are all obviously huge concerns when it comes to creating. Something that scale like this. Otherwise you end up just sort of landlocking yourself don’t you as a state if you’re in the operating using a 1 particular system and it’s completely different from someone else, they, they bring to you.
00:36:45.000 ⇒ 00:36:46.000 Any comments on?
00:36:46.000 ⇒ 00:36:47.000 Dr. Canny, can I answer this? You, but you ought to hear again, yet.
00:36:47.000 ⇒ 00:36:55.000 Yeah, that. The area. I’m, I’m, I’m coming to, like that’s where arrangement operates.
00:36:55.000 ⇒ 00:36:56.000 Okay.
00:36:56.000 ⇒ 00:37:07.000 And as a health and domain Well, S, what we size, you are available fire, right?
00:37:07.000 ⇒ 00:37:13.000 Yep.
00:37:13.000 ⇒ 00:37:14.000 Yeah.
00:37:14.000 ⇒ 00:37:21.000 For health, the intervable DHL. Bye. So we, our own architecture that will have companies with efforts here and that can also have Health and normal data.
00:37:21.000 ⇒ 00:37:23.000 Go ahead.
00:37:23.000 ⇒ 00:37:31.000 Yeah, so their bootings, attend. I think you started with GDPR, right?
00:37:31.000 ⇒ 00:37:32.000 Yeah.
00:37:32.000 ⇒ 00:37:39.000 You know, very strong. India’s data production law is still in still in I mean, it’s just a bill which is not past it.
00:37:39.000 ⇒ 00:37:56.000 So it’s being revised as we speak. So the concept of, the data will move from one state under the state is not a problem because The data owner is defined by at least the UHH.
00:37:56.000 ⇒ 00:38:04.000 UHH, UHH, Nice one, but health, health records, right? They have different.
00:38:04.000 ⇒ 00:38:05.000 Yep.
00:38:05.000 ⇒ 00:38:11.000 It has individual, right? So let’s say you are an individual and that you go to your doctor let’s take a hospital.
00:38:11.000 ⇒ 00:38:21.000 The hospital can or cannot don’t have there is no there is no mandatory for them to be part of the overall, it sure has to that the government is providing.
00:38:21.000 ⇒ 00:38:30.000 Okay. The government is saying like UPA. They say every doctor should. Or could be part of this network right.
00:38:30.000 ⇒ 00:38:40.000 But doesn’t mandate anybody or any hospital say hey every record that you have to you have to make it available for the interoperability they don’t do it currently Okay, that is the 1st problem.
00:38:40.000 ⇒ 00:38:42.000 Okay, yeah. Yes.
00:38:42.000 ⇒ 00:38:50.000 Second problem whenever your data is requested for example you went to hospital A And I am, I am hospital and Dr.
00:38:50.000 ⇒ 00:39:03.000 B is hospital B. You go from hospitality hospital B, right? If both of these hospitals are there in the network only then there can be intervalability of, the interability of the stickers.
00:39:03.000 ⇒ 00:39:09.000 It doesn’t matter whether you are in state A, state state state is not a problem here. As long as you’re in India.
00:39:09.000 ⇒ 00:39:10.000 Okay.
00:39:10.000 ⇒ 00:39:20.000 And both these hospitals are in the network. Then you will get a consent notification on your app. They have a beneficiary app.
00:39:20.000 ⇒ 00:39:26.000 They also who consent on SMS, right? So basically they will ask you to tell a UDP of source, right?
00:39:26.000 ⇒ 00:39:27.000 I see. Yeah.
00:39:27.000 ⇒ 00:39:35.000 One time password. And then, Dr. D can access your record, based on the condition that we have here.
00:39:35.000 ⇒ 00:39:51.000 But the problem there is also what record you share and what is the record I share is also not not specified by Aba for example you came to me you have 10 records with me I said, has a hospital I only provided 2 records.
00:39:51.000 ⇒ 00:39:58.000 The beneficiary has to go to the hospital and fight with them saying, why is only 8 8 of records is not uploaded?
00:39:58.000 ⇒ 00:40:04.000 Currently the user behavior is also not aligned to this. They have no idea. They don’t see that data as that.
00:40:04.000 ⇒ 00:40:22.000 That crucial as well currently. Right, they are not as educated about this as, you know, or, So that is the second problem that the ecosystems, faces, right?
00:40:22.000 ⇒ 00:40:23.000 Yep.
00:40:23.000 ⇒ 00:40:29.000 But the 1st question. The acceptance of AA take for prognosis not for diagnosis.
00:40:29.000 ⇒ 00:40:38.000 Yeah, we can call it a diagnosis as well. He’s there, but it will never. Say, I don’t need the doctor at all.
00:40:38.000 ⇒ 00:40:53.000 The LB is prognosis. Are there will be some kind of diagnosis but the doctor has to still Come and sign it off and say, yeah, this is this is correct.
00:40:53.000 ⇒ 00:40:54.000 Yeah, yup.
00:40:54.000 ⇒ 00:40:58.000 What the it is correct and then you still have a treatment right So the human in the middle loop is not going away for next.
00:40:58.000 ⇒ 00:41:04.000 Next 10 years in India is what we are seeing in the, in the industry currently, right?
00:41:04.000 ⇒ 00:41:14.000 But, there is, there is also a big being worked on. On the device in the lines of I think, point FDA, right?
00:41:14.000 ⇒ 00:41:23.000 The CFPS, US has this, agency which Which says, okay, this, this is, the base which you can use for a particular purpose, right?
00:41:23.000 ⇒ 00:41:24.000 Yeah.
00:41:24.000 ⇒ 00:41:33.000 A similar bill is being passed. Being worked on by the government, I don’t know how long it’s going to come, take for that to come to market.
00:41:33.000 ⇒ 00:41:34.000 Yup.
00:41:34.000 ⇒ 00:41:37.000 But there is there is a move and movement around it as
00:41:37.000 ⇒ 00:41:43.000 Yeah, I see.
00:41:43.000 ⇒ 00:41:44.000 Yeah.
00:41:44.000 ⇒ 00:41:48.000 Yeah, for now it is it is kind of a wild rest. If you have a Y, you can go to a doctor if you don’t need your friend, you will use it.
00:41:48.000 ⇒ 00:41:54.000 Of course he will give you the learnings to you. Yeah, there’s no no issues to love.
00:41:54.000 ⇒ 00:42:03.000 As a startup one thing that we are we are taking that’s what Dr. Was trying to explain right as a startup as well as When we did it in Tamil Nadu, right?
00:42:03.000 ⇒ 00:42:13.000 We took few things as proactively for example differential differential privacy right and then differential consent.
00:42:13.000 ⇒ 00:42:22.000 And then we spoke about data fiduciary, right? Setting up a data judiciary for the state, right?
00:42:22.000 ⇒ 00:42:25.000 The state, and, even, Tamilada has a data policy of its own.
00:42:25.000 ⇒ 00:42:34.000 Okay, which, it’s saying. The data is owned by the government not to the, but you are trying to change those things in the state.
00:42:34.000 ⇒ 00:42:47.000 But yeah, those are the, those are the things that we want to do wherever we go.
00:42:47.000 ⇒ 00:42:48.000 Sure.
00:42:48.000 ⇒ 00:42:55.000 We want to make sure that the owner of the data that is created He’s always the individual. The upgraded data can be with the decision makers right but the individual PII and they say dinos is data.
00:42:55.000 ⇒ 00:43:05.000 He’s his one. He can he has all the control over it is what we want all we want to propagate.
00:43:05.000 ⇒ 00:43:13.000 Thanks, Ivan. And I think in the interest of time, it’d be great to just hear from you guys, whoever is keen to talk.
00:43:13.000 ⇒ 00:43:28.000 But it’d be cv. Dr. V or or yourself, our vendor as to where you think myself a new time could benefit your team the most In terms of the next steps you want to take going for the project you mentioned already that you’re currently working on.
00:43:28.000 ⇒ 00:43:40.000 Are there any particular blockers or obstacles that you think, and I could really help? Given our backgrounds in modeling everything from you know basic statistical epidemiological stuff I’m sure up to these done plenty of that as well himself So, you don’t need my expertise at all in Empty Miological modeling.
00:43:40.000 ⇒ 00:43:53.000 But there’s a little bit more that I’ve done recently in my Dr. To do with calls, of inference modeling, using operational data sets that might be helpful for policymaking.
00:43:53.000 ⇒ 00:44:09.000 Which I guess you’ll be working with a lot more because frankly most of us do not have randomized control control trial data as you know and so the cause of inference stuff modeling coming out of the gela nan’s stuff, modeling, coming out to Megan, and Nan’s group over in Boston, something that we could help you model.
00:44:09.000 ⇒ 00:44:16.000 If you think they’re relevant to any particular problems you have. That will be pertinent for policy.
00:44:16.000 ⇒ 00:44:29.000 And in terms of obviously the AI modeling, whether you think that’s something that you want to explore further in terms of how that’s engineered everything from the guest data custodian ship.
00:44:29.000 ⇒ 00:44:53.000 To having the having update modeling. To the pipeline to everything that you can imagine on that front we can also assist with giving you access to I guess if you want to call it that you know the best and latest and in the models there that are open source ultimately they’ll be using you know I guess what comes out of Europe and other conferences.
00:44:53.000 ⇒ 00:44:58.000 As we’ve been keeping up today I’m sure it’s the same for engineers on your end.
00:44:58.000 ⇒ 00:45:05.000 But it’ll it’ll be cool to hear what you where you think. We might be able to impact your vision the most.
00:45:05.000 ⇒ 00:45:10.000 Yeah, Dr. I will take this and then you can add to those. Okay.
00:45:10.000 ⇒ 00:45:19.000 Yeah. So, Chang, as I mentioned, we, compete with the pilot in the, That that data is.
00:45:19.000 ⇒ 00:45:25.000 I think we completed it, in 2022 end of, September 2022.
00:45:25.000 ⇒ 00:45:32.000 So that data is kind of stay like we don’t want to start any modeling on that. Yeah, that’s something which we are clear on.
00:45:32.000 ⇒ 00:45:39.000 We are in the process of, regenerating few of this because PHR, PGR is a big transaction.
00:45:39.000 ⇒ 00:45:46.000 And we also looking at, building the digital solutions grounds. So that’s what is the currently the world that we are constantly doing.
00:45:46.000 ⇒ 00:46:16.000 So we are in the process of building the platform. Taking the learning from PHR as well as what we hear from the from the We took these 2 inputs and we are rethinking our our strategy for expanding it to multiple markets right so multiple nations as well as multiple states right That I would, where and when I would probably add 2 2 dimensions to it, then and bad, right?
00:46:18.000 ⇒ 00:46:44.000 So our collaboration will start, when we land, in a model site. And then we have completed the model site before expansion of a larger larger state.
00:46:44.000 ⇒ 00:46:45.000 Huh.
00:46:45.000 ⇒ 00:46:52.000 For example, if we are taking a taking a nation with say 20 million population, right? A model say it will start with somewhere around a hundred 1,000 1,000 people right We will establish a model site and that typically takes some landing to completing it will take anywhere between 3 to 6 months.
00:46:52.000 ⇒ 00:47:03.000 So you and Uth will come in. Once we complete the model site. And the next step from model, it is expansion across the, right?
00:47:03.000 ⇒ 00:47:15.000 So from 100,000 people to 20 million people, that’s where we will expand. Right. You were, you were, you were, you were, in, in, From model site to that, expansion.
00:47:15.000 ⇒ 00:47:40.000 So that’s where I see you coming in. And the areas of exploration. Can be as broad and wide as possible because So you mentioned about, using data points for public policy, using data points of data points for i would say Data, point of data analysis and, insects for public policy.
00:47:40.000 ⇒ 00:47:48.000 Defining a public policy for the for a particular group of people. How do efficient efficiently manage my operations?
00:47:48.000 ⇒ 00:48:03.000 That is also one of the, one of the models that I pay building at. Also looking at, bottlenecks and areas of concerns or lists in in the service delivery.
00:48:03.000 ⇒ 00:48:13.000 Right, operation service delivery. public policy. Then, I would say in public policy even final.
00:48:13.000 ⇒ 00:48:22.000 Can you even can kind of, can we build a model to suggest? Hey for this particular population, for this particular type of population.
00:48:22.000 ⇒ 00:48:40.000 Can you create a program of source? Clear a program from data, right? Currently the program is also fixed for all in any Right, it’s across across the world, they don’t create programs for a particular state or a particular.
00:48:40.000 ⇒ 00:48:52.000 The nation defends the program. And they will they’ll implement it right so can we tailor it for different states So these are the things that, that initial things.
00:48:52.000 ⇒ 00:49:02.000 It’s, it’s not the complete list. There could be several more as to go. Because I also have because PHR we know is something which you have done.
00:49:02.000 ⇒ 00:49:09.000 PCR, we are, we are still exited about what is the kind of data that is going to come in.
00:49:09.000 ⇒ 00:49:10.000 Oops.
00:49:10.000 ⇒ 00:49:11.000 Alright, the kind of data that is going to be coming in is not going to be one vertical.
00:49:11.000 ⇒ 00:49:23.000 So in pH it was just is going to have real, healthy, rural development, social welfare, disability, all of those things will come into picture, right?
00:49:23.000 ⇒ 00:49:24.000 Yeah.
00:49:24.000 ⇒ 00:49:32.000 So now that data can give us a wealth of, insights that we can draw from it and also models which can help us in expansion.
00:49:32.000 ⇒ 00:49:54.000 So that is where I feel, you and, can kind of come in and add value. We can continue to discuss how we are to start from say next week.
00:49:54.000 ⇒ 00:49:55.000 Hmm.
00:49:55.000 ⇒ 00:50:04.000 What we can do is I can press to you how we are architecting few of these. I can, our systems, for, digital systems to collect the data, provide service delivery and all of those things I can kind of, brainstorm with you.
00:50:04.000 ⇒ 00:50:14.000 But the actual work for you that it will be at basically kind of thing, right? So you will say, this is good, because, the, we discussed, right?
00:50:14.000 ⇒ 00:50:18.000 Yeah, yeah, yeah. Okay.
00:50:18.000 ⇒ 00:50:19.000 Huh.
00:50:19.000 ⇒ 00:50:29.000 You mentioned. I’m currently going with Apache link by the way. Right. So yeah, I’m sticking to Apache, Apache completely, the architecture is almost almost completed and we are trying to create a development.
00:50:29.000 ⇒ 00:50:38.000 So I was waiting for that to complete and kind of give you an some kind of overview as well.
00:50:38.000 ⇒ 00:50:39.000 Okay.
00:50:39.000 ⇒ 00:50:46.000 So the idea is for the next. A couple of them a few months. Still we land and we complete model sites.
00:50:46.000 ⇒ 00:50:55.000 We will be working on How to collect how to best store a data. For you to come in and do a model.
00:50:55.000 ⇒ 00:51:05.000 But you’re actual work on creating a model will start once you complete the model site. And before we expand to your state and that will that will keep on evolving.
00:51:05.000 ⇒ 00:51:14.000 For example, if you let’s say you have in a country A and site A right the work that you do inside A for country A will be different.
00:51:14.000 ⇒ 00:51:25.000 In a society as a country B. Right? Because countries and the population also is a big variable. It’s all your model.
00:51:25.000 ⇒ 00:51:31.000 Right, so you in you creating that model you have to also take those those parameters into picture. So I think that is what.
00:51:31.000 ⇒ 00:51:43.000 That is where I think you will come in. Dr. You can add if I have, or something.
00:51:43.000 ⇒ 00:51:56.000 It’s finding everything. So, Jan, and it’s, the idea is, you think about a burger, like what we are having is Millions of population in the basement.
00:51:56.000 ⇒ 00:51:57.000 Yep.
00:51:57.000 ⇒ 00:52:05.000 And what we are having within the stuff is like a hundred like 500 plus services. In the mode, it’s the national ID, it’s about health insurance.
00:52:05.000 ⇒ 00:52:06.000 Yeah.
00:52:06.000 ⇒ 00:52:16.000 It’s about the VFI program. The top red is what AG is talking about. If you are not architecting in a way that we can readily apply the top red.
00:52:16.000 ⇒ 00:52:24.000 Then the bottom bread and the stuff will be, go on multiple iterations. That is what we learn.
00:52:24.000 ⇒ 00:52:39.000 So in Chad, what we are trying to do here is during PHF we conducted I think some of the user experience research.
00:52:39.000 ⇒ 00:52:40.000 Hmm.
00:52:40.000 ⇒ 00:52:50.000 The data repurposing how it is going to happen all the privacy designed like that so now this this population governance since it is a fresh and new We are trying to add the layer of how the artphones analytics can come and play a role.
00:52:50.000 ⇒ 00:52:51.000 Yeah.
00:52:51.000 ⇒ 00:53:09.000 Can’t this be immediately? If you are going because this is coming from the experience if you are not doing it right now before right landing into the model site Then we will try to transform the data and then apply the analytics over it.
00:53:09.000 ⇒ 00:53:10.000 Yeah.
00:53:10.000 ⇒ 00:53:17.000 So that is where we are trying to solve the problem. So that, we give us 2, type of advantage.
00:53:17.000 ⇒ 00:53:27.000 Like one is the time taken for transformation and applying advance. I mean, this will be, maybe we can able to work on that.
00:53:27.000 ⇒ 00:53:37.000 That’s 1 part of it and second life. For example think about x ray DA right it’s being developed at the same time by multiple universities.
00:53:37.000 ⇒ 00:53:53.000 Also by startups also by the big tech companies. Can we ever to apply everything for the same record that CSI is not available for that’s sufficient is not available for much of the sites in general, including us.
00:53:53.000 ⇒ 00:53:58.000 Thank you, you are having a x-ray, we can upload it in one place. Can we use multiple, that is developed by Stanford University.
00:53:58.000 ⇒ 00:54:07.000 One can be developed by IPM, one can be developed by a startup from India Institute of Technology.
00:54:07.000 ⇒ 00:54:17.000 That is not available. So we are trying to solve these 2 problems. One is once we land in a model in the data model, that one’s analytics.
00:54:17.000 ⇒ 00:54:24.000 To readily apply something to draft in. That’s the 1st thing. Yeah, what AJ is trying to do.
00:54:24.000 ⇒ 00:54:34.000 Communicate the second part is can be able to use multiple analytics to standardize ourselves because we don’t have any benchmark.
00:54:34.000 ⇒ 00:54:35.000 Yeah.
00:54:35.000 ⇒ 00:54:39.000 For example, if you have 2 A’s, can be able to compare only we can compare with supervised learning.
00:54:39.000 ⇒ 00:54:47.000 Or we have to go for a goal standard test. And then can we compare among the A’s?
00:54:47.000 ⇒ 00:54:48.000 Yeah.
00:54:48.000 ⇒ 00:54:55.000 That is no, method available. So these are the 2 things we are interested here. Maybe I’m VIII reading what are and then told.
00:54:55.000 ⇒ 00:55:03.000 So, we, have to take it forward in a way like, this will work it out when we demonstrate in the model.
00:55:03.000 ⇒ 00:55:05.000 Hmm.
00:55:05.000 ⇒ 00:55:06.000 I think.
00:55:06.000 ⇒ 00:55:11.000 So that’s 2 months, 2 to 3 months away. Like, if we are landing there, it will take 2 to 3 months time.
00:55:11.000 ⇒ 00:55:12.000 Yep.
00:55:12.000 ⇒ 00:55:22.000 Meanwhile, we can able to, start the process. Right, is it? Anything you want to add or any voice, please, we can structure this.
00:55:22.000 ⇒ 00:55:27.000 How we take it forward because right now we don’t have a immediate output or a milestone.
00:55:27.000 ⇒ 00:55:45.000 What comes to my mind i think the recent researcher mind is if you’re able to come up with a methodology i think we can put it in a paper and we will publish it as So how this can be applied and what all the things we can able to cancel.
00:55:45.000 ⇒ 00:55:51.000 That may be the output of this discussion. Maybe in one or 2 months we can able to do it.
00:55:51.000 ⇒ 00:55:56.000 Then immediately once you landed and said we can demonstrate for a size of a hundred 1,000 records.
00:55:56.000 ⇒ 00:55:59.000 So what I will what I will also want to do is, as I mentioned, 2 things.
00:55:59.000 ⇒ 00:56:02.000 Okay.
00:56:02.000 ⇒ 00:56:07.000 That’s the one. Immediately what we can work from say next week or a couple of things from now.
00:56:07.000 ⇒ 00:56:14.000 It’s, as Doctor mentioned, if, you can, ally work on the methodology and try to publish it as a research paper.
00:56:14.000 ⇒ 00:56:22.000 I think Chang is just coming out of the PhD more, so he will be more interested in writing a research.
00:56:22.000 ⇒ 00:56:27.000 Right.
00:56:27.000 ⇒ 00:56:28.000 Okay.
00:56:28.000 ⇒ 00:56:29.000 I’m actually there for the record I am not interested in writing a research paper. I’m actually interested in getting a tech product in place.
00:56:29.000 ⇒ 00:56:34.000 Yeah.
00:56:34.000 ⇒ 00:56:35.000 Yeah.
00:56:35.000 ⇒ 00:56:36.000 I’m for I’m not interested at all writing research paper. And I think I used to be a just important disclaimer.
00:56:36.000 ⇒ 00:56:44.000 Yeah. That was that. I, we were just discussing today and doctors, doctors, mentors, senior mentor.
00:56:44.000 ⇒ 00:56:53.000 It’s a different view, Chang. Like, see. The journey starts as a operations person in the government.
00:56:53.000 ⇒ 00:57:01.000 In 14 landed in the. Then again, we got a research grant of 4 million to 10 million in 2,016.
00:57:01.000 ⇒ 00:57:02.000 Yes.
00:57:02.000 ⇒ 00:57:15.000 Then again came back into operations population scale. But if you want to have this as a sustainable way, there is a mix.
00:57:15.000 ⇒ 00:57:16.000 Gotcha, gotcha.
00:57:16.000 ⇒ 00:57:17.000 So we run a new matrix. And the mathematics will have all these 4 components. Oh, yeah.
00:57:17.000 ⇒ 00:57:28.000 Yeah. So, so what we want to do is I want to set up create the development. Okay, along with the schema that we are trying to .
00:57:28.000 ⇒ 00:57:34.000 This the amount of contemplability that has to be there for a different different countries like We don’t want to develop this for each country separately.
00:57:34.000 ⇒ 00:57:38.000 Yeah.
00:57:38.000 ⇒ 00:57:51.000 So the levels of ministries and what services they provide. And how they provide services also different from 1 1 country to other country.
00:57:51.000 ⇒ 00:58:01.000 It’s also different from one state to another city. So we want to build that configurability and dynamicity to the solution that we are building.
00:58:01.000 ⇒ 00:58:18.000 That’s that’s where the complexity of. Also. Modeling the internet address and it’s not always speaks about comes right How can I create a model for analytics?
00:58:18.000 ⇒ 00:58:19.000 Yeah, yeah, yeah.
00:58:19.000 ⇒ 00:58:24.000 Which is which is other dynamic Then the pipeline has to be, pipeline has to take care of the transformation.
00:58:24.000 ⇒ 00:58:25.000 Yeah.
00:58:25.000 ⇒ 00:58:43.000 To say whatever is the data that comes in. Based on few few parameters we will we will model it in a way right so the The configuration and the dynamicity at the front and will directly affect the model that you are building for advanced artists.
00:58:43.000 ⇒ 00:58:55.000 So that is going to be very interesting for us to arrive at. So once I have the, whole pipeline worked out, so I’ll tell you what, at the top of, I’ll tell you what, what I’m planning.
00:58:55.000 ⇒ 00:59:08.000 So we have Letter as the apprentice. Flatter will be the Flatter will have APs built on SAS TP Python for all its Create, read, update, right?
00:59:08.000 ⇒ 00:59:18.000 We typically don’t delete any data, we only deactivate it, right? And, from there, FASCIBA, it is, it is, posted to 2 places.
00:59:18.000 ⇒ 00:59:27.000 It is an event. Anything that is created by the flatter. Where it is I’m creating individual or updating an individual record or creating giving a service, right?
00:59:27.000 ⇒ 00:59:35.000 So what after this, create, read and update, right? So create an updates are sent back to the system as a event.
00:59:35.000 ⇒ 00:59:46.000 And the event is posted in Kafka for further processing. It is also logged in post-cress for, for looking at it as a time series, right?
00:59:46.000 ⇒ 00:59:54.000 What happened, wind changed and all of those times it is like So from Kafka, we pick it up in Flink.
00:59:54.000 ⇒ 01:00:02.000 Flink actually transforms the data for 3 systems. We maintain the complete data in Cassandra.
01:00:02.000 ⇒ 01:00:03.000 Okay.
01:00:03.000 ⇒ 01:00:14.000 Hello. Right. We made a relationships in, NFLJ, GraphP. There are few places where we have to have a documented with like fire, right?
01:00:14.000 ⇒ 01:00:19.000 A health desk fire. There is a standard, right? We are going to store it as a document in.
01:00:19.000 ⇒ 01:00:27.000 Mongodb. Right. Now, Slink will be used for real streaming streaming updates of this pipeline.
01:00:27.000 ⇒ 01:00:34.000 Slim will also be used for, bash process as well as street closes to iceberg as a data lake.
01:00:34.000 ⇒ 01:00:35.000 Okay, okay.
01:00:35.000 ⇒ 01:00:43.000 Right, iceberg is from peace where we will use for analytics. You’re interested, it’s, it’s sounds interesting.
01:00:43.000 ⇒ 01:00:52.000 So we, we do super as the. Visualization engine. But we are thinking of either spark or dual we are still thinking about it.
01:00:52.000 ⇒ 01:00:53.000 Mostly, druid, we do it and sparkle 2 agencies what I’m I’m thinking in my mind.
01:00:53.000 ⇒ 01:01:04.000 So this is the high-level architecture apart from this, there’ll be multiplier components. Okay.
01:01:04.000 ⇒ 01:01:18.000 For example, the log of. How the system is being front and busy the flatter is being used right a business being used That log will also be, fed into either, one instance of iceberg.
01:01:18.000 ⇒ 01:01:23.000 For analysis directly. It doesn’t have to go to Cassandra or anything. It can come directly the dataally for us to analyze it.
01:01:23.000 ⇒ 01:01:37.000 Right. So that is something which you are looking at. We are we are also looking at the help help desk thing right they might create a help desk so do you want to keep it in Kassandra or we want to have a separate thing so so this is the high-level architecture.
01:01:37.000 ⇒ 01:01:38.000 Sure. Sure.
01:01:38.000 ⇒ 01:01:45.000 For you to have a vision. But once they have the, they’ll set up. We, we can, we can have a look at it.
01:01:45.000 ⇒ 01:01:46.000 Okay.
01:01:46.000 ⇒ 01:01:58.000 I hope to do it by this weekend. Yeah, we have a few of my teammates are. Still joining us so it will take some time so but I want to complete it this week is what my target is that Yeah.
01:01:58.000 ⇒ 01:02:06.000 So why don’t we do a couple of things? So one is all the kind of like high level opportunities will kind of scope.
01:02:06.000 ⇒ 01:02:23.000 So Chang, we’ll take a look at each of those and basically cat about are even what you originally said, which was each of those items will take and say, okay, where are opportunities for us to plug in?
01:02:23.000 ⇒ 01:02:24.000 Yes.
01:02:24.000 ⇒ 01:02:33.000 We certainly should have another conversation. Maybe next that’s more. Technical focused that we can schedule. Whether next week or the week after, where we can schedule and walk through like technically where we can plug in.
01:02:33.000 ⇒ 01:02:39.000 We’ll also, talk a little bit about kind of the AI evaluation. So how can you try different models?
01:02:39.000 ⇒ 01:02:50.000 How can you have a test data set that you could run this AI on? That’s actually some of the stuff I’m doing right now is how can you compare Llama with the new GBT model on the same results and all that stuff.
01:02:50.000 ⇒ 01:02:53.000 Yeah.
01:02:53.000 ⇒ 01:03:02.000 And then I think we’ll just take, you know, this meeting, as like an intro and we’ll kind of scope out in a document each of those different opportunities and kind of like our feedback and.
01:03:02.000 ⇒ 01:03:07.000 We can pop on another call maybe next week or the week after to kind of run through that just based on how much time we have.
01:03:07.000 ⇒ 01:03:09.000 Yeah. Yeah.
01:03:09.000 ⇒ 01:03:19.000 And, you could also mention about the countries we have visited and, about the experiences we have gained there and what are the other places you are also looking forward to move upon.
01:03:19.000 ⇒ 01:03:23.000 And you can ask his opinion on where we can take this forward into other areas of his suggestions.
01:03:23.000 ⇒ 01:03:33.000 Yeah, yeah, sure, sure. So currently we are as I mentioned we are in 4 places where we have where we have kind of in touch discussions.
01:03:33.000 ⇒ 01:03:39.000 In India we have 2 states. One is quantity and provide how much you open those 2 places.
01:03:39.000 ⇒ 01:03:50.000 It’s country ways we are currently in which is closer to Indonesia so those are the 2 countries we are currently in discussions.
01:03:50.000 ⇒ 01:04:01.000 We are also in plans to go to Utan, Nepal and a few Couple of other is to European and .
01:04:01.000 ⇒ 01:04:10.000 And African countries so all of these are PGRs, PGR area of expertise, right?
01:04:10.000 ⇒ 01:04:18.000 So, we are trying to reach there, but if you have any suggestions already. Anything that you that you can suggest let us know.
01:04:18.000 ⇒ 01:04:30.000 One more thing that I wanted to add, is, see, while PCR is focused on Population health and we are speaking about orders.
01:04:30.000 ⇒ 01:04:44.000 We also have a another so, the, IS, a, Then also I believe, your expertise will come in handy.
01:04:44.000 ⇒ 01:04:47.000 So that is something which we will discuss. In the once we come, let’s, let’s go through the strips in PCR.
01:04:47.000 ⇒ 01:05:03.000 Let’s not confused too much. Okay, I know it is too much. Too much content in, in a 1 day call.
01:05:03.000 ⇒ 01:05:04.000 Oh.
01:05:04.000 ⇒ 01:05:10.000 So let’s go through this in the next call as well. They start looking through that then I will introduce you to a few more few more stuff that we are doing.
01:05:10.000 ⇒ 01:05:27.000 The developments are for both CGR as well as PGR is has started. And we are, while for PGR we are focusing on open source.
01:05:27.000 ⇒ 01:05:28.000 Yeah.
01:05:28.000 ⇒ 01:05:33.000 How much Commercially free, stuff. For CGR we might focus on GCP because it’s a SAS.
01:05:33.000 ⇒ 01:05:34.000 Okay, okay, okay.
01:05:34.000 ⇒ 01:05:35.000 Yeah. So. Sure.
01:05:35.000 ⇒ 01:05:38.000 Yeah, it’s a SAS, right? So we will discuss that as we go. Okay, let’s just do that.
01:05:38.000 ⇒ 01:05:45.000 Okay. Okay.
01:05:45.000 ⇒ 01:05:46.000 Yeah.
01:05:46.000 ⇒ 01:05:51.000 I’ll just share you the document of the CGR one the clinical governance and research. You could share it the child as well but if not just for let’s create a group and just keep trying in the loop as well.
01:05:51.000 ⇒ 01:05:55.000 So we can just share the documents and the group and you can just have a look at it all the time.
01:05:55.000 ⇒ 01:05:56.000 Okay, great.
01:05:56.000 ⇒ 01:05:57.000 Yeah.
01:05:57.000 ⇒ 01:05:58.000 So it would be easy for us to take forward the discussions in the future very easily.
01:05:58.000 ⇒ 01:06:08.000 Okay, I’ll create a group with everybody here and then check let’s we can catch up offline and I have the recording of this and we can kind of go through all the notes and then.
01:06:08.000 ⇒ 01:06:09.000 We can think about if we have time between now and maybe next week, but I’ll message you guys.
01:06:09.000 ⇒ 01:06:15.000 Yeah.
01:06:15.000 ⇒ 01:06:16.000 Sure.
01:06:16.000 ⇒ 01:06:20.000 Hopefully end of this week, early next week, about a next meeting. But any documents or PowerPoints you can send our way, feel free to prefer to send it.
01:06:20.000 ⇒ 01:06:21.000 Hmm.
01:06:21.000 ⇒ 01:06:22.000 That’s kind of drop on.
01:06:22.000 ⇒ 01:06:29.000 Would that would like if you share Shang’s number and what’s up, it would be great for me to keep on posted the few documents, it would be great for me to keep on posted the few documents you have to share.
01:06:29.000 ⇒ 01:06:31.000 Cool, I’ll set a group with everybody.
01:06:31.000 ⇒ 01:06:32.000 Yeah, sure, sure, sure.
01:06:32.000 ⇒ 01:06:33.000 Hmm. Sounds good. Sounds good.
01:06:33.000 ⇒ 01:06:37.000 And this is. Oh, it was a useful discussion. So please.
01:06:37.000 ⇒ 01:06:42.000 Oh, this is great. This is like one of the most dense conversation I had in a long time, so it’s good.
01:06:42.000 ⇒ 01:06:43.000 Yeah.
01:06:43.000 ⇒ 01:06:47.000 Usually it’s like it’s just boring normal stuff. So For me, it’s really great.
01:06:47.000 ⇒ 01:06:49.000 Yeah, I mean, sounds like you guys are trying to, yeah, sort of what, sorry.
01:06:49.000 ⇒ 01:06:57.000 Shag we have. Yeah. Gotcha, we are hope that we hope that we don’t bore you with stuff that you were not interested out of.
01:06:57.000 ⇒ 01:06:58.000 Yeah.
01:06:58.000 ⇒ 01:07:01.000 We hope like we are feeding you with stuff. It’s going to exceed you a lot.
01:07:01.000 ⇒ 01:07:02.000 Okay.
01:07:02.000 ⇒ 01:07:15.000 No, I mean, that the unique problems to almost every single region, even from one hostel to another, there are unique problems and what kind of solutions you can implement in a health care setting.
01:07:15.000 ⇒ 01:07:24.000 I think each and every one of them deserves their own sort of particular team like you guys are sort of have banded together.
01:07:24.000 ⇒ 01:07:33.000 I think the I think they’re hugely interesting for different reasons. I think they’re very educational each and every single one.
01:07:33.000 ⇒ 01:07:34.000 Yeah.
01:07:34.000 ⇒ 01:07:35.000 They’re gonna be recurring patterns, but every single product I undertake and certainly either for a research angle or from a tech side.
01:07:35.000 ⇒ 01:07:45.000 Have brought up very very . Yeah, inside for. You know differences no matter how much people talk about interoperability FIHR and the rest of it.
01:07:45.000 ⇒ 01:07:51.000 Actually, a lot of human factors that we don’t incorporate, for instance. And so it’s all very well.
01:07:51.000 ⇒ 01:08:01.000 And I being able to produce models, you know, left, right, and sense, and we can, we’ve, you know, we, to be showing that we can in our, in our work as well as our research.
01:08:01.000 ⇒ 01:08:18.000 But it’s that’s not what it’s all about. I think the thing I keep reiterating time and time again to my friends and to other people when I when I go and, the few times in my life I’ve been asked to, it’s not often that this happens by the way, is all about, so what?
01:08:18.000 ⇒ 01:08:26.000 Like I think that question that I always come back to and it sounds so facile but You, you can whip out the most advanced model, all you want.
01:08:26.000 ⇒ 01:08:29.000 Okay.
01:08:29.000 ⇒ 01:08:34.000 But if you don’t have the ultimate golden sight as to why you’re bothering to do it.
01:08:34.000 ⇒ 01:08:45.000 Why it matters why this particular model over something so much more simple algorithmically so much more simple, then you’ve lost sight of why you’re even bothering to ask people like who time and me to do this kind of project.
01:08:45.000 ⇒ 01:08:54.000 I think that’s something that I’ve also led the hard way by talking to individuals who work with unofficial health care providers in India.
01:08:54.000 ⇒ 01:09:01.000 As you know, the great. Unseen part of the iceberg of prescriptions, particularly antibiotic prescriptions.
01:09:01.000 ⇒ 01:09:15.000 You know best than me. Then in India that happens through unofficial health care providers people who are outside the remit and mandate of government or regulation and control.
01:09:15.000 ⇒ 01:09:20.000 And so for them, it doesn’t matter if I’ve got the best AI algorithm, whatever else.
01:09:20.000 ⇒ 01:09:33.000 But then what matters more is that they have enough stock. In a pharmacy. So is that the incentivization model we need to come up with for them to use a certain algorithm we come up with?
01:09:33.000 ⇒ 01:09:34.000 Yeah.
01:09:34.000 ⇒ 01:09:38.000 It may well be the case. It’s not simply a case of me giving you a you know, really super flashy app.
01:09:38.000 ⇒ 01:09:39.000 Yeah.
01:09:39.000 ⇒ 01:09:46.000 They don’t give a shit about that. Thank you. They give a shit about being able to give enough antibiotics to their local loved ones, their neighbors who are going to judge them every single day if they fail to do the right thing.
01:09:46.000 ⇒ 01:09:56.000 So, you know, these sorts of experiences I’ve had in, no matter how superficially, which seems, which are definitely relevant to the Indian context because they are directly from them.
01:09:56.000 ⇒ 01:10:06.000 I see to be. Both interesting and challenging all in one. I think where we saw my Abilities can fit into into your system.
01:10:06.000 ⇒ 01:10:14.000 Would love to be able to help obviously as much as we can. But I do, I would say 2 things to sort of mull over would be.
01:10:14.000 ⇒ 01:10:26.000 You know, during your 1st thing is during the process of thinking more where, Do please always, you know, reduce it back down to, okay, so we can get you, to, and chat to do this machine learning model.
01:10:26.000 ⇒ 01:10:31.000 We can pay them money to do this, but why are we doing it? What’s it actually for?
01:10:31.000 ⇒ 01:10:43.000 Is it going to make an impact? How can we measure that impact, etc? I think the second thing is, yeah, is to, it was one thing I wanted to ask you at the beginning of something else you can we can follow up on.
01:10:43.000 ⇒ 01:10:57.000 Is how readily, and I will be able to. Supervisor the high level versus having a team that will be able to do some of the So the basic, so cleaning and the analysis and stuff that we can.
01:10:57.000 ⇒ 01:11:04.000 Give advice on that front. I think they’ll be also very important just to clarify. Before we move on to the next project, next meeting.
01:11:04.000 ⇒ 01:11:12.000 Hope that’s, helpful at least just to frame. Where we go next. But yeah, super interesting stuff.
01:11:12.000 ⇒ 01:11:16.000 I mean, I think, I think there’s a lot of, you know, I mean, there’s clearly a lot of potential here.
01:11:16.000 ⇒ 01:11:20.000 There isn’t a solution that exists today that’s that’s trying to See, so many different angles as you guys are.
01:11:20.000 ⇒ 01:11:28.000 Yeah, there’s a, yeah, there’s a long way to go, but it’s gonna be good.
01:11:28.000 ⇒ 01:11:38.000 I’m looking forward to, you know, I had we had a very interesting It was a 5 6 h discussion on the day, right?
01:11:38.000 ⇒ 01:11:39.000 Yeah.
01:11:39.000 ⇒ 01:11:40.000 Wow.
01:11:40.000 ⇒ 01:11:56.000 So we had a very interesting discussion. Hopefully we will maybe continue that discussion literally and you know we get a chance we will be you person but looking forward to working with you I think we are going to learn from you a lot and I hope you will also be able to give you some bits to learn from us.
01:11:56.000 ⇒ 01:12:03.000 Right and that’s the fun in any work that we do right at the day. So looking forward to learning together.
01:12:03.000 ⇒ 01:12:09.000 Yeah and in the process we’ll create some impact. We will, will build a business, right?
01:12:09.000 ⇒ 01:12:14.000 So we’ll be able to get an invite and build a business. Okay.
01:12:14.000 ⇒ 01:12:15.000 Okay. That’s a
01:12:15.000 ⇒ 01:12:17.000 Okay, we’ll talk next time. Thanks, thanks everyone for your time. Lovely to meet you.
01:12:17.000 ⇒ 01:12:18.000 Yeah.
01:12:18.000 ⇒ 01:12:21.000 Okay, great meeting you today and kind of meeting with you today as well. Looking forward to met you.
01:12:21.000 ⇒ 01:12:22.000 Lovely meeting you guys.
01:12:22.000 ⇒ 01:12:25.000 Yeah. Sure.
01:12:25.000 ⇒ 01:12:26.000 Bye.
01:12:26.000 ⇒ 01:12:27.000 Okay.
01:12:27.000 ⇒ 01:12:28.000 Thank you. See you, I’ll just connect with you after a while.
01:12:28.000 ⇒ 01:12:29.000 Okay.
01:12:29.000 ⇒ 01:12:30.000 Yeah, bye, bye.
01:12:30.000 ⇒ 01:12:37.000 See, I’m, I don’t say. But.