89. Saving Lives at Scale: Part One

Greg: So with that background about THINKMD, how was it founded and what sort of thought process went into building this application?

Meg: THINKMD was founded by two Barrys. So we have Barry Finette and then Barry Heath. Both are clinicians themselves, pediatric clinicians; Barry Heath is a specialist in critical care, and Barry Finette is a specialist in emergency pediatric care. Barry Finette, specifically, had been doing some humanitarian work over his career, and he realized that there were some skills that really sit with physicians that could be transferred to basically anyone anywhere, and especially in places where healthcare access was extremely limited or perhaps non-existent. So he reached out to Barry Heath, who was a good colleague in friend of his, and pitched this idea about creating an AI platform that would basically enable those skills to be transferred to frontline health workers, or even just a lay person, maybe with no clinical background at all, so that they could provide primary healthcare access, and access healthcare themselves, wherever they were.

Greg: So can you talk a little bit more about how this application actually works?

Meg: It's quite simple. So I think one of our major goals is what we've just said, to really eliminate preventable death and provide access to quality healthcare for anyone anywhere, but we need to do that simply. So the UX/UI piece of our platform, so the way it looks and feels and behaves, is extremely simple and intuitive. What the platform starts by doing is, it guides a user, and depending on the end user, if it's someone at home, or if it's someone that's working through a program of one of our partners; it will collect some initial data, like a user profile. So you'll know, who is using it, where they are, and in what capacity they'll be using it; whether they're a nurse, a community health worker, or for a self assessment. The platform then goes on to take history of the person being assessed. The health worker or the person who is self-assessing will then do observations of key clinical conditions.

Meg: And then there's some physical assessment aspects as well; so things like, taking of heart rate, respiratory rate, things like skin turgor, where you simply pinch the skin on the back of the hand or on a child's stomach, as an example, to demonstrate dehydration. And then after all of the questions have been answered, it obligates the end user to answer every question, you'll basically get feedback, you'll get feedback on the overall summary of the assessment, based on the input that was provided by the user. And depending on the type of user, you will get a triage level, so standard care, immediate, or urgent. And then depending on the capabilities and the training of the person using the platform, you might be given the treatment and follow up instructions as well. So very simple, very straightforward, and it guides you through the whole process.

Greg: Full disclosure, I just pinched my arm and I'm dehydrated.

Meg: Yeah.

Greg: Well, I think it's more common than people realize.

Meg: Absolutely.

Greg: So it sounds like it can actually maybe provide the physician or the person taking the assessment with some further ideas on cause and treatment, is that correct?

Meg: Yeah, absolutely. The overall assessment will basically demonstrate and showcase all of the answers to every single panel or question; so basically every panel has a question. And then once you get that summary, it will basically give you an overall summary of what the causes could be. So the four key clinical conditions are dehydration, respiratory distress, malnutrition, and sepsis, or potential for infection. And then you've got multiple other diseases that are also being assessed for, things like the potential risk for malaria, skin infection, ear infection, UTIs, among many others, that would, based on what was inputted, would come up as the overall assessment for that child or that person.

Greg: Cool. So do you guys input some of that data then track it or monitor it, so you can provide more holistic or community data back on trends for diseases or issues like that?

Meg: Yeah, that's definitely a direction that we go. It depends a lot on our partners as well. A lot of the times we're working with partners directly within their programming, they're the actual implementers, so we're working with their cohorts of frontline health workers who are capturing this community level data. And then from that community level data, exactly as you were just saying, it's possible to visualize and analyze that data to demonstrate things like: overall population health, monitoring of health over time, and even looking program specific at, how well is the program doing? What are areas for improvement? Et cetera.

Greg: I think that probably would help with epidemiology. And that's probably fairly forefront on a lot of people's minds right now, with what's going on in the world. But would that help target resources to locations that might be having outbreaks of malaria or COVID or whatever they're facing?

Meg: Yeah, that's absolutely the hope and a lot of the intention. There's something that we're calling syndromic surveillance, where you can really break down the clinical pieces of each of those assessments, so that you can really look at the potential change in fever and other indications of potentially malaria, over a malaria season, or perhaps outside of malaria season. You can definitely look at things that help identify potential outbreaks, looking at trends of data over time. And definitely during times of either known epidemics or even now during something like a known pandemic, that can help to identify potential cases, potential clusters; and in that way, assist governments and health systems to manage transmission through knowing where to allocate things like tests, or labs, or lab equipment, or treatment.

Greg: That's awesome. And how many countries and languages do you guys target?

Meg: Currently, we're in upwards of, I think 13 countries. We've got multiple projects in some countries. We are in Sudan, Somalia, Kenya, Indonesia, Bangladesh, Zambia, Nigeria, South Africa; we have a new project in Togo, potential project in Jordan, and we also have some projects going on in the US as well; quite a bit of variation in geography, and in partnerships, which is great, and languages.

Greg: Is the team all in the US, or do you guys have developers and team members scattered across the world?

Meg: Currently we are US-based, we are not all... Well we have remote colleagues as well. So most of us are living in Vermont. And then we have new colleagues based out of Atlanta, Georgia, and parts of Michigan as well. Some of us do tend to, when we were able, have lived abroad and have intentions to do so again, that's a part of our life and something that we're grateful for in being in the tech space, for sure.

Greg: How has COVID impacted the program that you guys are doing? How have you seen COVID impact the overall health data you're getting? How has it really impacted your application, the data you're receiving and how you're giving back?

Meg: There's a lot of answers there, I think. So the platform that we've worked on, and the COVID assessment tool that we've created, basically allows for what was just mentioned, like a syndromic surveillance type of thing, where you begin to see trends of individuals that are demonstrating high risk of potential COVID. We can not confirm, obviously, through our application, whether someone actually has COVID, so that would require the testing part. So I would say, we are helping and assisting ministries, and health systems, and definitely our partners who are the actual implementers, in helping the populations that they serve, understand better the situation and risks of COVID, where they are. And speaking to, again, kind of the similar thing we've just mentioned, knowing where there might be a large cluster of high risk; understanding how to potentially focus or prioritize that region for testing resources, laboratory equipment, and things like that.

Meg: But definitely, on one end of the platform, we have definitely the education piece. So we are absolutely contributing to that prevention part as well, through our implementing partners, of course; and then the syndromic surveillance ties in the rest of that piece. In the programs in which we're working, to get to your question about how it's impacting, maybe the programs in which we're partners--it's a little bit concerning, based on the focus that is now in COVID-19, which is granted of course, but a lot of our partners work in maternal and child health. And there's a lot of concern about, basically going back on all of the progress that has been made for maternal newborn and child health in communities in low to middle income countries, where resources are very low. Campaigns, like vaccine campaigns that have people going--health workers going into the community and providing vaccinations to children, directly to their homes even, those types of efforts have been slowed because of COVID, fear of COVID, and the realities of COVID.

Meg: So we're trying to figure out how to maintain support of our partners in their normal activities; like their maternal newborn child health, and other health activities in the community and facility level, while also helping to support their efforts and understanding the COVID situation where they're working. So it's twofold, but it's not, without the partnerships, none of this is possible. And definitely with the ministries and health systems supporting the testing, to understand the factual levels of incidents of the coronavirus and regions where they live is essential to getting to the truth.

Greg: Yeah, I'm sure that each local probably has different cultures, different customs, different beliefs. And so I'm sure you have to really massage the message and target it for each individual local.

Meg: You definitely have to work with your on the ground partners, people who live in and work with the communities that they're serving; and even, obviously, members of the communities themselves to understand how to be most effective in that messaging, absolutely. Aside from just mere language, there are a lot of cultural context to really be sensitive to and aware of, so that you can effectively educate individuals on what is currently known about their virus, and the ways that they can prevent it, signs and symptoms of it, when to know to go get testing; and so to really slow, and or stop the transmission on the community level, is what we see as being a part where we can be helpful, for sure.

Greg: Yeah, and I know that in more resource starved, or low resource areas, that education can be a problem, or a lack of education, and then, of course that feeds into, maybe some superstition or something like that. Do you have any examples of overcoming that sort of a hurdle?

Meg: It's all about working with individuals who the community trusts; I would say that's really the biggest thing. And of course there are probably many examples of, and for many diseases, there's a lot of, maybe not misunderstandings, but like you were saying, superstitions or cultural beliefs or that have been generated and then remain due to trusted sources of those communities transferring that information. So we even still have a lot to learn about the coronavirus, and it's very easy to have misinformation. So I think it's really about transferring knowledge to the correct people who are trusted by the community and then can share that as well. Health workers are definitely a part of that system, and just making sure that, we as partners and as health workers ourselves in some sense, are really working to ensure that we have, and are sharing, the most up-to-date, relevant information as possible, so that the information that's being shared is not misinforming, but is rather keeping the communities up-to-date on what is known.

Greg: Yeah, and that's got to be a challenge especially with something that's moving so fast, like COVID where we are learning new things almost every day, it feels like, to be able to come back to somebody and say, "Hey, listen, what we thought was true two weeks ago, we know a little bit different now, and here's the new information." So there's got to be a real... I find it challenging personally-

Meg: Absolutely.

Greg: And I'm in technology and I keep getting all these updates and it's like, "Okay, well now, this is true, and I thought it wasn't but..." So it-

Meg: Exactly. Yeah, absolutely. It generates a sense of distrust, and not knowing. So when you don't know, you can assume a lot, we all do that. So that is definitely one of the biggest challenges for sure, but it's not impossible to overcome.

Greg: Anything that you've learned and any surprises that have come out of those learnings that you want to talk about, or that are top of mind for you that you want to share?

Meg: So I was in Bangladesh actually in March, when our own country was taking some measures to shut things down and understand things better before we made other decisions, type of thing. And while I was there, working on a project that was non-COVID related, obviously the news of COVID and the fears of COVID, and the misinformation about COVID and otherwise, was going around the globe. So my colleagues in Bangladesh and I, were working in a region called your Jessore, on this other project, and on our way back, you could see individuals, like the buses, the ferries, the car, everything was just loaded with people who were fleeing from Dhaka, which is the major city there. So it took them quite a bit of... It's interesting, because you don't want people to move too quickly, you don't want them, you want them to stay home.

Meg: But in situations in Bangladesh, you've got a lot of informal markets, individuals who are going to Dhaka, the central city, to generate income, and then bringing that back to their families in the rural villages; and so that's basically what was happening. So it's the realities of a country like Bangladesh, and many others, are our own included, in some sense, it's hard to know when to provide the messages of what people need to do. Also, it's necessary to consider, the realities for folks, wherever they are, in many places in the world, informal markets are the only form of income, and so social distancing and staying home is almost, not a reality for some folks.

Meg: So it was almost amazing to see. It was... The word traveled to the place where I was, and I almost got to see the initial reaction; people fleeing from Dhaka to go to their homes in the rural communities. And then very quickly almost returning back weeks later, regardless of instruction from their governments and ministries, because they needed to do what they needed to do to survive. We all know what we need to do and what we are advised to do, but sometimes it's really hard to communicate why and to understand the reasons for why people do what they have to do.

Greg: This was a great conversation with Meg who leads research and implementation at THINKMD. Again, this is part of a two part episode. Next week we'll talk with her colleague Alex, who is the CTO of THINKMD about the technical challenges they faced building this platform.