COVID-19 and the Perspective of Health Authorities: Optimizing for Human-Tech Partnerships

Q&A with Kyle Towle, Technology team at Safe Paths, former Senior Director of Cloud Technology at PayPal. In conversation with Deepti Pahwa, Communications team at Safe Paths and Innovation & Brand Strategy Consultant

As the pandemic forces us to reassess everything and lockdowns begin to relax, digital contact-tracing solutions are being worked on intensely. Digital interventions are emerging as a useful aid for health authorities to supplement their manual interviewing process. We at Safe Paths are developing supportive and reassuring solutions with a central focus on human needs. We foster the mindset that we need new types of partnerships and multiple entities to come together for the common good. More than ever, it is critical to understand the needs of health authorities worldwide. In this context, we had a chat with Kyle Towle, who has been working with the Safe Paths Technology team, on his interactions with public health authorities. We discuss how we can support building more resilient health systems through technology use in contact-tracing ecosystems in order to make manual processes more efficient. At the same time, we also explore whether digital epidemic control can deliver what it promises, and what challenges are ahead.

Deepti: Tell us something about yourself, and what made you join the Safe Paths movement?

Kyle: The spread of COVID-19 in my home state of Washington and lockdown made me do something about the situation. I have been closely following the way different countries have handled the response, including countries like South Korea and Taiwan that addressed their COVID-19 response with a heavy amount of technology for tracing. Obviously, the test, trace, isolate approach has been proven as the way to go, but there was this idea there of harnessing technology to do this. So, I started digging into the efforts being taken within the U.S. We’ve got big tech companies like Google, Amazon, Microsoft, Apple, but none of them had any solution close to what South Korea accomplished, who not only flattened their curve but stopped the virus in its tracks. So, in my quest to figure this out, I came across the MIT-led project Safe Paths. I retired a year ago, after a 14-year stint at PayPal. So, I joined the initiative with the intention of having some background and experience with the executive leadership of leading engineering teams and contributing. There’s a unique thing happening here with the focus on privacy first. And that’s significant. Some of the ways that this was solved in other parts of the world is coming at the expense of privacy. Privacy is deeply valued in the U.S. and other countries. So the ability to gain adoption for these technologies, and for them to be useful, requires this focus. Safe Paths from MIT is a highly trusted brand, and a really cool project that I am hoping has material positive effects on preventing loss of human life as well as reopening economies and getting folks back to normal.

Deepti: I have to say this privacy-first approach is one of the things that got me to this project as well. And I was so inspired when I saw the TED talk of Prof. Ramesh Raskar at MIT, which was more about the question of “do we really need to trade off privacy for the sake of utility?” And this exactly is why I think most of the people who are working on the Safe Paths team are volunteering their time, efforts, and resources, and this is exactly what is bringing them together. So what has been your personal / professional experience in this direction, and how is your interaction with health authorities in this aspect?

Kyle: My career at PayPal was all about building a high-trust, privacy-protecting, financial technology solutions platform. Here, I’ve been drinking from a fire hose, and in the last month we probably had 30 to 40 of these deep-dive conversations with healthcare authorities. Synthesizing all their inputs, and gathering their needs, have helped us to come up with a way to approach this. It’s really fascinating what, where, and how they’re thinking about the problem. Number one, they’re figuring things out, and learning just as all of us are. How does digital contact tracing address the pandemic, and the best way to approach it? There’s this divide on, from a healthcare authority standpoint, GPS versus Bluetooth technology. As well as this idea that from a pathogenic standpoint, we believe that the best scenario is mixing both together. This goes right into the just-released Google Apple Exposure Notification API (GAEN), where it’s very clear that you are not allowed to mix Bluetooth API and location services in any app. It’s a broad spectrum of concerns they’re dealing with. But, one of the things we’ve found is about the great value propositions of using GPS right now, and the way that it works in our technology is simply being able to go through a contact-tracing interview and having the GPS location history of the patient on a map. The contact tracer and the patient are able to look at it, work through that map together, and trace their recent whereabouts and close contacts while redacting any of their personal information.

Deepti: Now that you touched on the topic, could you talk a little bit more about Bluetooth vs GPS, in the context of how this technology is helpful for contact tracers?

Kyle: GPS is a technology that’s only precise up to about 30 feet without any interference. So it’s not great at saying who exactly I came into contact with. However, because it can be presented visually, it provides an amazing context for the contact tracer and the patient. It can tell you where they’ve been in the past 10 days and who they’ve been very close to. And seeing that laid out on a map is super helpful for driving what the healthcare authorities are trying to do, which is they want to know the closest contacts that patient has had. The majority of the spread in the community is happening through close contacts, as per the feedback that we’ve been getting from epidemiologists. And so that’s the most important thing. Are the patients able to recall their trips, their travel, someone in the car with them on those trips? These are some of the questions that the contact tracers would be asking, and looking for a residence that the patient visited that’s not their own. Who else was there? And throughout this process, the contact tracers are reacting and removing the sensitive points like personal residences and anything that could be used for reidentifying a patient from those points and saved on the healthcare authority side. They also want the Bluetooth capability, though if the patient may not be able to recall who they had come into contact with, then the contact trace cannot be followed. But the ability of Bluetooth to tell them about who they had encountered within six feet is super valuable as well. At the same time, employers want to know this information when it comes to folks who are in their facilities. So, there are a lot of competing concerns and rapidly developing thoughts on the best ways of using technology. We are in a unique position to build the tools, the solutions, and the technology that they need to help fight this pandemic.

Deepti: We understand a lot about the technology perspective from what you just explained. We are also trying to highlight the challenges that healthcare authorities face in terms of adoption. What are the concerns they have when they are thinking about a solution? And why Safe Paths? Also could you dig a little deeper into empathizing with their situation, not only from a technology perspective, and what you have from them as feedback?

Kyle: The primary mission of the contact tracer is trying to figure out those close contacts. When they’re doing it manually, it depends on the memory of a patient to share all those close contacts. But when it comes up visually in the form of GPS location history, they’re able to get that much-needed aid, during their contact-tracing interview, which significantly increases their recall. Another significant goal that they have is this engagement with the community. And they’re looking for how we can proactively reach out and connect with the community who may be worried about COVID-19, and which places are safe and the safe paths one can take within their community. They are eager to know if they came in contact with someone who had COVID-19. These are all things that we at Safe Paths are looking to address, with this combination of a mobile app and a web backend on the health authority side called Safe Places. The health authorities are thinking much more from the standpoint of how one can fight this, and pin it down, and address the epidemiology concerns. How can we get in contact with everyone that might be COVID-positive? And the nature of that disease makes it really challenging, as you’re not able to just use symptoms; you have to depend only on the close-contact information. So this technology at Safe Paths is positioned that way. There’s a lot of sensitivity placed around the human factor of what you talked about, and empathizing with them. These contact-tracing interviews are stressful for a contact tracer and also for the patient. There’s a lot of anxiety involved when someone has been told that they’ve been diagnosed positive, and what’s next for them. And they’re going through this interview trying to recall where all they’ve been to. So, here the idea of technology, which can make it a lot smoother experience, and more robust interaction, that takes some of the stress for the patients, and also the contact tracers. Thus making it really valuable.

Deepti: True, so we do agree at Safe Paths that manual contact tracing cannot be replaced. And we definitely need to work with health authorities to get digital interventions and smooth the process. So, in terms of aiding and accelerating this work that we are doing already, how can we take the adoption to the next level?

Kyle: What we’re talking about is the rapid adoption of a network effect that we’re building between the mobile app and the healthcare authority, from a technology point of view. But, it is a bit of a chicken and egg problem. And so there’s this need to have a strategy for getting adoption and be able to create that network effect. One of the amazing things that we’ve discovered in all of these conversations with healthcare authorities is that some of them are motivated and are already working with a technology partner or a system integrator, who’s either building something like this or has been looking at our codebase and looking at modifications, and want to join forces with us. They want to team up, and not duplicate each other’s work, and come up in a way that we’re building. This not only is for our local county, state, country, or part of the world, but for everyone. One of the ways we can accelerate is, as we’re going through these implementations, by having a formal way to bring on these folks who want to contribute. One of the greatest values is the healthcare authorities themselves wanting to be those technology partners with us, and contributing to the code base. The more powerful this coalition becomes, it will create products that are going to solve these problems for everyone. And we are a uniting voice in this conversation with Google and Apple, about what healthcare authorities need from a technology standpoint. Google/Apple cited that there’s a real concern among certain healthcare authorities around a Bluetooth-only approach, also that some populations may not be having a smartphone. People are concerned that the phone they have may be too old to run the version of iOS required for running this Bluetooth API, which is being released in mid-May. And when you consider the likelihood of the Bluetooth interaction, it largely depends on the adoption rate. Because if I’ve got my smartphone app and I’m using the Bluetooth API in that app, and interact with someone who’s COVID-positive who doesn’t have that app running at that moment, it wouldn’t register me as having an interaction. And so if you think about that, and just numbers of not only newer devices, but downloading the app and, in some of these populations around the world, the proposition isn’t helping solve the problem. But in other places, they look at it from the opposite view where populations with newer smartphones really want it. We don’t want a bunch of notifications from 30-foot encounters, but rather want the six-foot encounters from Bluetooth. Bluetooth also has false negatives, because you’re missing the fomite problem, like if someone sneezed on a door handle, and then five minutes later you touch that handle, it’s not going to pick that up, unlike GPS. So, there is an emerging divide. There are also two different types of healthcare authorities, depending on what part of the world they’re in. If they could use both together, that solves it all. And so we’re pursuing a strategy that is going to meet the needs of these different healthcare authorities with different dynamics and demographics in their population. As well as working with Apple and Google on that, to securely work in a privacy-protected way, since we are able to blend them both, and move forward, in order to offer the best possible solution, for fighting the pandemic, without sacrificing our privacy.

Deepti: Exactly. That’s what we’ve been focusing on since the beginning at Safe Paths – that we want to have the best approach possible. It doesn’t matter if it’s Bluetooth or GPS. We want to have a privacy-secure maximum adoption kind of ecosystem of various solutions. At the same time, there are two things that may be related right now as well. One, how are the governments feeling about this approach, as they have a big role to play in terms of the adoption in a certain country or region? And secondly, we recently learned that Google and Apple have said that they will only allow one app (per region) to run on their GAEN. Your thoughts on that?

Kyle: Yeah. Great question. The government’s influence is interesting. Everyone we’re talking to, the most motivated, agile, ready to move forward are the ones more closely connected with the clinics, who are doing these contact-tracing interviews, dealing with patients at a county level or city level. And they’re super motivated to get up and run. The other dynamic with those is that they literally have no resources to install things. And the reality of what we’re offering to them in Safe Places is a web tool. It’s an open-source, white-labeled approach. It’s not a software-as-a-service approach, where we say go to this URL and we’ll get you set up and onboarded. Because one of the things Safe Paths is being careful of is not to host any of the data that’s involved in this. We believe that there should be some contracts in trust between patients and the healthcare authorities. And we’re not trying to insert ourselves into that equation that forces patients to think of Safe Paths in a way — why do I have to trust them with my story, my medical information? We have pivoted to a trusted-partner network, where some of the companies who are actually contributing code on a volunteer basis on our open-source project, host them and are willing to do that for some of these smaller Health Services. The bigger health services, unfortunately, which are at the government level, are largely, except for a few, in analysis paralysis; the marketplace is so fragmented right now with different solutions and they may be looking at 20 or 30 different solution providers to solve the same problem, and trying to figure out GPS only, versus Bluetooth. There are big  CRM (Customer Relationship Management) software providers that are looking at bolting on some of these capabilities. I think that whenever we’re talking with the larger government, like a state or larger countries, they have so much on the line constituency-wise and try to make a good decision that they’re not able to move forward very quickly, as quickly as the smaller countries or counties and states are. When it comes to overcoming analysis paralysis, you need something on a small scale which proves that it works. And that is the most powerful thing which gets people out of that mode and trying to make the right decision, without much data, which is what they’re trying to do.

Deepti: Absolutely. I think prototyping is the way to go always. You show the results and then you make it, and then make the adoption happen, and you fail fast. Get something small; if it doesn’t work, you haven’t invested a lot in it. Being able to do that is very critical on how we move forward. But to the second part of the question – Apple/Google will only allow one app to run on their GAEN. What are your thoughts there?

Kyle: They’re worried about fragmentation. So if, let’s say, California as a state, and each county, decided that they want to solve the problem with a slightly different flavor of the app. That would be a real problem. There’s some value in modularity of what’s going on and the sharing of Bluetooth keys and interaction keys between neighboring counties. Some places, like we’re working with Kansas City, they have five or six counties that intersect in the Kansas City Metro area.  Now how valuable is it, if each county has a different app? So that’s what Apple and Google are primarily concerned about. They’re not saying you can’t have one app for multiple states or one app for multiple countries, as long as those apps are going to be able to address the concerns and needs of their healthcare authorities, localization for local languages, etc. They just want to avoid fragmentation in the marketplace where the efficacy of the app, and its tooling, and the ecosystem is impacted because every county has their own flavor of approaching it. Unfortunately, that could mean slowing down adoption, because now you’re asking an entire country or a state to settle on one way of doing it. Right now the debate is still quite active about the best way, and the one way it should be done.

Deepti: Touching on another point, there will soon be tremendous amounts of data which would be collected for the contact tracing, with respect to generating heatmaps and dashboards for instance. So, how do you see that part of it in terms of health authorities, and how Safe Places / Safe Paths is actually going to help that? Also the apps and the data would be judged on the value of their data output to be able to make better decisions. Your thoughts on it?

Kyle: I think that’s a great point. My background with the industry is building infrastructure, software and platform software, where there are a lot of players, solution providers, and tools that can be leveraged. One of the most important qualities of anything you build or pull into your ecosystem is this idea of composability: if you can mix and match capabilities and integrate with the ecosystem. It’s incredibly valuable, because each one solution provider usually has a few things that you know well. From a product standpoint, we want to focus on what we do really well, which is merging this ecosystem of the privacy-first data collection from mobile phone apps, mixing with the interpretation of that processing on the healthcare authority side. If we want to stay focused on that, realizing that we’re not the only piece of software in that ecosystem which the healthcare providers are using. There’s a tremendous tie into various approaches to do the CRM. And we’re hearing different views about different states going different directions with a suite of CRM software. There’s legacy CRM software that’s in the mix. And I think we’d be best served from the standpoint of approaching that, as we want to be able to integrate with any of the solutions that are out there. And the same goes with the dashboarding map analytics. There’s a tremendous amount of attention and energy in that space as well. And being able to feed into those things, in a way that preserves privacy to provide decision-making tools for public health officials. Public health officials are more interested in the decision-making side of the data and the software, and only then maybe the contact-tracing interview side. And that’s an important part of the ecosystem that we look at more from an integration, composability, compatibility, and interoperability standpoint. We’re not going to build the best heatmap dashboard software out there.

Deepti: I believe what you explained above is what we are doing now with Safe Places as I understand. We are also involving healthcare authorities when we are creating these things in terms of dashboards. Can you talk a bit about it?

Kyle: Yeah, I mean for Safe Places right now, the primary functionality is supporting contact tracers, going through a contact-tracing interview and redacting, ensuring that we can create public points of concern, a data set that can be downloaded and used by mobile apps, either with GPS, overlap algorithms on the phone, or Bluetooth. From a Bluetooth standpoint, we talked about a Bluetooth key-server on the Safe Places side. But you’re right, from a public health official standpoint, at a higher level, they are more interested in decision making, where they can get help in generating heatmaps. When we have conversations with healthcare authorities, that includes folks who lead the contact-tracing teams, it goes really well, because what we’re building is compelling for them, and they’re excited about it. When we meet at a state level, when we talk to the healthcare officials, they don’t always go as well because we don’t have that robust dashboard heatmap side of what we’re doing.

Deepti:  How do you see Safe Paths beyond this pandemic? I mean, right now we’re all fighting COVID-19 and toward flattening the curve. But, we all know that Safe Paths has a bigger vision when everything subsides, and when we have the vaccines, etc., how do you see Safe Paths shaping up?

Kyle: I’d answer it in two ways. I think number one, myself, and every volunteer that I’ve had this conversation with, and working closely with, isn’t looking beyond the pandemic. We are all laser-focused on building technology that addresses the pandemic first, foremost. Now, there is a serious question out there still around: will this ever go away? Is this going to be more like the seasonal flu? Is this going to be something that we deal with every flu season? Or maybe it never really goes away. It’s all about minimizing the damage and keeping everything open. That is one side of this. But I think if you take a step back, and if I answer your question from a more holistic point of view, I think one of the key lessons here that will apply to solve other problems is the ability to harness this type of technology and not make that privacy trade-off. I think there’s something valuable in proving that point of how this technology can work and not make that exchange. If I could just add to that too, I think there’s something significant going on here with the volunteer movement. And I don’t have a lot of background in anything like this. My deep career at PayPal sort of sheltered me from some of this, but the amount of work, the amount of momentum, and what’s coming out of a group of volunteers and partnerships is actually amazing. When you step aside and think about what’s going on, and how hard folks are working to put out some of these incredible technology solutions to fight a common enemy in this pandemic, it is really fascinating. I think it’s one of the coolest things and rewarding things about this experience. And I wonder if we’ll look back at some of what’s going on as a model of a path for solving other problems that don’t have maybe an immediate financial incentive attached to them but have an incredible amount of good for society.

Deepti: Yeah. This is exactly what we’ve always seen. You know, this is how humanity comes together whenever there is something like a common enemy such as this one. At the same time, what I have experienced with people, like 1,200 volunteers right now here, putting in so much of hard work, time, energy, resources is also because sometimes I feel that we always had a common value system, something at the back of our minds with the emergence of the Internet, which was like, why do we have to always give away our privacy for the sake of utility? And this solution just came up and sprung up in the COVID-19 crisis which talks about solving immediate problems for sure, but can become a model for something so much bigger. And that’s why we talk about trust as one of the key things that Safe Paths here is trying to build.

Deepti: Great, thanks. On that note, would you have any final thoughts on Safe Paths?

Kyle: It’s a very rewarding experience to work with all the people who are donating their time to make this happen.

Deepti: Brilliant. I echo the same thoughts as you about our team here. It’s wonderful that we are all working towards being a trusted partner for our communities, and also making tools easier for decision making for the people who can actually make the difference, like health authorities. Digital initiatives undoubtedly have the potential to play a crucial role in protecting public health and aiding the work that is being done to contain COVID-19.