PathCheck: Filling missing pieces for (future) pandemic response

‘Waze’ App for the Pandemic Journey

This is an edited transcript of the TEDxMIT Talk by PathCheck Foundation Chairman & Founder, Ramesh Raskar, Ph.D. The original video presentation can be viewed here: Waze for our Pandemic Journey | Ramesh Raskar | TEDxMIT

We continue to live in the fog of the pandemic. And for our journey through this pandemic, every day we wait for a beacon of trust and information from public health. But maybe we got this all wrong. Maybe the solution is right here amongst us, you know, despite billions of dollars and scientific achievements in vaccines, we seem to have missed one critical element in the pandemic response and that's citizen engagement. How do we get the information from all of us to public health and how do we get personalized, trusted guidance to each of us now? Who here uses traffic navigation apps like Google Maps or Waze and so on? 

It's kind of amazing, right? I mean, of course, the way this works is we willingly share our GPS data with Waze; Waze, analyzes traffic density and calculates potential hotspots and trouble areas, and gives us highly personalized information of how we avoid those parts to get from point A to point B. So this is a great example of self orchestration.

 

 

Learning from Asian countries without slipping into a surveillance state - So can we have an app that looks like Waze for the pandemic to go through this haze of confusing data? In some of the Asian countries, they were able to control their pandemic in less than three months by techniques that unfortunately involve some invasive surveillance and coercive social disciplines. Now, we need a slightly different solution to do the same thing in democratic countries. So my team at MIT, in February of 2020, started working with a lot of players with CDC and WHO and we proposed a privacy-preserving solution and launched our first app. There was a lot of excitement. So with that, we created a non-profit organization, PathCheck Foundation. We spun out of MIT, raised a few million dollars very quickly. And building the world's largest nonprofit open-source project for COVID19 solution. Very soon it became the official app in five US states and territories, but there's so much more we need to do to create a Waze for a pandemic response.

Bad Data Make Bad Public Health Interventions - Now consider Mike an essential worker in Boston, as he goes through his pandemic journey. Most of his interactions with the society, his treatments, when he gets vaccinated, remains invisible or at least disconnected from public health gene under the side, who's in public health and making very decisions about predicting spread or, when lockdown, the efficacy of treatments and vaccines, and whether all the operations are equitable, he's flying in the dark.

And this is the challenge we have been facing in this pandemic. Poor data that's available to public health leads to decisions and policies that often don't play out. And when all of us see those policies, not having a reasonable impact and being disconnected from reality, we start losing trust and there's less adherence. And of course, there's a spectrum of folks who would respond to these incorrect policies and their level of adherence. So is there hope when there's a missing citizen engagement, the classic loop of interventions predictions and providing alerts completely falls apart? Because when we say evidence-based, it's often not evidence-based and the trust that comes out of these evidence-based guidelines is completely missing. 

 

Apps can fill the gaps  -Apps can fill the data gaps by using clever mechanisms for capturing data, analyzing it with global AI, and engaging it. And if you can analyze the stream of data coming from citizen apps and then engage citizens with highly personalized guidance, actually we can go very far. And that's exactly what Waze does. Not only Waze, but the trillions of dollars of value in big tech companies is also possible because of one simple formula, they figure out a way to have very high consumer participation, and then they encourage us to share more data and they get the fly yield of the capture, analyze and engage with citizen participation. If the big tech can do this and along the way, make trillions of dollars, why can't we use the same idea for such an important problem as public health and pandemic response?

 

 

Adoption factors

Waze

Early Pandemic Apps

Privacy

Company promise

Govt Branding

Ease of sharing

Passive use

Disconnected apps

Tangible Benefits

Traffic ahead

Did not help

public health

Reliable

1% adoption

Urban area

For Waze, the mantra is “share data, get benefits”, but for all these crowdsourcing apps, there are four key criteria for grat adoption: privacy, ease of sharing, tangible benefits, and reliability. I mean, it's as open as just opening the app. There's nothing else you have to do. There's a tangible benefit, you know, immediately where the traffic density is, and it's very reliable. You know, if it doesn't matter which car you're using, whether you're rich or poor.  Waze will give you information that has high fidelity without any bias. And more importantly, these apps work, even if there's a very tiny fraction of the population using it. Even if one percent of the population on the streets is using the app. It's enough to understand that if a vehicle goes through an intersection very smoothly, that there's not much traffic on the other hand, even if one or two vehicles start slowing down compared to the traditional data for that intersection, we can start seeing some oranges and reds in that region. So this is very critical. Even 1% of participation can change the flow. 

The funding paradox: billions for vaccines, zero for apps

If you just focus on vaccine deployments, look at the maze of the vaccination journey. We spent billions for the science, but we provided no funding to figure out the citizen engagement.

And that's why we saw these heart-wrenching scenes of elderly people lining up for hours over multiple days to get shots and a lack of engagement for those who might be doubting or have a lack of information on their own vaccination status. And of course, we have no clue what the side effects are because the best systems we have in our country only have a few tens of thousands of entries when over 200 billion people have been vaccinated. So the tracking of side effects is missing disappointingly, the $65 billion plan that has been announced for the future pandemic response also misses the most critical element, citizen engagement, and citizen apps.

Patchwork of Apps: Disconnected, Disengaged

Now there are, of course, certain apps that appeared in the beginning, you know, for symptom checking, contact tracing, quarantine paths, vaccine credentials. And so on the challenge with these apps is that if you use the same four criteria as before, it's not clear if they got it right. The strategy for apps is just like Waze would be to share health data and get benefits, right? But for privacy, we ended up using branding as a government app for ease of sharing. Most of these systems for testing vaccination, vaccine credentials, and so on are all disconnected.

It is not clear that tangible benefits because some of these apps, although their privacy-preserving, did not stream any meaningful data to public health, so they can create a bird's eye view of what's going on. And in terms of reliability and bias, we have great challenges as well because some of the solutions only work in the areas that have good smartphones or in urban areas that they're slightly more powerful. What we need here is a single app for the whole pandemic journey, a kind of an app that you would use to request a Uber ride or click a button on Amazon, and automatically something shows up on your doorstep, and getting that right is not easy. You probably have two questions. What about adoption? When there are millions of other apps what about privacy?

Privacy Tech, Adoption and Fragmentation

You might say GPS data is low stake information. So Waze can do a very good job of crowdsourcing and providing us benefits, but health data is sensitive. And so we need centralization and that has the complexity of consent and anonymity, and you're absolutely right. Fortunately, our group at MIT and many organizations and companies like Apple focus on decentralized solutions with a technique called no peak privacy. So if you use apple maps, nobody not even apple knows your starting point, your ending point, and your route. And despite that, it's possible to calculate the traffic density and do aggregation. These new types of Computational algorithms are so powerful that now we can start doing crowdsourcing while maintaining privacy.

 

Another question is adoption. As I said earlier, all we need is a fraction of the population in some of our calculations, less than 1% of the population needs to start using this type of apps. So that gene in the public health office can start making some meaningful decisions. As I told you earlier, for the adverse reaction and side effects reporting, the numbers have been abysmal. Unfortunately, if you look at a lot of clinical data, including data for vaccines, most of the decisions are based on participants that are less than final outputs. That's based on results from less than a thousand people. So having 1% of 200 million is actually a luxury.

There are so many apps that are going to stay disconnected, what about all the inevitable fragmentation? The analogy here is the weather app, the national weather service, you know, kind of creates databases and all of us use different, you know, weather apps and, you know, hurricane and alert based apps. It doesn't really matter because they create a backbone and that's what creates this synergy. 

What are features of the app for pandemic journey?

Here is one created by our foundation PathCheck for pandemic journeys that need to have many pieces. And it's going to take us some time to get it going, but here's a quick teaser of this app that you can check on our website https://prep.pathcheck.org . To capture data from Mike, he has symptom diaries, his tests, his medication, his vaccination, his proximity of social activity, some sensor data, and even some queries that get streamed to public health in a privacy preserving way.

 So that Jane in public health can take very critical decisions. Of course, she can use other data sets as well, but the data that's coming from all of us is gonna be much more or meaningful. And then finally she can start creating predictions and alerts and send them in a highly personalized way, as opposed to each of us checking the news or talking to our friends or calling up our doctors and try to see what's the right thing for me to do. She can push the data into the system so that Mike can get extremely personalized information about early warning and risks and highly personalized nudges for behavior changes, symptom prediction, to see if he does get sick, or if he's feeling better, which way is going to proceed. And any credentials that can keep him in the loop.

Getting ready for self-orchestration for the next threat

And if you talk to the experts, this pandemic, unfortunately, is just a rehearsal for all kinds of future threats, including bio-engineer threats, and to get these apps to be optimized with a good user experience. And that points to citizen behavior are gonna be challenging. Just think about eCommerce or your favorite apps and how long it has taken to get them right. So we need to start working on these apps. Now, the beauty here is that a similar crowdsourcing strategy is also good for slow-moving threats. That could be an existing public health crisis, like diabetes, obesity, domestic violence, and so on. And we can do a drill or rehearsals, or we can deploy the systems on some of these, let's say low stakes and slow-moving threats for public health so that we are ready when we have to and come in the next public health crisis, which could be very fast-moving. So to deal with the haze of this pandemic, we need a Waze like app and only apps can fill the gaps in the data for this public health. And when we stream that data, we can finally look at this beacon that can analyze and beam to us trusted, highly personalized information. Thank you.

Contributors to the article - Ramesh Raskar, Mohammed Rizin, and Rohan Sukumaran