PathCheck: Crowdsourced Analytics

Problem: Can we fill the gaps in public health knowledge by crowdsourcing data from citizens and providing them with personalized engagement to improve the pandemic response?  Planning without rich data, or citizen engagement leads to inefficiencies, lives lost, and socio-economic costs. The unfortunate crisis in India in April-May 2021 is an example: it is impossible to coordinate in a top-down manner because health authorities cannot estimate (or nudge) the activities of millions of citizens, the resources they need, and the dynamics of the variants (symptom trajectories. impacted age groups).

Our solution is a  novel ‘NoPeek’ open-source computational privacy technique that helps capture crowdsourced health information, analyzes it for public and precision health, and engages users via personalized recommendations. 

Vulnerable populations: Our innovations include cryptographically secure paper-based QR codes for populations without smartphones. Further, for those concerned about surveillance, we avoid recording raw sensitive personal information.

Our goal is to fill the gaps in public health knowledge by crowdsourcing data from citizens to improve pandemic responses. 

In a pandemic, data is one of the most important tools for policymakers, public health officials, and health system agents. Citizens navigate exposure alerts, symptoms, testing, treatment, and vaccination, but much of this journey is not shared with public health and remains invisible to the health system.

These gaps in data gathering, especially about activities, contacts, and status of citizens are one of the key challenges to understand the health crisis and provide an effective response. The data-scarce planning leads to inefficiencies, lives lost, and socio-economic costs. 

Technological app-based solutions can be used to fill in gaps, but they currently suffer from two key problems that make citizens uncooperative and disengaged: 1) a lack of incentives for users to adopt and engage with the app and 2) citizen’s privacy concerns and the fear of the government becoming a surveillance state.

So how can we achieve crowdsourcing, privacy, and personalized engagement in a pandemic? And how can we simultaneously provide planning tools for public health by making citizen data available in real-time?

We create open-source crowdsourcing and engagement software that governments and campuses can deploy for their citizens. Thereby, empowering them to reduce the gap in public health knowledge.

We use the “NoPeek” privacy approach, partially developed at MIT. The NoPeek approach is one step ahead of consent-based or anonymized approaches because it removes the need to share any raw data with servers.

Our technology stack uses the “Capture, Analyze, and Engage” loop.

Capture: (i) background (age, health history, ...) (ii) status (symptoms/tested/vaccinated, non-user data: mobility data, previous disease data, vaccination trends, social media data) and (iii) Activity: (sensor trail e.g. GPS/BlueTooth trail, ..)  All using NoPeek protocols to preserve privacy.

Analyze: We use federated learning modules without looking at raw, individual citizen data and help form foresight based on our predictive models. Further, in hindsight, we understand the disease dynamics as well as provide insights on the current snapshot of the pandemic.  

Engage: This module delivers personalized Information  (risk score, cases nearby, precaution, test or vaccine credentials, creates nudges, supports gamification, incentivizes safe actions, and makes recommendations specific to each citizen. With NoPeek, the server never discovers what message was shown to any citizen.

Our crowdsourcing exposure notification application and crowdsourcing solution in India.

We have 3 audiences:

1. Public Health Agency (PHA) planners at both the country and regional level

For PHAs, we create a data-rich stream of citizen activity to fill the gaps and complement other data-gathering efforts to understand case rates, the severity of symptoms, equitable deployment, and effectiveness for vulnerable populations.   

2. Organizations with large campuses

For organizations with campuses (i.e. universities or large employers), the decision-makers usually take the lead in adopting innovation and expensive solutions to ensure safety, encouraging adherence to procedure (e.g. frequent testing), maintaining continued operations, and managing incentive programs (e.g. automatic sick leave). A significant frustration amongst these leaders is waiting for city or state solutions. Our solution is an easily deployable toolkit that enables intelligent forms of data capture and also has a way to make secondary use of signals people share every day for understanding the pandemic better.

3. Citizens (including vulnerable populations)

For citizens, early exposure alerts, personalized risk scores for activities based on their health conditions, and visualizing local pandemic spread are critical. 

For vulnerable populations, our solution works for citizens without a smartphone. We developed a paper-based QR code with cryptographically tamper-evident digital signatures for citizens to engage in the test-trace-treat-vaccinate ecosystem without disclosing any personally identifiable information.

Our new crowdsourcing solution in India. We are also making secondary use of signals like proprietary search trends of online pharmaceutical companies and mobility data, which are involuntary and unbiased signals extracted from individuals' day-to-day activities. With the help of these diverse data captured in a privacy-preserving way, we are able to provide 

(i) Hindsight - Understanding the disease dynamics and effects in India

(ii) Insight - Know the current situation of the pandemic in the regions (at the granularity of state/cities)

(iii) Foresight - Predict upcoming hotspots, effects of variants and different health policies adopted.

The currently deployed solution had peak users as high as ~25 million during the peak of the COVID-19 crisis in India and was endorsed by various government and health bodies.

1. Respond to emerging unusual symptoms and identify approximate heatmaps of the location of sick individuals. 

2. Respond to spread (encourage citizens to test after exposure, automated alerts to close contacts after positive tests) to break the chain of infections.

3. Respond to lack of early knowledge about treatment plans (we gather treatment and outcome info directly from citizens and crowdsourced via doctors).

4. Respond to vaccination needs.

5. Respond to the needs of vulnerable populations with paper credentials (monitor equitable deployments).

6. Respond to socio-economic pressures by making it highly affordable for the government and campuses.

We chose Growth because we have great momentum with our initial open-source offering for contact tracing and exposure notification.  This was developed in less than 6 months and deployed in Minnesota, Alabama, Hawaii, Louisiana, Guam, and Cyprus, as their official app. This app has been downloaded over 2 million times and is actively promoted by government leaders.  

We are a 1-year old nonprofit with a well-recognized name and proven ability to generate revenue and secure donations.

We are now expanding the open-source product to meet more needs.  This new part is in the Pilot stage. We have the initial software built but expanding the NoPeek privacy-preserving mechanism for these additional modules is more complex. 

Financial support from the Health Securities challenge enables us to further broaden our open-source software toolkit.

Importantly, as mentioned above, we already have 4 pilots identified: Nigeria, Brazil, India (already launched), and United States (Alabama).

Our strategy is similar to what some Asian countries used to combat the pandemic but with non-invasive data-rich decision making, and engagement. While most public health solutions try to create tools for PHAs and a top-down ‘command and control’ solution, we take the bottom-up approach. We solve the problems in technology, trust, and open innovation. We are inspired by the foundations of Mozilla, one of the few large-scale consumer-facing open-source software projects focused on privacy.

Our effort is innovative and unique for three reasons. 

Tech: Unlike other organizations that use consent and anonymity as a poor substitute for privacy, we use the NoPeek approach developed at MIT. The set of NoPeek algorithms are computationally hard to breach and are being rigorously reviewed with mathematical proofs for cryptographic guarantees. We believe consent and anonymity-based approaches are prone to data security breaches. Governments are unwilling to let private for-profit or non-profit players deploy a solution for their citizens with such security risks.

Trust: Unlike commercial solutions with vendor lock, sensitive health data and guidance requires transparent, unbiased, open-source solutions. 

Open Innovation: Similar to Mozilla foundation, our open-source software is an open and transparent innovation platform drawing talent and partnerships for growth.

Our Exposure Notification open source software has been downloaded over 5m times, our covid dashboard with the University of Alabama visited by over 100,000, 3,000 people from government, nonprofits, commercial vendors, and students. have attended our educational and networking webinars

For year one of this proposed solution, we will be in the pilot phase with 5-10,000 growing to over 50 million in year 5.

We deliver an open-source software toolkit for governments and campuses that is deployed by regional health-IT companies. We cannot influence in a quantifiable way the ultimate impact on citizens (i.e. saved lives) since adoption and integration with the public health ecosystem is handled by locals.  But, we can help with great technology.  Here are our goals and success metrics.

• Goal 1: Secure pilots commitments
  • Metric: 4 government entities agreeing to pilot (partially achieved)
• Goal 2: validate technology (NoPeek, crowdsourcing and other technology)  
  • Metric: recommendations and privacy validated by third parties in pilot with 100 users
• Goal 3: deploy successful pilot
  • Metric: success criteria pre-established for each pilot and met after 4 months.  Pilots rolled out to at least 50% of the citizens at the client campus with 20% monthly active users
• Goal 4: new clients
  • Metric: 10 large and 10 small campuses signed by year three
• Our internal metrics are:
  • Thought leadership (Monthly mentions in news media, subscribers to our social media channels)
  • Creating a worldwide network of health-IT companies: (numbers of companies using our software versus the number of companies attending our webinars)
  • Creating a pool of talented professional volunteers: (monthly active users on Slack)

We need 15 paid staff and contractors (mostly developers).  300 volunteers contribute to the open-source software and some use their own research grants to work part-time. 

Also, we are collaborating with various groups like coronasurveys, DeepCOVID, and more.

Full team: https://www.pathcheck.org/en/a...

Our team and advisors have deep expertise in public health, medicine, epidemiology/infectious disease, software development, bioinformatics, human-centered design experts, ethics and behavior scientists, policy experts, and geneticists. 

IP and innovations (see work above on NoPeek)

We have a formal contract with MIT for shared resources and technology transfer.

Our volunteer base is international and they strongly guide the needs, design, messaging, and prioritization of the solutions in those regions.

Prof. Raskar sits on the steering committee of the NSF PREPARE program which is focused on pandemic planning and resilience. Prof. Raskar testified as an expert witness for the US Congressional hearing on Contact Tracing and Exposure Notification. The project received seed grants from NSF and Schmidt Futures Foundation. We participate in all standardization efforts.

PathCheck is fully committed to diversity, equity, and inclusion in our leadership team and our organization. While we are a young organization (less than a year old), diversity, equity, and inclusion is infused into our culture and practices.

Specifically

• Our project teams comprise members of many nationalities, geographies, races, genders, and socioeconomic backgrounds
• For decision making at the leadership level, we openly question and discuss our unconscious biases, discrimination, racism, and organizational structures that may support these items
• Our research looks at underserved communities of the pandemic and the pricing of our services around open source software is sensitive to this news of lower and middle-income governmental agencies.

The prestige of winning the SOLVE Challenge strengthens our initiatives and helps with the barriers (financial, legal, and educational) discussed above.   Specifically

An MIT SOLVE Grant helps us generate matching donations.

Working with partnering SOLVE members helps government leaders and regional health IT companies gain confidence in our NoPeek privacy and open-source crowdsourcing software.

SOLVE members, MIT faculty and other interconnected parties can help us create education and outreach programs for government leaders as we explain the legal aspects of privacy, safety, and efficiency. We hope that many SOLVE member educational institutes will run workshops about the benefits of NoPeek and crowdsourcing.

Getting expert technical support from Google/Facebook/CubeIQ/Palantir and other tech members will ensure even higher quality software.

In our experience working with governments around the world with our Exposure Notification solution, we learned that financial and PR were the biggest hurdles. Governments, especially in lower and middle-income countries are budget-constrained, even in a pandemic.  To make the biggest impact in saving lives and reducing strain on medical systems, action needs to occur quickly. Financial resources to help fast track response problems with associated positive PR  gives government officials the support they need to act quickly

We are very impressed with the SOLVE partners and are specifically interested in working with the following to bring additional credibility and confidence to our 4 pilots.

• The Gates Foundation – We have had initial discussions with Himanshu Nagpal and CK Cheruvettolil for funding to support our effort. Gates Foundation funding brings funds and credibility.
• McKinsey would help with additional relationships at the government level. Their guidance on risk management of these 4 rollouts and then post-deployment case studies would be invaluable
• Google and Facebook for technical support and possibly mobility and symptoms data they have been releasing in this pandemic.
• CubeIQ and Palantir for telecom data and vaccination data, e.g. Palantir have managed vaccine rollouts in the US with software like Tiberius.

Several other SOLVE members can help support our 4 pilots in rollout marketing, adoption, and program design

The pandemic has disproportionately impacted low-income and vulnerable communities. A patchwork approach of public health interventions fails to address the problem because of the gap in understanding the health and behavior data of these communities. Efforts to invasively monitor at-risk communities and enforce compliance will lead to a surveillance state.  

Our solution at PathCheck takes a bottom-up approach to empower individuals. Vulnerable populations have low trust in consent or anonymization-based monitoring methods. Similar to Apple Maps, we use a NoPeek privacy approach which eliminates the risk of socio-economic profiling. It prevents misuse of data or data breaches because no raw data is ever shared from the user’s smartphone with anybody else, not even PathCheck. 

Similar to navigation map apps where a small number of drivers can indicate traffic problems, we need only 5-15% adoption to harness data and provide statistically meaningful insights to the public health authorities. Vulnerable populations can use this ‘Apple Maps for Health’: participate in crowdsourcing without giving up privacy, remain anonymous, get personalized guidance that is free of socio-economic bias, stay safe with a pandemic ‘radar’, and navigate the pandemic journey without the fear of interacting with law enforcement.

For the population without smartphones, we use a paper-card approach with QR codes that are cryptographically secure and tamperproof. We also support SMS. 

We have already deployed free, open-source solutions used by millions. The Prize will allow us to convert a pandemic response solution into a broader wellness solution that especially supports vulnerable populations.

The pandemic has disproportionately impacted low-income and vulnerable communities like refugees. A patchwork approach of public health interventions fails to address the problem because of the gap in understanding the health and behavior data of these communities. Efforts to invasively monitor at-risk communities and enforce compliance will lead to a surveillance state. Refugees often prefer anonymity and maybe refused access to services withut sharing name, date of birth and other sensitive information.

Public health interventions require a benevolent all-seeing all-knowing AI that acts as an impartial, honest broker. Today’s public health efforts are hobbled by the lack of access to data and the inability to engage individual citizens. The health surveillance and public messaging mechanisms are incomplete. Our solution at PathCheck is addressing these two main problems: how to fill the gaps in knowledge by crowdsourcing data while maintaining privacy and how to provide personalized meaningful nudges without creating a big brother or coercive system. 

To build AI for humanity’s benefit, we have made tremendous progress in all three areas: Capture, Analyze, and Engage. To capture data and engage citizens, we have data from our apps as well as external data through proprietary search trends of online pharmaceutical companies and mobility data, which are involuntary signals and unbiased signals extracted from individuals' day-to-day activities.

To analyze, our team has built several deep learning and analytics tools, e.g., 

1. Predictive modeling: Estimate the emerging hotspots, disease dynamics, effective resource allocation, identify the spread of variants and symptoms across demographics, risk stratification, and more (Glorioso et al., 2021; Patwa et al., 2021; Sukumaran et al., 2021).

2. Agent-based modeling for understanding the implication of different health policies (Romero-Brufau et al., 2021). Test hypotheses in Non-Pharmaceutical and Pharmaceutical interventions to devise effective response measures.

We are well positioned to use the Prize as our solutions are already deployed in 7 jurisdictions and have millions of users, we can ethically mine this data with NoPeek privacy.

For hospitals and citizens who want a decentralized platform to know where pandemic-related resources are located, and share availability about hospital beds, oxygen, food, and other resources, the PathCheck India Karuna system is a web and mobile platform that allows for constructive interaction between hospital administration, doctors and the population while remaining easy to use for all parties. 

Crowdsourcing using integrated surveys, trackers with real-time positioning of hospital resources available, vaccination slot notification portals, and broadcasting communication channels allow for a reliable and accurate optimization of resource allocation, where civilians get real-time updates on the verified availability and location of resources. With the Global Fund prize, we will be able to pilot the application as an official platform in collaboration with the Ministry of Health in major cities in India such as Bangalore and Mumbai, as well as maintain the system to allow for a larger user base. As the demand for the application continues to grow during the COVID-19 pandemic, the Global Fund prize will help us allocate funds towards the maintenance, expansion of the application to other countries, and integration of more features for optimizing on-ground delivery of resources.

This app not only helps fight against misinformation surrounding supply allocation, but it also builds a community where citizens help each other in preparing, predicting, alerting, and responding to pandemics.

Our live app can be found on karuna.pathcheck.org