CoVis: real-time disease risk assessment

CoVis brings real-time disease risk assessment to individuals and organizations starting with location-based risk assessment for COVID-19. The CoVis team will build data-driven solutions augmented with machine learning (ML) to track the progress of existing and novel diseases at a global scale.

Ayan Paul, PhD. Chief Scientific Officer and Principal Investigator for the CoVis Project funded by DESY, Hamburg.

One of the biggest challenge during the COVID-19 pandemic has been the delivery of pertinent information to individuals so that they can make informed choices. There have been a lot of uncertainties surrounding the level of risk the disease poses and what actions can be considered safe so that only the most necessary actions can be prioritized. While information is available from several sources, there has been a lack of centralization of all relevant data and their interpretation in a manner that individuals can use to assess their real-time risk from COVID-19. This is a problem that has affected the entire global population. The uncertainties surrounding the risk one faces from COVID-19 often leads to excessive strain in one's life. While directives are issued by local institutions, it is difficult to make informed choices by blindly following these directives which have been changing quite rapidly over the span of the pandemic. A quantitative measure that takes into account the real-time actions of a person (locations visited, protective measure followed) has not been developed for individuals to use from a device as accessible as their mobile device. Moreover, disease risk assessment at a global scale is still an open challenge.

The CoVis solution has two classes of audiences:

1. Individuals will have a mobile-based app to interact with which will update them with personalized information on the risk they face from diseases.

2. Organizations will have access to the CoVis datastreams through APIs and customized UIs

We have conducted tests with users (over 400 individuals) to design the mobile app and determine what functionality works best. Tests have been focused on UI design, UI functionality, feature selection and consumer needs.

To understand what needs organizations have in estimating disease risks we have spoken with companies like Deutsche Bank, Germany, PwC, Hoyer, Techniker Krankenkasse, Moogsoft, Management Sciences for Health, Allways Health Partners etc. We got quite useful feedback from several of these companies pertaining to how we should engineer our product to best serve their employee and client base.

The ongoing pandemic has taught us that neither private nor public enterprise can flourish without technology being effectively used to avoid natural crises like this pandemic or climate change or the disbalance of the natural order in the ecosystem. Hence, global collective action is necessary and we want to appeal to organizations to help us address disease risk management collectively.

The basic version of the CoVis app will be made available for free. It will provide real-time risk assessment and risk score trends with which the users can understand how their behaviour affects their risk score. Almost all functionality of the app is available for free. This will allow the user to get all the relevant information and make informed decisions.

CoVis will gradually ramp up to disease tracking on a global scale. This will include forecasting of disease and early-warning system for epidemics. This data, when combined with existing and emerging medical knowledge, can be used for reducing the human and economic cost of an epidemic by providing actionable information that can help in building local and national policies.

Informing users of their risk from a communicable disease will leave a positive impact on their lives by changing their behaviour in a manner in which they will expose themselves to lesser risks. Collectively this can slow down the spread of the disease and also potentially lower the cost of the pandemic in terms of lives lost. The information provided in the CoVis app is sourced solely from reliable sources and processed with a mathematically sound framework. Moreover, we put user data privacy as one of our top priorities in designing and developing our solution. These can help generate trust in the information and guide users to safer behaviour.

The CoVis app is built to operate at scale. The AWS backend is designed and built to support millions of users or more. The app will be made available in several countries in the near future. As CoVis moves to monitoring other diseases these will be made available in the app. In addition, the information stored in the CoVis backend (about disease prevalence) will be made available through the CoVis API. Covis Inc will not share or sell individual user data and limit its collection to the bare minimum. Moving to low and middle-income countries will require the development of alternative technologies for information dissemination like SMS based services. These will be localized based on existing infrastructure in each country. 

We measure success with two parameters. 

• the number of users that will use the CoVis app
• the number of users that will convert to using the premium version of the app.

The metric that can be used for the effectiveness of the solution is to see whether the distribution of risk score amongst the users reduces in a certain area adjusted for the disease progression in that area. This would indicate that the users of the CoVis app are changing their behaviour in a positive manner. We also plan to conduct dictionary surveys amongst users who will maintain a daily journal of how they feel the app is affecting their lives.

The primary barrier in implementing the CoVis solution is funding and workforce. While we have kept the running costs of the CoVis app extremely low with in-house innovations, data mining at the scale we are targeting requires a skilled and dedicated data science team along with a team of experienced software developers to maintain the algorithms that will process the data and to build the frontend UI for the consumers. In addition, we will need to hire ML/AI researchers and experts in medicine and life sciences.

CoVis also needs a team of consultants who will tailor services for government and private organizations based on the data processed by CoVis. Since CoVis plans to serve several countries, an international and culturally aware team is necessary as the perception of disease risk and the requirements for alleviating it varies a lot from region to region. Especially, operating in low and middle-income countries will require the support of local and international NGOs and national governments as solutions specific to each country need to be built.

Covis Inc. is classified as a DESY, Hamburg spin-off since much of the algorithm for the CoVis app was developed at DESY with DESY funding. All IP owned by DESY is licensed to Covis Inc. exclusively with very minimal restrictions and costs. DESY has no commercial interest in Covis Inc.

The biggest support we need now is support from a community that believes in our solution and is willing to promote it by getting the word out to the world and with funding. The Trinity Challenge focuses on building long-term solutions so that a healthcare crisis of this magnitude is avoided in the future. This is the sole goal of Covis Inc. taking the route of disease risk predictions and monitoring. The CoVis app is just the beginning for us. To go any further and to scale to global risk assessment we need funding and support from like-minded groups and the Trinity Challenge provides us with that platform. Our goal is to win the challenge and connect with institutions and organizations that will believe in our vision.

We are also looking forward to other modes of support from the Trinity Challenge including collaborations with partners and other teams and maybe sharing of useful healthcare data like those available in the open database. We also look forward to contributing with our expertise to open science.

Data: With vast amounts of mobility data related to locations and populations, Google can certainly help us in developing solutions for diseases that spread through contact and proximity like COVID-19. We are using data from Google in the CoVis app. For example, access to the popular times data collected by Google would allow us to very accurately predict risk at any location.

AI/ML: We can benefit greatly with AI/ML knowledge from giants like Facebook, Microsoft, Google and through academic collaborations with university partners in the USA, the UK, Singapore, Australia and China.

Human behaviour: In order to build a comprehensive risk assessment platform, we will need to understand human behaviour and their response to disease spread. The Behavioural Insights Team can help us with that.

Computational resources: We can also benefit from contributions in terms of computational resources from academic institutions or private organizations. Help with cloud services from Google and Microsoft will be highly appreciated.