Tracing COVID-19 Patients, Are we doing enough?

Amit Dubey and Prof. Triveni Singh IPS

The first case of the Covid-19 coronavirus pandemic in India was reported on 30 January 2020, and since then, in the manner that it’s increasing, it’s becoming a challenge for healthcare stakeholders and law enforcement agencies to cure and control the spread of such virus. The healthcare and law enforcement agencies are searching for new tools and technologies which can help them to manage the system efficiently with limited number of resources. As per Cabinet Secretary’s instructions, all the states are asked to effectively monitor 15 lakh international air passengers who have come to India from January 18 to March 22 and it’s not possible without an effective use of technology. Considering the socio-economic structure of India, finding travel pattern of those people is a tedious task even with an AI based Big data analytics tool. Meanwhile companies like Facebook, Google and Amazon are working closely with World Health Organization to use their data sets to prepare an AI based predictive engine to help various government organization. Indian innovators are also emerging with new kind of tools and apps to help the government bodies at a different scales The government stake holders are successfully using the tools like mobile apps, chatbots, predictive data analytics and logistics management tools to manage the system efficiently. There are almost 15 different types of mobile apps for COVID tracking and patient monitoring, launched by various state government and central ministry. These tools and solutions are developed by discrete bodies and private players and may have technical and legal issues linked to privacy concerns. Moreover, our existing healthcare systems and processes were not designed to handle such drastic situations and we are developing these new tools only to give us a breathing space. It’s like a fire fighting situation or to manage the system somehow till the vaccine is ready. But no one can guarantee that such situation will not occur in the future and that’s why we should build an eco-system which can handle any such pandemic in future also and can ensure minimum spread through humans and objects. Most of these solutions are built on data analysis aa nd this data can be sourced either through a mobile app or from telecom operators. In either of these situations there are clear data privacy concerns as the data will move across multiple stake holders and analysis will be done on a centralized server and that’s why a secure back end system is required with regulated data access to authorized users. We can divide this data analysis process in three parts, Historical data Analysis, Real Time Tracking and Real Time Predictive analysis.

Historical data analysis of infected persons to identify the potentially infected people by analyzing data received from telecom companies or through mobile apps. The data analysis helps to identify the areas where the person has been to like grocery store, food outlet, shopping mall including his office and society , The location data and activity trails can be received from telecom companies and can be analyzed to generate such heat maps. We should start this exercise with the location data and contact list details of already infected people obtained through their Call Data Records. However, we should also include data through hospital OPDs and also testing more and more people at every hospital and nursing homes, who are coming with cold, flu, fever and cough symptoms. However, to ensure the data privacy: The analysis can be done by keeping the data anonymous through a reversible masking algorithm where we can get back the actual identity of a person whenever it’s required and the big data analysis engine would process the data completely in an anonymous way. Real Time Tracking: Real time tracking and data analysis is required to achieve following things. Monitoring the quarantined people if they are violating the guidelines and doing any type of movement. Monitoring any newly infected person and alerting the people who could be a potential target due to the proximity with the infected person in the last 10-15 days. Alerting the relevant bodies, who could have been affected due to his last 15 days activities (e.g. society , office, friends, colleagues etc.) Real Time Predictive analysis: After lock down, regulating people to stop future spread by alerting them through real time proximity analysis The app should tell them if there is any possibility of someone getting an infection due to his/her travel history. The app should guide them in real time if they can go to a particular area or not. The app should alert them if they are in proximity with someone who could be a potential carrier.

The above 3 analysis techniques have limitations and the biggest challenge is of data accuracy and data completeness. The insufficient data can give us an extrapolated result but may not be always correct. Any error of the input data will also be extrapolated and will change the end result to a great extent. At present, there are two types of apps being used by Government bodies – Surveillance Apps and Contact Tracing Apps. Surveillance Apps are installed on the mobile phones of people who have been quarantined. A travel range is set, which is restricted to the size of their houses – When someone crosses the range, the server gets a notification and the person will then receive a call from authorities or will be penalized. There is a possibility that the person may leave his phone at home and then roams outside, so an additional feature is added to give a random voice or video call to the person where in his photo or voice is matched with a system to confirm his presence. Alternatively, the person may be asked to share his photos every few hours which will be geo-tagged and time stamped for location and time verification. The location details in the photos’ metadata will be used to make sure they are where they should be. The audio calling based verification of a person will work in case of feature phone as well as not everyone may have a smart phone and their live location can be obtained through operator. Contact tracing App is another app for predictive analysis to find out the potentially infected person due to his proximity with an already infected and identified person. This is to trace primary and secondary contacts of Covid-19 patients so that the administration can move fast Normally, the data from such applications will be transferred to an integrated platform for further analysis, reporting and alerts. At present, most of these solutions are designed assuming the spread due the movement of a person. However, there are risks of spread due to the movement of an object as well, that’s why, we need a much more holistic analysis and also to use IOT sensors and other source of continuous data inputs to provide a futuristic solution. To understand this we need to identify the vulnerabilities from a birds eye perspective, There are three possibilities for such spread. Location, Person or Object. Here Place is fixed but person and objects could be moving carriers. But actually infected place could also be a variable if it is dependent on other two factors. So there are two variables and third one is deterministic but its status is dynamic and completely based on other two variables. Now the solution should have the provision to alert someone in either of these three situations, If he is about to enter an area /Location where some infected person has visited already or if he is exposed to any such moving Object ( food, goods, groceries ) or if he is actually coming in proximity with a Person who could have a finite probability of infection. Location : Someone visited a place and touched something which was exposed to some infected person before.

Person : Someone is meeting a person, who has met some infected person before this and could be a potential carrier. Object : Someone is picking up an object which could have a probability of earlier exposure. Here the real challenge is to associate an object movement with a person’s movement and that’s where we would need a QR Code based packaged stuff with tracking possibilities. The solutions that are available today are designed to handle only current challenge and will not work for futuristic issues. Now it’s the time to build a technology ecosystem, where we can fight easily with any such infection or virus in future also which is not possible by merely tracking the people. We may need a unique identity and tracking feasibility of each and every object which could be a carrier. be it human, animal or machines.

DISCLAIMER : (Prof. Triveni Singh IPS, SP Cyber Crime, UP and Amit Dubey are renowned authors and cyber-security expert. Views are personal.)