Combatting COVID-19 with a city clone
Predicting pandemic patterns using digital twin technology to map the virus and determine interventions.
4
MINS READ
Pune is a bustling city in western India, close to one of the largest cities in the country, Mumbai. With a population density of 5,600 people per square kilometer (15,000 per square mile) and slums emerging as hotspots during the onset of the COVID-19 pandemic in 2020, Pune’s city administrators needed a smart and effective solution to tackle the developing situation.
To address the challenge, TCS Research collaborated with a Pune-based NGO, Prayas Health Group (PHG), to address the situation using Enterprise Digital Twin (EDT) technology.
The unpredictable spread of COVID called for a solution that could be effective at the societal level.
It also required designing and implementing region-specific interventions for effective damage control. Given the circumstances and limited resources at hand during the onset of the pandemic, EDT emerged as a possible solution for tackling the situation.
The EDT approach facilitates the simulation of what-if and if-what scenarios (analyzing possible outcomes of an event using conditions set by changing the values of specific parameters and measuring their impact) for insights-driven decision-making, using modeling and simulation (M&S) and artificial intelligence (AI). The technology has been successfully implemented in business and cyber-physical systems for industries like telecommunications and manufacturing.
“The digital twin solution helped model all the relevant factors—virus characteristics, demographic heterogeneity, and mobility patterns, among others—that influenced the spread in a city. This city digital twin served as an ‘in-silico’ experimentation to explore effective interventions without compromising public health safety,” says Souvik Barat, Principal Scientist at TCS Research.
The application of EDT proved useful not only in predicting the spread of the disease, but also in understanding potential outcomes of the strategies (if implemented).
The digital twin that was developed replicated Pune and its jurisdictions.
The simulation digitally represented the local population, along with information such as demographic heterogeneity, mobility, viruses and their lineages, vaccines, and non-pharmacological interventions (NPIs) as ‘agents’.
By incorporating various NPIs, the simulation allowed for comparative analyses over a period of time, as well as prompts for considering appropriate measures in a specific area of the city. Together, these helped to predict the spread of the virus and the possible load on the healthcare infrastructure for new variants, vaccine roll-out strategies, and strategies for the steps needed to relax pandemic protocols.
The simulation was compared with real-life data for the same period. The validated digital twin was then used to check the effectiveness of interventions through simulations using what-if scenarios.
The simulator represents a 1.5 × 1.5 km2 block in a municipality ward. Blue dots denote susceptible citizens while magenta indicates contact and the possibility of exposure. Those likely to get infected are indicated in orange or red dots, depending on the severity of symptoms. This is based on the duration and frequency of proximal contacts, age, gender, and comorbidities.
Below is a macro view representation of Pune in terms of infection:
The simulation generates the required data for computing key performance indicators:
Based on simulations with different NPIs and possible future variants, here are some key predictions and insights that aided Pune administrators in effectively managing and curbing the spread of the virus:
Ward-wise pandemic progression: A holistic picture of the demand for ICUs, ventilators, and oxygen beds in a specific locality was estimated.
Efficacy of NPIs: Predictions suggested that if 80% of people in a locality wore masks, it would reduce the critical caseload in hospitals by 25% during peak months.
Proactive testing: It was revealed that by doubling the testing rate, the hospital load could be reduced by 10% during peak months.
Impact of lockdowns: Comparative analyses showed that opening offices could lead to higher, active hospitalized cases, compared to the opening of shops; weekend lockdowns could lead to overcrowding of public places during the week.
The overall analysis cautioned that a significant section of the Pune population was susceptible to infection if appropriate precautions weren’t taken. It also prescribed that gradual lifting of the lockdown, in addition to the strict compliance to safety standards, would help mitigate the burden on healthcare providers.
The efficacy of the Pune digital twin was visible as the pandemic grew in intensity.
The number of deaths, infections, and load on hospitals in the months of September and October 2020 unfolded in the manner the twin predicted in August 2020.
“In addition to predicting the second wave in advance, all our predictions about the load on medical infrastructure, in terms of critical care and deaths, were very close to reality from June 2020 to April 2021,” Souvik highlights.
“The unique potential of the digital twin to run ‘what-if’ experiments about movement restrictions, behavioral interventions, and testing strategies could aid the decision-making process. The same was shared with the local authorities who planned the interventions,” said Dr. Vinay Kulkarni, Founder Trustee, Prayas Health Group.
In March 2021, a report based on the predictions and jointly prepared by TCS and PHG was shared with the Pune district administration to help formulate a response to rising cases in the city. Based on the forecast, the report suggested implementing restrictions in public places to further restrict the spread.
“In Friday’s meeting chaired by deputy chief minister (of Maharashtra) Ajit Pawar, much credence was put on the findings and recommendation (from) the TCS-Prayas report,” a news report said, adding that “the district administration announced restrictions on operations of restaurant and bars, public gardens, schools, and colleges, as well as function halls”.
“The ability of digital twin to visualize real-life complexities and capture dynamically changing parameters at a granular level was something of great value in predicting how the epidemic will evolve in the city and the role of different interventions in curbing the spread of the virus,” says Dr. Ritu Parchure, Senior Research Fellow at Prayas Health Group.
The Pune digital twin is an example of how collaboration can address complex problems businesses and society face.
Besides the spread of COVID, it can be used for other issues cities face, such as alleviating traffic congestion.
Pune is one of the most congested cities in India, with commuters regularly spending long wait times in traffic. Though the levels have come down due to pandemic-related factors, the challenge persists and is likely to escalate in the future. In 2018, a media report highlighted that Pune has more vehicles registered than the human population in the city, “a first for any urban area in the country.” Using mobility data, EDT could offer solutions and suggest interventions to more effectively manage the traffic, especially during peak hours.
TCS Research is also identifying new possibilities of addressing other societal and sustainability challenges using EDTs. These include improving river water quality, achieving a net zero carbon footprint, developing smart and sustainable cities/power grids, delivering smart healthcare, and developing e-mobility solutions. Further internalizing the concept, TCS has repurposed EDT for its employees to ensure the smooth transition from working from home to returning to the office.