NECSI-Behavioral Interventions

National and subnational lockdowns have been the most widely used measures for controlling the spread of a pandemic in the absence of medical preventive methods. However, lockdown policies often overlook human factors, health disparities, and human mobility, worsening health inequities for vulnerable populations. Mobility data facilitate informative surveillance of disease spread at multiple scales and yields multiscale mobility patches. The formation of mobility patches often correlates with socio-economic and ethnic characteristics of societies and therefore can serve as proxy indicators of vulnerable populations and their activities. Our evidence-based platform is designed to empower the lockdown strategies considering human mobility patterns and human factors accounting for societal needs, costs, and equities at various scales.

Vulnerable populations bear the brunt of the COVID-19 pandemic. The uncertainty of disease dynamics, hard-to-control human mobility, and competing objectives coupled with the nature of applicable data present obstacles to stakeholders charged with developing optimized lockdown policies and minimizing the disease diffusion. This is a national and global problem mainly due to the nature of today’s human connectivity patterns. Since the beginning of the COVID-19 pandemic, lockdown strategies in many countries ignored the connectivity of demographic characteristics of societies and movement patterns. In most cases, lockdown policies and data analytics have been based on administrative borders such as state, province, or county borders which represent neither the socially connected areas nor demographic spreading of populations. 

Our AI platform aims to examine geographical mobility patterns before/during/after COVID-19 outbreaks in the US. Our findings will demonstrate which populations are more affected, where they spend time, how intervention policies change their behaviors and how can be optimized in the future. These epidemiological data analyses are just emerging, and however health responders have been aware of the potential of the large-scale social data, they have not had any clear and well-structured insight about the possible solutions.

NECSI-Behavioral Interventions platform aims to develop, validate, and disseminate a novel technology-based methodology for optimizing lockdown strategies that match the needs of vulnerable populations. Availability of Big Data about human movements, city infrastructures, and demographic characteristics of locations, along with the georeferenced datasets of disease dynamics, well position us to find answers to two critical questions: 1) How lockdown policies have affected aggregate movement behaviors of different population segments that vary by income, race/ethnicity, socio-economic status (SES), and/or rural/urban residence?, and 2) How to adapt mobility restrictions to prevent inter-community transmission with minimal socio-economic impacts? To answer these questions, we will use and develop new complex systems, deep learning methods, and algorithms to extract the patterns and use these patterns for modeling adaptive behaviors of populations. The findings of this research will inform health policymakers about the social effects and limitations of the existing policies and lay out strategies for more effective planning in the future to leverage social dynamics with disease dynamics to achieve better health outcomes and minimize economic losses. While our analysis is done for the US, similar strategies can be applied to other countries upon data availability.

Due to data limitations, our current analysis is in the scale of US country. However, it has the potential to be applied to other areas and countries. We will investigate the patterns all over the US. So our findings will serve all the US residents and especially the vulnerable segments of the populations. While large-scale lockdowns have been the most effective way of controlling the COVID-19 pandemic, many people and businesses have been suffered from them. The current solution aims to understand the effect of lockdown strategies on populations by studying the dynamical patterns of movements and their changes since the beginning of the pandemic and offer solutions that better match with social and business needs.   

Since 2020, NECSI has been involved in influencing policy at the local, state, and national levels in the US and internationally in multiple countries. Our outreach work via posting on our blogs (https://necsi.edu and https:// www.endcoronavirus.org), social media like Twitter, Slack, and Linkedin, and interviewing with news channels will inform governors, corporate policymakers, disaster response teams, scientists, and data providers, and will increase public caution. The New England Complex Systems Institute, under the leadership of Prof. Yaneer Bar-Yam, has founded a volunteer community of over 6,000 participants of which nearly 5,000 have signed onto a Slack community coordination platform. Volunteers have developed the website endcoronavirus.org and are participating in multiple projects to use scientific analyses to communicate to policymakers, engage in community outreach. 

Our work is widely distributed through influential individuals on Twitter, including through the president who has retweeted two of our tweets tracking the outbreak dynamics. Multiple articles have been written about our efforts. Our efforts to influence policy based upon science began in January. On January 26, 2020, the publication of an article informed early US policy on travel restrictions. The researchers of our group have actively presented the findings in scientific meetings and conferences. Interviews on CNN, Wall Street Journal, New York Times, Washington Post, Boston Globe, and OpEds have covered a number of topics. 

Our offered solution and target population are exactly aligned with the goals of the Challenge. We are looking to decrease health inequities by empowering large-scale lockdown policies that better help underserved and vulnerable segments. Endcornavirus created since the beginning of the COVId-19 pandemic, is organized by NECSI specifically to accelerate the fight against the COVID-19 pandemic at various scales from subnational to national and international. Using up-to-date mobility patches, our platform can predict and model the spread pattern of pandemics and help strengthen disease surveillance.

The main important requirement for having large-scale solutions for socio-economic problems in societies is the accessibility of big up-to-date data. Thanks to many technology companies that collect data, we now have reasonable access to all data we need for scaling up our idea and build service for health policymakers to adapt lockdown strategies with current mobility patterns and social needs and predict the next steps of actions during the spread of a pandemic in different areas of society and at various scales.

NECSI-Behavioral Interventions approach has the following innovative features:

Dynamical patterns: The evolution of the patterns makes it possible to define the effectiveness of the policies over time. 

Solution approaches: a combination of known and new AI and machine learning techniques will be used to build and analyze the patterns. 

Improving equity and public health: This solution provides novel insight into how mobility patterns of vulnerable population segments get affected by the lockdown strategies associated with a pandemic and how this information can be used to improve equity and public health in various scales of society. 

Adaptive lockdown strategies: While some studies have reported positive correlations between the mobility ratio and COVID-19 growth ratio for administrative borders, no prior work has tested mobility fragmentation pattern and their applications in optimizing lockdown policies. This solution is novel in offering a human-centered, technology-based solution by extracting mobility fragmentation patterns from historical and recent movement preferences of individuals and defining socio-economic nature behind them to adapt lockdown strategies with actually connected areas and social needs of societies. 

Our current analysis is designed to analyze and offer solutions for the US country and will serve the whole US residents. In a year, we will develop the platform and try to reach the right audience and policymakers. In five years, we will develop our business idea and scale up our solution to be used for more countries and to offer international solutions. 

We are measuring our progress at multiple levels in social media and the real world. In social media, we are monitoring audience growth rate, post reach, the social share of voice, applause rate, average engagement rate, amplification rate, virality rate, conversion rate, bounce rate, cost-per-click, comment conversation rate, and some others. Our tweets gain thousands of views on @endCOVID19 and @yaneerbaryam accounts, as well as posting on LinkedIn, and ResearchGate accounts. Our viral tweets reach thousands of social media users. @endCOVID19 has 12.5K followers and @yaneerbaryam has 65.3K followers. Besides this, we have a large user base on our endcoronavirus.org website. Thousands of volunteers have signed up at our website during the past year. We manage communication with volunteers on Slack. 

The ultimate measure of progress will be through the adoption of our tools to inform policy by community and government leaders. Leading up to that ultimate goal, we have intermediate goals of publication and meetings with stakeholders as metrics of progress. 

2 full-time

2 part-time

Beginning in 2006, NECSI warned of how the increased ease and frequency of global travel may make the risk of pandemics more severe than previously thought. In 2014, NECSI advised a variety of organizations regarding the Ebola outbreak in West Africa, advocating for community-based monitoring. This strategy was what was ultimately adopted and effective. In January of 2020, NECSI began to publish its predictive modelings and guidelines regarding the COVID-19 outbreak. We were at the forefront of warning of the exponential growth and need to properly lockdown and isolate cases. NECSI has utilized its knowledge and research to mobilize endcoronavirus.org, which has provided guidelines, support, and policy advice to communities around the world in combating the pandemic. 

As part of NECSI, Leila Hedayatifar has led the effort to use mobility pattern analysis to identify true communities, rather than those defined by artificial state and country boundaries. This is instrumental in predicting spread and formulating a proper response policy. Olga Buchel with a background in Information science has developed multiple interactive visualization platforms. Elena Naumova's area of expertise is in modeling transient processes with application in immunology, ecology, and epidemiology of infectious diseases. Trained as a mathematician, she has been working with epidemiologists, immunologists, virologists, and public health professionals, as evidenced by joint publications. She has developed innovative analytical and computational tools to monitor and assess spatiotemporal processes of respiratory infection transmission and the intricate relationships between the manifestation of infection at a population level.

Three out of four of our core team consists of accomplished women originally from other countries. Through our partnership with endcoronavirus.org, we promote inclusion from all communities throughout the globe, accepting all volunteers with open arms in the effort to eliminate COVID 19 through grass roots efforts.

Our human-centered, tech-based, and innovative solution drives benefit from the correlations among various social patterns with deprivation and demographic characteristics of societies to improve public health and equity of societies at appropriate scales. We are collecting and combining mobility data, deprivation indices, and COVID-19 infection data and utilizing various AI and Machine Learning techniques to build a platform that can be used for lockdown optimization during a pandemic. This solution has the potential to be scaled to larger or smaller scales to be more global or local. This solution is now feasible by the accessibility of various sources of mobility data.  

As an educational and scientific-based institution, we need partners to help us develop our knowledge about how to better fit and develop our product with the business market. We need strategic advisors who have better knowledge about the market and relationships with social media and possible clients.

We believe that our project very well matches the mission of the MIT Sloan Health Systems Initiative that is to improve health and lower healthcare cost, and we are willing to partner with them. We are also interested to partner with researchers in MIT Center for Collective Intelligence initiative, as our project is also about collective behaviors of a connected society by the movement of individuals. To scale our product, we intend to leverage strategic partnerships with MIT members that can help us with data collection and product marketing, so integration can happen quickly and easily. We are willing to partner with Uber. 

Our solution and product aim to increase equity and public health by adapting lockdown policies with the social needs of societies, and so reducing the costs of a pandemic, especially vulnerable populations. The current solution is designed for the whole US country. We will build a platform that can offer solutions for more time and cost-effective lockdown strategies during a pandemic at various scales from city to national. 

We are leveraging multiple AI and Machine Learning techniques to automatically extract and detect social patterns and their evolution before/during/after a pandemic begins to spread in societies to offer optimized solutions at the right time. Our solution will improve public health, and as a consequence, will increase equity and sustainability for all segments.