Data-informed response to Covid-19: Costa Rica case study

Background

The coronavirus pandemic’s onset was rapid and chaotic and highlighted more than ever the importance of data-informed decision making. With its mission of providing decision-makers and citizens with the very best data to understand the state of their society and prioritize actions, the Social Progress Imperative, a US-based global nonprofit, pivoted its work to respond to the crisis.

In Costa Rica, in partnership with INCAE Business School and collaboration with national and city governments, we developed a multidimensional tool that captures the socioeconomic risk to help decision-makers understand the crisis beyond the immediate health and economic impacts: The Integrated Risk & Response (IRR) Tool. The Tool considers social, economic, health, and real-time population interaction tracked through Facebook’s mobility data (available via Facebook’s Data for Good program) to define location-specific risk levels. It is used by Costa Rica’s President and mayors across the country to monitor the pandemic, prioritize actions and resources, and guide the closing and reopening of communities based on a holistic understanding of risks. This has quickly become a cornerstone of the government’s data-driven pandemic response strategy, allowing Costa Rica to safeguard the health and wellbeing of its 5.1 million people and maintain one of the lowest numbers of Covid-19 cases and death in the continent.

Movement Range Data

The characteristics of mobility within a population are directly tied to virus spread – thus, mobility data can be particularly illuminating during Covid-19. To understand a population’s interactions, INCAE Business School and the Social Progress Imperative adopted Facebook’s Movement Range Maps data as a vital indicator of the Tool. As this data is publicly available at the subnational level, it provides credibility and transparency that ensures the data is tested and aligns with the Costa Rican government’s strategy of tailoring pandemic policies and actions at individual cantons or municipalities. The Movement Range Maps provide data about how populations are responding to physical distancing measures. Its Change in Movement metric looks at how much people are moving around and compares it with a baseline period that predates most social distancing measures, while its Stay Put metric looks at the fraction of the population that appear to stay within a small area during an entire day. With over 70 percent of the Costa Rican population being Facebook users, and a decent proportion having opted in to share their location information, this Facebook data acts as a reliable and robust proxy for mobility. Changes in mobility help predict increases in virus transmission and highlight behavioral patterns, such as low uptake of governmental guidelines (Chart 1).

Chart 1 | Mobility in Costa Rica since March 2020 (weekly average)

Source: INCAE Business School and Social Progress Imperative with Facebook Data for Good (dataforgood.fb.com)

“We built Movement Range Maps in the hope that these tools would support effective policy development related to lockdown orders and re-openings. In the case of Costa Rica and the work of the INCAE Business School and the Social Progress Imperative, the data has been used for exactly that.”

Laura McGorman, Facebook’s Policy Lead for Data for Good

From data to action to impact

Facebook’s mobility data complements other socioeconomic risk factors to form the IRR Tool, and has been adopted by Costa Rica’s Emergency Operations Center, an operative and strategic group leading the nation’s pandemic response and includes the National Emergency Commission, the Ministry of Health, the Ministry of Security, the Ministry of Economic Planning, the Office of the Presidency, the Costa Rican Red Cross, the Costa Rican Social Security System, and others. The Tool provides data and insights to support the development of policies and actions implemented through emergency committees in each of the country’s 82 municipalities, carefully balancing health, economic, and social factors around their different contexts to develop sustainable actions against the pandemic. A strategy coined the Shared Management Model reaches a balance between economic activity and health measures. The IRR Tool is helping national and local authorities contextualize the socioeconomic risks created by the pandemic, and by using Facebook data, it provides fresh insights about people’s behavior. Movement Range Maps data is updated daily on the Humanitarian Data Exchange (HDX), and presented at weekly intervals by INCAE Business School’s Jaime Garcia to the Emergency Operations Center. As a member of the Center’s data team, Garcia is responsible for ensuring decision-makers have the most relevant, up-to-date data available to make critical decisions to protect the population’s safety and well-being. “Facebook mobility data has significantly strengthened our ability to guide responses at the national level and at the same time measure the efficiency of the strategy in the short run. We are now facing very challenging times, and the insights from this data that are presented to different authorities in the country, starting with the President and the Ministers, are extremely valuable.”

To promote social distancing and limit the virus’s spread, the Emergency Operations Center turned to data to generate a response. Facebook’s mobility data showed that weekends had the highest levels of population movement, promoting the government to adopt a policy to reduce interactions by restricting vehicle movement on Saturdays and Sundays based on their license plate number (Chart 2). However, in some areas, despite restrictions, population movement remained high. Using the tool, we were able to cross-analyze this data against other socioeconomic indicators, such as population density, productive structures, propensity for home office, or the presence of informality in the job market, illustrating why a one size fits all response doesn’t work. For example, in rural and tourist cantons, mobility data is higher than in urban areas as individuals have to travel greater distances to access public and commercial services. The populations in these cantons also have less propensity to work from home given the nature of their agriculture and tourism-related jobs.

Chart 2 | Average mobility in Costa Rica before and after the weekend vehicular restriction

Source: INCAE Business School and Social Progress Imperative with Facebook Data for Good (dataforgood.fb.com)

Insights from the Christmas Holidays

During the Christmas holiday, Facebook data showed a mobility phenomenon. There was extremely high mobility in the days immediately leading up to Christmas Eve, comparable to the pre-pandemic levels. This fueled the rapid spread of the virus, overwhelming hospitals and creating a shortage of critical care beds not seen at any other point during the pandemic. With this insight from the data, the decision was made to tighten vehicle restrictions further and introduce lockdown measures in the highest risk areas. Population movement was dramatically reduced in the following 15 days to some of the lowest levels recorded during the pandemic, particularly in the greater San Jose area, Costa Rica’s capital and largest city.

Despite pandemic restrictions in place, many Costa Ricans opted to maintain the tradition of celebrating New Year’s Eve at one of the country’s 300 beaches. Facebook´s data confirmed that 12 of the 20 cantons with the highest mobility levels during that period were coastal cantons (see Map 1). This was particularly concerning as Costa Rica’s centralized health system has limited capacity to support large outbreaks outside of San Jose, so a sharp spike in Covid-19 cases in these coastal communities could threaten to collapse the country’s fragile healthcare system.

Map 1 | Average mobility in Costa Rica’s cantons during the 2020 Christmas holidays

Source: INCAE Business School and Social Progress Imperative with Facebook Data for Good (dataforgood.fb.com)

The government quickly introduced protocols closing beaches in the afternoons and implementing restrictions on specific social and economic activities, such as  closing 75 percent of all bars and casinos and limiting national parks to half capacity. This immediately reduced movement in the beach areas and adverted an outbreak in these communities; specifically, in the most visited coastal cantons, the reduction in mobility during the first two weeks of January was very strong, decreasing from an average positive mobility index of 0.0123 to a decrease in mobility of -0.0627, even though the impact was differentiated by canton (see Chart 3).

  Chart 3 | Average weekly mobility in Costa Rica’s most visited cantons

Source: INCAE Business School and Social Progress Imperative with Facebook Data for Good (dataforgood.fb.com)

“The fact that people stayed at home, sharing with their bubble, has allowed a containment of the virus and therefore this helps not to collapse the hospital capacity or compromise the Intensive Care Units”, argued Alexander Solís, president of Costa Rica’s National Emergencies Committee.

Looking ahead

Given what we have learned about this crisis, it is clear that the eventual allocation of vaccines in Costa Rica will also need to carefully consider local economic, health, and social contexts. As part of the IRR tool, Facebook mobility data will be critical in helping decision-makers understand the current situation across the country to determine where vaccine distribution will be most effective. However, in the current period, vaccines are unlikely to be widely accessible for Costa Rica until 2022, necessitating the continued use of the IRR tool to monitor the pandemic and adjust policies and actions to best manage the current crisis.

As a tested instrument that has had a significant impact in Costa Rica, saving thousands of lives and avoiding health care system collapse, this tool is now being scaled across multiple countries, including Brazil, Mexico, Colombia, South Africa, India, and beyond.

“The integration of Facebook data into the Covid-19 socioeconomic tool has guided innovative policy in Costa Rica, impacting millions of lives. We look forward to developing it further to promote data-informed decision making in response to the pandemic across the globe.”

Franklin Murillo, Social Progress Imperative’s Global Network Director

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s