Humanity has long been driven by a desire to foresee the future; today, science enables us to statistically predict certain outcomes. Be it weather, disease spread, or climate change, we can now access relatively accurate forecasts or scenarios.
Computational epidemiologists as Alessandro Vespignani are to epidemiology what physicists are to meteorology, creating models to predict impacts; not of weather but of public health.
His team at the Northeastern University Network Science Institute builds biological and social models.
Be it weather, disease spread, or climate change, we can now access relatively accurate forecasts or scenarios
“Models are not oracles – Vespignani explains during an interview with TrendSanità -. They are probabilistic, not deterministic. This is crucial to remember. In a wartime metaphor, computational epidemiologists are like intelligence officers, not soldiers”.
In times of uncertainty, people crave definitive answers, though this is not feasible. “The sooner we accept this, the better our science communication will be. Experts build models; decision-makers adjust expectations and policies based on proposed scenarios and their uncertainty”.
The beginning of the pandemic
At the start of 2020, while most of the world was settling into the new work year after the New Year celebrations, Alessandro Vespignani was alert at his computer. He was analyzing how to reorganize his lab in case the epidemiological situation in China failed to improve. Information on several suspicious cases of atypical pneumonia had been circulating from the East for weeks.
In times of uncertainty, people crave definitive answers, though this is not feasible
“Initially, we hoped it was something similar to the 2002 SARS outbreak – the epidemiologist recalls – a disease with very low asymptomatic transmission. At that time, we didn’t have a clear picture of the transmission mechanisms”. However, once the symptoms were better understood, a red alert was issued.
On January 17, a 6 am teleconference was convened in Boston to discuss case numbers. “It quickly became clear that this was a significant outbreak, and indeed, a few days later, the WHO classified the global risk as very high”.
Maps of the Future
“It’s often said that epidemiologists create scenarios, but that’s misleading; we model situations based on data from governments and Crisis Units that ask us to explore situations based on specific assumptions. For instance: what if everyone were vaccinated, or if work from home were imposed? – tells Vespignani, who is also president of the ISI Foundation in Turin, a non-profit organization that focuses on data science for social impact, including forecasting and scenarios for infectious diseases -. These maps of the future aren’t predictions but tools for health policymakers to understand what could happen under specific conditions“.
After years of advocacy, the US recently established a Center for Forecasting and Outbreak Analytics within the CDCs, akin to the National Weather Forecast Service but for infectious diseases. The CDCs also funds with more than 250 millions dollars a network of national academic centers advancing the analytic science to predict and control infectious disease spread. “I direct the Epistorm Analytical Innovation Center, located in the Northeast US”.
The centers operate under a five-year plan, updated regularly, and collaborate with academia, private sectors, state public health departments, and various public health territorial organizations. “This is a significant step forward that we hope continues to expand”.
The European situation
Unfortunately, no similar progress has occurred in Europe. “I believe predictive and analytical centers for infectious diseases should be established in every European country, with coordinated inter-European efforts“. Some countries perform better than others, but a shared vision is generally lacking. In the US, the predictive system is multi-model: to achieve accurate forecasts, it’s essential to create an ensemble of models.
While a Forecasting Center for Infectious Diseases has been established in the USA, a similar structure and a common vision for prevention are lacking in Europe
This requires an infrastructure where various groups work collaboratively: “The current Academic incentive system, which rewards competitiveness among research teams driven to publish, doesn’t foster a collaborative, multidisciplinary environment”. In Italy, updating the Pandemic Plan has been a topic of discussion for months. “Having a plan is great, but it’s not enough if it stays in a drawer – Vespignani notes -. Planning exercises, practices and constat updating are needed; similar to the drills civil protection agencies conduct for natural disaster response”.
Data Quality
“Data has been one of the major revolutions of the past 20 years. For a long time, detailed modeling wasn’t possible. Today, technology greatly assists us with this”.
To understand the global spread of epidemics, analyzing air traffic is critical. “Back when tickets were paper-based, it was challenging to obtain homogeneous, high-quality data. Today, we have a partnership with OAG data scheduler, which provides real-time traffic data that we can use for forecasting”.
It doesn’t stop there: today, we also have mobile data, which enables tracking of people’s movements. “It’s important to remember that we’re not following individuals, to protect their privacy, but rather use statistical patterns and average behaviors that we incorporate into our models, which became much more highly detailed”.
Data has been one of the major revolutions of the past 20 years
Looking forward, there’s a framework for a sentinel mechanism to monitor epidemic trends. “Just as satellites orbit in meteorology, we need observers constantly measuring virus evolution – Vespignani explains -. With the CDCs and the European Community, we’re working on Traveler-Based Genomic Surveillance, conducting metagenomic analysis on airplane restroom wastewater to track, for instance, Covid variants”.
Twenty to twenty-five sentinels in major airports would suffice to gather enough data. “There are certainly logistical issues to resolve, but we can make it happen. The challenge is to convince global decision-makers of the value of these initiatives. I truly hope these technologies can transform how we combat viruses over the next 10-15 years”.
© Photo by Matthew Modoono_Northeastern University
