https://covidactnow.org/
Led by former Yale student Rep. Jonathan Kreiss-Tomkins ’13, D-Alaska, a team of data scientists and epidemiologists created a new digital tool examining the potential stresses of COVID-19 on state health care systems given different scenarios of intervention.
Kreiss-Tomkins launched the website, covidactnow.org, on March 20 alongside Max Henderson, an ex-data scientist at Google, and Igor Kofman, the former CTO of Dropbox Paper. The platform serves as a comprehensive tool that public leaders and health officials can use to make quick and informed decisions about interventions that will effectively combat the COVID-19 crisis. Despite widespread fear that self-isolation measures will batter the economy, the website creators say that the model is a testament to how social distancing and sheltering in place are imperative measures for the foreseeable future.
“The economy is unfortunately ill-fated because of this pandemic,” Kreiss-Tomkins said. “It is just a question of whether we want to reduce the fatalities by acting early, or suffer massive hospitalizations and unnecessary deaths and still have the economy suffer.”
To persuade both citizens and policy leaders to take more drastic measures to curb the spread of the virus, the developers made a spreadsheet of the data and assumptions used to build the model available to the public. The team continues to update its interactive tools with the help of experts from the Center for Global Health Science and Security at Georgetown University.
On the homepage of the website, a map shows a multicolored mosaic of the United States, each hue corresponding to a different level of intervention currently implemented in the 50 states. Thirty-eight states are teal, which indicates that citizens have been legally ordered or strongly recommended to shelter in place while all nonessential businesses have shut down. Eleven orange states are currently implementing social distancing measures which include “voluntary shelter-in-place” for high-risk groups along with school closures. At the time of writing, only South Dakota is tinted red, indicating that its state government has mandated little to no action to stop the advance of the pandemic.
When a viewer clicks on any state, the model displays a graph of projected hospitalizations over the next five months. After clicking on the red state of South Dakota, the model reveals that a peak in hospitalizations will occur by May 2, with over 19,377 people requiring intensive care if limited intervention measures persist. By this time, the state’s health care system will have already been overwhelmed for a month. Superimposed onto the graph is a curve with the projections for South Dakota if the state implements shelter-in-place policies for three months. The number of hospitalizations projected for May 2 falls to just 157, and the state’s hospital capacity is not overwhelmed.
Avoiding an overrun of healthcare systems will be crucial in protecting the country, Kreiss-Tomkins explained. “If a health care system is overwhelmed, you have a sort of a cascade of negative effects,” Kreiss-Tomkins said. He explained that the effects include “collateral damage” of patients who are not sick with COVID-19 but suffer from conditions like heart disease or diabetes, who will be deprived of the medical attention they require if health systems fail.
Joseph Ensminger, an entrepreneur who helped with the launch of the website, added, “Our nation’s parents, kids, and individuals that happen to be in a car wreck, have an emergency … or who are already sick will need a place to go.”
Although the model gives detailed, data-based projections for the spread of COVID-19 in the United States, the website lists its limitations. The model uses a reproductive rate for the disease that does not control for variables such as population density or climate differences. The analysis also assumes each person spreads the disease at the same rate. In reality, some people are “super-spreaders,” passing the disease to a significant number of individuals, while others are almost entirely isolated from their communities and therefore not capable of infecting a large group. Finally, when calculating hospital capacities, experts accounted only for aggregate hospital beds, excluding ICU beds and ventilators from these projections.
To address these limitations, the website stipulates that the model intends “to drive fast action, not predict the future.” Kreiss-Tomkins explained that the overarching projections for peak hospitalizations that the model makes are “basically undisputed in the epidemiological and academic community.” He described that the model is not meant to definitively state that on a certain day, a certain number of hospitalizations will occur. Instead, he argued that the broad trends illustrating early intervention will help to save state health care systems and the lives of many people should be used to inform policy.
“The earlier and more aggressively we act, the better our choice in a week or a month or a year will look like,” Kreiss-Tomkins said. As he spoke, he was packing up his belongings, concerned that if he stayed in his current residence in Juneau, Alaska, he would be a vector for the disease, endangering some of the older members of his household.
“Every person who loves their spouse, parents, grandparents, and their friends, their life is literally in your hands,” Ensminger said. “You’re making an ethical choice in reacting to this virus. It will be seen as such when the chaos eventually settles down and we look back to how we reacted to this pandemic. Your character will be revealed.”
On Wednesday, the U.N. Secretary-General announced that the coronavirus outbreak is the greatest challenge for the world since World War II.
Correction, April 2: A previous version of this article referred to Kreiss-Tomkins as a Yale alumnus; in fact, he attended Yale but did not graduate, as he left the University to pursue a career in politics. It also referred to Ensminger as a Tennessee government official; in fact, he is a private sector entrepreneur.
Sydney Gray | sydney.gray@yale.edu