Forecast Map Predicts Counties at Greatest Risk from Surge in COVID-19 Cases

A map forecasting which counties will be worst affected by a surge in COVID-19 cases has been created. The map, developed by researchers led by Charles Branas at the Columbia University Mailman School of Public Health, suggests that even regions that have not been badly impacted by the novel coronavirus may face critical care bed shortages over the next six weeks. The U.S. has become the country worst affected by COVID-19. According to the Johns Hopkins University tracker, there are now almost 432,500 cases and more than 14,800 deaths across the country. Models predicting when the outbreak will peak in different states have been developed to help authorities manage healthcare. By understanding when they should expect an uptick in cases, they can try to prepare healthcare facilities to accommodate larger numbers of patients. One model developed by a team at the University of Washington's Institute for Health Metrics and Evaluation (IHME) suggests that as a country, cases will peak on April 11. However, this varies state-by-state. Kansas, for example, is not expected to peak until April 20, while Nebraska will not see its peak in cases until April 26. In the newly developed forecast map, the Columbia team devised a system to assess bed shortages in critical care facilities under a variety of scenarios. These included different levels of social distancing and hospital surge response capacities. The researchers used the number of critical care beds and the potential for hospitals to use beds outside of intensive care units. A mathematical model for how COVID-19 is expected to spread across the country was developed by simulating the movement of spatial and temporal movements of infected people. The study related to the development of the map has not been peer reviewed, meaning it has not been assessed by a panel of experts not involved in the research to judge its findings. The research found a vast number of deaths could be averted if person-to-person contact is restricted effectively. This includes social distancing and self-isolation of anyone with symptoms. Nationally, it is estimated that 100,000 deaths, part of an original estimate by the federal government, could be prevented if aggressive confinement measures are put in place.

"Over the four week study period, excess deaths from inaccessible critical care ranged from 24,688 in the very low response scenario to 13,268 in the high response scenario," the team wrote. "Northeastern and urban counties were projected to be most affected by excess deaths due to critical care shortages, and counties in New York, Colorado, and Virginia were projected to exceed their critical care bed limits despite high levels of COVID-19 contact reduction. "Over the four week study period, an estimated 12,203 to 19,594 excess deaths stemming from inaccessible critical care could be averted through greater preventive actions such as travel restrictions, publicly imposed contact precautions, greater availability of rapid testing for COVID-19, social distancing, self-isolation when sick, and similar interventions." They said even more deaths could be averted by preparing healthcare facilities effectively, such as repurposing non-ICU beds for critical care. This also potentially includes the use of a single ventilator for multiple patients. Study co-author Andrew Rundle, associate professor of epidemiology at Columbia University, told Newsweek: "I knew intellectually that social distancing was key to suppressing the infection, but to see visually how different the hospital demand scenarios were for different levels of social distancing was surprising. And having seen these differences on the maps, it makes me very concerned about calls for relaxation of social distancing." He said one of the surprise findings was that city centers appear to have the capacity to cope with a surge in demand, but the suburban ring counties appeared to be at greater risk of being overwhelmed. "In retrospect, these findings make sense—many of our major medical centers and academic medical centers are located in urban centers, so there is a lot of capacity there."

Read more

• We Predicted the Coronavirus Pandemic but Nobody Would Listen
• U.S. Entering 'Peak Death Week' in Coronavirus Pandemic

Rundle said they developed the map out of curiosity to find where populations at risk of severe COVID-19 were around the country. Over the weeks he received more feedback from health departments and planning offices and eventually other colleges got involved, leading to the developed maps. "The key limitation to our maps of counties that are at high risk of seeing their health care capacity overwhelmed is that we have not yet been able to factor in staffing levels," he said. "Also we can't factor in the ingenuity, grit and can-do aspects of our front line medical professionals. I expect that they are going to find ways to stretch our medical resources in ways that we were not able to anticipate in our models—and that is a good thing. "We as a nation did not have a plan as the pandemic descended upon us, and as we move through this period of mitigation we really need thoughtful plans on how we move back to a containment stance and ease the stay at home orders. It could be really easy to screw up the easing of social distancing and then for us to experience a bad second wave of cases." While the latest forecast could help offer guidance to healthcare providers on how to best prepare over the coming weeks, models like these are subject to a number variables that make them problematic. They rely on a flow of data. As time elapses, the forecasts can change. Robin Thompson, from the Mathematical Institute at Oxford University, U.K., works on predictions of the coronavirus outbreak. Discussing forecast models, he told Reuters: "In a situation like this where there are so many unknowns, it's fair to say it's impossible to predict with any kind of precision at all when the peak is going to happen." Harlan M. Krumholz, a professor at the Institute for Social and Policy Studies, Investigative Medicine and of Public Health at Yale University, spoke to Newsweek about the IHME forecast when it was published. He also said these models are subject to change: "We can take it as information, but we need to recognize that it's a very dynamic situation and it is about more than just the biology of the virus. It is about how we are conducting ourselves and the degree to which society is moving to slow the spread. "The situation is highly dynamic, fluid and changing almost every day. Not so much because of the biology of the virus, but because of the way we're responding to it."