May 19, 2020
Will our hospitals have the capacity to care for those who become infected? This question remains at the heart of COVID-19 planning and response discussions everywhere as public health officials and policy makers consider how lifting interventions such as stay-at-home orders will impact a potential resurgence of the disease.
In its ongoing work related to the COVID-19 pandemic, the University of Virginia’s Biocomplexity Institute has developed a Medical Resource Demand Dashboard to project when various regions across the Commonwealth of Virginia could potentially reach maximum hospital capacity due to the pandemic. The visualization tool couples the Institute’s COVID-19 epidemic outbreak simulations with projected hospital occupancy numbers to ultimately help public health officials improve medical resource allocation and avoid crises related to hospital capacity.
The dashboard features an interactive map alongside projections for hospital bed occupancy and weekly hospitalizations based on multiple intervention scenarios. Current scenarios explore projections with stay-at-home orders lifting on May 15 and June 10, social interactivity returning to 25 and 50 percent of pre-pandemic levels, and hospitalization stays of 8 and 14 days. The dashboard currently features data only for Virginia, but Institute researchers will soon expand it for the entire United States based on national Hospital Referral Region (HRR) data. The Institute team will update the dashboard weekly as more data becomes available and government-directed interventions evolve.
“Our goal is to help public health officials better understand the factors that could lead to a medical resource crisis in a certain area and provide them with the foresight to shift or add resources to avert that crisis,” said Mandy Wilson, research scientist for the Institute’s Network Systems Science and Advanced Computing (NSSAC) division and lead developer for the Medical Resource Demand Dashboard. “With regional visibility, officials can see where they might transfer patients, add hospital beds, or send needed supplies such as ventilators or personal protective equipment.”
“These projections allow us to explore how different interventions can influence epidemic behavior. The numeric values should not be taken as actual values because they rely on ever-changing conditions, such as levels of social distancing, mask wearing, and so on,” Wilson said.
The Dashboard features:
- Projections of weekly hospitalizations per Virginia Hospital Alerting and Status System (VHASS) region based on various model scenarios;
- How projections change over a six-week period of time, starting from the current week;
- The percentage of occupied hospital beds by VHASS regions based on weekly hospitalization projections and an assumed 80 percent non-COVID-19 occupancy rate;
- A heatmap showing regions that are likely to be stressed by the epidemic; and,
- A reactive chart showing how weekly hospitalizations and percentage of occupied beds are likely to trend over time based on regions selected on the map.
“The dashboard, along with the modeling environment, represents a significant advance in epidemic science,” said Madhav Marathe, director of the Biocomplexity Institute’s NSSAC division. “The data-driven mechanistic models provide current projections of medical resources, enabling decision makers to plan and respond to the outbreak in real-time. The dashboard features an integrated view of medical resource projections, including numeric data, scenarios, and spatial maps. We are not aware of other dashboards that provide this level of spatial and temporal detail and scale.”
The Institute has been at the forefront of epidemiological modeling for the COVID-19 pandemic since early 2020, and has worked in partnership with several U.S. government agencies as well as the office of Virginia Governor Ralph S. Northam, the Virginia Department of Emergency Management, and Virginia Department of Healthto inform evidence-based decision making. Institute researchers have developed a suite of COVID-19 Epidemic Response Resources including the new Medical Resource Demand Dashboard and COVID-19 Surveillance Dashboard that have been accessed by approximately one million users from more than 200 countries and territories in the world since February 2020.
“For more than 20 years, our team has supported the U.S. federal government and other countries through several epidemics, including planning for H5N1, the 2009 H1N1 pandemic, the MERS outbreak of 2012, and the Ebola outbreaks of 2014 and 2019,” Marathe said. “Our team has a long history of developing tools and dashboards to support epidemic planning and response, and remains committed to providing solutions to problems of societal importance."
Additional information such as modeling framework details, assumptions, and frequently asked questions for the Medical Resource Demand Dashboard are available at https://covid19.biocomplexity.virginia.edu.