Updated modeling from the University of Washington suggests the state has been effective at “flattening the curve” of COVID-19 cases through business closures and social distancing, though public health experts say those measures must continue to prevent a second wave of the disease.
The UW’s Institute for Health Metrics and Evaluation has created state-by-state projections of COVID-19 deaths, hospitalizations and ventilator needs to help hospitals prepare as they move through different stages of the pandemic.
In a news briefing Monday, the institute’s director, Dr. Christopher Murray, said tweaks to the methodology and an influx of new case data have generated a rosier outlook for Washington state, though the model continues to estimate a nationwide death toll of nearly 82,000.
In late March, the model indicated more than 1,400 people in Washington were likely to die of COVID-19. It also showed hospitalizations in the state would peak on April 19, with about 100 more patients than intensive care unit beds.
Now, however, the model suggests “peak resource use” at Washington hospitals occurred on April 2 without a statewide shortage of ICU beds. It projected daily COVID-19 deaths in the state would peak on Monday before dropping to 18 deaths per day through Thursday and declining slowly thereafter. And it suggests the state will reach its total death toll – 632 – on May 10.
While hospitals across Washington – especially those on the West Side – have had to make rapid changes to brace for a surge of cases, the state’s entire health care system has not been overwhelmed as the IHME model initially forecast.
The model still offers a grim outlook for other parts of the country.
For example, New York, the state hardest hit by the coronavirus outbreak, is projected to have a hospital bed shortage of nearly 12,500 when it hits “peak resource use” on Wednesday. The model indicates New York will tally 15,618 deaths from COVID-19 before the pandemic is over.
Murray said earlier projections were based on a relatively small sample of data from the federal Centers for Disease Control and Prevention, which showed there were about 11 hospitalizations for every death caused by the coronavirus. Newer data shows the ratio is about 7-to-1 in New York and even lower in other states.
Murray said researchers also were able to include new data from cities in China and Italy that appear to have reached their peak number of cases. He said “incorporating that information makes a very big difference in terms of our estimate of need.”
While the projections represent the best guesses of researchers, the modeling still involves a high degree of uncertainty. The range of uncertainty for the nationwide death toll, for example, has a lower limit of about 49,000 and an upper bound of more than 136,000.
Citing confirmed cases of COVID-19 in Washington hospitals, Murray said outbreaks on the West Side appear to be subsiding. But the statewide peak is an average, meaning hospitals in Eastern Washington still could see an increase in cases.
“I think what we are seeing … is pretty compelling evidence that at least King County and Snohomish County have peaked, and the rest of the state, not so,” Murray said. “But put them together, and that leads us to a more positive picture than we previously thought.”
Murray said that’s “tremendously good news at some level – that it’s not just the Italian and Spanish communities that are able to have social distancing work. We think it’s a U.S. example now of the effectiveness of social distancing.”
Unlike other models that attempt to predict how the virus will spread, the IHME projections rely on the number of deaths attributed to COVID-19. Murray has told the Seattle Times that’s a more reliable gauge of the epidemic’s course than confirmed infections, as the frequency of testing varies from place to place.
The IHME projections have been cited by federal officials including Dr. Deborah Birx, the White House’s coronavirus response coordinator, although the Trump administration has said it anticipates between 100,000 and 240,000 deaths nationwide in the best-case scenario.
“The reason our numbers are so much lower is that we saw and believed … that social distancing can lead to a peak and then a decline to near-zero transmission,” Murray said.
The IHME model assumes “full social distancing” will continue nationwide through May 20. As more data becomes available, the projections are being updated almost daily at covid19.healthdata.org.
“Like a weather forecast, they’re going to change as data comes in,” Murray said, adding that “the methods for forecasting are being improved and refined every day.”
Meanwhile, experts at Washington State University, the Spokane Regional Health District and MultiCare have been collaborating on another model designed to forecast local hospital burdens due to COVID-19.
In an email, Spokane Health Officer Dr. Bob Lutz said that model “is envisioned to provide a local/regional means by which we can determine our status and manage accordingly.” He referred questions about the project to WSU and MultiCare, which runs Deaconess and Valley hospitals.
WSU researchers involved in the project agreed to an interview last week, but the interview was canceled after MultiCare declined to discuss details of the local forecast.
“We are keeping those numbers internal while we determine the accuracy of our modeling,” MultiCare spokesman Kevin Maloney said in an email. “With models like these, there are multiple variables that come into play – some of them are unpredictable. To minimize these variables, we are continually tweaking our models to present the most accurate information possible.”
MultiCare and Providence, which runs Sacred Heart Medical Center and Holy Family Hospital, both have declined to say how many COVID-19 patients they are treating on a given day.
Subscribe to the Coronavirus newsletter
Get the day’s latest Coronavirus news delivered to your inbox by subscribing to our newsletter.