Study Projects 853 COVID-19 Deaths in Wisconsin, 81,114 Nationwide

UPDATE: The University of Washington has revised the study, and the projections have gotten significantly worse for Wisconsin.

A new study from the University of Washington estimates that 853 Wisconsinites will die after contracting COVID-19.

The study attempts to determine the number and timing of deaths related to COVID-19 in all 50 states. Released March 26th, it estimates 81,114 deaths nationwide from a broader range of 38,242 to 162,106 deaths.

Authored by UW Institute for Health Metrics and Evaluation Director Christopher J.L. Murray, the study forecasts the country is short almost 50,000 hospital beds and 15,000 intensive care unit (ICU) beds based on peak utilization and average hospital availability in each state.

But in Wisconsin the peak is estimated to be experienced later than anywhere else and the hospital bed shortage is smaller than any neighboring state.

Nationwide, the healthcare system will experience its greatest strain on April 14th, but that date is pushed back to May 22nd in Wisconsin. It’s the latest such date for any state and Wisconsin is one of only five states that have a peak date in May. No narrative is included in the study explicitly explaining why Wisconsin’s peak will happen so much later, but the model relies on recorded deaths as part of its estimates.

Unlike other studies, including one that estimated the Milwaukee area had the sixth greatest COVID-19 outbreak to-date, the university study relies on deaths rather than reported cases in an attempt to avoid issues with variations in testing capacity. “Deaths are likely more accurately reported than cases in settings with limited testing capacity where tests are usually prioritized for the more severely ill patients. Hospital service need is likely going to be highly correlated with deaths, given predictable disease progression probabilities by age for severe cases,” wrote Murray. “In this study, we use statistical modeling to implement this approach and derive state-specific forecasts with uncertainty for deaths and for health service resource needs and compare these to available resources in the US.”

The study predicts Wisconsin will not run out of general hospital beds, but will be short 30 intensive care unit beds. That’s based on data Murray compiled of the average utilization rate of hospital facilities and not the total beds in the system. Wisconsin is estimated to need 202 ICU beds to treat patients with COVID-19, but based on past utilization would only have 172 available. Wisconsin Department of Health Services Secretary-designee Andrea Palm said earlier this week the state has 2,500 ICU beds.

An interactive website offers charts visualizing healthcare capacity and total deaths in each state for a four-month window. Users can mouse over specific days to see the estimated range for hospital resource utilization and total deaths.

According to the study, the models and resulting visualizations will be updated as more data becomes available. “At this early stage, even 1 to 2 days’ more data for a state will improve the estimates of service need and expected deaths,” wrote Murray.

The study was not able to estimate ventilator capacity in every state. Wisconsin officials said earlier this week that the state has 620 ventilators. The study estimates 109 ventilators will be needed in Wisconsin just for patients suffering from COVID-19.

“Any attempt to forecast the COVID-19 epidemic has many limitations,” wrote Murray. “Our models explicitly take into account variation in age-structure, which is a key driver of all-age mortality. But these efforts at quantification do not take into account many other factors that may influence the epidemic trajectory: the prevalence of chronic lung disease, the prevalence of multi-morbidity, population density, use of public transport, and other factors that may influence the immune response. We also have not explicitly incorporated the effect of reduced quality of care due to stressed and overloaded health systems beyond what is captured in the data. For example, the higher mortality rate in Italy is likely in part due to policies around restricting invasive ventilation in the elderly.”

Officials at the Department of Health Services said on a media call this week that they are looking at models to predict the pandemic’s impact on Wisconsin and what resources will be necessary to fight the disease. Dr. Ryan Westergaard, Chief Medical Officer of the Bureau of Communicable Diseases, said the models are being consistently updated as new data becomes available. “The best-case scenario is that we would see a peak in the next two to three weeks and that’s what we’re really going for,” said Westergaard during a briefing Friday of the state’s models.

State officials have repeatedly encouraged individuals to stay home to combat the spread of the disease and the study’s models are predicated on individuals complying with social distancing orders.

How will both the state’s and study’s models change going forward? Wisconsin had recorded 17 deaths from COVID-19 as of noon Saturday, more than double the eight reported on the 26th. An analysis of the study model’s dataset shows it predicted only three deaths over the same time period. It did not have Wisconsin recording 17 deaths until April 3rd.

Urban Milwaukee contacted the UW Institute for Health Metrics and Evaluation with regards to the increased death total in Wisconsin, but did not hear back by the time of publication. We will publish an update if new data becomes available.