Based on the declining death toll estimates from a national COVID-19 model, Minnesotans are excelling at social distancing.
The model, from the Institute for Health Metrics and Evaluation (IHME) at the University of Washington’s School of Medicine, has gradually lowered the number of deaths estimated for Minnesota – attributable to the novel coronavirus – over the next four months.
What was an original projection March 26 of more than 1,900 deaths in Minnesota, the latest model simulation from April 5 estimates that 632 people will die from COVID-19 in the first wave of the pandemic by Aug. 4, with a range of 224 to 1,362 deaths.
As the makers of the model note, all estimates "assume the continuation of statewide social distancing measures," and that if "such policies are relaxed or not implemented, the US could experience a higher COVID-19 death toll and hospital burden than what our models currently predict."
Nationally, the model has has been updated five times since March 26, with the nationwide death toll estimates increasing in three consecutive updates before dropping in the latest simulation, which was released late Sunday night.
- March 26 update: 81,114 deaths (range of 38,242 to 162,106)
- March 30 update: 82,141 deaths (range of 39,174 to 141,995)
- March 31 update: 83,967 deaths (range of 36,614 to 152,582)
- April 1 update: 93,765 deaths (range of 41,399 to 177,381)
- April 5 update: 81,766 deaths (range of 49,431 to 136,401)
"Our model’s increase in nationwide deaths since the March 31 release is primarily driven by increasing death tolls in states that previously had very few COVID-19 deaths," the study explained after the April 1 update. "States with more COVID-19 deaths, such as New York and Washington, show far less fluctuation across daily updates. If the number of cumulative COVID-19 deaths rise by state, we expect increasing accuracy for predictions across states."
How accurate is the model?
Depending on where you look or who you listen to, projections vary wildly when it comes to fatality projections. For example, a Mayo Clinic expert recently said that predicting anything beyond two weeks is extremely difficult, while U of M infectious disease expert Dr. Michael Osterholm continues to warn (74-minute mark of this podcast) that it's not out of the question that 1.6 million Americans die from the disease over the next 12-18 months.
Again, the University of Washington model is only looking at the next four months, which it expects to be the first wave of the pandemic. It does not elaborate on how many waves there could be.
Doug Schultz, information officer with the Minnesota Department of Health, notes that "all models are intended to be used as directional indicators or roadmaps, not as precise predictors of future outcomes or as accurate indicators of current data."
The current data in the University of Washington model is not 100% accurate in regards to Minnesota-specific data.
For starters, it doesn't recognize that Minnesota has closed non-essential businesses, nor do the number of available ICU beds match up with the numbers listed on the Minnesota Department of Health's COVID-19 dashboard that is available for the public to see right here.
The model says Minnesota has 355 available ICU beds when in reality, according to the health department, there are, as of Sunday, 313 ICU beds available with an additional 1,623 that could be readied within 72 hours. Keep in mind that the number of open ICU beds varies daily and the University of Washington model keeps that number static at 355.
The model also estimates that Minnesota will need fewer than 450 ventilators by the peak of the outbreak, which it suggests will happen around April 23, at least a month earlier than what the Minnesota Department of Health has been projecting (mid-May to mid-June peak).
The state health department data shows that there are 1,428 ventilators available right now in Minnesota, with another 888 on order.
"The models are only as good as the data and the assumptions that went into them. The MN model used the best Minnesota-specific data available at the time, and is constantly being reviewed and updated. But it’s assumptions and data inputs do differ from other models, so it is not surprising to see difference occurring between them," said Schultz.
"Minnesotans should consult a variety of sources to reach their conclusions."
It's expected that the Minnesota Department of Health will update its model within the next week or so, which if revealed to the public, could give a better indication of what could happen in the coming weeks as the outbreak evolves.