Explaining Minnesota's coronavirus model and why its death toll estimate is higher

The model's findings were revealed to the public on Friday.
Author:
Updated:
Original:

The novel coronavirus model developed by the Minnesota Department of Health and the School of Public Health at the University of Minnesota has a higher death toll estimate than other models in part because of the timeline it uses, and because it assumes there will be multiple peaks over the course of the pandemic. 

There are numerous moving parts in an ever-evolving situation, and as the modelers learn more about the virus and how it is impacting Minnesotans, it should be expected that estimates in deaths, ICU beds needed, ventilator needs and every other variable in the fight against the disease will change. 

It has already, with MDH and U of M modelers already having revised down its death estimates from the initial model it was using last month. On Friday, MDH and U of M experts provided a media briefing about the model. Here's what we learned:

What the model is showing

For starters, this chart shows what percent of COVID-19 patients, based on their ages, require hospitalization, hospitalization in the ICU, and mortality rate. As you can see, the mortality rates for patients who are 80+ years old are much higher than other age brackets.

That is evidenced by the fact that the median age of those who have been confirmed to have died from COVID-19 in Minnesota is 87 years old, as of Friday. 

Screen Shot 2020-04-10 at 12.48.30 PM

The modeling uses a variety of variables to make assumptions about what could happen under different scenarios. Scenario 4 in the chart below is the scenario closest to what Minnesota is experiencing right now, with the "Stay at Home" order extended this week by Gov. Tim Walz to go through May 4. 

Scenario 4 essentially took the initial two-week stay-home order and extended it by four weeks until May 10. It helped influence Walz's decision to extend the actual stay-home order until May 4. 

Under Scenario 4, the peak of the first wave of the outbreak is estimated to come in 13 to 21 weeks. The deaths estimate of 9,000 to 36,000 Minnesotans is not what could happen during the initial peak. It is an estimate for the duration of the pandemic, which the model assumes could feature multiple peaks and last at least a year.

Screen Shot 2020-04-10 at 12.52.21 PM

The models have been revised down from last month's estimates, where possible death numbers ranged from 55,000 with social distancing to 74,000 with no mitigation.

The reason that some of the figures have changed is because more is known about the virus than was known a month ago.

Screen Shot 2020-04-10 at 2.10.32 PM

For example, the latest MDH model takes into account that the average ICU and hospital stays for severe coronavirus cases is shorter than a month ago, however it also notes that the rate of transmission is higher – with one infected person infecting 3.87 people, compared to 2.5 in the first model.

It also takes into account that more ICU beds are available now than there were a month ago, with patients who don't have an ICU bed available having a 1.5 to 16.5 percent higher chance of dying.

Here is the full presentation from Friday morning.

Death toll estimates cannot be compared to U of W model

A common reaction after seeing the high death estimates is shock due to the widely shared University of Washington model that only predicts 442 deaths, with a range of 107 to 1,566, in Minnesota. But the University of Washington model doesn't forecast what could happen beyond August 4 whereas the Minnesota model goes out for at least a year.

"We could spend an hour on why that model is different from Minnesota's model and why we think, though their researchers are as bright as our team, that they have made some incredibly optimistic assumptions," said Stefan Gildemeister, the Minnesota State Health Economist.  "Optimistic in regards to what mitigation is actually in place, with regard to the extent of death data that is actually accurate to have come out of China and other kinds of assumptions. 

"The big difference is they're predicting four months," Gildemeister said. "At the end of the four months, that team is saying 93 percent of the population won't have immunity. So what will happen at that time?

"Just because we're staying at home doesn't mean the virus isn't waiting for us to come back and infect people through its typical routes of infection who don't have built immunity."

Gildemeister said the MDH/U of M research team is "so tired" of answering questions about the difference that they may create a 4-month model so Minnesotans can see what the Minnesota model's death estimate is through Aug. 4. 

What's more is that Gov. Walz and other state leaders are basing their policy decisions on more than the Minnesota model. For examples, hospitals leaders around the state are actually more pessimistic about ICU overload during the peak than the Minnesota model is. 

"Their own models, as the governor has said, are much more pessimistic than this one," said MDH Commissioner Jan Malcolm. "The hospital models are suggesting that the time is really critical because they think that the peak ICU demand will be greater than these numbers."

Walz is also taking into account the ongoing impact on the state's economy from the Stay at Home shutdown, and the realization that this cannot continue throughout a 12-16 month period.

Multiple peaks during the pandemic are possible

While the Minnesota model's death estimates are for the duration of the pandemic, it doesn't reveal graphics for how many ICU beds could be needed if the outbreak features multiple peaks. 

For example, this might be New York's first peak, and it could be smaller than future peaks during the outbreak. 

"This is the peak of Wave 1 for New York. I've heard other epidemiologists very concerned that New York will see waves much larger than what they've already seen in coming months," said Malcolm. "This is not a one-time event. It isn't: you hit the peak and everything goes back to normal. This is going to be with us in a really challenging way, in multiple waves, until there are treatments and a vaccine. 

"That is just the hard reality we need to factor into the planning and into these policy decisions. How do we not only manage whatever size this first wave turns out to be, but if in this first wave not that many of us are exposed yet, that is not the end of the story."

Future waves of the COVID-19 outbreak being larger than the first wave is based on the assumption mitigation via business closures and stay-home orders could be relaxed in the future to help keep the economy running.  

"New York is seeing those numbers under mass social distancing. That's what's going on right now to bring those peaks down," explained Shalini Kulasingam, of the University of Minnesota School of Public Health. 

"What we're modeling out is over the course of a full year, assuming that social distancing measures end at some point."

The more researchers learn about the coronavirus, the better the assumptions will be and the more accurate the estimates will become. Updates to the Minnesota model will change in the coming days and weeks. You can read more about the model and see the slides from Friday's presentation here

Next Up

Related