Minnesota's coronavirus death projections are much higher than others. Why?

The nation's top infectious disease official, Dr. Anthony Fauci, cut his projection of the U.S. coronavirus death total by half, crediting social distancing. Fauci's new prediction -- that 60,000 Americans would die -- was dramatically different than the numbers Minnesota's health commissioner and governor are using.

Gov. Tim Walz said Thursday that Minnesota's latest modeling shows that between 6,000 and 36,000 Minnesotans would die. Walz said he relied on the projections when deciding to extend Minnesota's stay-home order through May 4. The order has effectively closed thousands of businesses and put 385,000 Minnesotans out of work.

"What I’d like to do is sit down for all of us to understand," Walz said. “These models were not meant to predict deaths. They’re meant to predicts trends and differences as they go."

Minnesota officials have never fully released their modeling, which was initially run in mid-March and led to Walz's first stay-home order. The second version landed this week, just before Walz extended the order.

State Health Department officials have scheduled an 11 a.m. Friday briefing with reporters to walk through the details.

The top Republican in state government, Senate Majority Leader Paul Gazelka, said Thursday that he no longer supported Walz's stay-home order.

"We have to get on with our lives," Gazelka tweeted. "We are ready for the surge now. Why shut MN business down for a NY sized surge?" 

The comments marked a significant change in tone for Gazelka, who had said Wednesday that the governor's decision to allow a few businesses to reopen was "welcome news."

Minnesota's model is wildly different than the University of Washington's projections, which show 456 Minnesotans will die of the coronavirus -- less than one-tenth of the Minnesota model's best-case scenario.

Professor Ali Mokdad, one of the University of Washington's researchers, cautioned not to pit one model against others. He said his model projects mortality based on state death counts that are updated daily.

"We’ve been doing very well predicting mortality," Mokdad said in an interview. "We monitor the data on a daily basis, what we have been predicted and the number of cases. That has been spot on. That is the strength of our model."

The University of Washington model has been revised twice since initially showing that 2,000 Minnesotans would die.

Minnesota officials have not provided as much detail about their model. Pressed for an answer, Walz told reporters Thursday "that model doesn't run that way."

Minnesota's projections take in more data and run 16 months, as opposed to Washington's model, which ends in August, Walz said.

"There’s different ways to approach this," Walz said. But I think focusing solely on (a projection), that it doesn’t line up, is not understanding what this data does."

Despite their differences, the researchers credited social distancing for having the biggest positive effect on the infection rate.

"We better be careful that we don’t say, OK, we’re doing so well, we can pull back," Fauci said in an interview on NBC. "We’ve still got to put our foot on the accelerator when it comes to the mitigation and the physical separation."