Dribbles of information

The governor extended the shelter-in-place order until May 4; it originally was to end today. I am working hard to be grateful that my family and I have everything that we need for now.

For the last two days of published numbers, coronavirus tests in Minnesota have been 94% negative; that’s not very useful without knowing the testing numbers and thresholds for the different groups involved. Nursing home cases continue to increase, as can be seen in the increase in median age, and in the breakdown by residence type.  It appears that there are increasing cases among health care workers.

Since we’ve already been shut down for two weeks, the continued spread is through “essential” activities.  Detected cases could be as little as 1% of actual cases.

The state has released a few slides giving some of the details of the scenarios being modeled. They are assuming that social distancing will do little to “flatten the curve”, and will at best delay the peak ICU demand by some weeks; possibly out into July, given the present shelter-in-place order.

Present ICU surge capacity is projected to be about 1000 beds short of what would be needed at the peak, based on the graph.

The scenarios include various degrees of social distancing through the summer, expressed as percentage reductions in contacts: 20% and 50%, compared to 80% for shelter-in-place. They claim to be modeling far enough out to see a second peak in later months, but no further information is given about them.

Digging further, they have put out some technical information on the model. For the most part, I have no technical quibbles with it; my computer modeling experience is in large atoms, which while being governed by extremely complex physics, will at least behave consistently within the given physics. Things like coronavirus spread and climate change involve entire systems changing over time, so the whole thing is a moving target (and is very, very likely to diverge from the model’s predictions).

That said, their range of estimates for R0 looks overly conservative to me. A single person could infect dozens of other people, even now. They are figuring R0 = 3.87, on average.

I’m doubtful now that there will be much of a lull in the summer; they certainly didn’t model one. In the Twin Cities, a lot of activities come to a halt in the summer, as people go off on summer vacations or up north to their cabins, but there is going to be less of that this summer.  So my guess is that the first big peak is going to include almost everyone; even efforts to isolate the nursing home residents are failing already.

The other thing I noticed is that their ICU survival rate seems extremely optimistic, with a mortality rate of only 0.111 per 10 person-days for octogenarians.  Or perhaps I’m misunderstanding that part; I would have guessed it was no better than a 50-50 chance of survival for that age and severity, and that  many of them would die in the early days of their ICU stay.

Projected deaths for the state are now 9,000 to 36,000.

Hopefully there is a corresponding model for the economic effects of shutting down. But I haven’t heard that there is.