Why is the state government hiding the questions its Covid-19 tracers are asking?

On November 20th, Gov. Walz shut down Minnesota’s bars and restaurants again despite their being linked to only 1.7% of the state’s Covid-19 cases since June. What makes this even worse is that the way the state’s Covid-19 contact tracers work makes it highly likely that this is an overestimate.

Asking the right questions to get the right answers

When I reported the Department of Health’s numbers for Covid-19 cases originating in bars and restaurants, I received a message saying:

The three questions that they ask as part of their contact tracing questionnaire are in the 14 days prior to onset of symptoms, did you (1) visit a bar/restaurant; (2) attend a large gathering; or (3) come in contact with someone who was positive. In other words, when they say that they have no data to suggest that other retail has high rates of cases is because they are simply not collecting such data.

If this is correct, then we would expect to get a higher number of cases linked to bars and restaurants because they are the only kind of establishments we are asking about. If somebody contracted Covid-19 at Target or Walmart say, the state’s tracers would never find out because they don’t bother to ask. The questions assume that the only place you can contract Covid-19 is at either a bar or restaurant or a large gathering. By excluding all other options, they make it more likely that these places will be over-represented as vectors of transmission.

Yet, even with this tilting of the table, Minnesota’s Covid-19 contact tracers are failing to generate large numbers of cases linked to either bars or restaurants. Figure 2 presents data on cases by likely exposure from the Department of Health’s most recent Weekly Covid-19 Report. We see that ‘Community (outbreak)’ – the category which includes bars and restaurants – accounts for just 4.3% of cases.

Figure 1: Share of Covid-19 cases by likely exposure

 Source: Department of Health

Furthermore, it seems that Minnesota’s bars and restaurants have become relatively safer over the course of the pandemic. Minnesota’s second surge in cases began in mid-September. Looking at the Department of Health’s Weekly Covid-19 Reports for September 17th and December 10th – the most recent – we can see where the new cases in this second surge have been traced to, as Figure 2 shows. Over this period, the share of Minnesota’s Covid-19 cases traced to ‘Community (outbreak)’ is down to just 2.9%. And this is with the contact tracer’s leading questions.

Figure 2: Share of new Covid-19 cases by likely exposure, September 17th to December 10th

 Source: Department of Health

Indeed, over the course of the whole pandemic we have no idea where 59.6% of cases originated (adding together the shares for ‘Unk./Missing’ and ‘Community (no known contact)’) and over the course of the second surge this number has risen to 65.2%. I’ve said before that Minnesota’s Covid-19 tracing regime is being overwhelmed because of the sheer number of infections. Its increasingly poor results might also be down to the exceptionally poorly constructed questions it is asking.

Why are the questions a secret?

This all assumes that what my correspondent was telling me is accurate. It would be nice to confirm that and one would think that it would be an easy matter for the authorities to clear up by telling us what questions the tracers do, in fact, ask.

But, for some reason, they refuse to:

There is no reason whatsoever why the Department of Health cannot provide the list of questions its contact tracers are asking. The Governor is closing business and wrecking livelihoods based, apparently, on this data. Those affected have a right to transparency.

John Phelan is an economist at the Center of the American Experiment.