Thread by @reichlab

Since mid-Jan, @cmyeaton, myself, et al have been sending out weekly reports about influenza-like illness (ILI) in the US. It started out as a curiosity, with us thinking that maybe #COVID19 was already here in the US and we just weren't seeing it yet.

https://twitter.com/reichlab/status/1221176521253146627

1/

These reports are not peer-reviewed (we will, when we find time to breathe again, submit a version of this to a journal). For now, we continue to find the report to give us useful intuition about this VERY rapidly changing situation.
2/

This latest report, which uses ILINet data through 3/14, provides, to my eye, the most visible signs yet of a possible signal of #COVID19 in ILI data. It's not conclusive, b/c this is not lab-confirmed data, but the signal is stronger than past weeks.
github.com/reichlab/ncov/…
3/

What is particularly interesting is the combination of ILI visits (as a fraction of all outpatient visits) going up AND the fraction of positive flu tests going down. See this pretty, if worrisome, graph from Evan Ray.
4/

You can download the report for yourself, or even, when the @CDCgov API comes back online (did everyone trying to download the data this afternoon break the data backend?) run the report for yourself. All our code and data is on GitHub.
github.com/reichlab/ncov/…
5/5