COVID19 good news on March 31

For the majority of the articles I've written, I've tried to stay focussed on signs of progress and I've held an optimistic view on where things are headed. In the last article, I broke from that a bit to introduce some troubling conclusions based on the data and the models. Today, I want to get back to highlighting some of the positives again.

A lot has changed since that last article. President Trump announced the 30 day extension to the nationwide lock down. Now April 30th is the new target date for opening things back up. We should keep in mind though that this will vary by state. Some parts of the country are in the flattening phase of their battle with COVID19 and some are just starting to experience explosive growth and announcing their own shelter-in-place directives.

So let's get to the good stuff! Scott Gottlieb published the first credible document about what it will take to get back to normal titled A Roadmap to Reopening.

Here are some important quotes to be aware of in the document, but the whole thing is a good and easy read:

Slow the Spread in Phase I. This is the current phase of response. The COVID-19 epidemic in the United States is growing, with community transmission occurring in every state. To slow the spread in this period,1 schools are closed across the country, workers are being asked to do their jobs from home when possible, community gathering spaces such as malls and gyms are closed, and restaurants are being asked to limit their services. These measures will need to be in place in each state until transmission has measurably slowed down and health infrastructure can be scaled up to safely manage the outbreak and care for the sick.

Phase 1 is very similar to what I wrote about 2 days ago describing the various phases.

State-by-State Reopening in Phase II. Individual states can move to Phase II when they are able to safely diagnose, treat, and isolate COVID-19 cases and their contacts. During this phase, schools and businesses can reopen, and much of normal life can begin to resume in a phased approach. However, some physical distancing measures and limitations on gatherings will still need to be in place to prevent transmission from accelerating again. For older adults (those over age 60), those with underlying health conditions, and other populations at heightened risk from COVID-19, continuing to limit time in the community will be important.

This would feel like a major accomplishment to get to this phase. So let's see what that takes:

Trigger to Move to Phase II. To guard against the risk that large outbreaks or epidemic spread could reignite once we lift our initial efforts to “slow the spread,” the trigger for a move to Phase II should be when a state reports a sustained reduction in cases for at least 14 days (i.e., one incubation period); and local hospitals are safely able to treat all patients requiring hospitalization without resorting to crisis standards of care4; and the capacity exists in the state to test all people with COVID-19 symptoms, along with state capacity to conduct active monitoring of all confirmed cases and their contacts.5

If we can greatly expand testing via the recent announcement by Abbot labs of the 5 minute test and we get 14 days sustained reduction in cases, we can expect states to re-open things. So of course, I'm now attempting to predict that 14 day period. A key question about the document: Does he mean sustained reduction in new cases or active cases? These are vastly different timelines. We will attempt to predict both using the data we have.

Another bit of good news is that Italy has now developed a pronounced downturn in new cases! One of the sources of my frustration last article was not having data that a lock down approach to mitigation of spread was working very well. Since that article, I now have several key pieces of data demonstrating the effectiveness.

Italy is still behind the S.Korea flattening and reduction of active cases trend when normalized around days from peak. However, this is now a clear downtrend and reduction in new cases starting on March 26th. Italy is now 9 days from their peak on March 21st, but has only reduced new cases by about 38%, whereas S.Korea had reduced by 80% 9 days after their peak. One of the potential reasons for the lock down method being less effective in reducing new cases may be described in this report about home quarantine not being the right path for the mild cases. In Wuhan, they concluded that those who only had a mild form of the infection would go home quarantine and spread it to their families and still go out and spread it in common places. Italy started involving the army to police streets a week into their national lock down to further reduce unnecessary movements.

Still, the Italy data points to significant progress with the lock down method, that we can expect to show up in the US data soon as well.

I learned about this site last night and the data it shows of COVID19 related fevers by county is pretty awesome. Apparently Kinsa Health has sold over 1 million internet connected thermometers (IOT FTW!) and built this database and trend map of the temperature readings. Now, what's doubly cool, is that while they built it to look at influenza trends, they've been able to isolate the atypical readings as COVID19 related by factoring out the flu numbers.

The above map basically correlates with all of the reported COVID-19 hot spots and the best thing about this data is that it is predictive! New cases and New deaths in the US lags reality by at least 1-2 weeks, which is what makes following the data challenging at times. It's a very long feedback loop. This fever data is near real time! They were even able to see the post-spring-break Florida Covid-19 explosion before the hospital admissions had started.

Now check that out! The legend above shows that blue indicates a decrease in abnormal temperature readings over a 7 day period, which means that around the country things are getting better and in some cases quite rapidly. If the 14 day downtrend of new cases is the criteria (vs reduction in active cases), we could already be in that downtrend.

From the previous articles, my model predicted we hit our peak number of new cases on March 28. Using Italy's data, the model predicts that we start a pronounced downtrend in new cases starting on April 3rd. 14 days later would be April 17th. However a downtrend in active cases will not start until April 19th according to the model, which would take us to May 3rd. I have not yet attempted to predict WA or any other state though. One of the reasons is that the noise in the data is significant because of lower sample sizes. I will work on this as again, the criteria for triggering a re-opening is a state by state evaluation.