Updated data sites (take them with a grain of salt)

Some places I'm watching...

• National Flu Monitoring for Symptoms here (not yet updated for week 11 through March 14)
• Optimism in the face of the Imperial study here.
• Uptick in NY ILI cases; by a lot in recent days here.
• ICU beds by county here.
• COVID test tracking; how widely are we really testing? State by state here.
• Predicting flu symptoms via connected thermometers here. 

NY Influenza Like Symptoms Symptoms Rising

I've been waiting to see when the NY surveillance website will update to show numbers from the weekend and then on Monday. The reported cases and percentage of ILI to overall visits is the highest it's been all flu season. The chart below shows the percentage by day.


On Monday the percentage was 11% of visits due to ILI. There was a surge in ILI presumably due to flu in late January and early February where the percentage got up to 9%. If we look at straight-up counts, the number of visits on Monday due to ILI were 1,257 which is also the highest it's been even during the presumed flu period where the high number of ILI visits was 1,174.

One might expect the normal cadence of overall visits to remain somewhat steady even if the COVID-19 results in much higher numbers of people coming to be treated. If that's the case then 10,170 visits might remain steady. If the actual numbers double the percentage won't double, so I think worth looking at both percentage and actual counts.

I don't know how visits translates to capacity, but that's the million dollar question. If the number of visits grows at 30% per day that report ILI when will it overwhelm capacity in the NY healthcare system? It will be interesting to see when it hits a doubling point.

Deaths in Italy vs China

I’ve avoided looking at the data on deaths because I think it’s highly retrospective. Also it is a super downer. And I think there’s enough data on the curve in China to ask what it looks like and what can we learn about what is happening in Italy and the US.


The number of new deaths in Italy last week was much higher than it ever was in China. Almost three times higher. The graph above shows the new deaths by week; the orange is the actual week and the grey is time shifted to show Italy and Hubei at similar times in their outbreak (even thought they are six weeks apart). Last week was devastating for Italy.

An important point here is when did social lockdown occur? In China it is reported to have happened in what I’m calling "week 0" (there were 40 deaths and Hubei was locked down January 23, 2020). Italy waited a week (relatively speaking) and locked down when there were 332 deaths in “week 1" (on March 9, 2020).

There are a million other factors from how patients are treated, demographics, etc. But it isn’t hard to imagine that one big factor in the huge numbers last week in Italy includes the additional week of socialization. In the US we are in “week 0” with 46 deaths this last week. If we can effectively shut down our interaction maybe we have a chance at looking more like China than Italy.  

Compared to the H1N1 Pandemic

I’ve been asking myself why the response to this COVID-19 is so much more than the response was to the pandemic in 2009. Honestly, I don’t remember the 2009 pandemic very well. It didn’t hit my radar in any substantial way. I certainly didn’t change my behaviors to limit the spread.

In reading through information on the CDC about the 2009 H1N1 pandemic I learned a few things that help me understand what is similar and what is different.

In 2009, the flu was widespread and from April 2009 for about a year it infected 61 million people in the US. There were 274,000 hospitalization and 12,470 deaths. The death rate was 0.02%.

The H1N1 flu had an R0 (measuring how contagious it is) of 1.2 to 1.6. The WHO puts the COVID-19 RO between 1.4 and 2.5.

If we assume similar contagiousness with the estimated death rate for this Coronavirus it becomes clear why this is so much different. If 60 million people get the coronavirus and the death rate is 1% (the low end of the estimates right now), we’re talking about 600,000 people in the US dying. 

This is really, really different. The stakes are much higher.

Symptom Reporting

Someone tweeted a link to a daily surveillance data site of NY “Influenza Like Illness (ILI)” here. When you click on the data point that shows “ILI” for 2020 it opens a dashboard that you can get daily reporting numbers. Seems like a great way to get inference of what's happening in these outbreaks given we don't have widespread testing.

As of today, the latest data is 3/12 and the trajectory is going up steeply. It seems like we’re seeing 9% of visits with ILI as of Thursday. It suggests an increase in visits with symptoms like COVID-19. I couldn’t find anything similar for Washington state.

I looked for National data as this seems to be something we’re doing across the country (ILI surveillance). The CDC reports weekly numbers here. https://gis.cdc.gov/grasp/fluview/fluportaldashboard.html It shows that Washington state has had elevated incidents of ILI. 

I think it’s also interesting to point out that we’ve had an active flu season and you can see we haven’t hit the levels we were at in December yet. It is also interesting to note the national data also tracks confirmed flu cases. While we’re seeing elevated ILI we are seeing confirmed flu cases still dropping. At least for the most recent week.

Looking at historic data we’ve seen rates as high as 10% during flu season. In 2018 it was a weekly average of 9% in NY according to these sources. It only lasted for two weeks. I think it will be telling if we pass 10% in our locales for a sustained period of time.