How Is North Carolina Doing Compared To Other States?

As a long-time resident of North Carolina, I wanted to see how our state compared to some other states in the growth of COVID. For comparison I chose the next largest state, Georgia, and the next smallest, Michigan.  I also added Florida as the third most populous state and also one of the warmest. Finally, I also created a black line showing either the US rate or North Carolina's "share" of the US number based on population. The collection of charts (which will be updated daily) is here.

The results were pretty surprising.  Here is the growth in cases (click any graph to enlarge). 

Because US cases pretty much arrived at the same time, I did not attempt to graph them on a common timeline, like I have with countries.  North Carolina is well below its peers in COVID Cases.  The black line shows that we would closely resemble Michigan if we had our representative share of all US cases.  We are WAY below where I expected us to be.  At first I thought maybe we just weren't doing as many tests in NC as other states. As I've written, much of the growth reported in the US is actually due to a spike in testing. 

It turns out that NC has been doing tests at a higher rate than Ga and Mi (even though the CDC headquarters are in Ga).  We parallel the black line so closely, it seemed that we may have been getting tests based on our % of US population, except for March 24th.  Florida, which is twice our size, has done about twice our number of tests.  

Even with the additional testing, perhaps we're just not finding the cases that Ga and Mi were finding, but all the other states have suffered many more deaths than NC has to date:

We are so much lower on confirmed COVID cases and COVID deaths, there ought to be some explanation.  Here are a few possibilities:

• Atlanta and Detroit both have direct flights to China where RDU and CLT do not.  I looked at cities with china flights, and the cities with the most China Flights -- NY, LA, SF, and Seattle  -- are all hot spots.  Las Vegas was one of the smallest cities to have a China flight and their current number of COVID cases are twice the numbers of Iowa and Arkansas, who have very similar populations.   
• It could be that we're on a different timeline.  I'm assuming COVID came to the states at about the same rate, but Mi and Ga had our case counts only 4-6 days ago.  If this is true, it should become evident in the next several days. {UPDATE: A friend who's close to health officials in NC says that states are definitely on different timelines.}
• It could also be that NC's population is less concentrated than the other states.  The greater areas of Atlanta and Detroit's each represent 50% of their state's population, while greater Charlotte is only 24% of NC's population (less because some of that is in SC), and RDU is around 12%.

The last graph I made was on the rate of positive tests.

I show this test only after March 13th because the testing before then was scarce so the numbers had major shifts.  Again, NC has a different pattern than the others.  Mi and Fl, like the US is trending up on the percent of positive tests while NC has gone down.  This could be a sign that other states are doing better at where they use their tests, but given NC's low death rate it seems more likely that we actually have a lower infection rate despite doing more tests than Mi and Ga. 

The graphs from this blog are all conveniently located in a tab in my spreadsheet and that data is on another tab.  Feel free to share https://tinyurl.com/NCCOVID with others who want to track this information. 

Morning Update for Friday March 27th

This has been quite a week.  While the growth in cases in the US is sobering, here are some things to consider:

First, last Saturday, I explained how our rise in cases is connected to our rise in testing.  Here is today's common timeline on new cases:

Click to Enlarge

And Look at the Number of New Tests Done

Click to Enlarge

As I wrote about last Saturday,  most of the growth we're seeing is from the rapid increase in testing we're doing.  Just look at the increase in daily tests we've had.

Click to ... you know

We've gone from testing underachievers to doing several times more tests than any other country in the course of 10 days, completing 100K tests just yesterday!

Another thing to look at is our number of cases relative to population:

While we're still climbing, when you consider how much larger we are than most of the other countries with high rates of infections, we're just in the middle of the pack.

About 4 out of 5 COVID cases are considered mild -- some coughing, some fever, but nothing that would put someone in the hospital.  Perhaps many of the incremental cases we're finding from more testing are these mild cases.  One way to test that idea is to look at the number of fatalities in different countries:

Note that this graph isn't adjusted for population, its just number of cases.  While our case rate is going up, and every COVID death is tragic,  this graph is in stark contrast to the first one showing case growth.

Finally, I want to show you one other interesting graph.  In this one, I'm looking at the rate of positive tests as a % of daily tests.

 If you look at South Korea, who now looks like they really did an incredible job of containing COVID,  by now, their positive rate was relatively flat. If the do the same number of tests per day, you would expect the % of positives to go down as the number of cases go down.

Our cases dropped for a bit, before we started really increasing the test rate, but now its climbing. We went from 6000 cases a day to 100,000 cases a day from the 14th to today, you would expect the number of positives to go way down as we got outside the critical cases at the hospitals and out to more people who just suspect they have it, but instead it's gone up. I can think of two explanations (and if you have others, please comment).  The first is that the number of cases we have is increasing even faster than our test rate.  If so, that would line up with some of the more dire predictions of what's in store for us.  The second is that we've gotten smarter about who we test.  Early on when testing was in very short supply, its possible we didn't send nearly enough tests to hot spots like New York and Washington State because it would mean not sending many at all to places like North Carolina.  We also have some great tech minds using the systems that get so many people to click on ads now using those systems to pick the right people to test.  I hope it's the second case.

At 100,000 tests per day, we're also now in a position to randomly test people in different communities to get a sense of what the real infection rate is and where COVID is most concentrated.

One thought to take into the weekend -- last month, most of our brightest tech minds were spending the day trying to figure out how to get you to click ads. Now their trying to see the path of an invisible virus.  Some of our brightest medical minds were working on drugs and devices that wouldn't see the market for 10 years. Now they're getting stuff done in weeks. I've always felt the US is a magnet for people who want to get things done.  I think we're beginning to see what we're capable of in a crisis like this, and I think the rest of the world -- particularly countries just beginning to be affected by COVID -- will be grateful we were part of this pandemic because of the mass of innovations they will benefit from.

Has Climate Had An Impact On COVID Growth?

There's been a lot of speculation about how much warmer temperatures might help us in the fight against COVID.  There's reason to believe it could.  The flu is a seasonal virus that grows rapidly in the fall and practically disappears in the spring.  When I started building my COVID Tracking Spreadsheet,  I made sure to select countries from the northern, southern, and equatorial regions so we could compare them.  Because it helps to compare cases on a common timeline where Day 1 for each country is where they have > 100 cases, it has taken a while to get enough data -- most of the equatorial and southern countries only recently joined the > 100 case club.

Now we can take a first look at comparing how countries in different regions did.

Click to Enlarge

To make it easier to compare,  all northern countries are Yellow, equatorial in Green, southern in Purple. The US is in Blue (while technically India is a northern country, I grouped it with the equatorial because of its similar climate for most of its population).So far, the Equatorial countries are on a very different growth path than the northern ones.  There is no northern country with a lower case amount than ANY equatorial country in my group - regardless of population or economy. 

When I first say it, I wondered if its possible the equatorial countries aren't testing as much as the northern countries so they just aren't discovering cases.  Let's look at the number of tests each country had competed by Day 12:

Country	Tests as of Day 12
USA	15,535
Italy	31,000	estimate
S Korea	125,851
France	21,620
UK	50,422
Canada	142,154
India	24,254
Philippines	2,147	as of day 11
Malaysia	14,000
Thailand	14,086
Vietnam	30,548	as of day 4
Australia	52,663

While there are quite a range of testing, the differences don't seem enough to explain the difference in trends.

Warm climate still isn't necessarily the only reason for the slower growth. Most of these countries had several more weeks to prepare for COVID and may have done a better job at testing and isolating early cases.  It's also important to note that these countries aren't acting like COVID won't be a problem for them.  Yesterday India ordered a major shutdown of their economy to hold off the virus (although they are only testing a couple thousand people a day in a country of 1.3 Billion!  Rwanda, another country on the equator has also shut down schools and businesses.  If the warmer climate will tame COVID, their leader's aren't counting on it. 

As the spring progresses, it will become obvious if that is enough to save us from COVID.  For now, its best to keep R0 low by practicing social distancing, self quarantining if you have a fever or cough, and washing your hands often. 

Our Case Number Keeps Climbing, Do We Need More Extreme Measures?

It's been 12 days since we started to impose extreme measures to stop the spread of COVID, and our number is still going up fast, even compared to other countries.

Click to Enlarge

As I explained previously,  we're seeing a big rise because we're doing many more tests.  We now have done more tests than any other country, even those who started dealing with COVID well before us. 

Notice how the slope of the test line matches the slope of our growth line?  The growth we're seeing is mostly the process of discovering people who had the virus before we started social distancing and other efforts.  To illustrate, take a look at this graph:

Click to enlarge

The red line shows a possible number of cases that could have been in the US at the beginning of the month (it could be much more or somewhat less, but it will serve for this illustration).  The blue line shows the number of cases we've actually counted based on testing.   The space between them represents the "hidden cases" that are in the US but haven't been tested yet so aren't part of the count.  Notice how the red curve flattens after the extreme measures are taken.  This is also an estimate, but as I explained yesterday,  the curve DEFINITELY significantly flattened after that day.    The thing is we cant see it yet because we're still working through the volume of hidden cases.  In time, our testing will catch up with the actual number of cases and we'll see a rapid drop in the number of new cases, but until then we really can't see the effect of the shutdown. 

So do we really need to add even more extreme measures to stop the spread of COVID?  Maybe, but we really can't know just by looking at the rise in cases now.  

One more thing:  Although the red line continues to go up through the end of this graph,  if  you could see it a little further, you would see it start to curve down sharply because on average someone recovers from COVID 22 days from when they have symptoms. So on the same upward curve of new cases starting March 5th will be a downward curve of recovered cases starting March 27th. 

To see the impact of Extreme Measures, look at R0

Its been a week for many of us since social distancing has closed schools, events, and many businesses.  In my last post, I explained how the current growth is about discovering cases that existed before these measures were taken. Even if nobody in the US caught COVID (and I'm sure many did) since last Monday we would still see the new cases spiking because of the amount of tests were now doing (up again by a almost 50K on a Sunday!)

So how will we see the effect of these measures and what are they likely to be?  To start with this question, you have to learn about how a figure Epidemiologists use to measure and predict the growth of a pandemic.  R0 (called "R naught") is the measure of the number of people someone with a virus will infect on average during the time they are contagious.  There's a lot of intense math involved in calculating this number, but if you knew R0, you would expect the number of cases in the future to be   Present Cases x R0 ^ Generations (^ means "to the power of") .  Generations is the time for someone infected to become contagious themselves.  The math to measure R0 based on actual data can be very complicated and can vary widely.  Fortunately, all you have to know for this post is that R0 has three inputs:

• The average number of people in a day exposed to someone who is contagious
• The average probability someone exposed to a contagious person will be infected themselves
• The number of days someone is contagious

R0 is these three factors multiplied together.  We can guess at these numbers, and large numbers of researchers around the world are working to discover them for COVID, but to know the impact of a preventative measure on R0, we just have to know we've reduced one of these factors.  

If we estimate that on average the number of people someone in the US is exposed to in a day has dropped by 75%, then R0 has also dropped by 75%.  To see the difference, let's look at the current number of confirmed active cases in the US and how the cases grow using an R0 with no measures and one with social distancing:

If you factor in people who are recovering from COVID, the red line should turn downwards relatively soon while the blue line would continue to move up.  Remember that we're seeing growth from cases before social distancing -- ones where people had symptoms and others where it could take 7-14 days to show up -- so we'll still see that growth.  It normally takes around 5 days for a test to even show positive once someone is infected, so the effect of the last week will only start having an effect now.  The real fruit of those efforts will take a while longer to show up on reports.

 Social distancing isn't the only thing that has probably changed R0 in the last week.  The factor for the number of days someone is contagious speaks to the number of days they are spreading the virus.  Hopefully, most people running the typical starting symptoms -- low fever, cough, running nose -- isolated themselves very early compared to a few weeks ago when it would be typical to assume its just a cold and continue mingling with the public.  The 14 days that typically figures into R0 for someone spreading their germs is probably only a few days now.  That's another big R0 discount. 

The third factor is the probability someone exposed catches the virus. Washing hands regularly and staying away from common areas where you may come in contact with droplets left behind from someone with the virus should reduce this probability.  There's also a plausible reason to believe that warmer weather reduces the amount of time the virus can survive in droplets and on surfaces, so warmer weather could reduce this factor all by itself.  

Public health workers see R0 as a threshold.  When its higher than 1, people are infected at a higher rate than they recover and get off the roles so the number grows rapidly.  When R0 is less than 1, you have more people recovering than are replaced by new infections and the number of active cases shrinks rapidly.   Estimates for R0 in the US are around 2.0,  so if the steps taken last week reduce it by 50% or more, it could have a major effect on on the peak number of cases in the US.  This view was echoed in the highly respected medical journal The Lancet in an article from March 9th  (back when we only had 700 cases and we still had sports).

Social distancing measures reduce the value of the effective reproduction number R. With an early epidemic value of R0 of 2·5, social distancing would have to reduce transmission by about 60% or less, if the intrinsic transmission potential declines in the warm summer months in the northern hemisphere. This reduction is a big ask, but it did happen in China. 

As we all know,  the Lancet got their "Big Ask" and we went to extreme measures to reduce transmission.  

Of course there will be some people who don't reduce their social connections, don't quarantine themselves when they have COVID symptoms, and aren't taking more care to wash their hands and take other common sense measures,  for R0, is only matters what has changed on average.  Specific cases either way aren't enough to change R0.

The bottom line is that if you think we've cut in half any combination of the number of contacts on average people have with each other, the number of days on average someone contagious is out in the public before self quarantine, and the chance a virus particle makes it into your body, then you have good reason to expect to see improvement as the reports continue. 

Data and Chart Update for Tues 3/24

Here's the data from this morning:

Country	Cases	Critical	Deaths	Cases/1M	Tests Done	Tests/1M
USA	43,734	795	533	132.1	294,056	888.39
China	81,093	2,274	3,270	56.3	320,000	222.33
Italy	63,927	2,498	6,077	1057.3	275,468	4556.05
South Korea	8,961	59	111	174.8	348,582	6799.08
Spain	35,136	939	2,311	751.5
Germany	29,056	2	123	347.2
France	19,856	1,122	860	296.4	34,570	516.05
UK	6,650	20	336	100.1	78,340	1179.11
Canada	2,091	1	24	55.4		0.00
India	499	0	10	0.4	19,974
Indonesia	579		49	2.2
Philippines	462	1	33	4.4	1,793	17.09
Vietnam	123		0	1.3	17,148	176.17
Thailand	721	1	1	10.4	11,855	171.71
Malaysia	1,518	15	14	46.9	18,416	568.99
Singapore	509	14	2	90.7
South Africa	402	0	0	7.1	645	11.37
Argentina	301	0	4	6.8
Australia	721	1	7	28.3	61,848	2424.44
Chile	746	6	2	41.3
New Zealand	102	0	0	21.7	584	124.10

I've also added some rankings within the group:

Country	Population	Cases per population Rank	People Tested Rank	Tests/1M Rank	Deaths as a % of cases Rank
USA	331,000,000	6	3	5	11
China	1,439,323,000	9	2	8	15
Italy	60,462,000	1	4	2	21
South Korea	51,269,000	5	1	1	12
Spain	46,756,000	2			18
Germany	83,697,000	3			7
France	66,990,000	4	7	7	16
UK	66,440,000	7	5	4	17
Canada	37,742,000	10			10
India	1,339,000,000	21	8	14	14
Indonesia	264,000,000	19			20
Philippines	104,900,000	18	12	12	19
Vietnam	97,339,000	20	10	9	1
Thailand	69,040,000	15	11	10	4
Malaysia	32,366,000	11	9	6	8
Singapore	5,612,000	8			6
South Africa	56,720,000	16	13	13	1
Argentina	44,270,000	17			13
Australia	25,510,204	13	6	3	9
Chile	18,050,000	12			5
New Zealand	4,705,882	14	14	11	1

On the common timeline, the US continues to show an acceleration of new cases compared to other countries, but please see my post on why the rise in testing is primarily us discovering cases created before the US took extreme measures last week. 

Our focus on testing is also encouraging.  See this chart on the common timeline.  

The spike in testing relative to other countries looks a lot like the spike in new cases in the previous graph, supporting the idea that the rise in cases has more to do with increased testing than actual new cases.  As of today,  we have also tested more people than any other country on the common timeline (ie 22 days since crossing 100 cases).   We're still actually accelerating the daily testing.  See this chart of tests per day:

We're conducting way more tests per day than any other country and have increased the lead every day this week.  With the release of new, easier tests where the samples can be taken by the patient, we should see this number keep rising -- and hopefully the other countries' testing rates will also rise as we share our technology with them. 

Later this morning,  I'll be posting about how the steps we've taken could effect the growth of COVID in the US.  All Charts and source data are available at my sheet here.