April 19 - COVID-19 in Westchester County, NY

The usual charts are shown below followed by a discussion on mortality rates and reproduction rates.

The New Cases in Westchester County are within the same range (700-1000) as in the previous 25 days. Note that most entire states do not have that many New Cases per day! The "spread" or duration of the outbreak is longer than predicted by most models.  It is much longer than the Wuhan event (if you believe their data).

But the duration of the epidemic in Westchester may not be that much longer than Bergamo....yet!  The orange line is a fit of the Westchester data using the Bergamo duration.

If there is any good new it is that the Percent Testing Positive appears to be decreasing despite high  testing volumes.

The death rate (Orange Line) also appears to be trending slightly downward.  Note that this is a logarithmic plot.

Mortality

The White House Task Force Briefing showed a slide that Dr. Birx referred to as "Case-Fatality-Rates", but it was really a Population Mortality Chart.  Case Fatality is how many diagnosed people die. If more people are diagnosed because you are testing more, then your Case-Fatality should go down, because you are catching more of the mild symptomatic or asymptomatic cases. In NY the testing has been at a per capita rate much higher than the rest of the world so the case fatality would be expected to be lower than a country where diagnosis was by symptoms rather than a test or where only the very sick were tested.  In other words, COMPARING CASE-FATALITY-RATES IS DECEPTIVE.  Below are the Case Fatality data for major impacted countries.

The Population Mortality Rate is the deaths per capita.  This is also deceptive because most COVID deaths have comorbidities.  The cause of death can depend on the practice of the national health authority.  Which of the comorbidities is the major one? Or is it counted as a COVID death if any one of the comorbidities is COVID? (That is the US policy).  The rate is also a factor of how the country was able to slow the spread. It is doubtful in any case that the virus is 'Contained', only slowed.  For many of these countries, including the US, the pandemic is far from over.  More people will die and the rates will go up.  The population mortality rates in NYC (106) and Westchester (70) are much higher than the country at large because of the extent of infection.  Once again, the data are interesting, BUT DECEPTIVE.

White House Briefing Chart of Population Mortality

John Hopkins Chart of Population Mortality

Reproduction Rates

Reproduction rates are a measure of the contagion of the virus. A rate of 2 means that one person infects 2 new people.  A rate greater than one is an epidemic. Rates change based on region, weather, social distancing and other factors. Like the case-mortality-rates and population-mortality-rates, this rate can also be deceptive.  New York Governor Cuomo referred to the "New York Reproduction Rate" several times in his Briefing yesterday. Many of his assertions are very shaky.  Here are some of his statements and my response:

1. "We brought the reproduction rate down to 0.9 by distancing measures" - This is a bogus mumber based on confirmed cases.  What about all of the asymptomatic cases that remain invisible? Even if the rate is 0.9, what measures really got it there?  Maybe it was just closing all of the gyms and public transportation? (and not all of the other mandated closures.)  Or maybe none of the measures had any effect? It is unlikely, but maybe this is the natural course of the virus? There is no study to confirm that this "correlation is causality." (as Dr, Fauci would say)

2. "A reproduction rate of 1.4 is where we were at with no distancing measures" - Once again this is based on the initial testing which totally underestimated the spread of the virus in Westchester and the NYC area.  Others estimate the reproduction rate to be closer to 6 in a dense environment like NYC and lower Westchester..

3. "Reproduction rate of 1.2 would be back where we were" - As above, these numbers are all bogus. Cuomo is really putting too much value in the wrong numbers. The objective should be to keep the hospitalization rate manageable no matter the reproduction rate.  For example, right now the new case rate (ie. reproduction rate) is still quite high but fewer are being hospitalized.  Has he tried to understand the reason for that?

4. "Testing is the answer" - This assumes that people have a reason to get tested.  If you have no symptoms, which is common with this virus, you will not be tested, but will be a spreader. For the time being (until every NYer can get an instant test, which will never happen), one must assume they are positive and that everyone around them is positive. Antibody testing might be a partial answer, but Cuomo did not talk about that.

5. "Contact tracing is the answer" - In the NYC metro area, where people use mass transit, contact tracing is nearly impossible.  If the virus is spread by touch of hard surfaces, contact tracing is not feasible. It is foolish to place hope is this method of containment.  See example below.

6. "We contained the cluster in New Rochelle through contact tracing" - New Rochelle and its vicinity in Westchester County blew up into the hottest area in the country.  There were more cases per capita than New York City. Contact tracing did NOT snuff out this cluster! In fact, it had NO apparent effect on the trajectory of the epidemic in Westchester, which was disastrous!

April 18 - COVID-19 in Westchester County, NY

The normal charts are below.  They are followed by more news on antibody testing.

Westchester County continues to have a steady flow of New Cases, as does NYC and the surrounding metro areas.

The Percent Testing Positive is slowly coming down, but is slightly correlated with testing volume (the more tests the lower the percent testing positive). Westchester was consistly over 40% in early April and has been below 40% for the last three days.

Deaths in Westchester dropped below 30 for the first time in 10 days (Orange Line).

The epidemic curve for Westchester is getting wider and wider! Compare it to the curve from 18 days ago, April 1!  Nobody has proved that 'Slowing the Spread' is really reducing total infections or deaths....It is probably just delaying them.

Gaussian Fit of New Case Data in Westchester from April 1

Antibody Testing and the True Spread of Infection

Knowing how many people have been infected has a profound impact on the calculated risk of this virus.  If many people are asymptomatic, the mortality rate of the virus is smaller than thought.  If hard-hit populations, like Westchester, are really 50% immune vs. the confirmed cases of 2%, then the risks of relaxing personal contacts are less.

The length of time it has taken to do a random antibody study in New York is a bit frustrating.  Westchester is the ideal location for such a study. The spread is wide (over 2% confirmed infected), very deep viral testing (amost 6% of the population has been tested), and Westchester is well past its 8th week of high-infection-rate when the antibody production should be maxed. 

Mount Sinai (NYC) has an FDA approved test to measure the level of antibodies.  They have not done a large random study yet, but their work shows that patients had antibodies 14 days after infection with yet higher levels of antibodies at 21 days when the viral load was 'pretty much gone.' 

A study in Santa Clara County suggests that about 50 times more people were probably infected in the county than shown by the confirmed cases.

A similar study in Gangelt, Germany found a 15% infection rate vs. less than 1% confirmed cases.

    

All passengers on the Diamond Princess were tested for the virus (not antibodies). 634 of 3711 (17%) tested positive, many of whom had no symptions.  Who knows how many had the antibodies? 

San Miguel, Colorado offered the antibody test to its 8000 residents to help United Biomedical evaluate its test. Less than 1% had antibodies and another 2% were "borderline."  At the time there were no known cases in San Miguel. One month later there are only 13 confirmed cases in San Miguel.  Why didn't they test it in Westchester where the results would have been more interesting?  Maybe they were really looking for a 'clean' population so they could confirm false positives, which appears to be a real problem with these tests.

Do a majority of people in the NYC metro area now have the antibodies?  Let's find out now!

April 17 - COVID-19 in Westchester County, NY

The White House Coronavirus Task Force issued their "Guidelines for Opening Up America" yesterday. In order to begin the phased opening, states (or regions) must meet several "gating" criteria. Depending how you interpret them, Westchester could be ready soon OR Westchester is still many weeks away.

I have two minor problems with the plan.  First, the plan places no value on antibody testing for reopening.  What if random testing in Westchester finds that 50% of the population carries antibodies? That would lower the risk of reopening Westchester.  It also means the virus is much less deadly than thought (0.1% mortality).

Second, it puts too much weight on a reduction in cases and too little on hospital utilization.  Who cares if cases go up if they are not clogging the hospitals? For example, people under 30 are seldom hospitalized with the virus.  If they contract the virus as they return to work or school, who cares? They add immunity to the community in preparation for the scenario where a vaccine is not found.  (Remember...there is no vaccine for HIV or swine flu or...)

Gating Criteria for Cases
The guidelines recommend: "Downward trajectory of documented cases within a 14-day period OR Downward trajectory of positive tests as a percent of total tests within a 14-day period (flat or increasing volume of tests)"

Westchester, NYC, and Nassau County do not meet the first part of that criteria. See below. 

New Cases Trajectory

If you plot the NYC data based on 'Date of Swab' (as NYC is now doing) the trend is not yet clear as reporting lags by up to 5 days.

NYC Case Trajectory - recent data is incomplete

The State of New York certainly does not meet the criteria.  Most areas outside of NYC are still on an upward trajectory.  But the major driver for the increase in cases is the huge increase in testing.

NYS Case Trajectory - Yellow bars

The second part of this gating criteria gives the option of looking at the trajectory of the percent testing positive (instead of new case trajectory).  Looking at the results below, you could maybe argue that Westchester has a downward trajectory for this criteria over that past 14 days.

Percent Testing Positive Trajectory

It will be interesting to see how the NY Governor interprets these criteria.  He has already stated we are closed for another four weeks without even giving a criteria for opening.  Arghh!!!  We might revolt like Michigan. Our Mayor (Yonkers) has stated that 40% of the Yonkers deaths were in nursing homes and that the majority of the deaths in Yonkers were of people over 70 with underlying conditions.  It seems too obvious what we need to do!!!

Bergamo, Italy

I continue to use Bergamo as a comparator to Westchester.  It is the same size and same distance from country's biggest city (it borders Milan).  Some of the images are grim such as this article on the procession of caskets to the crematory ovens. I have not seen that in Westchester. Lombardy (the Milan region including Bergamo) is targeting May 4 for 'return to work,' but the criteria appear to be based on testing, of which there is not enough.

Westchester vs. Bergamo

April 16 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 15, 2020. Click on a chart to enlarge it.

We are 30 days into the National 'Slow the Spread' initiative, 37 days since the containment zone in New Rochelle, 44 days (Mar 3) since the first case in New Rochelle (Westchester) was confirmed, and 54 days since that first case had symptoms (Feb 22).  This first case had not traveled abroad recently. He had not caught the virus from a traveler (Patient 0), but from his community. Patient 0, who started the New Rochelle cluster, has not been found. The day after the first case was confirmed, nine of his family and friends tested positive. The virus had probably been circulating in Westchester since early February. At a reproduction rate of 6, there were likely thousands in Westchester infected by the time the containment zone was in place, making it useless. It is time for some reflection.

There is no doubt that the spread has been slowed in Westchester. The exponential explosion of an epidemic has been reduced to a steady stream of about 700 new cases per day with about 70 of those needing hospitalization and 70 persons released. The red line of new cases is flat and the cumulative cases (yellow) can be approximated with a straight line.  About 40 people are dying per day.  All of this has been true for over two weeks.  In engineering we call that 'reaching steady state'.

The case data for NYC and Nassau have this same trend.  There is noise in the data, but the case rate is not showing a decline.  New cases have been between 5 to 10 per 10K for over two weeks. [The decrease in hospitalizations in NYS and NYC contradict this, as do the decreases in deaths in NYC.]

Perhaps we are just on top of a very wide bell curve as seen with Bergamo, Italy below.  Bergamo had about 350 cases per day for almost three weeks before they started dropping.  And then it was another 10 days before they consistently were below 100 cases per day.

Bergamo province has 1MM residents, as does Westchester.  I have assumed the difference in the heights of the curves is due to the excessive amount of testing in Westchester, not any difference in the real infection rate. 2000 people have died in Bergamo compared with 640 so far in Westchester.

Bergamo, who is more than two weeks ahead of Westchester (if you believe my green and orange curves above), has still not reopened its bars, restaurants, and cinemas.

April 15 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 14, 2020. Click on a chart to enlarge it.

A decrease in the number of daily deaths is the last indicator of a lowered infection rate. Unfortunately it comes about four weeks after the infection rates were lowered (New Rochelle 'containment zone' was announced on March 10, Westchester was essentially quiet by the 14th, NYS stay-at-home was announced on the 20th) , that is if you believe that the 'social distancing' measures were the cause of reduced infections.  Unfortunately the feedback loop is long. It would be great if your phone beeped at the moment the virus invaded your body.."You've Been Infected."  Even worse, for some people there never comes a "beep."  They never have symptoms. They have no fever, they have no runny nose.  They slip past crude detection measures into our nursing homes and groceries. They innocently and unwittingly carry the virus to others. As we look for ways to 'reopen' we must assume everyone is a carrier. Perhaps recovered patients, those testing positive with antibodies, and those recently testing negative for the virus could identify themselves in some way, but otherwise we must protect ourselve from others AND we must protect others from US.

The deaths in Westchester have gone down three days in a row (orange line) even though the deaths in NYS jumped up slightly..

Deaths in NYC continued to drop.  The increases in deaths are coming from Long Island and upstate.

Here is the model of IHME which predicts the deaths in NYS to drop below 100 on April 25.

The new cases in Westchester dropped again (red line) and appear to be following the downward slope of the gaussian curve (blue line).  The orange line assumes an epidemic duration similar to Bergamo, Italy.

Statistically, Westchester (and metro NYC) is still riding on a plateau where the flow of new cases has not ebbed much.  If the virus has an incubation time of only a few days (which I am beginning to doubt) then the virus is still smoldering in the population and its contagion is such that despite the social distancing it is continuing to spread.

The new antibody testing will not be of much help to many regions of the country, but here in Westchester it could be of real value in showing if we are getting close to herd immunity. Could it be that more than 80% school-age children here are already exposed and immune? Or is the fraction of the population who have been exposed closer to the the number who have tested positive (2%)?  Let's get some testing done!

April 14 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 13, 2020. Click on a chart to enlarge it.

Hospitalizations at the state level and in NYC continue to go down.  Unfortunately, Westchester does not publish their hospitalization rate. The history for NYC is shown below.

NYC Hospitalizations

Deaths in Westchester went down for the second day in a row (orange line). That is not enough data to declare that it is a trend.

New cases for NYC, Westchester, and Nassau all declined, but there was also reduced testing. Was that a result of the Easter weekend or a real change in the requests for tests?  The percent of results coming back postive is still over 40%, indicating that tests are still being rationed.

Bergamo, Italy is still struggling to get their new cases below 100.  They first dropped below 100 a week ago, but continue to hover at that rate. I could not find the death rate for that province, but the death rate in Italy overall is dropping. The blue curve is the best fit of the Westchester data to a gaussian distribution.  The orange curve fits the early Westchester data, but uses a duration similar to the duration of the epidemic in Bergamo.

April 13 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 12, 2020. Click on a chart to enlarge it.

There is something strange going on with the viral spread in the NYC metro area.  Hospitalizations and intubations are going down (see chart below), which means the death rates should be going down soon, but they have not yet (see Orange Line in Westchester chart). However, the rate of new cases has not gone down in greater NYC (includes Westchester and Nassau). Are all of these new cases minor symptoms with no hospitalization required? Has this been seen in other countries?

The new cases in NYC, Nassau, and Westchester are still stuck on a plateau and are at about the same rate per capita.  Test results are over 40% positive. This is much higher than other parts of NYS or the rest of the country.

The table below shows testing per capita for different counties in NYS. Westchester leads the list with over 5% of the population tested. (About 2% are confirmed infected).  Manhattan does not make the top 10. Westchester was the location of the first case in NYS so that may explain part of it. It is also the weathiest county in the state, so the county labs may have been better prepared? Despite the high load of infections, there have been few reports of chaos in Westchester hospitals.

Number of tests per 100,000 residents

Westchester also makes the top ten in deaths per capita. (The data is a few days old.  Westchester is now up to 50 deaths per 100,000.)

Number of deaths per 100,000 residents

There is little change in the comparison curves for New Cases between Westchester and its sister province of Bergamo, Italy.

The Swedish Model

A few days ago, I recommended a scenario for getting back to normal.  I recently found out that this is close to the strategy pursued by the Swedes (their approach is a bit more lackadaisical than I suggested). It is lauded by some, but deplored by others. In my view it is a great control group for the massive social and economic destruction going elsewhere in the developed world. To take that a bit further, I would like to see states follow different models of "reopening." This would generate more data on what might be the best response to future pandemics. For 'experts' to say that they know the best course of action is preposterous when so little is understood about this virus and how to minimize its long term effects on our world.