May 27 - COVID-19 in Westchester County, NY

Westchester County started to reopen yesterday.  That is 75 days since the 1-mile containment zone was erected around the Young Israel congregation in New Rochelle and 65 days since all businesses were shuttered. Back then I did not think this would last that long based on the neat little bell-shaped curve of the Wuhan epidemic that appeared to only last about 14 days.  So much for relying on Chinese data.

The percent-testing-positive in the New York metro area is at about 4% -- a big change since the days when almost half of everyone tested was positive.

Westchester had it lowest death count since the epidemic began -- 2 deaths.

The filtered case-count shows the asymmetric trajectory of the epidemic -- a rapid rise followed by a very slow decline.  The true case count is underreported due to both rationing of tests in the early days and the large percentage of asymtomatic cases.

My hometown of Yonkers (a subset of Westchester County) finally released some data!  The chart below shows an almost bell-shaped epidemic.  It is quite different from the general Westchester curve.  About 40% of the deaths in Yonkers were in elder-care facilities. The large spike on the right of the curve is the recategorizing of earlier deaths in elder-care to COVID-19-related.  

I have been in Michigan for the last two weeks.  Life is totally different here from Westchester.  In my Michigan county there are just over 100 cases and 6 deaths.  People are wearing masks, but there are no lines at stores.  Despite restaurants only doing curbside business, life seems to go on as normal.  People are on the beach and in their yards.  Construction and landscaping is lively.  Traffic is normal.

May 20 - COVID-19 in Westchester County, NY

I had a nice break from my blogging.  Nothing unpredicted occurred since I shared my plots 10 days-ago.  The number of cases continue to decline in a linear fashion.  Italy has started to reopen.  They continue to amend their case count, which is a bit silly as everyone knows that cases are underreported due to the large number of asymptomatic cases.  The blips in the Bergamo, Italy data are due to those amendments, not to a rebound since they started to 'reopen.'  The Wall Street Journal contained an interesting article about Bergamo's travails.

The New York City metro area appears to be leveling out at about 5-10% testing positive.  The hospitals can handle that slow rate of infection.  In other words, the Spread has Slowed.  Now let's move on!

Deaths in Westchester are now consistently in the single digits.  Unfortunately there is a NYS requirement of a 14-day period of daily declines in deaths prior to starting to reopen. The statistical noise at 5-10 deaths per day will not let this happen.  Our governor failed statistics 101.

May 10 - COVID-19 in Westchester County, NY

This is my 45th day in a row of blogging the Westchester County, NY epidemic.  It has been 68 days since the containment zone was established in New Rochelle (in Westchester). At that time, each day was an unknown. How would the number of cases or deaths or percentage-testing-positive change? Now we are on the backside of the curve.  The direction and shape of the curve are fairly predictable for the next few days.  So.... I am going to take a break and give my readers a break. I hope you can tell from the charts where the next few days are heading.  I'll be back in a week to see if we were correct.

New Cases are following a linear decay.

All of the NYC metro area is trending the same way.

The percent testing positive may be stabilizing at about 10% for all of the NYC metro area.

Deaths in Westchester are also following a linear decay.

May 9 - COVID-19 in Westchester County, NY

The Governor of New York almost did a happy dance yesterday as he declared victory over the corona virus.  It is true that New York has been reducing cases, deaths, and hospitalizations for several weeks, but the real impact of the governor's actions on those results is debatable.  Where is the control group to show cause and effect?  If the control group is Sweden, the total impact of this actions was negligible or negative.

The 7-day averages above agree with the shapes of the epidemic from my data for Westchester County, Nassau County, and NYC.  Note that Governor Cuomo finally dropped the 3-day average used for much of the epidemic.  As I have stated earlier, a seven day average is needed to filter out the weekend noise.

The data from Bergamo, Italy (my Westchester sister-province for this epidemic) is holding steady at about 20-80 cases per day. They are testing at a lower density than Westchester so their numbers must be corrected.  I estimate that they are testing in Bergamo at four times the national average in Italy or 1000 tests per million people per day.  That is one half of the Westchester rate of 2000 tests per day. Correcting their case rate to that testing density gives a flow of 40-160 new cases per day to be comparable with Westchester.  The current rate in Westchester is 150-250, but still declining.

Westchester had only seven deaths yesterday, the fewest deaths since the outbreak began.

Headline:  "Governor shocked that Healthcare workers in Westchester have lower prevalence of disease than rest of population according antibody: 6.8% vs 13.8%  Same in NYC and Long Island." This is how the media can distort reality.  How many stories did you read about the devastating effects of the virus on first responders, healthcare professionals, and food supply workers? Very early I did an analysis of sick police officers whose plight the media had highlighted.  The prevalence of disease amoung police was the same as the general population or less.  They were more likely to get tested and get better healthcare than the rest of the population, and in the end, they fared better.

Who is Dying?   In case you have forgotten amidst all of the other hype, this virus in NOT a threat to most people. The data in New York continues to show that 80% of the deaths are in persons over 60 and 90% of deaths have underlying conditions.  As I have said before, 98% of the population will do fine.  It would be scary if we did not know who the vulnerable 2% are. BUT WE KNOW EXACTLY WHO THEY ARE.

Contact Tracing    Finally someone has the courage to say what I have been saying about the futility of contact tracing with this epidemic.  Scott Atlas of Stanford has made almost the same arguments I did on these pages on April 19!

Methods of Transmission  Erin Bromage gives some great case studies of clusters that suggest methods of transmission.  Hint: It is not riding a bike outside!

May 8 - COVID-19 in Westchester County, NY

Hospitalizations in Westchester continue to drop.  About 700 are currently hospitalized in a system with 3000 beds. This should be the trigger to loosen some of the restrictions in Westchester.  Unfortunately, the formula for starting the 'reopening' depends on seven metrics, at least one of which Westchester cannot reach for another 14 days. More on htat later. I do not have the hospitalization data for Westchester, but it looks much like that for NYC and the state of New York shown below.

2020-05-07 Hospitalizations in New York State

New Cases in Westchester are still dropping on a 7-day average, but there is a lot of day-to-day variation (red line). The trend has been following a linear decay (orange line) over the last seven weeks! Yesterday, the number of tests performed was almost double the average of the prior two days.  The more tests, the more new cases.

The percent-testing-positive is fairly independent of the number of tests.  That is why there is less noise in the data shown below. That is why I have often argued that the percent-testing-positive is a crude approximation of the prevalence in the community.  This means that about 10% of the Westchester population currently has the virus.  We know from antibody testing in Westchester that about 20% have recovered from the virus.  Therefore 30% have been exposed to the virus so far.  [There is data showing some recovered patients that do not have the antibody, so the number exposed could be even higher?] The percent testing positive may be starting to level out at 10%.  This is a level which the hospitals can handle.  The strategy was always just to SLOW the spread, NOT to STOP the spread.  Unfortunately, Governor Cuomo uses the latter terminology on ALL of his charts.  This sends the wrong message.

The number of deaths in Westchester for the last five days were: 19, 15, 15, 51, 17. The fifty-one number was a recategorization of nursing home deaths that occurred earlier in the epidemic.  It could be that yesterday's number also includes some deaths that did not really occur yesterday.  One of the metrics to reopen is 14 consecutive days of decrease in deaths.  A three-day-rolling-average is allowed to remove noise, but that is not enough to remove the noise in these data.  It looks like end of May at the earliest for reopening of Westchester.

May 7 - COVID-19 in Westchester County, NY

Our leaders are finally starting to act on the data.  Yesterday the county started testing everyone in nursing homes in Westchester, residents and workers.  From the beginning it was obvious that the elderly with underlying conditions were most vulnerable.  Forty percent of the deaths in Yonkers (a city in Westchester County) are in nursing homes.  In fact, several nursing home deaths from weeks ago are being recategorized as COVID-19 related, so that percentage will go up. The number of deaths reported today is unusually high and is believed to contain these prior deaths. The last three weeks of death data can be approximated by a linear decay (not a Bell Curve!).

The number of active cases in Westchester, the number hospitalized, and the number of positive cases continue to decline. There is noise in the data -- for example yesterday experienced a data dump of test data -- but the filtered data are trending down.

It is a good time to revisit the suggestions I made on this blog four weeks ago on April 9.  I had only one condition for starting the reopening by region -- hospitalizations drops to near normal -- not the complicated 7 metrics required by Governor Cuomo.  Much of the data I used to come up with my suggested actions to reopen have not change in the last four months. 

May 6 - COVID-19 in Westchester County, NY

New cases in Westchester fell to their lowest levels since March 18 (red line).  Bergamo, Italy is into their third day of reopening.  Their new cases fell to 12! (green line)

Deaths in Westchester have been at 15 for two days in a row.  Hospitalizations continue to decline.

The percent testing positive continues its predictable decline.  There is very little noise in the data from NYC, Nassau, and Westchester. They are all declining in unison.  In the next two or three days, the percentage will drop below 10 which will be in line with with other areas of the country with mild epidemics.

All is preceding well with Bergamo's 'reopening.'  The restrictions on travel are still very high.  Police have set up roadblocks to inspect vehicles for proper papers for travel. Italians cannot travel outside of their region without a legitimate business reason. They cannot travel outside the region to family or second homes.  Only 3% were cited for "being outside the home for reason not among those granted."

As mentioned in prior blogs, Bergamo is a province of about 1MM residents and has much in common with Westchester. However, their population of elderly is very high.  As a result, their fatalities per capita are much higher than Westchester.   The number of deaths in Bergamo in March was more than five times the number of deaths in March 2019.

In March, 5500 Bergamo residents over 65-years-old died from COVID-19.  A normal year would have claimed only 750. (Westchester total deaths is currently at 1116.  About 80% of those are over 65.) Demographic data states that 30% of Bergamo is over 65, so the chances of dying from COVID-19 is still less than 2% for that age group.

Bergamo Deaths for Men and Women

May 5 - COVID-19 in Westchester County, NY

New York announced the 'Gating' criteria for reopening regions yesterday.

Strangely, Westchester was not grouped with the New York City region which shared a similar epidemic and with which Westchester is tied by heavily traveled commuter train lines.  Instead, Westchester is grouped with the 'Mid-Hudson Region' which includes the less populous and lesser effected counties of Rockland, Orange, Sullivan, Ulster, Dutchess, and Putnam.

Each region must meet the gating criteria to enter 'Phase 2' which allows some low-risk/high-value businesses to open. Because Westchester was the first county in New York with a cluster (New Rochelle) it is ahead of the other counties on the epidemic curve.  I have not studied these counties (yet) in detail, because they are not similar to Westchester.  My early guess is that they will delay the opening of Westchester.

The criteria and the current conditions in Westchester are summarized below.

1. 14 days of decline in total hospitalizations (rolling 3 day averages)
The hospitalization data for Westchester is not published. The county executive usually gives the number in his daily briefing.  From his accounts Westchester meets this criteria. As I have stated in prior blogs, I do not believe the '3-day averaging' does an adequate job of eliminating noise in the data.  Much of the noise comes on weekends, so a 7-day average is needed.

2. 14 days of decline in deaths (rolling 3 day average)
Westchester deaths have been declining for three weeks, but there are not 14 consecutive days of decline due to the noise.  Even a 3-day rolling average does not remove the noise (as mentioned above). The green line is the 3-day rolling average.  The last three weeks can be approximated by a linear decay (orange dashed line). That decay predicts Westchester could have a '0' day by May 17.  For practical matters Westchester meets this criteria, but the criteria are poorly writtern.

3. less than 2 new hospitalized patients per day per 100K population
For Westchester this mean less than 20 patients per day. Currently less than 10% of the new cases are hospitalized and there are about new 200 cases per day.  Westchester is borderline in meeting this criteria, but the hospitals are not taxed and the hospital population is dropping. This criteria is not meaningful if hospitals are not stressed! I am not sure why they threw in this criteria? It is redundant.

 4. >30% hospital beds available
Westchester has 3000 beds and only about 700 are being used for COVID-19.  That number continues to drop. Criteria met.

5. >90days of PPE supply
I am not sure what this mean? 90 days supply based on future estimated hospitalizations?  Or is this some unrealistic supply based on a worse-case scenario? This seems a bit like they are just trying to slow things down.  Who has a 90-day supply of anything?

6. >30 tests per 1K per month
This criteria means 1000 tests per day for Westchester.  Westchester has been testing about 2000 per day for the last month.  Criteria met.  If 10% of tests are positive, that results in 200 new cases per day in Westchester (10% of 2000 tests) leading to about 20 new hospitalizations (see Criteria #3). Currently Westchester is running ~15% positive, but the trend will put it at ~10% in one more week.

7. >30 tracers per 100K (300 for Westchester)
This just means Westchester needs to hire 300 lackies.  I have explained in previous blogs why 'Test and Trace' is a fool's errand for a virus where 90% of the cases are asymptomatic.  [Antibody tests show at least 15 % of  Westchester has been exposed, but there are only 2.5% confirmed cases.]  Critieria met.

May 4 - COVID-19 in Westchester County, NY

Westchester Daily Cases are compared with Bergamo, Italy below.  Bergamo begins ending their lockdown today.

The percent of the NYC Metro area testing positive continues to go down. At 15% it is getting close to the national average.

The deaths in Westchester are going down very slowly.

Strangely, NYC deaths, which should be lagging Westchester, appear to have the traditional epidemic curve??

NYC Deaths per Day

May 3 - COVID-19 in Westchester County, NY

I have been comparing the Province of Bergamo, Italy to Westchester County. The epidemic curves below may not appear similar, but there are some similarities.

Bergamo Province is about the same population as Westchester (~1MM). Bergamo borders the largest metropolis in Italy (Milan) as Westchester borders NYC. From the graph it appears that Bergamo had fewer cases than Westchester (confirmed cases are 11394 in Bergamo and 29626 in Westchester), but this is just a result in the difference in the availability of testing.  10% of Westchester has been tested, while only 4% of Lombardy (the Milan region that includes Bergamo) has been tested.  Luckily the deaths are not comparable....Bergamo has 11291 deaths so far vs. 1067 in Westchester. Almost 1% of their population has died vs 0.1% of Westchester! [In the city of Bergamo 30% are over 65, more than double the 15% in Westchester. Bergamo province is less dense...1 MM people spread over 1000 sq.mi. vs 450 sq. mi. in Westchester.]

The lockdown in Bergamo started on March 9.  Westchester had a Containment Zone around New Rochelle on March 12.  All of Westchester was essentially closed by March 14, but the stay-at-home order did not officially go into effect until March 22. It is hard to judge the epidemic starts and peaks from the chart above, but Westchester trails Bergamo by 2 to 3 weeks.

Bergamo moves into Phase 2 tomorrow. Even in Phase 2 the restrictions are very high.  It is not much different than what Westchester is currently practicing.

The 7-day average (this cuts out any noise caused by the weekends which I have noticed is substantial) of New Cases is showing a favorable downturn.

The stream of new infections despite social distancing confirms how contagious this virus is.  Antibodies have show that about 40% of the Bronx (borders Westchester on the south)  have been exposed. Westchester is about 15%, but I am sure that varies across the county.  I live in a town in Westchester that abuts the Bronx and it would not surprise me if my town was closer to 40% exposed.

Ninety Eight percent of those infected do just fine.  If we did not know who the 2% vulnerable were, this would be scary.  But we know almost excactly who the 2% are!!!  Why don't we just act on that!!!

May 2 - COVID-19 in Westchester County, NY

Westchester County reported 262 New Cases of COVID-19 yesterday. It was only the second time since March 21 that the number has been less than 300.

A better indication of the reduced transmission of the virus is the Percent Testing Positive. This number has been less than 20% for four of the last five days.

Deaths per Day have fallen below 20 for only the second day since early April.  

Remdesivir gets FDA Emergency Use Authorization for treatment of COVID-19 patients.

Raoult publishes full article on Hydroxychloroquine.

May 1 - COVID-19 in Westchester County, NY

The beginning of May signals the start of reopenings across the US and Europe.  I hope someone is collecting the data!  Quick analysis and feedback will be needed to assist countries that are just a few weeks behind the leaders.  Each state in the US is following a differenct path. Comparisons of approaches and results will be interesting.

I have been comparing Westchester to Bergamo, Italy (as well as NYC and Nassau County).  I will continue to try and do that, but I do not currently have access to the Bergamo metrics that I think are important.

What are the important metrics in 'reopening'?  Here is my list.
1. The hospitals should not be overutilized.  This should not be a difficult metric to hit as most of the hospitals in the US were nowhere near capacity during the worst of the epidemic.  Bergamo was  overrun as were some hospitals in NYC.
2. Protect the most vulnerable.  From the very start it was obvious that those over 65 and those with underlying conditions were at highest risk (although most states ignored this in their executive orders). Is there anything we can give them prophylactically? Where are the studies on prophylactic hydroxychloroquine? As most of the deaths are in this group, the metric can be deaths per day.  Keep it less than 3 in 10,000 on an annual basis (0.03%) .  This is about the death rate for regular flu.
3. Economic health metric. This could be measured in unemployment rates.  GDP has too great of a lag to be a metric.

As I have repeated almost daily, this pandemic does not follow a Bell Curve.  It has a very long tail.  See Bergamo below.  Does this mean the virus continues to spread, despite social distancing?  Or does it mean that actions on Day 0 have implications over several months?  The latter would make if very difficult to make adjustments to a faulty reopening plan.

Almost six weeks after their peak, Bergamo will be entering Phase II on Monday, May 4.  For the first time since their stay-at-home order on March 9 they will be able to see relatives if they maintain 6 feet and wear masks.  They can run and bike outside and go to the park, maintaining 6 feet and wearing masks.  Gyms, cinemas, restaurants, hair salons, museums, churches, and playgrounds are still closed. Travel to second homes or to other regions of Italy is still not permitted. Public or private gatherings are not permitted with the exception of funeral parties of 15 or less, preferably outdoors with masks and 6 foot distancing. People with fever or respitatory illness must quarantine regardless of diagnosis. Businesses will start reopening in a staggered fashion.  Schools are closed until September.

Both the New Case curves and the Death Curve shown below are not following the classical Bell Curve of an epidemic.

Bike Sunday on May 3 in Westchester appears to be a go (with masks).  Bike Sunday is a 6.5 mile shutdown of the Bronx River Parkway for four hours on Sundays in May, June, and September. I guess I should be grateful that our Phase I in Westchester does not appear to be as restrictive as Phase II in Italy.

A recent 'random' antibody test estimated that 15% of Westchester had been exposed to the virus. The actual confirmed prevalence in Westchester is about 3%.  About 10% of Westchester has been tested for the virus.  On a personal note, my antibody test was negative. I suffered for eight weeks from late February to early April with a sinus virus. I am convinced I got it from the pool at my local gym. My point is that there are all kinds of viruses out there, most have not been identified, named, and given media coverage. If only 20% of the people tested for COVID have it, what are the other 80% suffering from? Who knows when an infection is going to be life threatening? Should I avoid my gym? What risks will I be willing to accept in the future?