Tuesday, April 7, 2020

April 7 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 6, 2020. Click on a chart to enlarge it.

New Cases in Westchester have declined for 4 days-in-row. This is the first time that has happened.




Reduction in New Cases is a leading indicator of reductions in deaths.  Because death reporting in at the county level has been sporatic and delayed, the death reports shown below are not very informative.  The death rate for the state of NY appears to be at its apex.



The New Case Count for Nassau and NYC showed increases vs the prior day, but hopefully this is just the normal noise in the data.



What is the real level of infection (and hopefully antibody presence) in the population?
The answer to this question will vary by region and can only truly be seen with widespread antibody testing.

Chester County, PA is starting antibody testing.  They have received 20,000 finger-prick kits from Advaite.  Chester has 304 confirmed cases.  

Harder hit areas are expected to have larger percentages of the population with antibodies.  I have been comparing Westchester (over 1% known infections) with the Province of Bergamo in Lombardy.  Lombardy is the hardest hit region in Italy.  40 of 60 voluntary blood donators in Castiglione d'Adda, Lodi Province, Lombardy, Italy tested positive for the virus.  All were asymptomatic.   This suggests a very high level of antibodies in some areas. 

We know from the demographics of deaths that school age children are minimally affected. But they are definitely carriers as the modelers at IMHE found that "Our statistical tests suggest that early actions such as school closures may have a larger effect in some communities than expected."  School-age children are another big chunk of the asymptomatic population with antibodies.

Monday, April 6, 2020

April 6 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 5, 2020. Click on a chart to enlarge it.

There is quite a bit of positive news from yesterday:
1. Daily Death Count for NYS dropped for the first time since the beginning of the outbreak.
2. The daily increase in hospitalizations in NYS dropped by half. (This matches with my conclusion on April 3, that "new cases [are] not dropping because people with minor symptoms are now able to get the tests? If so, are new cases less likely to be hospitalized?")
3. New Cases for NYC, Westchester, and Nassau all saw drops with testing levels staying the same. (see chart below).  In other words, the percent testing positive declined in all of the NYC metro areas.
4. The updated IHME model now predicts the peak deaths in NYS one day earlier, April 9th (see chart below). Westchester, NYC, and Nassau should peak prior to that, and in that order


IMHE Model for New York State updated on April 5, 2020



Bergamo Italy continues to see a reduction in new cases. This was an early and hard-hit area outside of Milan with population very similar to Westchester, so I have been using it as a comparator.



Death reporting at the county level in NYS is not believed to be reliable. (Green line at bottom of graph.) The 1.5% death rate is similar to the state average.



Odds and Ends

  • A NYC family doctor said she is only given 10% of the daily test-kits she needs for patients with symptoms. This confirms my guess that infections are being undercounted and lag the actual infection date.
  • Google has released their "mobility" data. Westchester's reductions in public transportation use, working, playing...is evident.
  • In the first three months of 2020, about 500,000 people died in the US. CDC Statistics.  About 2500 from COVID-19 or less than 1% of total deaths. If the total reaches 100,000 that will be 5% of US deaths for 2020.
  • CDC has begun issuing a weekly Surveillance Report
  • CDC says the surveillance study of the general population for antibodies has already begun. This study looks at blood samples from people never diagnosed in some of the nation’s Covid-19 hot spots.


Sunday, April 5, 2020

April 5 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 4, 2020. Click on a chart to enlarge it.

The number of Westchester New Cases has reached a plateau.  How long will the plateau last? It looks similar to Italy, so I decided to add some Italian case data to the plot.  Bergamo Province, just NE of Milan, was one of the earliest and hardest hit areas in Italy.  It has a population of 1.1MM so it is comparable to Westchester in many respects.  DataSource.  Bergamo reported a max daily case rate per capita about half of Westchester.  This could be due to less testing. Of more significance, the duration of the Bergamo event is about twice the duration estimated by the fit of the Westchester data. Should we brace for a long haul?




The Westchester New Case Count has reached a plateau. The number of deaths more than doubled to 197. (Did they just solve the reporting delay?)



Nassau has passed Westchester in number of total cases. The Daily New Cases per Capita also exceeds Westchester.




IHME has not updated their model since April 1. The model had been updated daily.  Is something going on there? The date of maximum deaths in NY is still predicted to be April 10.




Saturday, April 4, 2020

April 4 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 3, 2020. Click on a chart to enlarge it.

The curve of new cases in Westchester County is flattening, but signs of a true drop are not yet visible. The percent positive test results is still about 50% which indicates a rationing of tests.  In other words, new cases are being undercounted.  In comparison, the national average is about 15%. New York City and Nassau County are still struggling with test results coming back almost 60% positive.




The FDA approved the first serology test for COVID-19 antibodies. Cellex Antibody Test  It is a lab test (not a point -of-care test) and detects both IgM and IgG antibodies. This particular test is relatively easy to scale to the millions of kits, but it requires blood draws and lab technicians.  The testing of millions of subjects will take awhile. However, it should just take a few days to test a representative sample of New Yorkers.  For example, if 50% of a sample of 1000 people in New York City had the antibody, it would explain the trajectory of infection in the city....the infection was too widespread prior to social distancing to stop it. If only 2% have the antibody, it would mean that either the virus has a much longer incubation period than thought, or it is extremely contagious and defied the social distancing.

There are other antibody tests (as mentioned in prior blogs) that can be done in the doctor's office.  They are 'finger-prick' tests where a drop of blood is placed in the well of a plate. The presence of antibodies activates a color as in the test above.  These can also be made easily by the millions and have the advantage being done in real time.  Apparently they are not as robust, and only have an FDA waiver.  SD Biosensor and RayBiotech have said they are already shipping these items.





Nassau County (on Long Island, east of Queens) continues its very high rate of infection.



No change in the IHME forecast for the peak in NYC.




Death Demographics in NYC
Unfortunately the state has not published demographics of the deaths in Westchester County.  New York City, on the other hand, publishes a daily update.  Here are a few conclusions from their data:
1. Men are 50% more likely to die than women.
2. Only one person under 18-years-old has died out of 1867 total deaths.
3. Only 6% of the deaths are of people under 45.
4. 70% of the deaths had underlying conditions.

This report does not elaborate on the underlying conditions, but other sources list diabetes, pre-diabetes, and obesity as the major conditions.  In fact, the higher death rate in Louisiana may be due to the higher prevalence of these health conditions in that state.

Friday, April 3, 2020

April 3 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 2, 2020. Click on a chart to enlarge it.

The new cases in Westchester remain stubbornly high. Governor Cuomo used the word "disturbing." The cases per capita are at an unexpected level.  More than 1% of the population of Westchester is confirmed positive. If you add the asymptomatic and those not tested due to minor symptoms and scarcity of the test, it is probably more like 10%.  Are our expectations of the infection rate incorrect because of bad reporting by China? Do we really believe China contained the virus to the Hubei (Wuhan) area?  That would be like the US containing the virus in St. Louis.

Cases per 1MM residents (Total Cases)
Westchester ------------ 11,567  (11,567)
New York City ----------  5,955  (51,809)
New York State ---------- 4,159  (83,172)
New Jersey -------------- 2,473  (22,255)
California ---------------     239  (9,568)
Michigan -----------------    933  (9,334)
Hubei Province ---------- 1,159  (67,802)
China -----------------------    59  (82,465)

Perhaps Westchester is testing more than other areas?  (Yes, in Westchester we are testing at about 10x the level in South Korea that was "the best-in-class" model). The number of new cases correlates with the number of new tests (the more you test, the more you will find) so it appears that the demand for tests still outstrips the supply.

Is the reporting of new cases skewed due to the delay from swab to test report? Is it still several days from swab to public release of the data?

Are new cases not dropping because people with minor symptoms are now able to get the tests? If so, are new cases less likely to be hospitalized?

Is this data saying that social distancing is not working as well as it did in Wuhan? If the median incubation period is only 1.5 days (latest from CDC) and the duration from symptoms to test results is about a week, we should have seen a rapid drop off in new cases by March 25.

The gaussian fit of the data is getting worse as the "New Case" data now appears to be bimodal?



Westchester County officials announced they will not be releasing data for the municipalities any longer.  That means no more data on Yonkers.  The data was misleading because of a several-day-lag in reporting.  Was that the reason for discontinuing?  A better solution would have been to eliminate the time delay in reporting. The death data for Westchester is reported sporatically and also appears to have a reporting delay. 






The UW Model is still predicting a peak in death count on about April 10 for the state of New York....7 days away.  Governor Cuomo is using a model that predicts 1-2 weeks after that.



More on hydroxychloroquine: 

Hospital researchers in Wuhan published a randomized study using hydroxychloroquine vs. standard-of-care.  The hydroxychloroquine group had a larger reduction in fever, cough, and pneumonia vs. the control group.   https://www.medrxiv.org/content/10.1101/2020.03.22.20040758v2.full.pdf

Of more interest is their meta analysis of 80 patients taking hydroxychloroquine for lupus who were admitted to the hospital for reasons other than COVID in the same timeframe. None were positive for COVID-19. Contrarily, none of the COVID patients at this hospital had a prior history of hydroxychloroquine use. The evidence for prophylactic use of hydroxychloroquine is mounting. It may end the pandemic before a vaccine is ready. Unfortunately, several states have banned off-label use of this drug. Perhaps that is to preserve supply for healthcare workers, but the cynic in me tells me it is because they think they are smarter than the public.

Thursday, April 2, 2020

April 2 - COVID-19 in Westchester County, NY

Here are the updated charts with data through April 1, 2020. Click on a chart to enlarge it.

The gaussian-fit of Daily New Case Count for Westchester County was updated with the April 1 data point. The fit parameters were not affected greatly by the latest report.



The Total Cases in Westchester per-capita place it near the top of all regions in the world.  The infection rate based on confirmed testing is now over 1% of the population.  It has twice the cases per-capita vs. NYC. Why is this? One explanation is the virus was spreading in Westchester for days longer than in NYC before the mitigation measures were put in place. Yet, the hospitals in Westchester seem to be doing better than in NYC?  Why is that?



The comparison of New Cases between Westchester, NYC, and Nassau County are now plotted on a per-capita basis.



The IHME daily forecast for NYS has not changed much since yesterday.  It is still predicting a Daily Death Count peak on April 10.  Governor Cuomo stated yesterday that the IHME model predicted "high death rates through July." This is clearly not true from this graph.




This table below contains demographics of NYC deaths.  Out of 1139 deaths in NYC only 14 did not have contributing underlying conditions.



Wednesday, April 1, 2020

April 1 - COVID-19 in Westchester, NY

This is a new chart for today.  I am not a modeler, but I have fit the Daily New Case Count from Westchester to a Gaussian curve. There is a bit of noise in the New Case Count, so beware of drawing conclusions.   "30" corresponds with March 30th.

Later in today's blog I will refer to the modeling done at Univ. Washington using Gaussian curves of the Daily Death Count. The UW model is used by governors for local hospital utilization planning. Dr. Birx referred to it in the March 31 White House briefing.


The latest Westchester data from March 31.


New Cases compared with NYC and Nassau County with new data from March 31.


Here is the UW model for Daily Death Count for NY from March 27.  The peak of Daily Death Counts lags the peak of Daily New Case Count by ??14-21 days??.  Newer projections (available on their website) still predict a peak on April 10 of about 800 deaths.



More on the University of Washington Modeling

The full report is here:  http://www.healthdata.org/sites/default/files/files/research_articles/2020/COVID-forecasting-03252020_4.pdf
Daily Updated Projections by state are here:
https://covid19.healthdata.org/projections

The Institute for Health Metrics and Evaluation (IHME), an independent global health research center at the University of Washington, has constructed a state-by-state model for COVID-19 disease spread. It appears to be the model of choice for governors preparing for the next few weeks. It has been mentioned by name in a few of the White House Task Force Briefings.

The model predicts a daily death count for each state.  That number is used to make other predictions, including daily hospital utilization.

The daily death count for each state is a 'bell (gaussian) curve' where the height and width are determined by the date and extent of implementation of 4 mitigation strategies. The curve for New York is shown above.

Below are the authors' disclaimers.  My comments are in brackets.

"Any attempt to forecast the COVID-19 epidemic has many limitations. Only one location has had a generalized epidemic and has currently brought new cases to 0 or near 0 [in other words, only Wuhan has a fully shaped curve for model validation], namely Wuhan. Many other locations, including all other provinces in China, have so far successfully contained transmission, preventing a general outbreak [hence those provinces were not used to test the model because they did not follow a Bell Curve?] Modeling for US states based on one completed epidemic, at least for the first wave, and many incomplete epidemics is intrinsically challenging.
    The consequent main limitation of our study is that observed epidemic curves for COVID-19 deaths [?there exists only one complete epidemic curve for COVID-19] define the likely trajectory for US states. In this study, we do include a covariate meant to capture the timing of social distancing measures to take into account that Wuhan implemented 4 out of 4 social distancing measures within 6 days of reaching a threshold death rate of 0.31 per million. Our models explicitly take into account variation in age-structure, which is a key driver of all-age mortality. But these efforts at quantification do not take into account many other factors that may influence the epidemic trajectory: the prevalence of chronic lung disease, the prevalence of multi-morbidity, population density, use of public transport, and other factors that may influence the immune response. We also have not explicitly incorporated the effect of reduced quality of care due to stressed and overloaded health systems beyond what is captured in the data. For example, the higher mortality rate in Italy is likely in part due to policies around restricting invasive ventilation in the elderly. The model ensemble used does suggest that locations with faster increases in the death rate are likely to have more peak case load and cumulative deaths, but our uncertainty intervals are appropriately large."

One source of error sticks out to me; the model assumes that mitigation measure #4 is not important.  Wuhan implemented this measure, but no US state has implemented it. Measure #4 is severe limitations on travel, such as shutting down public transportation and restricting travel in and out of Wuhan to essential needs. From the paper: "Days with 1 measure were counted as 0.67 equivalents, days with 2 measures as 0.334 equivalents and with 3 or 4 measures as 0."