NY STATE | County Cases, Deaths from COVID (3)

April 6, 2020—Nassau County's death rate from COVID-19 has nearly caught up with New York City's 29.3 deaths per 100,000 population as of early-morning data.

New York City's five counties and five other counties – Nassau, Rockland, Westchester, Orange and Suffolk – dominate the figures.

Deaths in Large Counties as of April 5. The large population of both New York City's five counties and Nassau County have raised the statewide average death rate to 21.2, as shown int he first table. The only other county in the state to be above the state average is Rockland. Westchester is now just below the state average, followed by Orange County and Suffolk County. No other county in New York State comes close to these ten counties.

This is an update of two earlier posts on NY County data, April 3 and March 27. It utilizes the interactive database of the New York Times. See earlier posts for references and notes.

Deaths by Age Group and Gender. The second table shows how a pattern is already forming in the data, with the largest number of victims being over 70 and male. 11.4 percent of the fatalities were patients 90 and over, 26.9 percent 80 and over, and 53.4 percent were 70 and over. By gender, 61 percent of the deaths were men.

Source: Fatalities from NYS Department of Health. Population data from U.S. Census Bureau (2018). American Community Survey 1-year estimates. Retrieved from Census Reporter Profile page for Suffolk County, NY. Also used data from Towncharts.com and combined data to estimate the population by age ranges.

East Hampton Town. The COVID-19 cases in the hamlets and villages and other designated census places within the Town, as of April 4, 2020, show that the largest concentration of cases are in the Village of East Hampton. So far there are no reported deaths in the Town of East Hampton.

Update, April 8, 2020—The number of cases rose to:

Suffolk County 17,444 (Deaths, 323).

EAST HAMPTON TOWN, 75. (No reported deaths.)

Northwest Harbor, 17Springs, 17East Hampton North, 18Montauk, 10East Hampton Village, 10Sag Harbor Village, 1Napeague, 2Wainscott, 0Amagansett, 0

COVID-19 STATUS | U.S. Summary in One Table

April 5, 2020—The largest number of Americans have been suffering from the COVID-19 virus in eight states and the District of Columbia. As of April 4, the states were, in order of death rates: New York, New Jersey, Louisiana, Michigan, Connecticut, Washington, Vermont, Massachusetts. Their death rates all exceed the national average death rate to date of 2.6 per 100,000 population. The list of the worst-hit states underscores the role of social interaction in spreading the disease. New York City was the first big U.S. city to suffer from the pandemic, and its hospitals have quickly been overwhelmed by its spread. Governor Andrew Cuomo provides a daily briefing on his handling of the public health crisis.

Confirmed Cases and Deaths

Two main measures are being used to gauge the progress of the disease: confirmed cases and deaths. Both measures have built-in problems.

Confirmed cases provide an earlier estimate of the extent of the spread of the disease than deaths. It can take a week or more for symptoms to appear, and then it may take another week or more for recovery. Patients can be hospitalized for weeks before recovering or succumbing to the disease.

However, confirmed cases depend on the availability of testing equipment and personnel, and in the early days of the pandemic in the United States tests were in short supply.  

For this reason, a low rate of cases to population may simply mean that fewer people are being tested in a state rather than that the disease is not spreading in a state. For the same reason, the case fatality rate is a less reliable number than deaths relative to population.

Deaths by cause would seem to be a simple enough number to keep track of. In a democracy it is hard to hide a dead body for long. Deaths are closely observed and recorded and whistleblowers historically have not been, as in dictatorships, punished for speaking out. A cause of death must be listed. The course of the COVID-19 disease is easy enough to spot. There are in fact some problems with the data, but let's look first at the reported numbers.

New York State tops the ranked list of deaths per 100,000 population, with 18.3 recorded as of today. Within New York State, the most densely populated counties are suffering the most.

New Jersey and Louisiana both have half the severity of New York State, about 9 per 100,000. Michigan and Connecticut are 5 per 100,000. These numbers are rising daily.

As a comparison, the average number of deaths in the United States from traffic fatalities in a year is 12.4 per 100,000. (Note that the traffic fatalities number is the sum of events happening over a year's time, whereas we have only three months of pandemic data.) 

The overall U.S. death rate to date from the coronavirus is 2.6 per 100,000 population.

Issues with Cause of Death Data

Two kinds of issues with the data on death rates are emerging: 

• Underreporting of COVID-19 deaths. Some jurisdictions require that a test for the presence of the virus have been conducted in order to list it as a cause of death. So the absence of testing kits both lowers the number of cases and the number of deaths.
• Lack of data about victims. Even when the deaths are properly recorded, the hospitals or doctors signing the death certificates may be so pressed for time or staff that they are not recording demographic data about the victims. Ibram X. Kendi, in “Why Don’t We Know Who the Coronavirus Victims Are?” The Atlantic, April 1, 2020, suggests that African Americans are disproportionately the victims, but in many states we don't know. He asks why other states don't report this information—the problems including multiple illnesses and the lack of a test to prove that the victim had the virus; the coroner may be insisting on such proof (here is the Suffolk County, NY procedure for reporting a death). 

Who Are the Victims?

The evidence so far is that more blacks are victims than their proportion of the population, and men are much more likely to be victims than women.

Kendi cites data showing that African Americans are 14.6 percentof the Illinois population, but are 28 percent of confirmed cases of the coronavirus (Latinos, however, are a smaller percentage of deaths than in the population). An even more striking disparity has occurred in Milwaukee, where blacks account for 26 percent of the population but are half of the reported cases and 81 percent of deaths. In Michigan, blacks are 14 percent of the population but 41 percent of the victims.

Support for Masks, Shutdown

Support has been growing for closing meeting places like restaurants and wearing face masks in public. The support for these actions has bubbled up from the states rather than having been led from Washington. The White House at first did not give credence to the pandemic's arrival in the United States and then expressed concern about interfering with economic activity to slow the progress of the disease.

Among the top nine victims of the virus, blue states predominate. Some of the governors of these states have clashed with the President on getting support in the form of supplies, equipment or a national shutdown or mask-wearing policy. 

Democratic New York City and State, with more than 18 deaths per 100,000 population in the state, have supported closing public meeting places. Four other states and D.C. have Democratic leadership. These leaders of these states were early in supporting closing public places until the apex of the disease has passed and in calling for an immediate end to tariffs on medical supplies from China.

The other three states (Louisiana, Vermont and Massachusetts) are purple. Two have Republican governors and one has a Republican senator. Among the next nine states, ranked by death rates, three are red states and two are purple (Colorado and Pennsylvania).

In states with Republican or mixed party leadership, leaders have often appeared to wait for direction from the White House. This has sometimes led to disagreement between Republican Governors and Democratic Mayors over delays in locking down or advice on social distancing.

Note on Data: Confirmed cases and deaths by state are as of April 4, 2020 from the Johns Hopkins University Coronavirus website. A convenient tabular form of the JHU data was posted by The Guardian later on the same day. Total number of U.S. cases and deaths added by JT Marlin. Population data are from the U.S.Census Bureau, Population Division, Table 1. Annual Estimates of the Resident Population for the United States, Regions, States, and Puerto Rico: April 1, 2010 to July 1, 2019 (NST-EST2019-01), December 2019. The Census Bureau documents its estimation methods. Calculations of cases/population, deaths per 100,000 population and case fatality rates by JT Marlin.

METRO DEATHS | NY area likely to pass 10,000 deaths

Italy's lockdown.

April 3, 2020—On a Zoom chat this morning, someone asked about the future for the U.S. and New York City economy. I suggested that  it depends on how we manage the pandemic at the NY City, State and Federal level.

The number to watch, I said, is the death rate, because by now everyone is aware that the "confirmed cases" number is dependent on how many tests are done. Shortages of testing kits reduce the number of confirmed cases but also, by delaying proper care and isolation, contribute to higher death rates. The number of cases is not a good measure of the outcome of the disease.

Chart 1. Virus Cases by Country, David
Leonhardt, NY Times, March 31, 2020.

Whichever measure we use, the  prognosis for the United States is for the disease to worsen before it gets better.

Let's look first at the number of cases. Did the United States get ahead of the curve between the time that the disease started in China and then spread to Italy? The chart at right, by David Leonhardt in the NY Times on March 31, suggests not.

In fact, by this point in the progress of the virus, China and Italy and even Spain were both doing better based on cases. But this chart is subject to testing bias. If the United States is "not flattening," the reason may simply be that the United States has stepped up the rate of testing.

Deaths in a democratic country are hard to fake or hide. People are paid to track them. A death from the coronavirus follows a gruesome pattern that should now be not hard to diagnose. Population numbers are closely watched. So, horrible though they are, the death rates are "good"  measures of the spread of COVID-19.

Table 2. Top lines, NY Times Upshot 
table, March 27, 2020.

The table at right shows the top seven lines of a table published as an Upshot in the New York Times. As the accompanying text in the story points out, the death rates lag cases, so they are not the best measure of the challenges being faced at the moment in the hospitals.

Because China was ahead of Europe, and Europe has been ahead of the United States, in facing the coronavirus, the Lombardy and Wuhan death rates are worth looking at as a guide to what might happen in the New York City area.

If there is a bias in the death rates in Lombardy and Wuhan, it is probably on the low side:

• Yes, Italy has an unusually elderly population, second in the world after Japan. This would tend to raise the death rates relative to New York — the median age in Lombardy is about 45, whereas it is 38 in New York. But Lombardy numbers may be considerably understated because many deaths that did not occur in hospitals were not counted by the health care reporting system — i.e., they might not have been counted if they occurred where people live, even if they were in nursing homes. 
• The Wuhan numbers may also be undercounted. A classified U.S. report suggests that C.I.A. insiders believe that Chinese cases and deaths from the virus may have been systematically underreported.

What do the death-rate numbers tell us? It is not in the Upshot story, but if the progress of the disease in the New York area matches that of Lombardy, we can expect a death rate of approximately 0.5 per thousand applied to a New York City metro area population of 20 million, or 10,000 deaths.

Chart 2. New York deaths higher than
Lombardy at same point. Upshot,
NY Times, April 3, 2020.

The math is not hard to follow. Lombardy suffered 5,000 deaths, in round numbers. It has half the 20 million population of the New York City Metro Statistical Area.* So one would expect, all else equal, that the New York metro area would end up with double the Lombardy deaths, or 10,000.

Today, April 3, the New York Times published another Upshot chart showing how we are doing on death rates relative to the Lombardy region. It should not make us complacent.

Upshot shows on the chart that the NY City death rate until the last observation has been doubling every two and a half days. It has been on a trajectory well above Lombardy's for a week, until the latest day. Governor Cuomo has noted in his 11 a.m. daily updates a marked improvement in the State trajectory in the last few days. If he is right, it is a breakthrough and he should get credit for it. The Governor's recent moves have improved the odds that the State's hospital system will be able to handle the coming peak load of patients. Meanwhile, we should watch the numbers carefully. It still seems likely that the New York metro area will exceed Lombardy's death rate.

* The NY metro is the statistical area (MSA), not the Combined Statistical Area (CSA), which would add another 3 million people. The NY MSA is composed of four Metropolitan Divisions; one of them is Nassau-Suffolk, two counties that together constitute Long Island as a Metropolitan Division.

PANDEMIC | 2. What Affects the Spread of the Coronavirus?

This is the second of a series of posts on the subject of the coronavirus death rates by country. See also posts on March 19 (#1, Why is Germany's death rate so low?) Virus Deaths Tracker (FT).

March 20, 2020–The New York Times has an informative article today showing the growth of COVID-19 cases in each country. A sample of the charts is at right, for the United States. It is illuminating to see the same exponential pattern occurring in virtually every country that has a significant number of cases. (It would be easier to compare countries' data if all of the countries were on the same chart, preferably logarithmic, so that the rate of change is easier to gauge.)

The charts show a series of daily snapshots of the progress of the novel coronavirus in each country. The chart shows the number of confirmed cases of infection by the virus. What are the factors that determine the spread among countries like Italy on the high-mortality end and Germany on the low end? Here are some early theories:

1. How quickly a country created social distance. Hand-shaking (until this pandemic) has become a world-wide custom. But some countries are used to even closer social greetings, such as hugging and kissing. Amsterdam is proud of its three-kiss greeting (that is the meaning of the three Xes in its shield). Self-isolation may be harder to enforce in some countries. The closeness of families in Italy is viewed as a problem in getting more vulnerable age groups in the population to keep their distance from socializing younger generations.
2. A city's density, as an indicator of how much of a magnet it is. If, as most people believe, the virus started in Hubei, China, then the virus somehow made its way from there to each country, probably carried by human beings. Travelers on busses, trains and airlines are likely carriers, and one can be a carrier with no symptoms. Hubei is center of transportation and manufacturing system in China, which made it an easy place to spread to virus. New York became an epicenter because of its central role in communications, finance, fashion and technology.
3. How much testing is done, when it is done, and when it is reported. New York State has authorized a range of manual, semi-automated and automated testing approaches and the numbers of confirmed cases of the SARS-CoV2 has been rising rapidly. The State says it is now testing 10,000 people a day, has tested 32,000 people in total, many more than any other state, and has found 7,100 cases of positive readings, nearly 40 percent of the total for the nation (19,624). 

The Times story references the need to "flatten the curve," i.e., the need to take quick action to test people with symptoms of the virus and quarantine them and provide hospitalization for the very sick, who will suffer from exacerbation of respiratory, heart or other immunity problems they have.

Additional Data to Get. The charts show the "curves" themselves in each country, indicating confirmed infections, so that we can better understand the number of people who are hospitalized or quarantined, and the coming demand for medical facilities. Now we need to put these numbers together with other data:

• Hospitalization and death ("case-fatality") rates. It would be useful to know how many are hospitalized in each phase, since the greatest concern is about a likely shortage of intensive-care-unit (ICU) hospital beds, ventilators and respirators. New York State reportedly (as of the Governor's press conference on March 21) has only 3,000 ICUs, 80 percent of which are occupied. It will need at a minimum twelve times that number based on the experience in Italy and other countries. A survey of more than 200 U.S. cities by the U.S. Conference of Mayors shows great concern about running out of supplies of masks and ventilators. The cities need collectively 28 million surgical (N95) masks, 8 million test kits and 139,000 ventilators. Reminds me of a play by Shaw that I first saw at the Abbey Theater in Dublin, The Doctor's Dilemma; read it here free–https://www.gutenberg.org/files/5070/5070-h/5070-h.htm.
• Names of those who have developed immunity. This could be the base for treatments that could save lives and slow spread. Dr. David Posnett says: "If we only knew who had developed immunity to the virus.  Immunity may occur in someone who had the virus (either with symptoms or without), got better and developed IgM and IgG antibodies.  We know that happens.  If you are declared immune, you can safely go back to work which might help the economy.  You are safe as a healthcare worker.  You are safe to go buy groceries. You are safe to be around your loved ones.  You could donate plasma at the local Blood center and plasma with lots of antibodies could be administered to patients dying in the ICU (passive immunotherapy used previously against the Spanish flu and against Ebola for example)."