Month: August 2020

Following the Data: Covid-19 and the State of New Mexico – August 26 Update

August 26, 2020

by Devin

This article was written to provide an update to the original, which was published on August 14 in the Los Alamos Reporter. If you haven’t read that article, we highly recommend you go back and give it a read. Now we have 13 more days of data to add to the previous data set and given the impending expiration of the Governor’s latest orders, it seems prudent to give interested New Mexicans the most up-to-date information on the status of COVID-19 in our State. The data in this article, as in the original, are sourced from The COVID Tracking Project(https://covidtracking.com), sponsored by The Atlantic. Drilling down, the historical data for New Mexico can be found here: (https://covidtracking.com/data/state/new-mexico#historical).

If you read the original article, you will know that for New Mexico the main takeaways were: 1) the so called “spike” or “second wave” of the virus was not real because it was simply the result of a very real spike in daily new testing; 2) there was no corresponding “spike” or “second wave” of daily new hospitalizations or deaths; 3) the probability of a random New Mexican having died from COVID-19 by August 11, 2020 was approximately 2.1 in a million, and; 4) the numbers across the board were trending downward. The questions we want to address in this article are: 1) after 13 more days, do these conclusions still stand; 2) are the numbers still trending downward; 3) if so, by how much, and; 4) what else do the data tell us? 

The great news is, the answers to these questions appear to be yes, yes, and see below.

What the New Data Tell Us

The Death Count

The historical data for New Mexico show that the daily number of deaths from COVID-19 peaked around mid-May with an average daily value of approximately 8 deaths statewide per day. Daily deaths continued to decline after that until early July when the average daily value bottomed out at approximately 3 deaths statewide per day. The daily average rose back up to around 5 by the end of July and lately has been trending back down. The overall average daily number of deaths since the first case was recorded in New Mexico is approximately 4.5. The overall trend through August 22 is basically flat, with a very slightly downward trend, as can be seen in the graph below.

Previously we talked about the number of deaths from COVID-19 in New Mexico for the purposes of comparison with the most recent two influenza seasons. Since those numbers are absolutes, it seems more useful now to talk about trends. Using the US Census total population data estimate for 2019 for New Mexico of 2,096,829, during the two-week period of peak average daily deaths, from May 3 through May 16, on any given day a random New Mexican had approximately 4.2 chances in a million of dying from COVID-19. For the most recent two-week period ending August 22, 2020 on any given day a random New Mexican had approximately 2.1 chances in a million of dying from the disease. Half the chances of the peak period. Breaking these numbers down further, by age group, consider the following chart and graph:

Clearly, older age groups have a much greater chance of dying from COVID-19 than younger age groups in New Mexico. In fact, in New Mexico, for all practical purposes, the chances of someone under the age of 25 are almost zero (since the data set is incomplete, we can’t say it is zero and we know that it isn’t). Those aged 25 to 64 years account for approximately 29 percent of deaths in New Mexico from COVID-19. Approximately 70 percent of deaths from COVID-19 are among those aged 65 and older. What this means is that on any given day during the most recent two-week period, if you are a New Mexican aged 65 or older, your chances of dying from COVID-19 were approximately 1.5 in a million. If you are aged 25 to 64, your chances were 0.6 in a million, or significantly less than one in a million. If you are under the age of 25, your chances were almost zero in a million. Not to diminish the value of every human life, but these are amazingly low odds. 

There is also an abundance of evidence now strongly associating at least one or more comorbidity with dying from COVID-19 (https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/evidence-table.html). What this means is that healthy young people are, for all intents and purposes, in no danger at all from COVID-19. Furthermore, healthy middle-aged adults are also in practically no danger from COVID-19. Finally, the highest risk people are those over 65 years of age who have one or more established pre-existing health conditions that have been strongly correlated with mortality from COVID-19. According to the CDC, these are: serious heart conditions, such as heart failure, coronary artery disease, or cardiomyopathies; cancer; chronic kidney disease; COPD; obesity (BMI>30); sickle cell disease; solid organ transplantation, and; type 2 diabetes. So if you are healthy and over the age of 65, your chances of dying from COVID-19 on any given day during the last two weeks were less than 1.5 in a million.

New Cases and Testing

Moving on to daily new cases, we continue to see a sharp decline in new cases being detected on a daily basis. Since the peak on July 28 at 460 new cases, the average daily number of new cases for the most recent one-week period, ending August 25, 2020 is 122, with 73 being reported on August 25. This is an average rate of decline of 12 new cases per day. If this rate of decline continues, in approximately 11 days, sometime around September 5, New Mexico may no longer be recording any new cases.

From our previous analysis, we found a directly proportional relationship between daily new testing numbers and daily new case numbers. The graph below presents the data for daily new testing. As you can see, daily new testing has slightly declined from near the end of July through August 25 (with a 2-day spike in new tests near the end of the period).

The following graph shows the weekly average new cases per test (in percentages) since testing began in New Mexico.

The peaks on this graph are artifacts of the data set that resulted mostly in the beginning of the pandemic when data collection was still being organized and the initial challenges that go along with any major data collection effort is undertaken. As you can see, over time, as the data collection process became more streamlined, the line becomes more stable. The trendline, shown as a dotted line, provides a more accurate picture of reality. And the reality is, since the week of July 11 this average has either been flat or in decline. For the most recent week recorded, that average number of cases per test was 2.1 percent. What these two graphs mean when taken together is that daily new cases have been declining at a significantly higher rate than daily new testing; therefore, we cannot attribute all of the decline in new cases to reduced testing. In fact, it shows that very little of the decline in new cases can be attributed to reduced testing.

Hospitalizations

The following graph shows the daily hospitalizations of COVID-19 patients in New Mexico since the first patient was admitted in April. Once the reporting procedures stabilized around the last week of June, the average daily number of new COVID-19 patients being hospitalized was approximately 17.4. This number trended up until the end of July, where it was averaging around 28.1 new admissions per day. Since then it has been trending down, with the average daily admissions for the most recent week at 8.1. If this trend continues on its current trajectory, hospitalizations may also reach zero in New Mexico within the next week to two weeks.

Our final graph of the data that includes new cases, deaths, and hospitalizations combined is shown here:

Although the lines for new hospitalizations and deaths are somewhat difficult to see, it is clear that all the trends are either flat or down.

Conclusions

Over the last three calendar years, fatalities from automobile-related accidents in New Mexico averaged 1.1 per day (https://gps.unm.edu/gps_assets/tru_data/Crash-Reports/Fatality-Reports/2019-fatalities.pdf). As a community, we have accepted the risks associated with driving on our roads and highways, knowing that every day, someone is going to die in a car accident. If that number were 2 people per day, or even three people per day, would we radically change the way live our lives? Would we reduce speed limits to half their current limits if that brought the number down and, if so, how much would it need to bring that number down before we would accept such a change? These are very difficult questions and everyone is sure to have a different opinion.

The current public health order is set to expire this month. The daily numbers are rapidly approaching zero. Given this information, the question before us is, how much longer do we need to keep the various restrictions in place, if at all? Given that there was no second wave or spike in cases in New Mexico to begin with, a very strong argument could be made for ending them right now; however, an extension of the restrictions still would not be justifiable beyond an additional two weeks. If, after an additional two weeks, the trends reverse, we could, of course, consider extending some of the public health orders further, but clearly, no public health orders are even now warranted by the data. What is certain is that if our leaders are truly allowing science and data to drive their decision-making, another full month of restrictions is completely unjustified.

Endangered Species: Political Independents, Why Pre-Election Polling is Useless

August 18, 2020

by Devin

For as long as I have been paying attention to politics, namely, my whole life, all I have ever heard from the Political Intelligentsia, is that politicians who want to get elected must lean to their party’s extreme to win in the primary and then move to the center to win the general election. If this is news to you, you definitely are new to American politics. This is because it is typically the party die-hards who show up for the primaries, who want to see a candidate get nominated that best reflects their party’s values, but a lot more people show up for general elections and those people are less idealistic than the party die-hards. To win these less idealistic voters over, it is said that politicians must soften their views so as not to sound so extreme and thus convince these so-called “Independents” that they are the best choice for the job. Hence all the negative advertisements during the general election painting opponents as extremists, with lots of pull quotes from their primary campaigning to prove it.

“…anyone who pays any attention to politics, enough to participate in elections by voting, in 2020, and even in 2016, cannot possibly, by now, not know who they are planning to vote for in the Presidential Election.”

When the pollsters conduct a pre-election poll, they prefer to poll “likely voters.” After all, what point would it be to poll people who aren’t planning to vote? But to do so, they have to select a representative sample of voters from the population. Typically the pollsters will try to estimate the percentage of Democrats, Republicans, and Independents that are going to vote and then make sure the polling reflects those proportions (or not, if they have a different objective). The farther off these estimates are from reality, the lower the accuracy of the poll will be, so they say. Exactly how they come up with these estimates is not a topic I have spent any time trying to figure out, but what I do know is that undersampling or oversampling a particular group will definitely throw off the results. That is how they got the 2016 Presidential Election so stunningly wrong.

One of the explanations for inaccurate polls that I have heard many times over the years is that the people who are polled cannot be relied upon to tell the truth. The theory goes that they may not want anyone, even a pollster who doesn’t know them, to know who they are really planning to vote for. Then, of course, there are those who just want to mess with the pollsters by giving them false answers. But I have a new theory as to why the polls, as this article is titled, are simply useless. Their premise about Independents, is false.

It occurred to me recently that anyone who pays any attention to politics, enough to participate in elections by voting, in 2020, and even in 2016, cannot possibly, by now, not know who they are planning to vote for in the Presidential Election. In fact, the USA is so incredibly polarized now, and has been since at least 2015, that it is impossible for me to imagine anyone who pays sufficient attention to want to vote, to be sitting on the fence between Joe Biden and Donald Trump. Am I right?

But if this is true, then who really are all these so-called “Independents?” I have an idea.

“…there really aren’t any voters that are truly Independents.”

For many years I called myself an Independent, but then I realized at some point that even though I called myself that, I always wound up voting Republican. (In middle school, I stood on the corner of a bypass holding up a Dick Riley sign in Greenville, SC, because he went to high school with my father. That’s my only brush with support for a Democrat.) So why did I call myself an Independent? Because I wanted to signal that I was open to voting either way, depending upon which candidate seemed to be the best choice for the job. Of course, that was mostly back in the days when there were still a few centrist Democrat politicians around. My wife, Amanda, also called herself an Independent for many years, even though she almost always voted Democrat (she calls herself a Libertarian these days). So it occurred to me that maybe people who bother to take the time to go to the polls and vote really don’t include a set of people who might vote either way. At least, not in significant numbers.

But if that’s true, assuming that party die-hards always get out and vote and do so for their party’s candidates, how to explain the shifts back and forth in various elections from one major party to the other? And the occasional landslide?Here is what I think. I think that there really aren’t any voters that are truly Independents. Okay, maybe a handful. A statistically insignificant few. But in fact, everyone, for all intents and purposes, knows what party they most closely identify with. Sure, this year it appears that there may be a fairly significant number of people who are changing their party loyalty, but these aren’t Independents. These are people who are fed up with the direction their party has taken.

“Enthusiasm completely controls the outcomes of general elections.”

So what does this mean? First, it means that the current polling methodology is USELESS. Second, it means something else is driving the outcomes of general elections. Something other than politicians moving to the center to capture the “all important” Independent vote. So what is driving the outcomes of general elections?

ENTHUSIASM.

PERIOD.

FULL STOP.

I wrote a couple of blogs back in 2016 that touched on this that are still up on The Gatherer. I went back and re-read them the other day and everything in them still holds true. I wouldn’t change a word. Enthusiasm completely controls the outcomes of general elections. It may also be true in the primaries, but each party has so much power to control the outcomes of their primaries, that it probably isn’t as big of a factor. Think about it. If your party’s candidate isn’t firing up the base, how on earth do you expect the less idealistic voters in your party to get motivated to vote? And isn’t that exactly what we see going on right now in the Presidential campaigns?

“Enthusiasm for Trump is sky high. Enthusiasm for Joe Biden is practically non-existent.”

In my city, Santa Fe, New Mexico, you could potentially drive around all day and not see a single solitary campaign sign or sticker for Joe Biden, and this town is super left-wing. Of course, you don’t see that many Trump signs or stickers either, but then again, Trump’s voters don’t want to get beat up or have their car or truck get keyed. And they generally like their liberal friends and don’t want to be ostracized by them. Which is a very real possibility.

But I can assure you, there are plenty of Trump fans in Santa Fe and they will be coming out of the woodwork to vote for him on November 3rd. There may not be enough of them to overcome the Democrat die-hards, but their numbers won’t be insignificant. One thing is certain, there will be many, many more voters for Trump in New Mexico this time around than there were in 2016. When I speak privately with my conservative and centrist friends who plan to vote, they say they cannot wait for November 3rd to get here. Enthusiasm for Trump is sky high. Enthusiasm for Joe Biden is practically non-existent. It wasn’t that long ago that New Mexico was considered a “purple” state. If my theory is correct, it likely still is. And if that’s true, then Trump and the Republicans may very well just flip this state red this fall.

So what are the main takeaways from this idea that there really aren’t any Independents to speak of? First, to get accurate polling, pollsters should be polling enthusiasm. Period. Anything else is going to give a poor result. Second, politicians who want to win, need to stop swinging back to the middle to try to capture the so-called, but really non-existent, Independents. Does Trump ever walk anything back and apologize for anything he says? No. He doubles down. If there were any truly independent voters out there, doubling down would be the kiss of death for a candidate. Instead, what we see is his supporters getting even more fired up when he does that. That should be all the proof we need.

Author’s note: My apologies to Libertarians out there, whom I have left out of this analysis, but they so rarely have an impact on elections, that I couldn’t see how to fit them in.

Following the Data: COVID-19 and the State of New Mexico

July 13, 2020

by Devin

When a new SARS-type virus from China began to make the news this past winter, like so many others before it, my first inclination was to ignore it. It sounded just like all the others to me, much to do about nothing. But then it became much more apparent that this one was going to be different. As momentum started building and the disease we now call the all-too-familiar name COVID-19 was formally declared a pandemic, states and countries started mobilizing their health agencies to gather data on this novel coronavirus. Various organizations began compiling all this state and international data, collating it, and posting it on their websites. At this point I began taking a much greater interest in it. Looking around online, I found?The COVID Tracking Project(https://covidtracking.com), sponsored by?The Atlantic. Drilling down I was able to locate the historical data for New Mexico (https://covidtracking.com/data/state/new-mexico#historical).

The first thing I noticed is that the data are presented in a less-than-optimal way to understand what they mean. While the numbers for New Tests are given as a daily total, the numbers for Cases, Hospitalized, and Deaths are provided as cumulative totals going all the way back to the beginning of the data set. Cumulative totals are deceptive in that they are much larger than daily totals and they don?t tell the real story. In order to find the real story, you have to subtract the previous day?s total from the current day?s total to get the current daily total, for every single day of the data set. So I did this, and the results are very interesting.

As I did the math to get the daily totals, I typed them into three spreadsheets. The first spreadsheet tracks the daily deaths from COVID-19. These numbers are among the most important, because most people agree that the state of being dead is not particularly debatable and because ultimately, death is the worst and most feared potential result of becoming infected with this disease. The second spreadsheet tracks daily new cases of COVID-19. The key metric in this data set is not so much the number of cases, but the daily number of new cases per daily number of new tests. This value is expressed as a percentage since there are far fewer cases than tests. This value is so important because it removes the effect of variability in the daily new tests numbers. In other words, what we really want to know is not how many new cases there are, but are more people really catching the disease? The third spreadsheet combines the data for daily new cases, new deaths, and new hospitalizations, so that we can see how they compare.

What the Data Actually Tell Us

The Death Count

Starting with the daily number of deaths the first thing we notice is that deaths from COVID-19 peaked around mid-May with an average daily value of approximately 8 deaths statewide per day. Daily deaths continued to decline after this until early July when the average daily value bottomed out at approximately 3 deaths statewide per day. The daily average rose back up to 5 by the end of July and lately has been trending back down. New Mexico is currently averaging approximately 4 deaths statewide per day. The overall average daily number of deaths since the first case was recorded in New Mexico is approximately 4.5. According to the U.S. Census Bureau, the most recent estimated population of New Mexico (for 2019) was 2,096,829, or roughly 2.1 million.

As of August 11, 2020 the total number of deaths from COVID-19 in New Mexico was 688. Given a population 2.1 million people, that means that we have experienced approximately 32.5 deaths per 100,000 residents for the entirety of the pandemic. By comparison, during the 2018-2019 influenza season (Oct.-May), New Mexico experienced 11 deaths from pneumonia and influenza per 100,000 residents and in the prior season, 14 deaths per 100,000 residents (New Mexico Epidemiology, Vol. 2019, No. 10). So far, that means that in New Mexico, the chances of a random resident dying of COVID-19 are roughly two and half to three times their chances of dying from pneumonia or influenza in recent years. Not to diminish the value of every human life, but these are not large numbers. On any given day, a random New Mexico resident has approximately 2.1 chances in a million to die from COVID-19. Those are pretty good odds. A graph of the daily new deaths from COVID-19 is shown here:

The dotted trendline approximates the running average and notice that generally speaking, the line is relatively flat and there is no spike in deaths corresponding to the ?spike? in cases. A very slight uptick, yes, but nothing anyone could realistically call a spike.

New Cases

Moving on to daily new cases, what we see in the data are two things. First, from the beginning, we see a rise in daily new cases, which leveled off in early May at around an average of 140 new cases per day and declined slightly until the beginning of June, settling at around an average of 130 new cases per day. Second, starting in early June, we see another significant rise in daily new cases, peaking most recently in late July at around an average of 300 new cases per day. For the most recent two weeks, daily new cases have been trending downward with the most recent 3-day average as of August 11 being around 160 new cases per day. At first glance, these numbers would seem to support the claim that New Mexico experienced a ?second wave? of infections from COVID-19, which has now passed (see the graph below); however, a closer look at the data reveals the truth.?

First, as testing began, our testing capacity was extremely limited. As time progressed, our testing capacity increased; however, in order to make sure we had enough tests to go around, we limited testing to those who were experiencing actual symptoms of COVID-19. That made sense. If you?re not sick, you?re not in immediate danger of death from COVID-19. Second, in late April, our testing capacity increased dramatically and we opened up testing to anyone who wanted to be tested. A ?case? was redefined in the process from being someone who exhibited symptoms?and?tested positive for COVID-19 to?anyone?who tested positive?for COVID-19,?even if they were entirely asymptomatic. To find out the impact of increased testing and the new definition of a ?case,? we have to remove the effects of these variables from the data. To do so, we simply divide the number of daily new cases by the number of daily new tests performed. The result is startling. As you can see quite clearly in the graph above and the graph below, the daily number of new cases correlates very closely to the number of daily new tests performed since the beginning, with very minor small-scale variations.

The next thing to notice is the difference in the scales of the two graphs. The scale for New Tests is approximately 23 times as high as the scale for New Cases! To see what that means visually, the next graph, below, shows the two data sets on the same scale. It is difficult to see a spike in new cases in this graph.

And now, the most important graph of all, the graph that tells the real story, is the graph showing the weekly average number of new cases per test, below.

The peaks on this graph are artifacts of the data set that resulted mostly in the beginning of the pandemic when data collection was still being organized and the initial challenges that go along with any major data collection effort is undertaken. As you can see, over time, as the data collection process became more streamlined, the line becomes more stable. The trendline, shown as a dotted line, provides a more accurate picture of reality. And the reality is, since the beginning of June, when the ?spike? in new cases began, the weekly average number of cases per test has not exceeded 5 percent and for the last 4 weeks this average has been in decline. For the most recent week recorded, that average number of cases per test was 2.9 percent. The bottom line is, the ?spike? in cases was caused by the spike in new testing and the testing of people who were and are asymptomatic. The reality is, fewer and fewer residents of New Mexico are coming down with COVID-19 and at some point no amount of testing will give the appearance otherwise. This also means that there isn?t going to be a third wave, because there never really was a second wave. If you start hearing about a third wave, follow the data. Therein lies the truth, but that?s not the end of this story. We have one more data set to look at: Hospitalizations.

Hospitalizations

The third spreadsheet includes daily new hospitalizations from COVID-19 along with daily new cases and daily new deaths. A graph of these data is shown here:

This graph clearly shows that the number of daily new hospitalizations and deaths did not dramatically increase with the number of new cases. This is further evidence that the ?spike? in new cases was caused entirely by the extreme increase in new testing.

So what does all this mean? The number one justification for all the public health orders mandating the closure of ?non-essential? businesses, social distancing, mask wearing, self-quarantine, and staying at home except for emergencies and absolute necessities, was the oft repeated mantra of a spike in cases, the ever growing number of deaths and hospitalizations, and the rationalization that if these measures ?save just one life? they will have been worth it. The question before us is, is it worth it? Has it been worth it? I?ll leave that for you to decide for yourself, because as New Mexicans, that?s what we’re going to do anyway!

If you would like to have a copy of the Excel spreadsheet file from which these graphs were derived, you can download it here:

WUFLU-NMDownload