How to see through bad Coronavirus statistics

Over the past few months, we have seen many coronavirus statistics. We have also seen many statistical errors, which have made some of us skeptical. My complaint is that not enough of us have been skeptical. From the point of view of someone who uses statistics as part of his work, these errors are simplistic. Allow me to name a few of the errors and show how they can deceive people.

Choose which attribute to measure (deaths versus cases). The COVID-19 pandemic is an experiment with seven billion guinea pigs: us. At the beginning of the experiment, the cases and deaths were counted. The cases were determined with a doctor carefully examining the patient and deciding if this is from COVID-19. The deaths were determined with a doctor following the patient until the patient's unfortunate demise and again deciding if this is from COVID-19. The latter decision is likely more accurate, so it is a more reliable statistic. Counting the cases is still useful because it leads to a probability of death given infection. In the first few months, this probability varied a great deal. This means that one of these statistics is poorly measured. The inaccurate statistic is undoubtedly the number of cases.

Measure the attribute the same way throughout the experiment (avoid changing the way "cases" is defined). Initially, cases were defined with a doctor's examination. Later, a test for the virus was used. Later still, a test for the virus or antibodies was used. Currently, the patient and anyone with whom he had recent contact are called "cases." The CDC is responsible for much of this. For example, the CDC has been recently asking states to change the way they define cases. If that does not sound confusing enough, only some states are making this change, so, state by state, statistics will differ.

Look for problems with data integrity. There have been accusations of people fudging the data. For example, the CDC has been accused of changing old statistics.

Normalize the data (take into account different population sizes). Often, COVID-19 statistics are displayed state by state. Does this link say New York and California have a lot of COVID-19 deaths or just a lot of people? It turns out that as of 17 July 2020, New York has 167 COVID-19 deaths per 100,000 people, and California has just 19.

What are the consequences of these statistical flaws? Because of the change in counting the number of cases, that number has greatly risen over the past month. It looks so different from the COVID-19 death trend that it is obviously flawed. From the point of view of the typical reporter, looking at cases makes better news, and it beats up on President Trump.

Now Florida mistakenly looks like a leper colony, so even though Florida has only one eighth as many deaths per 100,000 people as New York, New York governor Andrew Cuomo declared that people visiting Florida and returning to New York must quarantine themselves for 14 days. Cuomo, whose state is tied with New Jersey as the worst state in the union, as determined by counting deaths, is jealously blaming a better run state for supposedly being worse than New York. Cuomo should put his emotions aside and try to improve his job performance.

Currently, the patient and anyone with whom he had recent contact are called "cases."

That’s just not true. The CDC definition of cases doesn’t include contacts.

Because of the change in counting the number of cases, that number has greatly risen over the past month.

Funny that he doesn’t offer any evidence to back up this claim.

Trump says it’s the increasing number of tests. Others point to the increasing test positivity rate to say actual cases are increasing. This guy just pulls a notion out of his ass.

There are real problems with the reporting on and understanding of COVID statistics but this idiot’s only making it worse.

But it will buy him a few clicks from the credulous.

The COVID-19 death count is completely fraudulent when compared to the influenza death count.

The CDC counts presumed COVID cases as confirmed cases. Nothing like that applies to influenza.

Also, the CDC counts 100% of pneumonia + COVID deaths as COVID deaths.

In sharp contrast, the CDC counts less than 10% of pneumonia + influenza deaths as influenza deaths.

COVID-19 is a severe common cold virus that is life threatening for the elderly and the infirm.

Protect the elderly and the infirm.

Everybody else - please return to your normal life.

Eventually that needs to be called out for what it is, but apparently Trump does not feel that he has the time to fight that battle right now.

He has all the time there is. He needs to introduce a new medical advisory team filled with actual, practicing physicians at a surprise press conference. Doing so will make Fauci and Birx come out in support of Joe Biden - but that will just be a formalization of an existing reality.

I can presume revealing the fake statistics - like many other topics - is on hold until Biden becomes the formal nominee of the Democratic Party. It’s already too late to change the minds of the Karens - many of whom will now wear their masks to the grave.

"The CDC counts presumed COVID cases as confirmed cases. Nothing like that applies to influenza."

Actually, just the opposite. Influenza statistics are ALWAYS estimates. Look at the data for the Disease Burden of Influenza: 2016-2017 season there were somewhere between 25 million and 45 million cases. 2017 - 2018, 39 million to 58 million cases. They can't tell you within 5 million how many actual cases there were because it's impossible to know. Few cases escalate to medical treatment. It's all statistical analysis and estimation.

Even hospitalizations and deaths are estimates with wide ranges. 2016-2017, hospitalizations for Influenza were between 380,000 and 860,000. Deaths were between 29,000 and 61,000. Even those ranges are only within a 95% confidence level. And even those Influenza statistics take two years to go from "preliminary" to final. The 2017-2018 Influenza season numbers are still listed as preliminary.

This happens with every disease. CDC never gets perfect statistics. Somehow with COVID-19, there's an expectation that they'll know - immediately, to a person - how many people are infected, hospitalized, and have died. It doesn't work like that. It never works like that. Not for any disease. People seem to think our government is usually doing a much better job than it actually is. It's the same government that conservatives like President Reagan have said was too expensive and ineffective for many many decades. It hasn't gotten better with age.

Thanks for your comments.

I am painfully familiar with the influenza data chaos on the CDC website.

They have at least two sub-agencies collecting data on respiratory infections, plus one more non-government monitor that also collects data.

The data I referred to in my Comment comes from the National Center for Health Statistics (a CDC agency).

NCHS files a weekly report on “Pneumonia & Influenza” deaths.

All NCHS P&I data comes from USA death certificates, which each state must file in a timely manner (about 80% electronic filing now).

Naturally, death certificates have their own accuracy issues, but we will postpone that subject for another time.

More importantly, at least half of pneumonia deaths are preceded by influenza or the common cold.

However, according to the NCHS data, there are 25 times more pneumonia deaths than influenza deaths!

Compare that to COVID + pneumonia deaths, where there are zero pneumonia deaths and 100% COVID deaths.

Bottom Line for me...

We need to immediately start counting influenza and common cold deaths EXACTLY the same way we count COVID deaths.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.