Featured image: Covid-19 case from JHU’s Center for Systems Science and Engineering (CSSE) in New York, USA (9.04.2020). Photo: KOBU Agency / Unsplash, (CC BY-SA)
Without analysis of life statistics, false information about COVID-19 statistics on social media can easily be incorporated, especially without the proper context.
For example, we may choose statistics that support our view and ignore statistics that indicate that we are wrong. You also need to interpret these statistics correctly.
It’s easy to share this false information. Many of these statistics are interrelated and can quickly lead to misunderstandings.
Here’s how to avoid five common errors and impress your friends and family by getting the statistics right.
1. What is scary is the infection rate, not the mortality rate
However, these posts overlook the infectivity of COVID-19. To do this, you need to look at the case fatality rate (IFR). This is the number of COVID-19 deaths divided by all infected individuals (estimated only at this stage, see point 3 below).
I’m really tired of monitoring the horrors of MSM. Coronavirus is only a mild upper respiratory tract disease. It’s less deadly than the flu. Don’t believe everything you see.
— Stephen Harrison (@ Stephen51037986) February 29, 2020
Next, and more importantly, we need to find out the basic reproduction number (R₀) of each virus. This is an additional number that is estimated to infect one infected person.
Influenza R₀ Is about 1.3.. The estimated value of COVID-19 is different, but its R₀ is Mid value of 2.8.. The jump from 1.3 to 2.8 means that COVID-19 is much more infectious than influenza because of the exponential increase in infection (see below).
Combining all these statistics reveals the motivation behind the “limit spread” public health measure. Not only is COVID-19 very deadly, it is deadly And Highly infectious.
2. Exponential growth and misleading graph
A simple graph may plot the number of new COVID cases over time. However, statisticians are interested in the rate of increase in total cases over time, as new cases can be reported on an irregular basis. The steeper the upward slope of the graph, the more you need to worry.
For COVID-19, statisticians consider tracking Exponential growth In case. Simply put, unlimited COVID cases have the potential to continually increase even more cases. This gives you a graph that tracks slowly at first, but curves sharply upwards over time. This is a flattening curve, as shown below.
However, social media posts routinely compare COVID-19 numbers with other causes of death numbers.
Even if researchers talk about exponential growth, they can still be misleading.
Israeli professor Widely shared According to the analysis, the exponential growth of COVID-19 “fade in after 8 weeks”. Well, he was obviously wrong. but why?
“Israeli professor provides another coronavirus prediction – Yitzhak Ben-Israel believes that Covid-19 spread will be nearly zero after 70 days.”https://t.co/6OcxevfRmI Stephen Bryen: “I don’t believe the global approach of performing lockdown is the right solution.” pic.twitter.com/alLUEDHxDZ
-Richard Falknor (@highblueridge) April 21, 2020
His model assumes that the cases of COVID-19 increase exponentially over the days, rather than a series of transmissions, each of which can take days. This allowed him to plot only the irregular growth in the early stages of the outbreak.
A better visualization truncates these unstable first cases, for example by starting with the 100th case. Or use an estimate of the number of days it will take for the number of cases Double (About 6-7 days).
3. Not all infections are cases
Next, there is confusion about COVID-19 infection and cases. In epidemiological terms, a “case” is a person who is diagnosed with COVID-19 primarily by a positive test result.
However, there are far more infections than cases. Some infections show no symptoms, some symptoms are minor people think it’s just a cold, tests aren’t always available to everyone who needs them, and tests Not all infections are picked up..
Infection is the “causal” case, and testing finds the case. US President Donald Trump was close to the truth When he said In the United States, the number of cases was high due to the high examination rate. But he Other It’s still completely wrong.
The more tests that are good (we are the most in the world) are comparable to the more Cases, Fake News Gold. They use the case to blame the incredible work done by the great men and women of the United States fighting the Chinese plague!
— Donald J. Trump (@realDonaldTrump) August 11, 2020
Do not test more result Often it is More accurate quote Of the true number of cases.
Epidemiologically, the best strategy is to test as widely as possible, rather than less, to minimize the discrepancy between the case and the overall infection.
4. It is not possible to compare cases with the same date as death
Therefore, deaths recorded on a particular day reflect deaths from cases recorded weeks ago, recorded weeks ago. Less than half Current number of cases.
Rapid case doubling time and prolonged recovery time also create a large discrepancy between counting numbers. Active case and recovered case.. Let’s look back at the real numbers.
One of the things I’ve noticed about this COVID-19 madness is that the media is very obsessed with spreading a few dead and not reporting on the mass recovery. Why stick to spreading fear and panic?
— Orby (@Nutty_Lulu) March 17, 2020
5. Yes, the data is messy, incomplete and subject to change
— 🌈🔶 Christian Martin (@CAdamMartin) August 12, 2020
Cases and deaths may vary by country and state. It also takes time to collect data. That is, a retroactive adjustment is made.
Looking back, we can only see the true numbers of this pandemic. Similarly, the modeler was fooled, so the early model wasn’t necessarily wrong.
Welcome to the world of data management, data cleaning and data modeling. Many armchair statisticians do not always appreciate this. until now.
Jack Laubenheimer is a Senior Research Fellow in Biostatistics at the University of Sydney.