not aware of.
Specific claim
As of April 15th, 2020, in Italy there are over 22000 deaths caused by coronavirus.
The problem here is that we cannot talk about causality because we don’t have experimental
data. The experiment with controlled input would require taking a random sample of healthy
people and contracting them with the virus. Since nobody is going to run such a cruel inhuman
experiment, we are compelled to take observation data and use common sense judgement to classify
which cases of death are related to coronavirus and which are not.
Let’s evaluate the modified claim then:
As of April 15th, 2020, in Italy there are over 22000 deaths related to coronavirus.
There is no error in this statement however, there are inevitable errors in defining related death
criteria, because it is based on fallible human judgement. Since the way we define coronavirus
related death is fraught with pitfalls, we are going to do it in steps.
1. The starting point is that we test a person for coronavirus. If the test is positive and a person
died within one month we classify it as a coronavirus related death.
2. The problem is that we include suicides, car accident casualties and all kinds of unrelated
incidents. We need to narrow down to specific likely death cause, i.e. ARS (acute respiratory
syndrome).
3. There is still a problem with the death cause. We include people who had preexisting diseases
that would independently lead to death.
4. ARS has many causes. In order to include the case as related we need to test negative for
other pathogens.
The list goes on. There are still plenty of problems with these criteria, but let’s say these four
steps would make a reasonable filter.
Reports show that steps 3 and 4 are not even taken into account, “Case-fatality statistics in Italy
are based on defining COVID-19–related deaths as those occurring in patients who test positive for
SARS-CoV-2 via RT-PCR, independently from preexisting diseases that may have caused death.
This method was selected because clear criteria for the definition of COVID-19–related deaths is
not available.“ [45]
”Electing to define death from COVID-19 in this way may have resulted in an overestimation
of the case-fatality rate. A subsample of 355 patients with COVID-19 who died in Italy under-
went detailed chart review. Among these patients, the mean age was 79.5 years (SD, 8.1) and 601
(30.0%) were women. In this sample, 117 patients (30%) had ischemic heart disease, 126 (35.5%)
had diabetes, 72 (20.3%) had active cancer, 87 (24.5%) had atrial fibrillation, 24 (6.8%) had de-
mentia, and 34 (9.6%) had a history of stroke. The mean number of preexisting diseases was 2.7
(SD, 1.6). Overall, only 3 patients (0.8%) had no diseases, 89 (25.1%) had a single disease, 91
(25.6%) had 2 diseases, and 172 (48.5%) had 3 or more underlying diseases. The presence of these
comorbidities might have increased the risk of mortality independent of COVID-19 infection.“ [45]
It’s a faulty judgement to classify the death of anyone who tested positive as related to the
coronavirus, because it includes cases with underlying illnesses that could have independently led
to death.
Again, Italy is not unique in this respect. Data coming from other countries is skewed in a
similar way - because of media and medical circles attention, the positive test result often makes
the fatal case to be automatically classified as caused by coronavirus without examining the real
cause of death.
7