COVID-19 - A SKEPTICAL VIEW
VLADIMIR KOLIADIN - 22 MARCH 2020
OBJECTIVE: A simple default hypothesis is analyzed that the "COVID-19 epidemic" has originated from a re-classification of usual seasonal flu cases, and that there have been no increase in the death rate from flu-like diseases.
METHODS: Comparison of the apparently new COVID-19 disease to the usual pneumonia cases in respect to the death rate and fatality rate for given population.
PREDICTIONS OF THE NULL-HYPOTHESIS: If the default hypothesis is correct, then the following is expected: (1) the death rate from influenza and pneumonia (including COVID-19 cases) is at its normal levels for given population, (2) fatality rate in COVID-19 cases would vary and depend essentially on the selection criteria for screening, but does not exceed the baseline level for pneumonia; (3) the distributions of the death rate and the fatality rate from "influenza and pneumonia" over age-groups would be of the same shape as in previous years.
BASE LEVEL FOR COMPARISON: The death rate from "influenza and pneumonia": USA (2017) - 17 (1/100,000/year), or 55,000 deaths; Italy (four recent seasons) - 11.6 to 41.2 (1/100,000/year) or from 7,000 to 25,000 (the average is 17,000) deaths per season. Case fatality rate: USA - 4.5%-5.0% per case (1.0-1.3 million cases of pneumonia and 55,000 annual deaths), Italy - 8% for community acquired pneumonia. The age distribution of usual deaths from pneumonia is essentially skewed towards the older ages: 93% of the deaths are in patients older than 55 (USA, 2017).
RESULTS: (1) The picture of the putative "COVID-19 epidemic" is in agreement with the predictions of the null hypothesis. (2) There are not any signs of an unexpected increase in the death rate and fatality rate among the patients diagnosed as COVID-19 as compared to the levels observed in usual cases of pneumonia. (3) The age-distributions of the death rate and the fatality rate correspond to their normal shape, essentially lopsided towards the older age-groups.
A central tenet of data based scientific inference is comparison of the observations to some base level, or the null-hypothesis. Let us try to apply this simple and clear principle to known facts about the so called "COVID-19 epidemic". The null- or default-hypothesis here is that COVID-19 epidemic has no essential differences from usual epidemics of seasonal flu, at least in respect to mortality.
The case-selection mechanism for diagnosis "COVID-19 disease" works as follows.
(1) If an individual has had fever for more than 4-5 days and has cough, he/she is tested for COVID-19 or/and asked about contacts with potentially infected.
(2) In some countries, like Italy, population is asked to stay at home with flu-like symptoms as long as possible. To be "socially responsible", they often seek for medical help only after 7-10 days of fever and cough, when their state gets unbearable.
(3) The case-selection mechanism is radically different in different countries. In some countries even individuals with relatively mild flu-like symptoms are tested. In other countries, only severe enough symptoms trigger the testing procedure.
What should we expect if to select individuals this way from a general population at a flu season IRRESPECTIVE to any particular virus?
PREDICTIONS OF THE DEFAULT HYPOTHESIS
Prediction 1: High prevalence of pneumonia. Normally, seasonal flu, if not complicated by pneumonia, shows positive dynamic (decline of fever) on day 3-4. Persistent fever, if observed after days 3-5 from the beginning, is a strong statistical indicator of pneumonia.
Prediction 2: Maximal fatality rate is expected in the countries where only patients with severe enough symptoms are tested, like Italy. The rate would approach the fatality rate specific to influenza and pneumonia at the country. On the other hand, if testing is performed for general population or for individuals with light flu-like symptoms, then the fatality rate would be essentially lower.
In the USA, for example, fatality rate in pneumonia cases is about 4.5 - 5.0% (1-1.3 million pneumonia cases and about 55,000 deaths per year). Only in some countries or separate provinces this level has been reached for COVID-19 cases. In many countries, the fatality rate reported is essentially below this level. For example, at March 21, 2020, the deaths-to-cases ratio was the following: Italy (8%), Germany (0.34%), France (3.5%), Switzerland (1%). Such a radical difference in fatality rates between countries is hardly compatible with the virus as the main causal factor.
Prediction 3: High prevalence of abnormally severe cases of pneumonia in the patients who had not received adequate diagnostics and anti-pneumonia treatment at the earlier stages of their flu-like disease.
In some countries (e.g. Italy), patients are given (and follow) instructions from medical authorities to seek for medical help only after 7-10 days of persistent fever and cough (see the quotation below). It is natural to expect severe two-sided forms of pneumonia with a much higher mortality rate. If pneumonia normally starts at day 3-5 of a flu-like disease, then, in these cases, the patients are devoid of adequate anti-pneumonia treatment at least for 5-7 days. This is more than enough for pneumonia to become two-sided and fatally severe.
Prediction 4: Mortality rate should strongly correlate with higher age. Most deaths from pneumonia normally occur in senior age-groups. For the USA, for example, the age-dependence is as follows.
Table A: Death rate and the number of deaths from "influenza and pneumonia", USA, 2017
AVERAGE DEATH RATE PER 100,000/YEAR: 17.1
TOTAL CASES: 55,671 (100%)
(See Table 5 on page 31 and Table 6 on page 35 in the following document from the Centers of Disease Control and Prevention of the USA: https://www.cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_09-508.pdf)
Prediction 5: If to project US data onto a country like Italy (population: 60 million), then approximately 10,000 annual deaths (60 million * 170 deaths/million) from pneumonia are expected. And most of them are expected to be concentrated within a 1-2 months, usually at late winter - early spring.
The following is known from the latest epidemiological research for Italy, e.g. from a paper published in the "International Journal of Infectious Diseases" at November 2019:
...Italy showed a higher influenza attributable excess mortality compared to other European countries, especially in the elderly...
...We estimated excess deaths of 7,027, 20,259, 15,801 and 24,981 attributable to influenza epidemics in the 2013/14, 2014/15, 2015/16 and 2016/17, respectively, using the Goldstein index. The average annual mortality excess rate per 100,000 ranged from 11.6 to 41.2 with most of the influenza-associated deaths per year registered among the elderly...
(Source: see https://www.ijidonline.com/article/S1201-9712(19)30328-5/fulltext)
If to use winter season 2016/2017 as the baseline for Italy, then the expected number of influenza-related deaths for one season would be about 25,000. Right now (March 22, 2020), the number of deaths attributed to COVID-19 is 4,825 (https://www.worldometers.info/coronavirus/).
Reported fatality rate in community acquired pneumonia cases: Italy - 8-9% (https://www.sciencedirect.com/science/article/pii/S0954611105001782), Spain - 8% (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0200504)
The highest fatality rates for COVID-19 positive cases never exceeded the baseline level for influenza and pneumonia expected for given population (5-8%). In many countries and regions the rates are essentially lower. The greater the strictness of the criteria for COVID-19 screening, the higher the fatality rate: just because a higher proportion of pneumonia cases got into the sample. And the reverse: if screening for COVID-19 is carried out in a broader population, having no serious symptoms, then fatality rate is much lower (again, simply because a fewer fraction of pneumonia cases gets into the sample).
The picture of the putative "COVID-19 epidemic" is in perfect agreement with the predictions of the simplest null hypothesis: artificial selection of individuals with seasonal flu-like diseases at their later stages and their reclassification under the new name is the main origin of the "epidemic".
There are no signs of even a slight unexplained increase in the death rate and fatality rate among the patients diagnosed as COVID-19 as compared to the levels observed in usual cases of influenza and pneumonia.
The shapes of the age-distributions of the death rate and the fatality rate correspond to their normal shape, essentially lopsided towards the older age-groups.
The putative "epidemic" causes diversion of huge medical resources, both human and technical, from the normal treatment of a broad spectrum of various diseases and conditions, including dangerous ones. This shortage of medical attention would definitely result in extra deaths, not that noticeable as the COVID-19 related deaths widely publicized by media and medical authorities, but much more numerous.
As for the economic consequences of the aggressive and panic-driven "anti-epidemic measures", they may be really fatal for the global economy. And each 1% of the economic decline is followed by millions of new extra deaths worldwide, especially in developing countries.