COVID-19 is still here and, as the song goes, getting stronger every day. Here is today's update from Dr. John Campbell, after which a few comments from me.

As of today, March 6, 2020 the official number of confirmed COVID-19 cases in US (read: Government approved disclosure via CDC) is 133. Washington State 70 cases with 10 deaths, the majority of which came from a single nursing home in Kirkland. This includes the quarantined passengers from the Diamond Princess, transferred from Japanese quarantine to U.S. for quarantine in Kansas.

Dr. Campbell makes the point that the time for proactivity rather than reactivity in the UK, and perhaps elsewhere, is now. Good point--where draw the line to begin school closure, banning large crowds and emphasized self-quarantine.

For self-quarantine to be effective EVERYONE in the household needs to self-quarantine, even if only one person self-quarantines on the presumption of COVID-19 illness. That means in many cases loss of income. It also implies loss of ability to shop for food and supplies. Americans, being the "rugged individualists" which conservatives prefer, are unlikely to effectively self-quarantine when economic survival, hence domestic survival, is at stake.

WHO reports death rate of 3.8%, based on about 3804 deaths amongst 100,000 confirmed cases. And here's the flaw in the estimates of disease infectivity and R0 (the number of contacts made ill by contact with a patient with COVID-19).

Despite all the fine medical minds involved in dealing with this disorder, dreadful ignorance still abounds.

The heart of the problem, which I have not seen adequately addressed either by Dr. Campbell or other experts, is the difficulty posed by asymptomatic cases. An asymptomatic person is defined as a "person infected with the virus, showing no obvious symptoms, who is confirmed by testing at some stage to have the disease".

The class of asymptomatics contains two subclasses.

Class A asymptomatic are those who eventually show disease signs and/or symptoms.

Class B contains those who, although infected with the disease, are never ill, demonstrating neither symptoms or signs of the illness and who subsequently recover without external evidence of having had the disease.

It is the Class B group which fouls up all epidemiological estimates.

Until wide-spread testing of large population groups, symptomatic and asymptomatic is carried out, we will have no idea of true infectivity. Yes, counting the "confirmed cases" and then counting deaths in those confirmed cases yields a rather easy calculation. For instance, with today's figures, per Dr. Campbell of ~100K infected and 3700+ dead (i.e., confirmed to have died to COVID-19) is a simple calculation.

But what of the Class B patients who will never be counted until all people in any given geographic area have been tested, regardless of symptomatology? Excluding China, known to have falsified the disease for months and likely continuing to do so, no nation has yet begun widespread testing. South Korea and perhaps Japan are soon going to be doing almost universal screening due to adequate supply of accurate test kits.

For purposes of this essay, I consider (incorrectly of course) that all test kits used are both sensitive (i.e., detect all cases with disease) and selective (detect only people with the disease). So in that perfect, yet unrealistic scenario, testing a whole population would give realistic figures about true infectivity and true R0.

Based on the rate of growth of diseased patients identified, outside China, the disease seems to be spreading exponentially. But is that so? It is quite likely that there are many, many more Class B cases than suspected.

For instance, why is the infection rate in children, up to early teens, so low? They are usually disease factories with undeveloped immune systems and crowded class rooms. The prevalence of confirmed disease in children is amazingly low. Influenza, for instance, is notorious for afflicting not just the elderly, but especially also children.

Here I propose that there is an enormous reservoir of Class B patients. This proposal can only be contradicted when large scale administration of accurate tests are administered.

An ideal example of such a situation would be to test elementary and middle school students living in an already effected country but none of whom show any signs of disease.

Consider the passengers and crew of the Diamond Princess, all of whom were quarantined for 14 days after the first diagnosed COVID-19 case. Did they all join Class A? The number of people aboard the Diamond Princess is probably 100%. The number of people confirmed infected, as of February 17, 2020 was under 500. Quarantine ended February 19. I don't find figures for the case ascertainment for the date. Let us assume, as I did in yesterday's essay that in two days, the number of affected passengers and crew doubled. That would mean less than 900 people infected. But let us consider a truly geometrical progression in disease incidence> Thus in two days, the case number would quadruple, involving 1800 people.

Two factors would then need to be considered before proceeding further.
1. Test kit accuracy--it won't be 100% sensitive and won't be 100% selective.
2. If R0is 2.6 as claimed, then each affected person should infect 2.6 other people.

Assume the worst case scenario, that the 2.6 case increase all happened on the last 2 days of quarantine. In other words, between February 17 and February 19, when quarantine ended, then and only then each infected passenger would infect 2.6 other people.

Feb 17--454 cases
Feb 18--454 X 2.6 cases = 1180
Feb 19--each of the new cases infect 2.6 more people. Remember that in this worst case scenario, all 454 patients have already infected 2.6 patients (unrealistic obviously--but still worst case). So new cases would be due to spread from those infected Feb 18. Again assuming that in the next 24 hours, all new cases infected their theoretical maximum of 2.6 then we have

(1180 - 454) X 2.6 new cases Feb 19. This equals 1888 cases acquired in the last day.

Add all this up: Feb 17--454 cases
..........................Feb 18--1180 new cases
..........................Feb 19--1888 cases new since Feb 18.
__________________________________________
Total COVID-19 confirmed cases = 454 + 1180 + 1888 = 3522 cases

The number thus calculated, including calculation requiring maximum immediate infectivity is still less than the total initial at risk population (3700). Which means using worst case scenario of rapid and immediate infectivity (R0) equal 95% total infection rate.

But this number is extremely excessive relative to actual disease-ridden people, despite being quarantined together in a limited space for 14 days. A perfect disease hot house.

But as of Feb 23, Diamond Princess count was 705 ascertained between Feb 19 - 21..

Something is amiss here. Test kit accuracy, determined by false negatives and poor selectivity, underestimated or over-estimated infection. We don't know the figure.

If the R0 figure is accurate, then obviously the rate of spread from one individual to new patients is far slower than the 24 hours postulated in my worst case example.

My feeling is that:
1. Widespread test kit availability has not yet occurred
2. Test kit standardization has not yet occurred. This means acceptably reliable sensitivity and specificity of all manufactured kits used in population testing.
3. R0 calculation of 2.6 is either wrong or the rate of infecting the additional 2.6 contacts is much slower.
4. Far from 100% of people were infected despite being confined to a disease hot house.
705 / 3700 = 19%. Pandemics require 30% community infection. This did not occur on the
Diamond Princess

Conclusion (mine)

There is a huge underestimation of disease prevalence. Prevalence means the proportion of people with currently active disease to cohort (sample) size.

Death rate will prove much lower than the postulated 3.8%

What will it take prove or disprove my assertions?

1. Widespread usage of accurate test kits
2. Adequate provision for prodromal (i.e., both presymptomatic and symptomatic cases) to test positive for the disease.

Is infectivity present after patients have "recovered"?
Is the incubation really only 14 days?
What is the average duration that symptomatic patients remain in contact with uninfected people?

Barriers to Accurate Case Ascertainment

Not only do we have the aforementioned issues, but we have governmental inactivity or malactivity. Denial of local authorities to manufacture, distribute, use and report test kit data existed until only 2 days ago (when Trump removed an Obomba-era order precluding non-federal testing usage).

Guidelines authorizing physician and health service testing is overly restrictive. In part, this is unavoidable due to current lack of test materials. But the requirements for testing are much too narrow.

Here, in California, the Department of Health in the county where my clinic is located (San Luis Obispo County) restricts testing to:

"Individuals who are symptomatic AND have a history of contact with a confirmed case of COVID-19 within 14 days of symptom onset.
Individuals who are symptomatic AND have a history of travel to China, Italy, Iran, South Korea our Seattle within 14 days of symptom onset...

In the case of COVID-19 symptoms consist of
Temperature Greater or equal to 100.4 F AND
Cough OR Shortness of Breath or Fatigue"

The first Health Department specifically excludes ALL Class A presymptomatic contacts and ALL Class B patients.

Note: designation of Class A and Class B is my creation, but the principles are valid.