Danny's Weblog

Recent suspected SARS cases in Asia

The recent cases more or less correspond to my expectation: ie, SARS is now observed in a mutated form, corresponding to the standard mutability of the common cold, another coronavirus.

In contrast to my earlier doominess, I have to say that the news coming out of China makes me feel much more optimistic than I did. The reason SARS was able to take hold last flu season was because the Chicoms actively attempted to suppress news about it. (It's funny, this year news reports are casually referring to that whereas last year they pussy-footed around the issue.)

My impression is that the infectivity of SARS is actually not all that high. Very probably previous models of SARS have surfaced in the human population on numerous previous occasions but died out naturally at a level below observation. (Hmmm... I wonder *how many times already* the Chicoms suppressed a SARS epidemic? Or *other* epidemics?) On the other hand it was amusing to hear a WHO spokeswoman on TV today admitting that "our best weapon is screening and contact tracking" (not an exact quote) – what she was trying to avoid saying is "we still have no vaccine or any other effective treatment".

This low infectivity combined with aggressive countermeasures by the Chicoms will quite possibly prevent a major deathtoll or breakout of SARS into other vulnerable populations, eg India or Indonesia.

WHO releases new report on SARS epidemiology

www.who.int [http://www.who.int/csr/sars/en/WHOconsensus.pdf]

This should be read with reference to my previous article on SARS: www.panix.com [http://www.panix.com/~dannyw/weblog/2003/10/11#sars01]

Btw, in the WHO's document, "R0" corresponds to my "Tv"" and "R" corresponds to "Iv".

I particularly noticed the following (I quote from WHO's .pdf):

3. Case-fatality ratios

3.1. Simple methods for calculating case-fatality ratios (CFRs) from aggregate data will not give reliable estimates during the course of an epidemic.

That's very true, and it's always been very true, and it's something that WHO certainly was *not* saying during the epidemic. Quite the reverse. Don't panic!

Incidentally, I've *never* seen an estimate of the *survival* rate from SARS. Just how many people have actually recovered yet? How do we *know* that they aren't carriers?

And that's just *one* issue. The WHO admitted it actually had no useful estimates on practically all the other parameters!

Many recommendations themselves are catastrophic if they are to be taken seriously:

4.3. WHO to recommend that persons who have an acute febrile respiratory illness should not travel until their symptoms have resolved.

What that means is, because SARS can kill an unknown number of people from a single case, we'd better not let anyone travel who *might* have SARS. But how many people had "an acute febrile respiratory illness" last year? How are they, in practice, to be stopped without a huge system of scanners at the airports etc? And what happens if one is *not* detected at the departure airport, and detected at the arrival airport?

Here's another thing that paranoid worrywart ol' me was grizzling ineffectually about in the previous article:

5. The presence and significance of subclinical infection

5.1. Centres to complete serological testing of cohorts of contacts of probable and suspect SARS cases to determine the proportion of contacts who developed symptomatic and asymptomatic infection.

5.2. WHO to synthesize the results of serologic testing of SARS contacts at a global level.

5.3. WHO to facilitate Centres pooling data and experience on unusual laboratory findings (for example isolated SARS-CoV positive serology or positive polymerase chain reaction (PCR) in individuals with no or minimal symptoms) so as to determine the public health significance of these events and the action they should trigger.

Asymptomatic infection: isn't that just a little scary? Isn't it more scary now that WHO admits it needs to know?

On the other hand, it stoops to "suggestio falsi" elsewehere in the very same document:

There are currently no reports of the transmission of SARS from asymptomatic individuals.

And the incubation period... I remember various stories about how no new cases had been observed for a week, so a certain location was certified clean. Now WHO admits the evidence for those statements was never collected:

Outliers beyond a 10-day maximum incubation period are few in number and have not necessarily been subjected to rigorous and standardized investigation. However, it was noted that other mammalian coronavirus infections have long right-hand tails for incubation periods, so a long tail is also biologically plausible for the SARS-CoV.

The "death rate" figure has edged up again. It was 9.6%: www.who.int [http://www.who.int/csr/sars/country/table2003_09_23/en/]

(2.1) A global case-fatality ratio of 11% was recorded at the end of the outbreak (see also III.3). These figures may be revised again following a process WHO has begun with all centres that reported cases to close off the historical data set of the outbreak.

Btw, is there anybody in the world who understands the string of weasel words and fog in that statement?

The following seems pretty clear:

The data set will be set up initially to refine estimates of the incubation period.

Of course, we were being reassured way back in April that the outbreaks had been contained because no new cases had been found after the incubation period had elapsed... so that's all right then.

New study on SARS explains inconsistent statistics

When the first information started to appear about SARS I did some quick calculations and became extremely gloomy about the outlook. I was guessing one million cases by the end of the year.

Well, boy did I look stupid. My only argument was to claim that the chicoms must have been secretly shovelling truckloads of corpses into limepits.

At the beginning of October 2003 however I saw an interesting report on new research into the stored samples which explains why my forecast was not fulfilled.

Before I explain what that report said, I want to explain my basis for expecting a very high casualty figure.

In an epidemic, two numbers are very important in producing a forecast of casualties. One is the "transmissivity" (Tv) – how many new victims are infected by each existing victim. Another is the "infectivity" (Iv) – the likelihood of an infection being passed on between two particular people under particular circumstances. Fairly obviously, for an epidemic to start the Tv has to be greater than 1 at the beginning. For it to finish, we have to adopt measures which reduce Tv below 1.0, by examining Iv under various circumstances and trying to eliminate circumstances in which Iv is high.

It's important to be clear on the concept that Tv and Iv are not *fixed* numbers, constants for a particular disease. Tv is a number we are continually striving to minimize, and Iv is a function of conditions, and probably the race of the victims, a particular strain of the disease, etc.

The problem is that in the early stages of an epidemic of a *new* disease such numbers are impossible to estimate accurately from the only figures you have, ie the symptomatic victims. One reason why I was so confident that the estimates being produced by public health authorities were completely preposterous was that it was impossible even in theory to know the values of the basic parameters on which such estimates need to be based! In addition, I was flabbergasted that the authorities made the elementary statistical error of confusing the death rate with 100%-(survival rate), when we still had no clear data on the variation in the survival rate as a function of time from onset in the individual. When I first saw that such estimates were being made I vaguely thought "oh well, they must have used some advanced statistical technique to model, erm...", but no. The fact that this was admitted without heads rolling at every level of WHO is appalling.

WHO is now reporting that worldwide the fatality rate is 9.6 per cent: www.who.int [http://www.who.int/csr/sars/country/table2003_09_23/en/] and they have removed earlier pages which suggested far lower fatality rates (around 2.5%). They do not try to explain the difference.

Likewise, we were repeatedly assured that the epidemic had been contained at some location – only to find new cases. In other words, we simply did not know all the conditions under which Iv was high.

Even today, I am not confident that the death rate for SARS victims is being reported accurately. At the peak of the epidemic, I was very struck by the fact that the death rates being reported seemed to be directly related to one's degree of confidence in the truthfulness of the country's regime! Ie, high Tv and Iv in Canada, low in Red China.

Accordingly I was predisposed to make two assumptions which had a huge effect on my forecasts:

1. The numbers of infectees were rising fast in Hong Kong. Because I allocated Hong Kong, based on its open traditions and high medical standards, a high credibility value, I took those numbers as the most reliable values of Tv for forecasting.

It was particularly interesting that the rate remained relatively constant for a long time. This suggested to me that the measures being taken had absolutely no effect on Iv, in other words a technologically advanced society was still totally vulnerable. And the figures for victims were increasing at a rate which translated as *doubling every month*.

2. Now consider Red China. It was quickly evident that the source of the infection was there, and it quickly became more and more obvious that the Red Chinese had been covering it up for a considerable period. Victims had been grabbed and taken to closed military hospitals, with no information to their families. As the epidemic wore on it was admitted that signs of the new disease had been apparent as long ago as the previous year.

Their credibility was particularly suspect because their numbers both for new infections and for death rates were far lower than Hong Kong.

Based on this, I adjusted my personal estimates of the situation in Red China way upward. My feeling was that the Reds were announcing numbers which corresponded only to the minimum numbers they could get away with. I further felt that the Reds, while ruthless, were highly unlikely to have efficiently detected and isolated victims, especially the extremely significant, and largely ignored in discussions, group of asymptomatic carriers.

I therefore made a handwaving estimate that Red China was suffering not the 1000 or so victims it conceded, but probably 10000 or more: in other words, that the epidemic was completely out of control in Red China already.

Putting together those two factors meant that the doubling in victim numbers would continue indefinitely: so by the end of the year (about 9 months) one could expect 2**9, ie 512, times as many victims...

Another reason for gloom was that SARS had been identified as a coronavirus, ie essentially the same as the common cold. The problem about this is that the common cold mutates very rapidly; at any time there are several strains detectable, but within 6 months to two years a new set of strains is present. In other words, SARS was likely to mutate faster than we could create a vaccine (a process which normally takes at least 2 years). Also, victims of previous infections would have no immunity in later outbreaks!

Now let's consider the new research I mentioned at the beginning.

It has now been discovered that the strain of the virus which affected most victims in Red China was *far milder* than that in most of the rest of the world: both in terms of Tv and death rate. That's what the numbers were saying, but for me at least the Red Chinese had no credibility whatsoever, and I had no idea that there was a real genetic difference between the two strains.

This has its good and bad sides.

One is that I feel vindicated in my original analysis!

Another aspect is that it confirms my original view that SARS is, as a coronavirus, highly liable to mutation. Just as there are normally several strains of the cold virus "in circulation" at any time, in this case there were two (at least). The human race is extremely lucky that China, with its teeming masses of impoverished slaves held at close quarters, was largely exposed only to the milder version.

On the other hand, how is SARS liable to mutate *next*? It seems to have a reservoir in animals which would be impractical to eradicate, even if we were sure we had really isolated every asymptomatic carrier. So what do we have to expect in the next cold season? You can reasonably expect me to be an inexhaustible fount of pessimism, so here it is: I think the *original* form had to be the mild one, and it rapidly mutated into the almost-catastrophic one as soon as it entered human hosts.

I don't like that trend line.

Btw, having now done a websearch for this concept of "different strains", I see it was in the news as long ago as 2003 April: www.washingtonpost.com [http://www.washingtonpost.com/ac2/wp-dyn/A39029-2003Apr25?language=printer]

I guess when I saw it them I must have written it off as speculative happytalk, but it has now been confirmed. Of course things could still be worse (Aug 28): www.wired.com [http://www.wired.com/news/medtech/0,1286,60218,00.html]

Debug: hittotal: 29 startban: 0 dancookie: endbandate: banned: 0 tempdate: tert: jse: jsno jsh: 29