_logo_phere . . .Takin’ the “BS” out’a the BlogoSphere

In my last post, July 25th, I ran a few numbers from the CDC and the UK Ministry of Health and came to the conclusion that the next 6 months or so could see 800,000 US deaths from swine flu. This would be the largest fatal event in US history. The prediction was based on an infection rate of 40% and a kill rate of 0.7%. If either of these numbers is off, the guess is off.

Since then I have been scanning the data-horizon to see if any information is coming to light that would suggest my conclusions are way off. I get data primarily from the European CDC, which gives a daily update of European and world data. http://www.ecdc.europa.eu/en/Default.aspx

I also check out the US CDC for weekly reports. http://cdc.gov/h1n1flu/update.htm . And I find that SwineFlu.org has some very worthwhile on-going discussions that often lead to other sources of data.

Putting the bottom line here near the top, I cannot say that I have found any compelling reasons to modify my predictions, either up or down, but I am having trouble believing that 40% of the population will become infected.

You only need 2 numbers to follow what’s going on — one number represents how efficiently the virus spreads, and the other represents how efficiently the virus kills the people it does infect. Unfortunately, these numbers are all over the board.

The first number is generally presented as the percent of people in the population that get infected. The complication is that this can be expressed as the total number of people ever infected divided by the total population, or the total number of people presently infected divided by the total population. Mostly, the total infection rate is given. At the present, the number that seems to be the most accepted is that H1N1 will infect 40% of the total population over the entire course of the pandemic.

The killing efficiency of a virus is the second number of interest. One would think that the most dangerous virus is the one that kills with 100% efficiency. Not true — sort of. Influenza is a disease that runs its course very fast – days or a couple weeks, as opposed to, say, HIV, which takes years to kill the person. So if a flu virus kills 100% of the people it infects, it is killing off people faster than it can spread. It ends up as a flash in the pan. From a Darwinian point of few, such a virus is a loser.

A far more dangerous flu virus is one that kills only a moderate fraction of people it infects. The walking infected then spread the disease around. That’s H1N1.

There are two ways to quantify killing efficiency. One is the number of dead per capita. The other is the number of dead per infected persons. This second number is called the Case Fatality Rate (CFR). It is the most widely used. A CFR of 20% for instance says that the virus is killing 20% of those it infects and the remaining 80% are still able to transmit the disease. Sometimes CFR includes a time period, such as 25% per year. I don’t see that being done much with H1N1.

In my previous calculations I used a CFR of 0.7%. I got this by dividing the total US H1N1 deaths by the total cases reported as of July 25th. (Based on the latest available data, Aug10.09, the US CFR has increased to 1.2%. The CDC stopped reporting cases as of Aut10.09, so it’s no longer possible to track the number for the US.)

The European CDC has been reporting a much lower CFR for Europe – about 0.15%, or 10x lower than the US. I think the best explanation for the discrepancy is that the US is underestimating the number of cases, which results in an overestimation of CFR. Currently in Australia the CFR is running about 0.4%.

At any rate, we have a CFR range of 0.15% to 1.2%. This is interesting because estimates of past H1N1 pandemics range from 0.1% (1957 — a mild pandemic) to 2.5% (1918 — a deadly pandemic).

Compare the H1N1 figures to the estimates of avian flu (H5N1) CFR of 14% to 60% and you’ll understand why avian flu is such a scary prospect. http://jech.bmj.com/cgi/content/abstract/62/6/555

The 1918 H1N1 pandemic ran its course in three waves or peaks. In terms of deaths, the second peak, P2, was by far the worse. And so it is very significant that the CFR for P2 was 2.5%, while the CFR for P1 was only 0.5%, which is not a bad estimate for our present CFR. For some reason the killing efficiency of the 1918 virus increased 4-5x over the course of a couple months. The CFR for P3 dropped to about 1.5%.

In these terms, the goal of Tamiflu is to reduce the CFR. Tamiflu doesn’t keep anyone from getting the virus, it just shortens the course and, presumably, prevents deaths.

Of course, killing efficiency doesn’t tell the whole story. If the CFR of a virus is 50% but only 2 people get infected, then the situation is very dire for one of them, but not for the whole population. The first big question as to H1N1 is what percentage of its victims will die. The second big question is how many victims will there be.

I have no idea where the currently proposed number of 40% infection rate comes from. If you look at the current reported numbers of infections in different countries, none of them are above 1%. While I have no problem with the prediction that H1N1 could kill 2% of its targets, I just can’t get my mind around the possibility that the infection rate will go from 0.05% to 40% over the next few months.

BTW, vaccines work to decrease infection rate. They don’t help people who are infected to survive, but they reduce the number of infected people. So the pharmacological strategy is that Tamiflu lowers CFR and the vaccine lowers the transmission rate.

And so the world waits . . . with fingers crossed.

Tags: H1N1 //