Re: CDC: H1N1 could kill hundreds of thousands in the USA
Re: CDC: H1N1 could kill hundreds of thousands in the USA
Thanks for the kind words - still a lot to do. I wonder how many deaths were also not counted. Look at the world stats, they are basically the same as USA.
? 158,139 H1N1 (swine flu) infections in 141 countries and territories
? 956 H1N1 (swine flu) fatalities in 33 countries and territories
? The worldwide H1N1 (swine flu) fatality rate as of July 26, 2009 is 0.60% or one death per every 165 confirmed infections.
There are a lot of members, posters, and readers here who follow and tabulate the worldwide numbers on infectious diseases; that is one of the reasons this site is called FluTrackers. All of this tracking over the years has demonstrated that the "numbers" are not fixed and immutable. Rather the numbers and statistics are always situationally dependent on the reporting agency and country, case definitions, political oversight, and a variety of other factors.
The hardest numbers to hide or manipulate are deaths attributable to a specific disease, especially infectious diseases. Repression of information on a large numbers of deaths, particularly clusters of deaths from infectious disease, requires forced collusion of government officials, public health specialists, and news media outlets. That is why the likelihood that the CDC is withholding information on large clusters of H1N1 deaths in the USA is extremely low.
So if we can assume that the death count for H1N1 in the USA is reasonably accurate, then the only other number we need to determine the CFR (Case Fatality Rate) is the actual number of H1N1 infections. Note, we need the actual number of infections, not just the
reported number of infections. You calculate a CFR of .69% for the USA based on the number of
reported infections. But what about all of the unreported infections?
In a
June26th briefing, the CDC states:
As you may have heard yesterday, we at CDC are estimating that those reported cases are really just the tip
of the iceberg. We're saying there have been at least a million cases of this new H1N1 virus in the United
States so far this year. That's really not a perfectly accurate estimate. It's just a number, a ballpark figure,
that we think for sure there's been more than a million of these new infections. We know we're not tracking
every single one of them.
This seems to indicate that the number of H1N1 infected individuals today is probably many more than the million cases estimated by the CDC a month ago. Conservatively, using 1 million as the denominator for the CFR equation, the CFR for the USA would be about .03% based on 302 deaths. The difference between your estimate of .69% and .03% represents a scale factor of 20+ times.
Is a CFR of .60+% more correct than a CFR of .03+%? Who knows. The point of this discussion is that all estimates and probabilities about the CFR (and also the AR, R0, etc) depend on starting assumptions and definitions. The equations stay the same, but the numbers we plug into those equations vary based on suppositions, assumptions about infections rates, definitions of infected cases, etc. Until the pandemic is over and the number of deaths and infected individuals become fixed in history, we can not know the CFR with any certainty.
