cases. You should adjust the figures based on actual figures quoted by
Mexican officals but make sure you take into account aging. It took 2
weeks to die from this Swine flu.
"The first numbers out of Mexico were something like 175 fatalities
among 500 cases of swine flu. Those numbers were not correct, but they
made their way into the light and came to the attention of public
health officials and government authorities.
If these data had been accurate, they would be telling us 35 percent
of people infected with swine flu die of it. Fatalities rates are
usually expressed as deaths per 10,000 cases. Mathematically then, the
only available data yielded a swine flu fatality rate of 3,500 per
10,000 infected persons.
Somewhere along the line, authorities informed us seasonal flu, on
average, accounts for 36,000 deaths in the United States every year.
If the purpose of the number 36,000 was to ease our minds about swine
flu, it could only do so in the absence of denominators.
According to the Centers for Disease Control and Prevention (CDC),
seasonal influenza infects between five and 20 percent of the
population each year. If we use this range, and apply it to a ballpark
estimate of the U. S. population, 300 million, we would conclude the
fatality rate for seasonal flu is between six and 24 per 10,000
infections.
In other words, first numbers from Mexico indicated the fatality rate
from swine flu (3,500 per 10,000 cases) was between 145 and nearly 600
times as great as the fatality rate for seasonal flu. That would mean
swine flu could cause 5 to 20 million deaths in the United States
alone.
Worldwide (assuming the United States has five percent of the world's
population), the math would yield 100 to 400 million deaths, but that
is just the math. The lack of medical and health assets in the
underdeveloped world would likely push the death toll higher. Is it
any wonder public health around the globe was alarmed?
"
Robert Moran: Flu-ing around with denominator data
By Robert Moran/ Thinking about Salem
Fri May 08, 2009, 11:17 AM EDT
* Tools:
* Print
* Comments
* ShareThis
* StumbleUpon
* Newsvine
* del.icio.us
* Digg
Salem - Watching the contortions earth went through in defining and
responding to the swine flu outbreak, we are once again reminded of
the importance of denominators. In the scheme of things, denominators
are the great revealers of truth.
As we saw with swine flu, denominator data are often impossible to
identify. Numerators are usually easier. We simply count something.
Even with a slew of data, however, it can be a genius challenge to
figure out which are appropriate for a denominator. And, if there are
appropriate data, are they good enough to be the denominator? For
instance, a denominator that does not include all the numerator data
as well as the relevant universe is not a good denominator.
Baseball — bless its heart — is a great illustrator. Say for
instance, you and your friend (who shall remain nameless) read in the
paper the Sox are trying to decide whether to offer a contract to
Johnny Glove or Billy Bat. The paper reports Johnny Glove had 40 hits
in his high school career. Billy Bat had 25.
Based on these data, your friend, who is a better beer drinker than a
statistician, says the Sox should draft Johnny Glove because "he's the
better hitter." When your jaw drops open, your friend — whom you want
to keep because he buys the beer more often than not — says, "What's
your problem, fool? Forty's better than 25 any day."
You are left to explain denominators in a kind way. You ask the
definitive question: "How many at-bats did these guys have?"
Translation: what are the denominators?
It turns out Johnny Glove had 120 at-bats. Therefore, over his high
school career, Johnny Glove's batting average was .333. Johnny is no
slouch hitter. However, Billy Bat only played ball in his senior year.
He had 30 at-bats. His batting average was a phenomenal .833.
The first numbers out of Mexico were something like 175 fatalities
among 500 cases of swine flu. Those numbers were not correct, but they
made their way into the light and came to the attention of public
health officials and government authorities.
If these data had been accurate, they would be telling us 35 percent
of people infected with swine flu die of it. Fatalities rates are
usually expressed as deaths per 10,000 cases. Mathematically then, the
only available data yielded a swine flu fatality rate of 3,500 per
10,000 infected persons.
Somewhere along the line, authorities informed us seasonal flu, on
average, accounts for 36,000 deaths in the United States every year.
If the purpose of the number 36,000 was to ease our minds about swine
flu, it could only do so in the absence of denominators.
According to the Centers for Disease Control and Prevention (CDC),
seasonal influenza infects between five and 20 percent of the
population each year. If we use this range, and apply it to a ballpark
estimate of the U. S. population, 300 million, we would conclude the
fatality rate for seasonal flu is between six and 24 per 10,000
infections.
In other words, first numbers from Mexico indicated the fatality rate
from swine flu (3,500 per 10,000 cases) was between 145 and nearly 600
times as great as the fatality rate for seasonal flu. That would mean
swine flu could cause 5 to 20 million deaths in the United States
alone.
Worldwide (assuming the United States has five percent of the world's
population), the math would yield 100 to 400 million deaths, but that
is just the math. The lack of medical and health assets in the
underdeveloped world would likely push the death toll higher. Is it
any wonder public health around the globe was alarmed?
Over the next several weeks, public health massaged addition data as
they became available. As I write this column, public health is
concluding the current strain of swine flu is not even as virulent as
ordinary seasonal flu. The initial panic is fading. Federal
authorities, for instance, have rescinded recommendations to shut down
schools in the
face of an infection.
What the future holds for this strain of flu is unclear. The virus is
new and unique. Never before has science identified a flu composed of
four different influenza genes (two pig, one bird and one human). How
the virus will evolve is unclear.
One thing for sure, denominators will be at the bottom of the search
for answers.
Salem resident Robert Moran's column, "Thinking about Salem," appears
frequently in the Gazette. Contact him at SalemThink@aol.com. Readers
are invited to comment on Thinking About Salem columns on the
Gazette's web site, www.WickedLocalSalem.com — click on the
'Columnists' icon located next to the Blogs.
No comments:
Post a Comment