Estimating death rates from influenza in Mexico

With swine flu in the news I decided to make some very rough estimates
with available public data to see if what’s being reported really
reflects something unusual. My conclusion is that 200 deaths per
month is around what you might expect for the Mexico City area in

According to the CDC, there are about 36,000 deaths in the US each
year which are related to influenza.[1] The population of the US is
now about 307 million, while that of Mexico is a little over a third
of that, at around 111 million (according to the CIA World
Factbook[2]). If the death rate in Mexico is similar to that in the
US (it won’t be, but I’d expect it won’t be less, and it could be
close) then we can estimate at least one-third the number of deaths in
Mexico — about 12,000 per year.

If these were spread out evenly throughout the year there would be
about 1,000 deaths every month. They won’t be, but we are just making
rough estimates here. Besides, if the infection rates in Mexico
fluctuate over the year with the same pattern as in the US, as shown
shown by Google Flu Trends[3], then the infection rate in Mexico in
April is probably close to the average over a year. (Google just
added an “experimental” version of Flu Trends for Mexico, and from
that graph I’d also call April “average” — except for this April.)
From these infection rates I’d expect the death rate in April to be close to
the annual average. So as a rough approximation, we’d expect the number
of flu related deaths in all of Mexico to be about 1,000 per month.

The population of the greater Mexico City area is 22 million, about
20% of the population of Mexico as a whole. If the infection and
death rate are independent of location (they are not, but it’s a good
first approximation), then 20% of the 1,000 deaths would be in the
greater Mexico City area. That’s 200 expected deaths in April.

According to press reports, and the Google Flu outbreak map[4],
as of April 30th there have been 7 confirmed deaths and 168
unconfirmed deaths in the greater Mexico City area due to swine flu.
That sounds like a large number when it appears in the news, but compared
to 200 expected deaths from some sort of influenza it sounds less remarkable.

Of course this is a very rough estimate, and the CDC and WHO will
make the same calculations using better data and more refined
assumptions about such things as how the infection spreads in high population
areas and how many infections result in death. Their concern is also that
the virus may spread easily from person to person, perhaps more easily
than other forms of influenza.

Also keep in mind that the averages I’ve used are based on data that
have been collected in the past, over a larger time interval, while
the news reports are based on data collected quite recently. It’s
only a rough approximation to compare the two. There may be more
deaths in April than have been reported now, but those reports will
take some time to get into the public record. Whether the count goes
from 175 to around 200 or well above 200 remains to be seen.

To summarize, based on rough comparison to the US death rate for
influenza, I’d expect 1,000 deaths per month for all of Mexico due to
any type of influenza, and 200 deaths per month for the Mexico City
area due to any type of influenza. The number of deaths reported in
the news match this, at least in order of magnitude, so those numbers
should not be considered unusually high.

There may be other reasons to worry about this new strain of the virus
(like the fact that it’s new, that it’s type H1N1, and that it seems
to spread rather easily), but just the number of reported deaths
should actually not be cause for greater alarm.

Even so, it’s still a good idea to wash your hands, and cover your
coughs (and not with your hands!). But that’s always true for any flu

[4] and search for “2009 H1N1 Flu Outbreak Map”


We saw (and heard) the space shuttle go by

We saw the space shuttle Discovery wiz by last night, right after it
was launched from Florida. It was very cool.

I had read that we might be able to see it from here in New York, so
as soon as it launched we bundled up and headed out into the freezing
cold with Rusty in tow (actually he had us in tow, as is always the
case). We walked up the hill a bit to get away from the street lights
and to find a place where we could get a better view the southeast
horizon. It was cold with no clouds but not completely clear, with
only a few of the brightest stars dimly visible through an indefinite

Nancy saw it first, as a blink through the trees. It was a bright
white light, brighter than any star, pulsing slowly with a flashing
red light. My initial reaction was that it might be an airplane, but
as it rose higher, moving fast from southeast to northeast, you knew
it couldn’t be a plane based on the speed, which was more like that of
a satellite.

After it had passed, which took maybe a minute at most, we went in
to warm up. But it turns out it wasn’t over yet. About 15 minutes
later (I checked the time) we heard a loud, rumbling roar outside,
slowly growing louder. I went out to the front porch to see if I
could determine a direction to the sound, but it was all around us,
kind of how the sound of a noisy freight train going through a small
town at night is all around you (but this was louder than the trains
here). The rumble was more like when our furnace kicks in, only
louder and spread out. Actually, it was just like the rumble of the
space shuttle engines which we had heard on NASA’s video feed when
we’d watched the launch. So we actually heard the shuttle go by as

Scientist that I am, I had to do a quick cross check. Sound travels
at about a mile every 5 seconds, as everybody knows from thunder and
lightning. It was 15 minutes after we’d seen the shuttle wiz by,
which is 900 seconds. So the distance to the sound source would be
around 180 miles. And that is just about right, given what I’d seen
on a map for the expected trajectory.

So not only did we see the shuttle, but we heard it go by as well.
Very cool!