So I'm deeply honored to be the
distinguished speaker for two thousand and

thinks it's great to be back.

You're seeing some old friends and
some people I haven't met before.

So I know of five ways that could kill.

Say one hundred thousand people on U.S.

soil and
since the September eleventh attacks.

I've been looking at these five scenarios.

Mathematically modeling
them analyzing them and

then developing recommendations going
to Washington and trying to get these

recommendations followed and that's
what I'm going to speak to you about.

Today.

So what are these five ways
the first way is not on this slide.

It's an attack on a chemical plant
facility such as an oil refinery and

when I started looking into this I found
out that the research that I thought

was interesting and relevant had already
been done it was done quite well at

Lawrence Livermore Laboratory so
I didn't do research on this.

I did write an op ed in
The Wall Street Journal on this topic and

something in issues in science and
technology which is National Academy of

Sciences publication and for
President Obama's campaign.

I was on his advisory group for
homeland security and within that group.

I was the chair of the chemical plant.

Security.

The second one is a nuclear weapon going
off in a major city of the five This

is probably the worst one.

And there I go right now I'm working on
consequence may although it sounds like

an oxymoron kind of consequence management
after a nuclear attack is there something

you can do in the answer's yes I'm doing
that now I won't talk about that today.

I've mostly been looking
at prevention there.

How do you prevent nuclear weapons or

nuclear material from getting into
the country which is this port security.

Things have a paper and
risk analysis in a paper in Science and

Global Security on that.

And also look at the last line of
defense of putting sensors around

a US city to catch a truck coming in with
the bomb and I have three papers and

they're the main ones and I triple
the transactions on nuclear science and

then one in risk analysis and
one in operations research.

I'm not talking about that either
the other three of the catastrophic

events are biological
The first one is smallpox and

we have a paper in the proceedings the
National Academy of Sciences on that and

that was instrumental in
changing the the main question.

There was if there was an attack
say in New York City or

deliberate release of smallpox What's
the best post-attack vaccination

strategy in our work this is a joint
where that Kaplan who many of you know.

Yale.

We were told was influential in moving
the government from kind of surveillance

containment policy to mass
vaccination policy paper in

proceedings National Academy of Sciences
and this was covered in New York Times and

was on the Today show and what not and
I'm not talking about that either.

So what am I talking about the things
in the yellow arrows here so the other

two attacks are anthrax and botulinum
toxin in those two I am talking about.

And then I'm going to squeeze
in this one right here

pandemic flu that there's
not a terrorist releasing

this is more Mother Nature being the bad
guy and I'm going to talk some about that.

Not for most of these attacks
except possibly a remote detonation

of a nuclear weapon or something.

Everything else you need terrorists
in the country to pull off an attack.

So I've also done a fair amount of work on
border issues of how to keep terrorists

out of the country so terrorists
could come into the country two ways

one way is at the ports of entry
that is like at an airport in there

there's a fingerprint program and time
permitting I'm going to talk about that.

It's the U.S. Visit program and
then they can also come.

I'm in between the ports of entry and

that's like a sneaking
across the US Mexico border.

I have a paper and risk analysis it's on
line now it hasn't showed up in print yet

but I'm not going to talk about any of it
to really of everything here that's by far

one of the most complicated challenging
problem for the government in my view.

Have a paper I just finished this
week on counterterrorism and

I'm not going to talk about that either.

So I'm going to try to spend about ten
minutes on each of these four arrows here

because in this talk will be extremely
diplomatic I'd say it's a broad talk but

it's really a broad and shallow top you
know I'm just going to have ninety minutes

of slides on each of these topics and

I'm going to give you kind of ten minutes
worth on these four topics that encourage

you to instead of saving questions to
the young just ask questions while we go

because we'll be kind of
skipping from topic to topic.

OK.

This gets a little scary.

So hang in there.

OK So the first one is anthrax so
why is the anthrax a big deal.

So about nine months ago I was talking
to the person whose business card says

Special Assistant to the president
bio defense and I was asking him what

are the biggest bioterror challenges or
concerns that you have and

he said our top three concerns
are anthrax anthrax and anthrax and so

you know question is Why is that it's
several things one is it's highly deadly

the inhalation form of anthrax would
kill over ninety percent of the people.

If you don't get any medical treatment
in the other key thing is that

it's almost unique in its Dura billet even
in that if there was an attack in Atlanta

some kind of aerosol attack then these
spores It's almost like a gas would kind

of seep in through the doors and windows
and right before she got in the door.

Some of this would steep into the door and
up on.

The ground here and you come back
thirty years later in just kill.

Up the carpet in these spores would and
would be viable and

if you inhale them you could get infected.

And we have had a proof of principle.

So someone who may or
may not be who committed suicide.

Someone put two grams
of anthrax spores into

each of five envelopes in two thousand and
one and killed five people in this.

So our main goal was to
develop a skeleton of

what's a response plan for anthrax.

And so we developed a mathematical model
the first part is an atmosphere dispersion

model so this is a model that tells
you if we looked at a one kilogram so

there are two point two pound release
of anthrax spores from a thirty story

building in New York City in the plume
model tells you how many spores

you've been hailed given that you're
a certain distance from the release.

And then you would have a dose
response curve that tells you.

Given how many spores you inhaled
What's the probability that you would

get infected.

So in that would be quite age
dependent would be people.

Like my parents age or
you know quite a bit older.

That would.

Would be more likely to succumb to this so

that this is a spatial
part of the mountain.

There's the temporal part that
people who get infected then have

an incubation period.

That's the time between when
you inhale the spores and

when you exhibit symptoms and
that's quite variable from say.

A couple days or
even less to a couple months but

then if you don't get medical treatment.

Once you develop symptoms
you don't get treatment.

You're going to be dead
in a handful of days.

And so what we might have the treatment
by not of a treatment by cues.

Waiting lines here.

And so the main treatment is antibiotics
they can be used both as a prophylactic

that is to prevent disease given that
you have not yet developed some.

Terms or as treatment so if you get
it right after you develop symptoms.

It may be say forty percent or so
effective in preventing you from dying.

It's thought to be one hundred
percent effective as a prophylactic.

And then we look at aggressive hospital
care for the people who get sick.

And the punchline.

From our analysis is rapid
antibiotic distribution so

we found the key management lever here is
to distribute antibiotics to people as

quickly as you can in our base case we
estimated that roughly one hundred twenty

three thousand people would get
infected from two point two pounds

of anthrax spores and this was precooked
Trina assumptions that you know looking

back I think we're a little naive I'm
a little rosy in our assumptions about how

quickly government would kind of figure
out what's going on and respond to this.

And we found that roughly for every day.

It takes to distribute antibiotics
you'd save about ten thousand lives.

So Ed and I had Kaplan this work as
with the smallpox work was with that

Kaplan wrote an op ed that appeared in The
Washington Post in two thousand and three.

And we proposed that we use the postal
system to deliver antibiotics.

So we already had the the typical way
people were thinking about doing this

are points of dispensing kind of like how
we vote you go to your local high school

and you pick up you're in a biopics and
you know we have an infrastructure in

place that is it's almost everybody's
house in a six to eight hour period and

so we say well let's just have a snow
day even if it's Atlanta in in

you know September something just keep.

I heard you got a snow day recently right.

So you and everyone stay home.

You'd have.

Postal workers going around
with a police escort so

you know they don't get mugged and start
a black market with the antibiotics and

this was picked up on by people at
the Homeland Security Council so.

F. you probably know about the Nat.

Security Council after nine eleven
President Bush created a parallel

organization called the Homeland
Security Council and someone there.

Read this op ed and
thought This is a great idea and

he really took the ball and ran with it.

He got the C.D.C. You got H.H.S. Health
and Human Services on board and most

importantly and probably most difficult
if you've got the postal union on board.

So within one year in two thousand and
four Washington D.C.

was ready to use this program that postal
workers in a voluntary basis they'd be

they and their families would get
antibiotics before the attack and

then they would go door to door
then they run since then they run.

The field tests in three
cities Boston Philly and

Seattle that have all
been very successful.

This is been implemented many apples
basically it's in the middle of getting

rolled out there is twenty million
dollars in two thousand and nine.

For this and then.

Last year the the White House hired me
to do further study of this issue and

I can't tell you too much about it but I
do have an op ed in The New York Times on

this from last October that
summarizes this study that I did and.

Part of it is kind of
pushing this postal idea but

another part is that there's something
called the Strategic National Stockpile

that holds all of these antibiotics and
other medications and

we found that with the postal option
the bottleneck kind of switches from

dispensing this stuff to
actually getting the drugs.

The antibiotics to the postal people and
then we said there was roughly enough for

six hundred fifty thousand people that
were in these push packs that could

get out in twelve hours probably even
quicker like eight hours or something and

I've been told at least from people
at the White House that as a result.

You know after the op ed that they've now
switched from six hundred fifty thousand

to ten million that they
have in this pushback form.

Which is enough to.

To help with a large anthrax attack.

I've done other work.

Vaccination studies and

cleaning up after anthrax which is
a really big problem and a big deal.

I will say if anyone's interested
in any of these things I have

if you just go to my website.

There's like twenty or twenty five
papers on these topics and six up and

so you can read stuff that interest you.

OK so

that's anthrax second one is that I want
to talk about today's botulinum toxin.

So Richard Danzig who's a former secretary
of the Navy in the second Clinton

administration is one of the nation's
top bio terror experts and

I've been very lucky that he's really
been my main champion in Washington.

Both in suggesting a few problems
that I work on in both and

then after I do the work in
the kind of setting me up with

a few people who can help me
learn more about the problem but

then after doing the work Hill
post me in Washington and invite.

Influential people to my talks.

Anyways he wrote an influential paper in
two thousand and three called catastrophic

bio terrorism and he laid out the three
big threats to human bio terror threats

from humans smallpox anthrax and botulinum
toxin in a cold drink at that time I

had already and I had already done
the smallpox work in the anthrax work so

I started working on the botulinum thing
and actually looked at it in milk.

So why is botulinum toxin such a big deal.

So this is the is the poison.

It's the deadliest substance known to man.

So the what's called the id fifty
the infectious ghosts that if each of

us took this amount about half of us would
succumb to the disease the infectious dose

is a millionth of a gram that
is a really tiny amount.

And so you couple that with the stuff we
do operations in particular the large.

Centralized storage and
processing and rapid

distribution and consumption of milk which
we consume at the rate of seven billion

gallons per year and when you combine
the deadliness with the large scale

processing you get really a scary scenario
so I'm just going to walk you through.

This scenario Richard Danzig sent me up
with the guy who is head of California

food and safety and we took a tour
of a product very processing

facility in Sacramento California so I
grew up about four miles from Vermont and

I always thought growing up is like
Vermont Wisconsin you know these

are the places that do the milk it turns
out twenty three percent of the nation's

milk is processed in California and not
surprisingly these are kind of more modern

larger scale facilities than
the the ones in Wisconsin in Vermont.

So the middle block here
is the processing facility.

So one thing I learned is that cows are
milked twice per day when I do this work.

So there's sixty four thousand how those
associated with this one supply chain and

the milk on the farm is put
into a tank on the farm and

then a milk truck these fifty five hundred
gallon kind of sideways cylinder trucks

do a little traveling salesman tour around
the local farms pick up the milk and

then go to the dairy processing facility
and then they dump this milk into.

A fifty thousand gallon silo.

And then.

The milk gets processed and
some of the milk get some I'm sorry.

Some of the toxin gets inactivated during
the heat pasteurization process and

we get back to the later and then the milk
quickly goes through the supply chain and

if we don't figure out why.

It's going on with you know if you think
of say the peanut butter issues you got

or not you.

We collectively have now or
you know the tomatoes in

California than spinach and California
were were just as guilty as Georgia.

And other things that it takes them a
while to really figure out what's going on

and if it takes that long.

Here you're going to have.

Five hundred thousand people
drinking contaminated milk and

one part I left out is that these silos
these fifty thousand gallon silos by

law need to be emptied and
cleaned every seventy two hours and

by a little little formula you know
calculation you can figure out roughly how

much milk gets contaminated and
I didn't mention the take a step back.

Scenario here is ten grams
of botulinum toxin which

is not very much in a four gallon
jug of an acid mud mixture.

So someone could either put
this in a tank on a farm put it

on a tank truck while it's resting at
a truck stop or put it in the silo by law.

None of these things have to be locked.

And regardless of where
it is where it's released

upstream of the supply chain that
ends up well mixed in these silos.

OK So the bottom line is that half
a million people drink this in

about four hundred thousand
of them would be poisoned and

would come down with botulism Now right
now one of us had home canned goods and

came down with botulism our chance of
survival is roughly ninety percent because

they put us on a ventilator they give us
antitoxin on a scale of this would be no

ventilators there'd be no antitoxin in the
death rate would be about fifty thousand.

So we're talking about two hundred
thousand deaths in a particularly

cruel part of this scenario is
that it would disproportionately.

Hit children so
children have smaller bodies.

Need less tax and if you look at the data
they drink more milk than we do and

the third thing is that they'd
be there early canaries and

that most dairy processing facilities.

Shipped direct.

You know the little half carton
containers direct to the schools

which gets there quicker than just going
through the regular supply chain and

getting it from your supermarket.

Yeah.

Right.
And

you know the I'll get back
to that's a good question.

I'll answer it in about five minutes.

It would be easier to answer than yeah.

Right.

So we have an incubation period and
we use order statistics and we use the On

limit theorem on order statistics
kind of assuming if N. people.

Show symptoms and we find that K. thought
of then what's the you know when are we

going to actually ten people are going to
get sick and we catch the case of them in

terms of the key thing to base in
period when does this occur and

then we you leave that as
a variable to say you know we catch

this that one hundred person or
the thousand person or whatever.

OK So what are our recommendations here.

Three things first is
invest in prevention.

The F.D.A. has guidelines that says hey
put locks on the tanks in the trucks and

things like that we're saying
well gee of a dairy processing

facility can be as dangerous as a nuclear
facility in the hands of a terrorist.

Maybe these should be laws not voluntary
guidelines that actually had quite low it

here and raids.

The second thing was investigate second
thing was investigate the pasteurization.

Formula Is there a way to either
make that pasteurization longer or

higher temperatures so that we could
inactivate more of the toxin in the event.

That someone released botulinum toxin

in there in the third one which is
really our main recommendation is that

there is a it is to do a highly
specific in process test so right now.

Something I didn't describe is that
there is a forty five minute time window

from when a truck arrives to
the dairy processing facility and

when the milk is emptied out of the tank
truck and during those forty five minutes

they're doing an antibiotic residue test
which is testing to see if the cows that

created this milk have been pumped up with
too many antibiotics by the farmers and

so if we could come up with a test
that's less than forty five minutes for

botulinum toxin you can
pretty much take this.

Nightmare scenario off the table.

So what happened here.

So.

The whole story that goes along
with this this paper at this point.

So Richard Danzig hosted a talk for me.

I gave it in the fall of two thousand and
four.

And it was about two hours long.

The assistant secretary of public
health preparedness was there he was

the person who is in charge in
the government of managing this problem.

There is probably five people from the
White House people from every branch of

government F.B.I. military the chief legal
counsel from the dairy industry was there

and it was a very civil talking
we had lots of good discussion.

It was clear they were learning.

They found the educational at that
time everyone in the room me and

everyone else all thought that
the only cast that existed for

botulinum toxin was in Ally's a test which
would take three to four hours which is

not within the forty
five minute time window.

And so we all agree this really
is impractical this is not we're

you know kind of slowing down and
disrupting the supply chain for.

So and then the assistant secretary.

Took me out to lunch that day and said.

The the him I'm sure this is the only
reason he took me out to lunch.

He said the dairy industry was nervous
that you're going to publish this paper

they asked me to ask you.

If you're planning to publish his paper
and like when you're in seventh grade and

someone passes you know my girlfriend like
you do like you know one of those kind

of things and you know so I said yeah
that's what I do of planning to publish

his paper and he didn't you know you
just kind of nodded and didn't say much.

I didn't think much of it at the time so.

Later in the spring.

So what I want to do is some follow on
research and look at analyze that test

because I realized there was about an a
factor of one hundred greater sensitivity

in the existing allies a test than was
needed for this application and so there's

a way to trade off time insensitivity
I thought it might be interesting.

Math problem in an important.

You know real problem there.

And as I was looking around I
stumbled upon a fifteen minute

commercially available custom you know
magnetic electrochemical luminescent

tested me a couple days to learn how
to say that cuts for bottom line.

I'm in milk and in fact that the Pentagon.

The Department of Defense was using this
so that no one poisons their soldiers.

Our soldiers in the field.

And it was too late.

The galleys had already gone through at
the proceedings the National Academy of

Sciences it's called P.S.
because I'll be saying this over and over.

And so I couldn't really put this in
the paper so I tried writing an op ed and

I sent it to the New York Times of the
main things saying there's this fifteen

minute test.

Let's use this.

And so the way papers at this journal and
a few others

like New England Journal medicine and
science and nature and so forth.

Work is that there's a weekly cycle papers
are released on Mondays at five o'clock.

The previous Wednesday.

They're posted on a secure Web site for

the media to access and then the media
can write their own saris and

interview people as long as they
don't publish their article.

On Monday at five o'clock.

So Wednesday night C.N.N. you know
downloaded this article really as a D.

a couple hundred thousand people die they
give it to the government ask them for

comments.

That's Wednesday night Thursday morning
there's a letter from the assistant

secretary of public health preparedness
same guy that took me to lunch and

the commissioner of the F.D.A. written to
two people the editor in chief of P.N. a.

The president of
the National Academy of Sciences

saying please do not publish this
paper it's a road map for terrorists.

And to the National Academies credit

they did the right thing they withheld
publication of the paper they

took it off the secure Web site which
itself was not a media sensation but

certainly within an hour I was getting
calls from N.P.R. And C.N.N. Another

one saying what's going on here because
they realize there's something fishy.

Anyways they looked at the paper for
about a month.

So the handling editor what we would
say the associate editor something of

the paper was Barry Bloom who is the dean
of the Harvard School of Public Health

the referees read it carefully.

Barry Bloom read it carefully.

Bruce Alberts of present National Academy
of Sciences read it carefully because he

sent me an email finding three grammatical
mistakes I made and after about a month.

They met with.

The government people and
basically said we disagree with you.

There's nothing in here that
isn't easily and quickly googled.

And they published the paper and
when the paper came out.

Bruce Alberts the present
National Academy of Sciences wrote a two

page editorial that came out with
a paper that said not only is this

not a road map for terrorists.

But this helps make the nation safer and

the most interesting thing of this whole
tale is this fourth bullet point here.

So in Albert's op ed that it is
that authorial he wrote that

the government after nine eleven realized
how uniquely catastrophic botulinum

toxin in milk was and they started doing
experiments to intensify the heat pastors.

Ation process to see if it would.

Inactivate more of the toxin.

And that indeed to the government's great
credit they were closely with the dairy

industry and they have Cheney have
intensified the heat pasteurization

process they did this over about a three
month period slowly changing it did a huge

number of taste tests actually found most
people like the slightly caramelized

flavor better after the more
intense heat pasteurization and

it was until about a year later
where two people I met on separate

occasions one was the main person at
the White House working on this problem.

The other one was the chief scientist to
the assistant secretary of public health

preparedness they both said pretty much
the same thing to me and I didn't prompt

the conversation and I was meeting with
them on totally different topics and

they both said I don't know if you realize
that or not although the government had

run the experiments on the pasteurization
your paper the thread of your paper

coming out was really the catalyst for
them quickly adopting this but

in any case the government deserves
a great deal of credit for

doing this and now getting back to
your question why not orange juice.

This issue is really not just about
botulinum toxin in milk it's about moving

our country from food safety which despite
the peanut butter in this fire the spin

engine canned tomatoes were pretty good
at two food security so food safety is

preventing the accidental contamination
of food food security is preventing

the deliberate contamination of food and
except for this isolated incident.

Of intensifying the heat pasteurization
process the government has done almost

nothing on food security since
nine eleven in particular they

cannot answer your question
of why not orange juice and

so more recently we have a paper that
came out last year in the Journal of food

science that basically takes
a somewhat simplified version of the.

Coxon model does a mathematical analysis
of and comes up with a single equally

Asian That tells you this is
roughly the number of casualties

as a function of the key characteristics
of any biological or chemical agent in

the key characteristics of any food supply
chain and what I think should be done is

they should create a matrix that
has you know the the columns being.

Orange juice milk catch up on and on and

in the rows being bottom line of toxin
anthrax rice and on and on and on and

then fill out these entries just to get
a rough estimate of is botulinum toxin and

not the unique catastrophic scenario or
is this merely the tip of the iceberg.

So of course after the bottom line thing
no one was going to give me data to fill

in the matrix themselves.

I gave this we created a very Here's
the op ed got toxic note that

appeared in The New York Times which was
really in the end with the government knew

the paper was coming out what they really
were upset about was this this op ed.

But at any rate we
created a very low tech.

Interactive website for
the government to use so

they could just put in the characteristics
and we would spit out this and

they played around with I don't know much
else if they're using it or anything.

OK So that is the second of the four yet.

Nothing so
I really kind of kept pushing I wrote.

A smile.

OK so the new pasteurization formula and
the old one.

The as far as how effective it is how much
of a toxin it would inactivate that's not

public information and so we use data
on creamed corn which had a similar

ph would seem to be a key driver of
the new process there's no data out there.

I wrote a little paper and
also N.P.R. on SCIENCE FRIDAY.

He gave this formula right here.

I'm not going to go into detail but

basically said if if it's
this good then this is.

The you know we may or
may not need to do this test and

I wrote like a three four page thing and
sent it to bunch of places no one would

touch it you know no one would publish
it and then I just stopped and gave up.

Yeah absolutely.

That's why you need to do this in the
forty five minute window to believe it or

not the gold standard test for
about a line of toxin in milk is a mouse.

You literally give it the milk to
a mouse and see if he freezes up and

dies over a four day period.

And so this is the the test and.

In there could be false.

That's why you'd have to hold
the truck that's why you can't.

If it's longer than forty five minutes if
you release the milk in the supply chain

you've got a huge problem on your own
even though even from base rule even

after you get a test the probability it's
a real event is point zero or one or

something but.

OK flu.

Paper mathematical biology another
companion paper will probably

come out risk analysis is kind of at
the minor comment stage at this point but

the main one is really the book
in mathematical biology.

Here is a graph of death rate.

For each year over the twentieth century
a big spike in one thousand nine hundred

that is not world war one
that's The Great Influenza.

Most kind of public health experts

believe you know over the next
say five years or something.

The chances of a pandemic flu of the order
of nine hundred eighteen is probably more

likely than say a large scale
botulinum toxin or anthrax

attack in Certainly if it looks like one
thousand nine hundred the number of deaths

worldwide would be war if even a nuclear
weapon going off in New York City and

certainly the other things
that we're talking about.

This problem was posed to me by.

Richard Danzig who I had mentioned before.

So what's going to happen in the first
wave of a pandemic we're not going to have

a well matched vaccine.

We're probably not going to
have any anti-virals that

work even right now the seasonal flu this
year seems to be resistant to Tamiflu.

Hospitals we totally overwhelmed if
you listen to the White House they're

basically telling you to stay at home and

don't go to the hospitals because
we're not going to be there.

In the federal government's
also telling the cities and

states we're not going to be there for
you.

This is your problem.

So what Richard Dantzic posed
to me was are there any

non-pharmaceutical measures that
we can take to kind of mitigate

the effects of a pandemic flu and
so as I thought about this.

So things like washing your hands or
wearing a face mask or

something like that as
I thought about this.

I realized well it really depends on
how flu is transmitted so contagious

diseases like flu are transmitted
in one of three ways first one is.

Transmissions that would
be I would cough or

sneeze the particles would stay airborne
they drift to the back of the room.

You would inhale them and
you would get my flu.

The second one is contact transmission.

I would put my fingers in my nose or
eyes or mouth come over shake chips and

even though Chip never put his fingers in
his nose or eyes or mouth you would do so

he would catch my flute.

Third one is droplet transmission
those who have had young kids.

This is when you know your five year old
is right next to you give me a hug and

then they just spray you right.

Your face all over you and
don't even say excuse me.

So these are the three modes and

so you would think you know given
this picture here that the public

health community would have worked
out well gee how is flu transmitted.

But if you look at say two thousand and
four two thousand and

five the reports written by the government
on flu it says the relative importance of

the roots of train.

Mission for pandemic flu for
influenza is unknown.

And so we set about trying
to figure out what it is so

it created a very complicated most of
this is much more detail than the model

that I've been talking about for so this
looks at the virus coming out of your body

through a sneeze or cough looks of
the probability distribution of the size

of the particles given the size of these
particles how quickly are they drifting to

the surfaces how long do they remain
viable on various surfaces and in the air.

You know your inhalation rate you know
basically just modeling all three modes of

transmission.

That's what this big IT each
other is trying to convey.

And this is the most complicated slide I'm
not expecting you to take all this in but

so we can the spaghetti chart
turns into an infinite number

of differentially cannot turn into but
is formulated as an infinite number

of differential equations which we
can get an explicit solution for.

And then you can write that
the probability of infection

is one minus erased to the minus and
then three terms here and

these are meant to be pneumonic for
the three modes of transmission

now east each of these terms is itself
a kind of a messy mathematical expression

with lots of different parameters in
it but we have these expressions.

Ideally what you'd like to do is have
all the parameters of models estimate

the parameter value stick them
into these three things and

see which one is biggest whichever one
is the biggest of these three things is

the important mode of transmission we
could not do that because there just

isn't good enough data
to kind of see well.

If I put my fingers my nose.

How much virus is coming out my
fingertips if I shake chips hand how much

of the virus of my
fingertips is transmitted

to him how much is going
to go into his body.

How often is my five year old sneezing
right in my face and things like that and

we just it's hard to do when
there are it's unethical to.

To run controlled
experiments on influenza.

So you know these things are basically
unknown it's hard to even estimate

some of these things when you think about
it even within an order of magnitude.

So what did we do instead.

And instead of comparing the three modes
of transmission for one disease for flu.

We looked at the ratio for each of
the modes of transmission the ratio of

two diseases of flu in the common cold.

So the common rhinoviruses the common
cold and when you take the ratio of these

mathematical expressions magically all
the parameters that are hard to estimate

cancel out in the only parameters that
remain are these seven parameters that

are in the literature and easy to
estimate and when the dust settles and

you look at these ratios they vary by
ten orders of magnitude remarkably

You would think some of these things would
be the same for flu or rhinovirus and

they're remarkably they've evolved
in remarkably different ways and

these ratios strongly suggest they
don't imply that they strongly suggest

that aerosol transmission is
the main mode of transmission.

It doesn't imply it because we really
don't know what's going on with because

we need to know where everything about
running a virus then the second part of

this is rhinovirus you can do controlled
experiments and there was really wild

experiment done at University of Wisconsin
in the one nine hundred eighty S.

where they took a bunch of people playing
poker around a table half the people were

susceptible so they were healthy the other
half have been injected with rhinovirus or

inhaled rhinovirus their sneezing and
coughing all over the place and

they had in the susceptible is have
contraptions on them so they can't put

their fingers near their nose and right
there out so they're not going to get sick

like chip would have and they can you
know measure who gets sick and who's not.

I don't have time to go through the
details but basically analyzing that data.

I knew we found that clearly
in the rhinovirus case the sum

of this maybe a significant amount is.

New to aerosol transmission when you
combine that with the ten orders of

magnitude difference in the ratios we
think this is a pretty compelling evidence

that air aerosol transmission is
the key mode of transmission for.

Influenza.

So what are the what's the practical
implications of that this is

bad news would be much better if it was
contact transmission is easier to just

not shake someone's hand or not pick
your nose than it is to not inhale.

You know my.

My coughing and sneezing from far away.

So the key thing is face protection and

there's two two types of things you can
wear one is an N ninety five respirator.

This is like what is best workers where
they cost maybe two dollars apiece

in a hardware store indeed if you
take anything away from this hour

is to go to your local hardware store
pharmacy and buy a bunch of these.

Respirators and for you.

Faculty out there not the starving
graduate faculty I know you can afford

the more expensive to buy
the most expensive type you can

find is that usually at least
from my limited testing of them.

Those are the more comfortable ones and
if you're really going to go through this

that they're very uncomfortable and
you want the comfortable ones and

I'm not getting paid by any three M.
or any of these mass companies so.

Anyways that's respirators
the other one are surgical mask

this is what your dental hygienist wears.

So there's three ways to measure
the effectiveness of these face.

Protection things first one is
the filter inside these mass.

You know how much passes through it
in every one knows the end ninety

five works great.

One of the key things in our risk analysis
paper second paper is that most of this

field thinks surgical mask filters
aren't any good because for

some particles everything passes through
without actually doing the math and

finding that something like
ninety eight percent or

something or
ninety five percent of the virus is in.

In particles that are three microns or
more because the the bigger the particles

the more viruses in these part of it
grows is the volume of the particles and

we found that the filters inside surgical
masks are ninety eight percent effective.

They'll keep ninety eight
percent of the flute.

OK So that's the first issue
the second issue is the face fit.

So these respirators fit
fairly tight around your face.

They.

Let about ten percent of the air in
the virus sneakin surgical masks

are quite loose they lead in
about thirty seven percent.

This is no joke.

This this thing right here.

There's a study out of Harvard in
the eighty's putting pantyhose over

there actually over respirators and they
found even though the virus can actually

pass through pantyhose quite easily the
pantyhose will keep will eliminate all of

the face leakage and so if you took
a surgical mask in put pantyhose on top

of it you would have ninety eight percent
effectiveness we'd all go around looking

like it's Halloween or something or bank
robbers or something but this would work.

In We're not necessarily
suggesting that but

we are suggesting in you read the op ed.

So I had not been in the York Times on
this call Face facts from a couple years

ago and suggesting a a design that for
surgical mask that eliminates.

Surgical masks only cost like ten
cents apiece rather than like

a dollar to a piece.

OK So our main recommendation is that the
government stockpile in the government's

done a great job stockpiling smallpox
virus we have enough for everybody.

They've done a great job
stockpiling antibiotics and

other pharmaceuticals
they have not stockpiled.

Respirators or a mask for
as even though it's a lot cheaper to do.

There's absolutely no technological risk
involved we know these things are going to

be resistant to these massive amounts are
going to cause side effects that kill you

or something like that and it just
a matter of finding the will to do this

right now the lead time to build
enough mass respirators to be by each.

Months which is way beyond the first
wave or even the second wave and so

that's really the main
thing that I was pushing.

As far as what happened.

I'm just going to really just put
up a bunch of quotes here from

the various government documents.

Whoops.

As I said before that before
my study came out in early.

There was you know these are some of
the quotes that they basically don't know

what's going on the White House
plan in two thousand and six.

You know it's like a two hundred page
document not a single mention of face

protection.

I briefed the various parts of
the government mainly a one one on one for

one hour with Rajiv who was the head
person at the White House who you know

basically was in charge of writing the
White House plan and then slowly coming

around to the idea that there is short
range transmission and suggesting

that health care workers where in ninety
five respirators during a pandemic flu.

If they're available if not where
surgical masks then six months later

actually saying surgical masks should
be considered by anyone entering

a crowd and
indeed if you read the New York Times.

The press release on this.

They said there was quite a bit
of infighting at the C.D.C.

about whether or
not to include the word considered.

And I talked to various people Richard
Besser included who's now the acting

director C.D.C. and you know I think
it's I mean one thing is our papers.

It's all aerosol physics and it's
originally was a ninety five page paper

and it's not that easy
to understand really and

you know I think just not everyone is
convinced that this is is true but

anyways that they're making progress
on this in the last point is

probably the most
important that the F.D.A.

has cleared to and
ninety five respirators from the three M.

company to be sold and used for a pandemic
flu nonetheless they have not take.

The last and most important
step of stockpiling this and

giving it to people now which to me is is
a real no brainer I've been trying to push

you know the Obama people pushing on
the people I know to to get this done.

But to no avail.

Thus far.

OK So we're about eleven minutes to go.

OK one more topic in this
is a different topic.

Which is keeping Paris out of
the country through the airports.

This is also a paper in the proceedings
of National Academy of Sciences.

So I don't I'm going to go
fairly quickly through this but

those who are not U.S.
citizens who come through the country.

You have to give over your fingerprints.

When every time you come to kind of
show hands anyone done this before

anyone want to.

OK So you know I'm talking about so
originally.

So what are they doing.

First of all so they have a set of
fingerprints from six million people

who are known criminals and
suspected terrorists and

they're comparing your
fingerprints to those six million

sets of prints to see if you
are one of those people.

And the performance of the system

just look at the last two lines here is
pretty good the detection probability that

is the probability that if I'm
on that list I'm a bad guy and

then I come in and I give a new set of
prints and they compare it to the old

set of prints probably that they
get a hit is ninety six percent.

So that's good.

And then the FOS positive probability so
chances you or I or

any of us come through the airport they
mistakenly think we're one of the six

million people is only three in
a thousand which it's clear they kind of

chose this for
three sigma you know kind of accuracy.

So that seems pretty good too.

And so I should say there is they this
is based on the original system that

use your two index finger prints.

So what's the problem.

This is.
Pretty good performance certainly for

the government.

The problem is is that these software
systems that do the matching They also

compute the quality of your fingerprint
image and the system here that's these

are eight curves like detection probably
versus FOS positive probably curves

and there's a image quality is used
by the current vendor for the U.S.

Visit program and
about five percent of U.S..

U.S. population about ten percent
of the people on the watch list

have the worst quality fingerprint
image in for those people.

This is this curve way down here and

the detection probability
is fifty three percent.

So it's more like a coin toss than
ninety six ninety five percent.

And so terrorists could certainly exploit
this they usually would have a number of

people who are kind of willing to
come in they could choose people in

plus if you go to
the National Institute of Standards and

this website at least if you
used to one of the repercussions

of my papers that they've yanked
a lot of stuff off their website.

You know after I published my paper but
they used up pictures they Here's an image

quality one fingerprint use of
the image quality fingerprint.

So they would you know choose people with
the worst fingerprint image quality or

you could deliberately deface your
finger with you know chemicals and cell.

You know same peepers like that.

OK so since we don't have much
time I want to open it to Q.

and A I'll just cut to the punch line.

And it's not on the slide
of just leave you on the.

More fun picture there.

So the punch line is question.

They have a ton of data they have
an army of computer scientists and

statisticians working on this and
I have do is put it through you know they

have fingerprint images that go through
this software that tells you image quality

one two three four five six seven
eight and they just show you pictures.

Where you go OK totally missed on.

Question various places so the government
if you've got a read following the U.S.

Visit program they have bigger
problems in a sense than this.

So they've been trying for a number of
years to combine two databases one is

owned by Department of Homeland Security
which is the one I described the F.B.I.

owns one with forty eight million people
that has ten fingerprints one of them is

rolled one the other one is flat and
they're trying to combine these and

they hired McKinsey or some big consulting
company and it's not even done yet

they hired him like five years ago.

So that's one issue but for
example in the Taleban war.

You know any one they captured
they took their fingerprints and

although it's not widely disseminated the
technology they use essentially means that

all those fingerprints are image quality
a Unfortunately the worst fingerprint

image quality the other ones they get
from especially once this is murder.

They have lots on various criminals but

they'll get them from you know
people they catch at the border.

You know professors at Georgia Tech
who are you know suspicious and

if so does that answer
your question better than.

OK so
the punch line of this study is that for

the people with the worst
image quote print quality.

There is not enough information
on two fingers but there is

enough in information on ten fingers and
simply if we use all ten fingers.

Then we can get the detection probably
up to ninety five percent for

everybody in this is a lot of probability
and statistics and data analysis and so

you know I mean I'm making it sound
really easy but it's a one paper.

And then what happened here.

I have another follow up paper
coming out in operations research.

Soon it's already been accepted that
does some follow up stuff on a text

you're matched to rather than matter
if that means anything to anybody and

we found it to finger system
that actually works better.

But anyways what happened here was I
briefed a bunch of people I briefed the.

People the White House I briefed them
manager of the U.S. Visit program G.A.O.

others Anyways that was I
testified before the House of

Representatives Homeland Security
Subcommittee and two subcommittees in

the home with security committee
that were in charge of these things.

This was five weeks before the two
thousand and four presidential election.

I could spend ten fifteen minutes telling
you how politically charged this and

what kind of circus it was in it was same
time as Cat Stevens thing that was all in

this testimony but I don't have
time to tell you about that but

the bottom line was the ranking
member of this committee so

that time the leading Democrat on
that committee Jim Turner from Texas.

He wrote a poem he wrote a letter to
then Homeland Security Secretary Ridge

saying this study from Stanford came out
we really need to switch from two to ten

fingers Washington Post picked
up on this and wrote an article.

They didn't do anything but in ridges
outgoing press conference he did make some

noises about we should switch from two
to ten fingers but of course he didn't.

You know he wasn't he didn't do anything
but then in June so nine months after my

testifying the new secretary announced
the switch to to give ten fingers.

It's supposed to be done
by the end of this year but

most airports in we can
see a show of hands as far

as Atlanta airports are you know giving
your thumbs and forefingers slap here.

Anyone go through recently.

Is it ten fingers.

OK so now they've switched it.

Basically switched to two to ten fingers
on this and what I've been told behind

the scenes it's just interesting because
you know not being in the government.

You know you kind of remove you only
see partially what's going on but

I was told Representative dicks.

From Washington State just was
relentless and just kind of pounding on

Ridge hounding him you know nonstop
until finally they OK Will do.

It will do it.

So you need to find a champion here.

I have three or four minutes I have
two slides of just kind of reflections

on implementation factors but
why don't I open it up to questions and

said if there are questions
I'll spend the three minutes.

This.
Yeah.

Right.
So they give out so

I you know I didn't go into that
detail if you look at the OP that it

talks about it more so the idea is you're
not going to give a sixty day supply.

Which is what a person
would need to everybody.

You're going to give
basically ten pills which.

You know is maybe in enough for
five people days or

something like that in the idea is that
this will be done in eight hours for

everybody and then after you're done
with that then you open up the pods and

you open up the high schools in send
people back to get their resupply for

the rest of what they need and also for

there's going to be some people
falling through the cracks.

There's the pediatric thing that's more
special in people who might be allergic to

it and people who are out of town who
who had their schooling or something.

Yeah.

Yeah

so I of.

The last part effectiveness.

Like cost effectiveness.

I mean certainly we're looking at the main
measure here we're looking at our

lives saved.

You know on all the graphs we have you
know a number of deaths versus the number

grams released A versus how quickly
we can distribute antibiotics.

So we do you know I haven't
shown you any grasp it.

Yeah.
So there's a good question.

You know I do have a little here.

So remind me of your first question I
was I was going to respond to that I.

Your first question was right.

So I don't do I have a slide right up
front that I skip kind of more for

the risk analysts in the world.

I don't do threat analysis.

OK So you know I didn't tell you
what's the relative likelihood or

what's the just the likelihood
of a smallpox attack or

you know I mean each of these I
can tell from the government or

from people who have secret top
secret clearance you're getting.

You know things on there.

You know the latest reports
every day on their desk that all

of these scenarios I gave you are all
within the realm of probability.

I mean realm of possibility
I should say that you know

these are could happen these
are just totally in the sky and

I kind of leave and my view is we
have a very poor threat analysis and

we couldn't figure out if Iraq
had weapons of mass destruction.

How were really going to know what's going
on in the caves of Afghanistan impact

Pakistan and all that and secondly
the vulnerability analysis part is you

know you know we saw it with nine eleven
terrorists really want to do something.

They're going to be able you
know they really want to go off

the World Trade Center.

They're going to do it.

We're not going to stop them in so
my view is this whole thing comes down to

consequence analysis which is why I
started my talk by saying I know five ways

terrorists could kill one hundred thousand
people and that's what I'm focusing on

in my view is if the government could
take these five scenarios off the table

then they would have Department Homeland
Security would have been hugely successful

and they would force you know terrorists
to to hit us with attacks that

would maybe kill a couple thousand people
not a couple hundred thousand people.

Your second question about
data your absolute right it

really depends on problem the problem.

So for it.
Sample the U.S.

Visit thing there was a wealth of data
in you just do statistical analysis and

fit distributions and all that stuff and

we could say fairly confident
we what's going on and

some of the problems like the botulinum
toxin thing there is two orders of

magnitude uncertainty of the dose response
I said a millionth of a gram it might be

one one hundred millionth of a grammar
might be one ten thousandth of a gram and

you know I've synthesize and collected all
the literature including Russian studies

from one hundred fifty nine that people
didn't even know about in this country and

you still don't know you
can't do a test on people and

they've done some primate studies but
they're just dreamily expensive and

so what you try to do is frame the problem
and I didn't focus on that but

about really talked about there's
a fair amount of uncertainty of whether

terrorists could pull this off
if they could pull it off.

We have a very good idea of how many
people were dying to be hundreds of

thousands but

there's certainly a lot of uncertainty in
smallpox what's exactly the going to be.

How are people going to behave
during any kind of a pandemic or

any kind of terrorist attack and
you know my view you want to do is

develop policies that are robust to these
uncertainties you want to acknowledge

these uncertainties and to the extent
possible quantify them for policymakers.

Do you want to close it out there is one
I don't know if you have a question or

if you want to get.

OK.

So.

Yeah.

Yeah I mean it's interesting because our
new secretary you know is a governor

from a border state and presumably she
was chosen to take this out my view is

this is a hugely difficult I
mean was adept difficult math.

I'm just going to send you to
the risk analysis paper and

partly why I don't talk about this is that
it takes more than ten minutes just to

kind of describe what I'm doing and
also it's hard to give it across without.

It's such a politically charged issue does
that the other ones it's difficult to

get across that I'm just trying to be
objective here and look at the numbers.

You know I've briefed you know I will say
one of the bottom lines is that the bottom

line you know the to
catch people the border.

It's a several step process first
you have to apprehend them first.

You have to catch them sneaking across the
border then you have to detain them and

remove them the bottleneck has been the
detention and removing It's like a cooling

system where there aren't enough servers
with the servers of these detention beds

and they just kind of catch people
released them into the streets of the U.S.

and then they disappear and so
at least I pointed this out to them.

So if they want to you know
given the amount of resources

they want to spend they should
spend more on detention beds and

less on border patrol agents
particularly because during the Bush

administration they were taking National
Guard who had already done a tour or

two in Iraq and
sticking them on the border.

I think for photo op purposes and
I just thought it was Ron.

But this is you know Homeland Security is

a small part of the bigger picture
of immigration in you know.

My grandparents are were immigrants and
you know but

my view is that we have to
bring everyone in legally

who we want in this country who
you know there's a lot of jobs.

You know maybe not as much in this
financial environment a lot of jobs people

who live here don't want and once you
clean up that system then the people

who are actually sneaking across the US
Mexico border are probably bad guys who

are probably more up these days to be
involved with drugs than terrorism but

then it would be a lot easier
if you just have a few people

crowd coming across the border rather
than you know thousands and thousands and

thousands so
that's kind of my short take on it.

Yeah.

Yeah.

So I would say and so I have.

Refrain from giving a whole lecture on
these last two slides but just point out

this thing about high profile outlets
I would say journals like Science and

Nature which I haven't we had one kind
of rebuttal in science but N.A.S.

which I have like five papers or something
and I think these papers carry enough

reputation that most policymakers would
look at that and say OK this has the stamp

of approval from a serious scientific
journal and so we kind of believe it.

You know I mean if if they don't
have the people under them or

their Ph D.'s that kind of read the paper
carefully they're more apt to believe if

than if they just got some report from
Stanford University or something.

So I think that plays a role.

I will say I would agree with you.

I think the op eds I wrote so
I have you know six or

seven you know you know five in

the New York Times one in The Washington
Post one in the Wall Street Journal these.

In my perception is that
these have a huge impact.

You know the one thing you're educating
the public which you know for

something like telling them to buy mass
can really help them there was you know

a small surgeon if you got
a million people or something.

Reading this.
Some people will go out and do it.

I think more importantly is it put
puts pressure on the government to

act when they see an op ed in
a major newspaper like that and

I will just say you know as a former
editor of operations research.

You know there are just a lot
of luck involved getting stuff

in to various journals
whether it's be in a yes or.

The New York Times I think you'd.

You know the hardest thing like a lot of
things in life is getting the first one in

and now at this point they know me and

my like the one I had on post-traumatic
stress disorder a month ago.

You know them.
The last round the editor.

You know sociate or
whatever wrote back and

said Until next time you know
it's kind of they know me.

Now they will not that will
publish everything I send them but

they will read what I send them whereas
you know the first one or anyone you send

out of the blue probably the chances that
they're going to read it is pretty small.

Unfortunately so but I will say I try I
mean I think the two things that I do

that have been most effective one is is
this one which has a fair amount of luck.

The other one which is more
of the blood sweat and

tears just casting a while
as wide a net as possible.

And I go to Washington and
I will sometimes spend five or

six of you know a good five or
six one hour talks.

You know all these different
branches of government and

you just hope one of these hits.

In the hit rate overall is low but

if you're casting a wide net
something you know hits.

I think you should thank

him for

a little humor to endure is a tough
one to help you for your education.

Thank you.

Yes.
How do you think that things.