>> From the Library of
Congress in Washington D.C.
[ Silence ]
>> Welcome everyone, thank
you so much for coming,
today's lecture is something
[inaudible] more appeal
to younger people, but the other
age for this institution is 58.
So, that's -- well maybe, I'm
[inaudible] Science Technology
and Business Division of
the Library of Congress,
this is event is sponsored by our
[inaudible] Science and Technology
and Business Division and
also [inaudible] fund we have.
And we have a very
interesting speaker today,
I wake up with checking my Tweet and
go to bed with checking her Tweet --
and which my husband
doesn't like, but it is --
the Tweet is now and
on her Facebook,
social media became a
very more serious tool
for the various reasons
for [inaudible] to.
For example, this epidemiology
and infectious disease monitoring
and today's speaker Dr. Madoff is
the editor for a lot of people work
in infectious disease following
this information from ProMED mail.
And we have information
hand outs outside,
if you're interested
please pick it up
and Dr. Madoff did his medical
degree at Tufts University
and then moved to -- his fellowship
to Harvard Medical School,
and he served as faculty
there for many, many years.
And stayed in Boston and now he is
professor at UMass Medical School,
and while he's editor of this
ProMED mail and also he served
for the Massachusetts
State Department,
Department of Public Health.
So, he has three [inaudible] and
very busy, but kind and he came
to share with you for the
very interesting effort
to monitor disease worldwide.
Okay, before further do,
Dr. Madoff, please welcome.
>> Lawrence Madoff:
I'm going to touch
on several things during my talk
including some disease outbreaks
and it would be hard not to be aware
of emerging infectious diseases
in the last few months with the
Ebola crisis raging in West Africa.
I understand the President is
going to address that issue today,
and we've also heard
about Enterovirus d68 just
in the last couple of weeks
and it really is very frequent
that we hear about these outbreaks.
I'm going to do a few things,
I'm going to give a little bit
of a historical perspective
on our understanding of trends
in emerging infectious diseases,
talk to you a little bit
about how do we learn about
infectious disease outbreaks,
why it's important to learn
about them quickly and early.
And I'm going to do a lot of
this through the lens of ProMED,
the program for monitoring
emerging diseases,
which uses non-traditional sources
including media and social media
and sort of first hand
reports as a means of reporting
and detecting outbreaks of
emerging infectious diseases.
So, I'm going to start historically,
this is the CDC's publication
of morbidity and mortality weekly
reports and this is from 1981,
so we're moving back in time here,
and this is reporting on an outbreak
of Pneumocystis pneumonia
occurring in gay men in Los Angeles.
And this was the first public
report of a huge outbreak of course
that we all know today
as the HIV AIDS epidemic,
which has spread globally,
affected millions
and certainly is a major
public health and medical issue
which continues to this day.
And it was detected because
this cluster of young men
in Los Angeles got sick and needed
a drug that was only available
from the Centers for
Disease Control [inaudible],
and an astute observer
noticed, you know,
this is not the usual population
that gets this illness,
what's going on here.
And they astutely reported it and a
couple years later the cause of AIDS
and HIV was discovered and,
you know, I'm not going to go
into a long detail about that but
this of course was not the beginning
of HIV the disease, this
was not the beginning,
these were not the
first cases of AIDS.
AIDS had probably been
around for at least 30 years
when this outbreak was detected,
HIV had been found in tissue samples
in the U.S. going back to 69',
plasma from the Congo in 1959
and an evolutionary model
from looking at the sort
of biological clock
of the HIV virus.
It suggest that it emerged or
transferred to humans in the 1930's,
so it had been with us for
quite a long time and yet --
and as we all know it
is not a subtle disease,
this is not a disease that you
could see and easily miss or forget,
and yet we knew nothing about it.
And why is that and why was that and
why wasn't it detected, you know,
years earlier, so I'm
going to be a little unfair
to William Stewart the
1969 U.S. Surgeon General,
this quote has often
been attributed to him,
but I can't find it
written down anywhere.
And in fact many people
say that he never said this
and Dr. Stewart did
many good things,
he started the anti-smoking
campaign and I don't want
to [inaudible] his
name, but I used this
because this reflects what was a
prevailing attitude of the time.
That infectious disease
did not really matter
so much anymore we had
really conquered them.
This is a real quote, this
is from one of the, you know,
fathers of infectious
disease, the clinical specialty
of infectious disease, President
of the Peter Van [inaudible]
hospital, you know.
A learned man of medicine who said
that he really couldn't figure
out why we would need to
have infectious disease
as a medical specialty
anymore because,
again this was going to go away.
And this was into the late 1970's
now, and if you look at the history
of infectious diseases it's easy
to understand their optimism
at that time, that, you
know, going from a case --
a crude case raid of 800 per
100,000 infectious disease mortality
at the turn of the last century.
Down to nearly vanishing
by the 1970's and 80's,
you can understand why they thought
that infectious diseases were
going away and this is, by the way,
this is when I trained and
went to medical school,
and this was certainly an attitude
that was not uncommon
even at that time.
So, there's one other
quote is that "you can't --
what has happened does not
always predict what will happen"
and what -- by the 1990's
there was a recognition
that in fact microbes were with us,
that emerging infectious
diseases were going to continue
to be an issue and a problem,
and weren't going to go away.
The first institute
of medicine report
on emerging infections was published
in 1992 with a lot of input
from Josh Lederburg
who I was talking
about with Dr. [inaudible] earlier.
And other people who recognized
that microbes were very hardy
and very able to mutate, and
Dr. Lederburg was fond of saying
that it was a battle of our
wits against their genes,
because of their ability
to evolve and mutate
and have short life spans.
And by the 1990's the -- we have
sort of come around and realized
that infectious disease weren't
going to go away, but it was easy
to understand how an infectious
disease could be missed,
if you're not looking for something
you're not going to find it,
you're not going to see it.
So, by 1993 Ruth Berkelmen and
Jim Hughes were writing, you know,
the conquest of infectious diseases
which we people seriously talked
about the conquest of
infectious diseases
as if this was a battle we could
win, and who are we kidding
and I think that that's a
much more realistic view.
So, people who are smarter
than me, this is Jack Woodall,
Barbara Hatch Rosenburg and
Steve Morris, we're very clever
in recognizing two things that
were going on in the 1990's,
one was this rise in the recognition
of emerging infectious
diseases helped by HIV
and Legionnaires disease and
some newly emerging color.
I'm moving back into Latin America,
other things that had happened
that showed them that infectious
diseases weren't going away
and also recognizing the use
of the growth of the internet.
And you've got to remember
that in the 1990's, you know,
regular people were just starting
to have email and it would moved
out of kind of national laboratories
and a few research universities
into the main stream, and
people could actually get
and receive email.
And it was also moving into other
countries including the developing
world, so around that time they
started ProMED and it's a follow
on of a bio weapons conference
actually that was sponsored
by the Federation of
American Scientists,
the World Health Organization.
And they started this internal
mailing list with 40 people just
to tell each other about news of bio
weapons and they quickly realized
that bio weapons could not
really be distinguished,
biological attack can't
really be distinguished
from a natural outbreak.
And that outbreaks of emerging
infectious diseases were just kind
of a natural act of bio terrorism
and they found ProMED mail,
the program for monitoring emerging
diseases, they founded it in 1994
and it was and still is a
system of moderated email lists.
It has and had a website and we now
also branch out to social media,
so I'm going to talk about
social media, some both as a way
of disseminating information
and also about learning
about disease outbreaks, but we'll
talk about that a little bit more.
ProMED is free, we have about 70,000
subscribers essentially representing
every country in the world, and
it's a mailing list in some senses
but it's really more
like a social network.
This is a way that infectious
people who are interested
in emerging diseases can
communicate with each other,
there is an emphasis on one health
that I'll talk about briefly later,
the notion that animal and
eco-health are important
in emerging diseases, and
I'll also talk about a system
of regional ProMED networks
that now are out there.
But I'm going to use this
really as a way of talking
about some disease emergences
and how we learn about them more
than I'm going to talk about it,
this is the ProMED website,
it's a very useful tool.
For those of you who don't use it,
it's a way of getting
information daily, it's a way --
there's an archive of going back
to 1994, maybe 50, 60,000 reports
of emerging diseases that can be
searched by free text and by date
and disease and species
affected and so forth.
So, you can find out historically
about emerging diseases as well,
and all of this is linked, it's
also as I mentioned available
on Facebook, and Twitter
and other ways.
ProMED has about 50
people who operate it,
they're located throughout the world
and for the most part ProMED
functions virtually, we use email
and web tools to generate
our reports.
And as I mentioned there are
a system of regional networks
of ProMED, these are kind
of semiautonomous regional based
systems that are similar to ProMED,
they're meant for communicating
disease outbreak information
of interest in the
regions where they exist,
which are often less well served
in terms of infrastructure
and information resources.
But are often hotspots for disease
emergent's, so Latin America,
Southeast Asia, the African
continent both in English
and French, the former Soviet Union.
Middle East, and North Africa and
South Asia are two new networks
which we just brought on
board and are growing,
so these don't just re-broad,
cast or translate ProMED reports,
these are of and for
people in those regions
who operate these networks sort
of with our editorial guidance
and infrastructure support.
But they exist autonomously
and carry often different
information then ProMED,
so I'm going to fast forward about
20 years from the first outbreak
that I showed you to another
outbreak, this one was an outbreak
of pneumonia that was
happening in Southern China.
And we learned about this from --
as we learn about much of
our information from one
of our subscribers, as I say
ProMED is a two way network --
a social network, if you will.
And he asked if we had heard
about an epidemic in Guangzhou,
someone from a teachers chat
room lived there and reports
that hospitals have been closed
there and people are dying,
and this came to our
attention, we opined about it.
At the time there was a lot of
interest in Avian Influenza,
H5N1 had recently been
discovered and we wondered
if this could be the long awaited
human outbreak of Avian Influenza,
but we didn't know and
we put it out there
and soon after everybody learned.
By the 12th of March the W.H.O had
issued a global alert about cases
of A typical pneumonia, and
this was the Sars outbreak,
this was the beginning of the Sars
outbreak but prior to the 12th
of March -- a week earlier than the
W.H.O had announced this outbreak
to the world.
There were already people
appearing in Canada,
there had already been deaths
due to Sars in Toronto,
and how did the people there know
about Sars, well they knew about it
from informal sources of
information, sources like ProMED
and other sources that I will
tell you more about later.
But not through the formal
public health system,
but through informal recording
systems, now I put this
up for a couple of reasons,
the first is to just show what the
epidemiologic curve or epi curve
of Sars looked like,
there were cases go back
to at least the fall of 2002.
Human cases retrospectively
recognized in rural Guangdong
and then there had been this
peak in February which was
when CROmed first posted on it,
let's see if I can bring that up,
the index case back here,
the first CROmed report here.
And as you can see it went on
to affect thousands of people
in 20 countries, about 20% of the
cases occurred in healthcare workers
which is probably reminiscent
of some
of the other diseases we've
been hearing about recently.
And here's where the W.H.O --
I'm sorry, the first Canadian
case back here in March
and then the W.H.O report,
but why do we care about --
why do we want to learn
about outbreaks early,
and I think this is
important and I hope
that this curve can
help illustrate that.
What is -- because Sars
like Ebola, like Mer's,
like many diseases there's
no drug, there's no vaccine,
this is contained using traditional
public health means like isolation
and quarantine and keeping
people from spreading the disease
between each other's traditional
public health measures.
So, if you can intervene -- if
you could have intervened here
or here then perhaps you could have
prevented much of the later part
of the outbreak, you could
control, contain, minimize
or even eliminate the epidemic.
And maybe instead of an epi
curve that looked like this,
we could have seen an epi
curve that looked like this
and prevented thousands of cases
of this disease from occurring,
so that's why it's important to know
early, that's why it's important
to have an early warning
so that we can intervene
with public health interventions.
How do we learn about
outbreaks, so I --
as Dr. [inaudible] mentioned I'm
-- I sit somewhere around here,
I'm a state public health
official in Massachusetts
and we have a very robust
public health system,
in many parts of the world, not
everywhere, but in many parts
of the world there's a very
robust public health system.
Where information about
disease comes from the public,
from practitioners, from private
and public health laboratories
and this filters through to
local health and regional
and state public health up to the
ministries, CDC level and eventually
to world bodies like W.H.O
but as you can imagine --
so this is a great
system in many ways.
But you can also imagine ways in
which the system isn't perfect,
for example what happens if somebody
forgets or has a disincentive
for reporting an outbreak, and there
are many kinds of disincentives
for reporting an outbreak.
It can damage industry and
tourism and agriculture,
it can make you look bad, you
know, if something happened
on your watch it's not
good and also it's slow,
you can imagine the time
it takes for information
to filter through these layers.
In addition, if a disease
is not a reportable disease,
it's not recognized as
something that needs
to be reported then
it isn't reported,
so if something is happening
but there's no box to check
on the case report form
or it's not on a list
of notifiable diseases
it may never be reported.
And maybe this was the case with
HIV, there was a mysterious disease
that was killing people,
but there was nothing --
who did you tell, how did
you tell people about it,
there wasn't a case report form for
HIV because there was no such thing.
So, these are the advantages and
the disadvantages and I've sort
of [inaudible] to them already
of traditional public health,
it's robust and sensitive,
it's accurate, it's validated,
it's quantitative but
it also can be slow.
There are incentives
for non-reporting,
there can be broken links
in the chain that result
in something never being reporting,
it can miss uncharacterized
or novel disease and of
course it's expensive.
Traditional public health
in this country, you know,
it's cheap compared to
the health care system
but it still costs billions of
dollars, so what has developed
in the 20 or so years since
ProMED has been around,
ProMED is now celebrating
its 20th anniversary.
Is a system of informal surveillance
or what the W.H.O now calls
event based surveillance
or bio surveillance,
and this is a system
that monitors non-traditional
sources of information,
so things that we don't always think
of as health data like the media,
like social media, like
school absenteeism,
purchases from pharmacies.
Things that are not necessarily
part of the health system and
yet can inform us about
public health events
and can help us detect
things rapidly and early.
So, informal source surveillance
systems like ProMED talk to,
and you notice these -- the
bidirectional arrows here talk to
and receive information from
the media, from laboratories,
from local health officials, from
anybody with an email account,
from health care workers and of
course from ministries of health.
And they in turn disseminate
this information widely,
so that people learn about it,
so the advantages of this kind
of a system are that its
rapid, that its transparent
because there aren't any
disincentives for reporting things.
You can use any kind of
a source of information,
you can identify any kind of event
regardless of whether it has a name
or a diagnoses or [inaudible] agent.
And it's relatively inexpensive
to operate these kinds of systems,
now there are disadvantages,
these systems can potentially
be inaccurate, right,
you can amplify mis-information
or dis-information.
They're not necessarily
quantitative, this is not a system
for counting how many cases
of Influenza there are
or how many cases of Salmonella
occur in a jurisdiction in a year,
this is a way of finding out that
there's an outbreak of Influenza
or Salmonella but not of
knowing how many cases there are,
and whether their laboratory
confirmed.
And there are biases inherit
in this kind of a system,
so places that have a
lot of information tend
to convey this information
publicly and get it re-broadcast
and re- -- and amplified.
Until pretty recently the
internet was largely in English
and that's something that has
changed, you know, in the last five
to 10 years but it's still
changing, and it's --
there still is not the
richness in many languages
that there is in English.
Obviously things can get
sensationalized as well,
one thing that happened
in 2005, probably --
now the World Health Assembly meets
and develops what are called the
International Health Regulations,
periodically, and these had
not been revised in many years.
Starting back in the
90's there was an effort
to rewrite the International
Health Regulations
which had actually required
countries only to report
on Colora, Yellow Fever and Plague.
And if a country did not have one
of those diseases to report --
Small Pox was also on the
list, but was eliminated,
then there was no requirement
for that country to report it
to the World Health
Organization, many did
and the World Health Organization
of course responded to many things.
But these revised international
health regulations,
which I think were hastened
by the Sars epidemic,
so that the World Health Assembly
actually adopted them in 2005
and they took effect in 2007,
for the first time countries
were obligated to notify W.H.O
of any public health event
of international significance
or international concern.
And not a particular set
of diseases but anything
that could be an international
problem,
it also for the first time
allowed W.H.O to think
about unofficial sources
of information.
Now W.H.O is composed of
its member countries and so
under its existing regulation it
was really only allowed to think
about something, to
learn about something
from a member country
telling it about an event.
Now it's allowed to and encouraged
to use unofficial reports,
and as I mentioned they are now
encouraging member countries
to adopt informal bio
surveillance systems,
not unlike ProMED and similar.
So, just want to mention that,
another thing that I want to talk
about briefly in regards to
emerging infectious diseases is the
importance of One Health.
So, One Health is a concept that's
been around for at least 100 years
at this point, was certainly
something that Virchow one
of the fathers of pathology and
classical German pathology talked
about and Olser talked about.
He used the term One Medicine and
we've more recently adopted the idea
of One Health, but basically it
thinks about disease without regards
to species boundaries and
it recognizes that we --
a more ecological approach to
health, that we depend on the health
of the environment and the
animals around us and vice versa.
There really is a -- you can't look
at human health without looking
at animal health and vice versa,
and so people have called
this Eco Health or One Health
but it's become an
increasingly important movement
and certainly is something
that ProMED has recognized
and embraced since its founding.
So, I'm going to -- a little bit of
research I want to present to you
for a minute, this is work
that was done by Mark Woolhouse
in the United Kingdom and
he actually looked at all
of the human pathogens
he could find,
all of the named human pathogens he
could find, he found 1,407 of them.
And he noted that 58% of
them are zoonotic meaning
that the microorganism that --
is shared by at least one
other species, not just humans
but if you looked at the
recently emerged diseases,
130 of 177 recently emerged
diseases were zoonotic.
So, there was about a
two-fold relative risk
of a zoonotic disease
emerging, and if you think
about the important
emerging infectious diseases
that would fall off your
tongue Ebola, Mer's, Sars,
HIV, Salmonellosis, Rabies.
You know, virtually many,
many of the diseases, Anthrax,
that we think about are zoonotic
diseases, they are diseases
that occur in animals,
importantly in animals.
And often humans are just
a kind of a side root,
and not even the most important host
of these pathogens, but if you --
and those pathogens it turns out
are much more likely to emerge
and this looks at the host range of
a pathogen and it's subdivided here
by different types of pathogens.
So, if a microorganism has only
human pathogen, we say that,
you know, that's it's host range, if
it's shared with one other species,
two other species or more
than three species you can see
that it's risk of emergence
increases.
This is particularly true
for viruses because --
probably because of their capability
to mutate and evolve rapidly,
but it's really true of
almost everything and I think
if you had enough examples it
would be true of everything,
that zoonotic pathogens are
much more likely to emerge.
And that's why ProMED spends so
much of its time looking up at
and thinking about animal
diseases, to the point where a lot
of our human readers say things
like, you know, I really don't want
to hear about foot and mouth
disease or blue tongue or, you know,
this wheat rust because we
also report on plant diseases.
But it's really important to stay a
tuned to those things because that's
where the next disease
emergence is going to be from,
ProMED was the first to report
on the Novel Coronavirus
isolated in Saudi Arabia.
And I'm going to talk about
this for a couple of reasons
because this was an example of a
disease that was reported to us,
not from media or social media,
but from a laboratory actually.
This was an astute laboratorian, Dr.
[inaudible] working in a hospital
in Jeddah who identified a Novel
Coronavirus from a patient there.
And he had to be sequenced in
conjunction with a laboratory
in Europe and put the
result out there in ProMED,
frankly because he was frustrated
with how long it was taking
for an official report
of this to come out.
And so this was a way of letting
people know about something
without having to go through
all of the official processes
that are sometimes necessary,
and so I put up this next slide
because it was a good thing he
did, put it out when he did.
If you look back here, he put
this out on the 20th of September
and by the 23rd of
September physicians
in the United Kingdom were
reporting on a similar case.
They read this report and they said
hey we've got a very sick patient
from the Middle East and we
don't know what he's got,
all of our laboratory
tests were negative.
And when they read his report they
tested him for Coronavirus and found
out that indeed it was Mer's,
now another side effect
of the informal system
of reporting is
that it often encourages official
reports to follow quickly.
Because once -- and
I work, as I say,
I work in a state health department,
I understand the barriers
to telling everybody everything
right away but once something is
out there it's a lot
easier to tell people.
So, once something is already public
information it's then possible
for the Ministry of Health to
confirm that and to say yes,
yes we know about that,
we're on top of this
and we've also seen some cases and
this is what we're doing about it.
And that's what happened
in Saudi Arabia,
quickly after the initial ProMED
reports the Saudi government issued
public reports acknowledging
the existence of Mer's,
what came to be known
as the Mer's Coronavirus
or Middle East Respiratory
Syndrome Coronavirus.
One other thing I want to say
about ProMED is it's a forum
for people to ask questions.
So, when the Mer's incident
became public one of our readers,
a woman name Irene [inaudible]
who works for SOS International,
said I'd be interested to
know whether this outbreak
of severe respiratory disease
that we saw in Jordan in April
of 2012 could have been Mer's.
And she was right, in fact, that
was the beginning -- that was the --
now we think this is the earliest
known human cases of Mer's occurred
in a hospital cluster in Jordon
back in 2012, so this is another way
of illustrating how informal
sources of information outside
of the government sector
can be important.
I mentioned the growth of the
internet and why that matters,
but -- and it has also
allowed this kind of ecosystem
of informal event based bio
surveillance systems to flourish.
And I didn't want -- I certainly
don't want you to come away
from this thinking that ProMED is
the answer or is the only system
that does this but there
are a number of systems.
That [inaudible] was the first
web-mining tool developed
by the public health agency of
Canada, Health Canada at the time,
and it actually still
exists and goes out
and it scans the web many times
a day and finds outbreak reports.
Health Map uses a similar system but
uses a geographically based method
for finding out where outbreaks are
and mapping them in near real time,
and these are a number of other
systems that are both official
and unofficial, and formal and
informal that do the same kinds
of things that I'm
telling you about.
That look at informal sources of
information, information outside
of the health sector for trying to
identify early public health events,
the problem with this and with
many sources of information is
that there's a lot of
information out there.
This is a graph from the
economists showing just the amount
of information, I believe --
yes, this is a unit that I can't
even tell you what it means.
Exabyte's of information
and just telling what the --
looking at the volume in Exabyte's
of information, and the effect
of this is a fire hose of
information, how do you make sense
of it, how do you find in that
stream of information the things
that are of interest to you.
Now, ProMED uses a combination
of methods, we work very closely
with Health Map, as I mentioned
before, Health Map mines the web
for information on infectious
diseases, ProMED uses people,
human beings to look at all of
this information and filter it
and decide what's important and
group things together and comment
on it and put it out there.
Other systems use other ways
of looking at large amounts
of information, so I couldn't give
this talk without at least talking
about Ebola, this was from the New
York Times I guess two days ago
with CDC giving an
outside estimate of one
and a half million cases of Ebola.
Again, this was probably a
missed opportunity and one
if you think back to that epi curve,
if we could have intervened
affectively at an earlier time
on this epi curve, we would
not be seeing the kind
of exponential growth in the
epidemic that we're seeing now.
This -- I put this up and I was a
little hesitant to show this cartoon
because I certainly don't want
to make light of Ebola in any way
but this -- [inaudible]
published this cartoon
in the New Yorker in 2001.
And this is a women
dealing with her insomnia,
what things she was thinking
about and her inability to sleep,
and number two is Ebola and number
three says why am I obsessing
about Ebola, there's no
Ebola on 33rd street, yet.
And it is just to tell you that
I think was very [inaudible],
this is a -- she was ahead of her
time in recognizing how much worry,
and again speaking with my
state health department,
had on how much time we've
spent worrying about Ebola
in Massachusetts is a
testimony to how right she was.
This again was the first media
report that we found about Ebola,
this was back in March of 2014, we
think the index case was probably
in late 2013 but it came to media
attention somebody first used the
word Ebola in association with
this outbreak in a media report.
So, I also want to acknowledge
that much of what ProMED reports is
from traditional media, from
newspapers and magazines
and other material that's
out there on the internet,
journalists are doing a
lot of good work for us
and they're uncovering
outbreaks often, you know,
before they're named or identified.
And this was, you know,
could this be Ebola,
one of the important attributes of
a ProMED report is that it comes
with moderator comments, all of
our reports are read and reviewed
and commented on by a
subject matter expert.
And this particular subject matter
expert noted [inaudible] in this
that most of the victims had
been in contact with the deceased
or had handled bodies
was very characteristic
of hemorrhagic fever
viruses in Africa.
And there were -- she mentioned
other Hemorrhagic Fever viruses,
our deputy editor [inaudible]
who wrote this commentary,
and it was very astute for
her to pick up on this,
now ProMED has gone on, it has
published, this is now two days ago,
so we're probably up
to report number 175.
So, it's a good way of capturing
on a daily basis a summary
of what's going on and Ebola,
this is Health Map website
that I mentioned, I'm going
to talk about a few other ways
in which informal source information
is mined for public health use.
Health Map is very good at this and
as I say they're close collaborators
with ProMED, and they also picked
up very early reports of Ebola
or Hemorrhagic Fever -- this
Hemorrhagic Fever outbreak
and reported on it, and you can
use their website, healthmap.org,
to follow the outbreak again
in more or less real time.
Twitter is a very good way and one
that I often hear about outbreaks
for the first time, and again
attribute to good journalists,
Helen [inaudible] is one
of my favorite reporters,
[inaudible] another great
reporter but also official.
As you know, organizations like
W.H.O and Massachusetts Department
of Public Health and ProMED
also have Twitter accounts
and it's a convenient real
time way of following us.
But in addition to that,
what happens on Twitter can often
be an early warning of outbreaks
and so this was work by [inaudible]
at the University of Iowa looking
at Twitter as a predictor,
so these were just --
this was just a very simple --
this was reported versus
predicted weekly [inaudible]
in the 2009-2010 flu season.
And he was looking at predicted
reports based on Twitter
versus reported disease,
and you can see that even
that the Twitter reports mentioning
Influenza very closely followed
or perhaps even preceded
the reports --
the official reports of
Influenza like illness.
Our colleagues at Health Map used
Twitter in Haiti, Twitter reports
from Haiti to follow
the Colora outbreak
that has been happening there,
that unfortunately
continues to occur there.
And showed that in fact Twitter
reports preceded the actual case
counts, so that there was
actually a peak in reporting
of diarrheal disease
that preceded by a couple
of weeks actual public
health reports of disease,
so again this was simply using
publicly available information
from Twitter looking for
terms that were suggestive
of Colora from people in Haiti.
And they were able to use that
information to predict something
as simple as -- without
even monitoring the content
or the telephone numbers
that are called from a phone,
just looking at the number of hits a
cell phone tower is having can tell
you that something is happening,
I can't necessarily
identify what that event is.
But patterns of cell phone
use can identify in real time
or near real time an
event that's happening,
and this again this was work well
by Nathan [inaudible] and colleagues
on cell phone usage patterns
in Kenya, and just showing
that hotspots of cell
phone use could tell you
if an event was going
on in that area.
Google flu trends is something
that you've probably seen
or are familiar with, is another
very clever tool that looks
for search terms and it turns out
that when people are searching
for Influenza or other terms that
might suggest that they're sick
with Influenza, this is very good
predictor of Influenza activity.
And they showed that and showed
that the Google flu trends activity
very closely paralleled CDC volume
of Influenza, could also
be applied to [inaudible]
and the same trend has been used.
So, that using the so called search
stream, what people are searching
for and where they are can be a very
informative piece of information,
Marcus can also predict events,
so when a stock market crashes
or booms it can indicate that
something is about to happen.
And people have exploited
that phenomena by looking --
by using market based
approaches to predicting
or monitoring emerging
disease outbreaks.
If you can empanel people who
know a lot about Avian Influenza
or Influenza and ask them for
money to bet on, essentially,
to buy and sell shares in a
market on what they think is going
to happen to the flu,
it can predict,
it can be a leading
indicator of what's likely
to happen with Influenza.
This is a colleague [inaudible]
who demonstrated this in Iowa
for the Iowa prediction market,
the Iowa prediction market you may
have heard of are frequently used
to predict election outcomes
and were often followed during
the last Presidential election.
But they can also predict flu
outbreaks, if you ask people,
it's a way of aggregating expert
opinion in a way that's more
than asking somebody what do
you think is going to happen
but how much money are willing
to put on the proposition,
and that's a way of
aggregating expert opinion
and able to monitor and exploit.
So, does this work, do all
these tools have they helped,
and the answer is they seem to, this
is work that I did with colleagues
at Health Map looking at the
time to outbreak discovery
and public communication of
outbreaks, and we looked at this
for widespread infectious disease
outbreaks that had been reported
to W.H.O. And what we found is
that the time to outbreak discovery
and the time to public communication
of an outbreak had declined steadily
from 1996 to 2009, the study
period which we believe corresponds
to a period when this kind
of activity was going on,
can we prove the relationship,
no but I think it's a
reasonable [inaudible].
You'll also notice
what I think is the --
either the Sars effect or
the IHR effect that seemed
to accelerate the trend towards more
rapid public reporting of outbreaks.
As I mentioned there are a
number of systems out there
that do bio surveillance, this
slide makes a case for the benefits
of redundancy, the benefits
of synergy if you will.
That if you look at any given system
the time to outbreak detection,
and in this case higher is better,
the number of days before
official notification is plotted
on the vertical axis here, so if
you combine systems you can find
out about outbreaks faster than
any individual system is capable
of doing.
So, this makes the case for synergy
from multiple surveillance systems,
finally I just want to say that
it's important not to focus too much
on any one thing, at
the moment, you know.
We're paying a lot of
attention for example to Ebola,
but it may be that the next thing
that we need to worry about is
over here and so you need to --
we need to keep our eyes on the
horizon, not just in the telescope.
So, I'll really stop here I've
said all of these things and I want
to acknowledge our supporters
who have supported ProMED
and its work, and I'll stop there.
And there's a meeting that ProMED
is involved in, for those of you
who are interested in
this, it happens on --
beginning on Halloween in
Vienna, so thank you very much.
[Inaudible] answer questions here.
>> Any questions?
>> Lawrence Madoff:
Yeah, please, yes.
>> I noticed that you
didn't use the [inaudible],
I was wondering why [inaudible]?
>> Lawrence Madoff: So, the question
was how does big data play into this
and I didn't use the term, but I
certainly alluded to the huge volume
of data, the huge and accumulating
volume of data and yes, absolutely.
These kinds of things can
be useful and I think some
of the examples I used like search
stream is certainly an example
of using big data, you
know, excuse me, one second.
Of hundreds of millions of Google
users entering search terms is form
of big data, cell phone, aggregate
cell phone usage is a form
of big data.
So, absolutely big data can
be mined, but again it alludes
to the slide that I showed of
the fire hose of information,
the trick in using big data is
using it intelligently and using it
in a way that is selective and
can provide useful information.
Otherwise it's just lots of data,
and yes absolutely, if ProMED wanted
to put a report out on every time
somebody had a case of Salmonella,
just to keep harping on
Salmonella, but, you know,
there are thousands of
cases of Salmonella.
It's not -- it wouldn't be
particularly interesting,
the trick is recognizing when
there's when there's an outbreak,
when there's an aggregation,
when there's something
that ties those cases together.
And so the magic in using
big data, if you will,
is in trying to use it intelligently
and that's not easy, and it's --
there are a lot of people who are
smarter than me that spend a lot
of time thinking about that.
And I think that it's very
likely, certainly the kinds of work
that ProMED does we
get a lot of data
and we report only
a little bit of it.
But human -- a human centered
approach is really only one approach
and there have got to
be other approaches,
we're working with some partners
that are using these
kind of approaches.
There's a -- I mentioned
that my acknowledgement [inaudible]
has a program called the emerging
pandemics threat -- emerging
pandemic threats program
which is attempting to
use lots and lots of data
and aggregating it
in intelligent ways.
Including things like looking
for people who are in contact
with animals to see if they're
sick, looking at their serum to see
if there are signatures and
blood and that sort of thing.
We're doing some work with the
defense threat reduction agency
in trying to combine into a single
platform lots of data sources,
to try to make them available
again to human analysts
so that they can try to
interpret the vast amount
of data that are out there.
[ Inaudible Speaker ]
Yes, so the question is, are there
analytic systems that are trying
to do this, trying to combine
data and real time, yeah and not
yet that I've -- at least that
I heard about it successfully,
but there are certainly many
people working in that space.
Please.
[ Inaudible Speaker ]
So, I think your question is how
can you use data to make predictions
about what's going to happen
or to analyze what's happening.
[ Inaudible Speaker ]
Right, you know, I think -- I would
say two things, predictions are hard
to make, especially about the
future, I think that was yogi bear
who said that and I think that's
obviously true, you can look at data
and two people can look at the same
data and make different predictions.
I talked a little bit
about predictive analytics,
things that look at
data as at least --
perhaps not a prediction
but a leading indicator,
something that happens before
our other systems, you know.
So, maybe saying that the flu
is going to happen a week ahead
of when lab tests show
that there's flu,
it may not be because you're
predicting the future,
it may be because you're using
something faster than a flu test.
You know, because it takes time
for someone to go to the doctor,
it takes time for someone to
do the test, so one of the --
I think one of the very
innovative approaches is called
participatory surveillance.
Where people actually monitor
themselves and say, you know,
I think I'm getting the flu,
and can do that on a
website or a cell phone app.
And that data -- those data
are aggregated in a way
that that's intelligent,
the other thing to say is
that I think those -- a lot of this
needs to be demonstrated, I mean.
You know, I think that's what
science is, right, is that we need
to be able to -- you know,
we can have a hypothesis
that data will help us predict
or analyze or see things earlier.
But we have to show it, it's
on us as researchers to be able
to demonstrate that and I think
that's happening, but less than --
you know, there's a lot more
enthusiasm than there are good data.
>> You had about 50 spots in your
word, those people were exposed
or positioned or who are they?
>> Lawrence Madoff: So, Dr.
[inaudible] asked who the personal
that were on the ProMED staff map
are, so they're different kinds
of people, for the most part
they're subject matter experts
in different things,
some of them are a part
of our regional network system.
So, you know, consciously I
think that if they're people
in a region they're going to be
more aware of what's happening
in that region than people
outside of that region,
so some of them are
regional experts,
they're all health specialists
in some form or another.
Some are [inaudible], animal health
people, they're different kinds
of experts out there, but they're
all experts and I should mention
that ProMED staff are
all paid a little --
they get an honorarium for
their work, but we don't pay --
they're really volunteers,
they don't earn --
they don't make what
they earn in a sense.
They're underpaid, but
they're all over the world
and they all participate
in this network.
[ Inaudible Speaker ]
Yes, we're very selective
about who we bring on board
as a ProMED moderator, yeah.
[ Inaudible Speaker ]
Right, yeah, so really the goal
of the regional networks is
to improve surveillance for
the people in those regions,
it's secondarily to improve
surveillance for the rest
of the world, but really its
primary goal is cross boarder
and regional surveillance.
For people to be more aware of
what's going on in that region,
I have numbers somewhere, I don't
have off the top of my head,
most of our subscribers are
in the north, if you will,
they're an English speaking,
in the U.S. and Western Europe
and other English speaking
developed countries.
But we have a large number
of subscribers, you know,
I think there are over a
thousand I believe that subscribe
to our African specific services,
now not all of them are in Africa
and some of our subscribers to
the global list are in Africa.
We do have data on where our
subscribers live, I should point
out that a lot of people don't
necessarily get our emails,
they come to our website or they
follow us on Twitter or Facebook
and get the same reports.
But they don't necce -- we don't
necessarily have their email
address, but there are thousands
of people in those areas
that are making themselves
make use of it and we know
that from email addresses that many
of those people are in government
or regional or local
or international public health
in some form or another.
And many are not, many are, you
know, clinicians or individuals
in those regions who are just
interested, journalists for example.
[ Inaudible Speaker ]
Right, so that again is a reason
for our regional networks is to try
to build interest in a region,
to try to let people in a region
because these people are there
and they're going to meetings.
And they have colleagues
and they're on Facebook
and they can tell their
friends and so forth,
their co-workers about ProMED.
So, that's really our
effort, you know,
beyond that all we have is email.
[ Inaudible Speaker ]
So, just to speak of ProMED,
ProMED is using primarily --
is using social media as
a means of dissemination,
so we have about three or four
thousand people who like us
on Facebook and that's a
way of getting our posts.
All of our posts are pushed to our
Facebook page, so someone could
like us on Facebook and it'll show
up in their newsfeed on Facebook,
the same with Twitter,
we primarily use Twitter,
we have about 4000
followers on Twitter
who just receive our
posts by a Twitter.
And each tweet has a link to the
full report, it just has a headline,
you know, all of our reports have a
kind of stylized headline, you know,
Ebola, West Africa, Guinea
and then if they want
to learn more they can click on
it and it takes them to a webpage.
So, we -- so, through social
media we reach about another eight
or 9000 people but it's -- for
us it's mostly passive, we do --
you know, it's something
we'd like to build on
but we haven't had the
bandwidth to do that yet.
We are not trying to use it to
monitor disease, in the way some
of the other work that
I described is doing
but that's not something
I'm currently working on.
[ Inaudible Speaker ]
So, I think this whole question
is about DNA and using DNA
to track outbreaks or
something like that.
The closest thing that I can relate
to what you're saying would be the
use of DNA sequences from pathogens
and so, you know, that
can be very important,
that can help you establish
a connection in an outbreak.
If you have multiple people
with the same pathogen
and they have the same DNA
fingerprint that can establish
that those cases are related.
In the Ebola outbreak in West
Africa there were sequences done
of the outbreak strains in
West Africa and sequences done
from the strains in Congo, and they
were sufficiently similar to say
that they were related, but
sufficiently different to say
that they weren't directly related.
The outbreak didn't spread from
West Africa to Congo or vice versa,
that they were -- it
was the same strain
but it was a different
outbreak, and so DNA information.
And I think it also speaks
to the need for transparency
because there's a lot
of scientists who want
to keep private their DNA
sequences until they've published it
or until everybody knows about it,
and I think that may not be
the best way of doing it.
So, ProMED has always been
in favor of transparency
in DNA sequencing of pathogens.
So, if somebody knows an Avian
flu strand's genetic signature,
that's important information, that
might be important to someone,
you know, that DNA from someone in
Egypt might be important to someone
in Thailand, say, you know, who
both have the same strain of H5N1.
That can tell how it got
there or where it went,
so I don't think I exactly answered
your question but hopefully I was
at least in the same universe.
[ Inaudible Speaker ]
I guess only if it were
an infectious disease,
if [inaudible] were found for
breast cancer which is certainly
within the realm of
possibility, but that would be --
we're focused on emerging
infectious diseases.
>> Please join me to thank the
speaker, thank you very much.
>> Thank you for your attention.
>> This has been a
presentation of the Library
of Congress, visit us at loc.gov.