April 2, 2025 • 37 mins
What if your Google searches could predict disease outbreaks before traditional surveillance methods? That's exactly what happened during the 2009 H1N1 pandemic when researchers discovered online search patterns matched CDC data – but delivered results much faster.
Welcome to the fascinating world of infodemiology, where digital footprints become powerful tools for public health. In this eye-opening conversation with experts Dr. Heather Duncan and Dr. Patrick Murphy, we explore how researchers analyze everything from tweets to search queries to understand health trends, track disease spread, and even identify mental health risks.
The implications are both promising and concerning. While infodemiology offers unprecedented speed and insights for public health response, it raises critical questions about privacy, ethics, and the responsibility that comes with identifying health risks online. If AI flags someone as potentially suicidal based on their social media activity, what obligations exist to connect them with resources?
Perhaps most shocking is the revelation that just 12 individuals were responsible for 60% of the anti-vaccine content circulating on certain platforms. This precision mapping of information flow demonstrates infodemiology's potential to target interventions effectively.
As social media increasingly becomes Americans' primary source of health information, understanding these digital dynamics becomes crucial for public health. Dr. Duncan shares her vision of creating accessible, automated tools that would allow even small health departments to harness these powerful insights without extensive resources.
Subscribe to Infectious Science for more fascinating conversations at the intersection of technology and public health, and share your thoughts on how digital surveillance might shape the future of healthcare.
Thanks for listening to the Infectious Science Podcast. Be sure to visit infectiousscience.org to join the conversation, access the show notes, and don’t forget to sign up for our newsletter to receive our free materials.
We hope you enjoyed this new episode of Infectious Science, and if you did, please leave us a review on Apple Podcasts and Spotify. Please share this episode with others who may be interested in this topic!
Also, please don’t hesitate to ask questions or tell us which topics you want us to cover in future episodes. To get in touch, drop us a line in the comment section or send us a message on social media.
Instagram @Infectscipod
Facebook Infectious Science Podcast
See you next time for a new episode!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:09):
This is a podcast about One Health the idea that
the health of humans, animals,plants and the environment that
we all share are intrinsicallylinked.
Speaker 2 (00:17):
Coming to you from the University of Texas Medical
Branch and the GalvestonNational Laboratory.
Speaker 1 (00:21):
This is Infectious Science.
Where enthusiasm for science?
Speaker 2 (00:25):
is contagious.
Speaker 3 (00:30):
Everybody, welcome to this episode of Infectious
Science.
We are excited to be back forthis week's episode.
So we're going to be talkingabout infodemiology.
So we are joined by two gueststoday Dr Heather Duncan and Dr
Patrick Murphy.
Heather, would you introduceyourself and tell us how you got
into epidemiology?
Speaker 4 (00:49):
Sure.
So I originally started mycareer as a literary studies
scholar and decided a couple ofyears ago that I was really
interested in public health andwanted to be an epidemiologist.
So I went back to school andgot my master's in public health
and I am now finishing up myfirst year of my PhD in this
(01:10):
field, and I initially had noidea that infodemiology existed.
So I came to the field becauseof my former interest in
studying digital media as ahumanities scholar, and then,
when I realized that was alsohappening in epidemiology, I
became pretty interested in it.
So the work that I'm doing nowis more focused on infodemiology
(01:33):
.
Speaker 3 (01:33):
Very cool, Dr Murphy.
How did you come toinfodemiology?
Speaker 4 (01:37):
Surprisingly through Heather.
I had never heard of it untilshe brought it up and I think
it's fascinating.
It has just so manypossibilities and we're all
about linking things together.
This is a great way to do that,yeah, and Patrick and I co-own a
science and healthcommunications and medical
writing business, so we alsocame at it from that angle as
(01:59):
well, and one of the things thatwe are interested in working on
together is developing usingthe field of infodemiology that
can be hopefully one day in thefuture utilized by smaller
health departments to do butbasically to do more with
surveillance than they'recurrently capable of doing.
Gotcha.
Speaker 3 (02:20):
Very cool, I'd say.
When I first heard ofepidemiology, I immediately
thought of epidemiology and kindof the study of how diseases
are moving through populationsand case numbers, things like
that.
So I know, when I originallywas talking to you about the
episode, I was like oh, isinfodemiology just basically how
misinformation can spread inthe virtual world, just the
pathogens do in reality.
But it's actually broader thanthat and the definition that you
(02:42):
sent back to me was that andjust I want to just make it
clear to our listeners that way,as we go forward, when we're
talking about infodemiology,we're talking about the science
of distribution of determinantsof information, particularly via
the internet, in a populationwith the ultimate aim to inform
public health and public policy.
So can you get into?
Because there's different termswithin it.
(03:02):
Can you get into?
So that's like whatinfodemiology is, but can you
get into, because there'sdifferent terms within it?
Can you get into?
So that's like whatinfodemiology?
Speaker 4 (03:08):
is.
But can you get into what isinfovalence?
Or like, what's an infodemic?
Yeah, that definition that youjust gave.
I just want to give creditbecause he's known as, like the
father of this field, I guess,if you want to use parental
terms, but Gunther Eisenbach isthe one who came up with that
definition in the early 2000sand it's essentially the exact
classical definition ofepidemiology, which is the
(03:29):
distribution and determinants ofdisease with an eye towards
fixing them, basically.
But we're specificallyinterested in digital spaces and
you're right, there's a hugebreadth, I would say basically
everything that epidemiologydoes as well, and actually even
a little bit more.
I think it's just that we areinterested instead of human
(03:52):
beings.
Our unit of study is digitalcontent, basically, and that
could be anything from a tweetto a Google search or a unit of
information essentially, andwithin infodemiology, you
mentioned these two terms info,which is a play on the word
surveillance.
So I'll start with infovalence,because this is where the field
(04:16):
started out.
Well, to give you a little, totell this more as a story,
basically, the earliestinfodemiology studies were
people who were in the mid tolate 90s were starting to notice
that websites were beginning topop up that were providing
health information, and so someof the first questions people
(04:37):
had were like is this goodquality information?
How do we decide whether awebsite is trustworthy?
And nowadays, that's such a, Iguess, a simplistic question
compared to the way our digitallandscape looks.
But that's all it was initiallyso.
Gunther Eisenbach I don't knowexactly how he entered the field
(04:57):
, but he became interested inthese very early studies.
Except he wanted to ask adifferent question.
He wasn't just interested inevaluating the quality of health
websites.
He wanted to know can wesomehow use the internet to do
disease surveillance?
Because there's a lot ofdifferent things that
epidemiologists do, butsurveillance is one of the most
(05:19):
crucial because it's how we knowthat a disease outbreak is
occurring, and especially ifit's something infectious.
Obviously that's one of thecore functions of our public
health departments.
So he wanted to know can we usespecifically Google searches to
find out if there is anoutbreak occurring?
(05:40):
And he and his colleagues had achance to really try this out.
I don't know if you guysremember the swine flu outbreak
of 2009, 2010.
So that was the first time thathe really got a chance to apply
this idea that maybe we couldlook at people's search behavior
and use that to figure out,faster than our sentinel
(06:05):
surveillance networks, whether aflu outbreak is taking hold in
a specific community.
So that's where the idea ofinfovalence was born, and those
early studies by Eisenbach andhis colleagues were pretty
successful.
They validated their dataagainst the CDC's data and they
found that it very closelymatched the predictions that the
(06:28):
CDC was making with their data.
Except it was much more timelyBecause when we're talking about
the internet this stuff doesn'ttake particularly long to
process.
It doesn't have to go throughall of the bureaucracy and pass
from person to person the waythat traditional surveillance
does.
So that was a huge advantage,and at this time also this is
(06:51):
shortly after 9-11, that wasalso when, like the anthrax
attacks took place.
So the Department of Defensewas also really interested in
this technology and insurveillance specifically.
Speaker 3 (07:03):
Out of curiosity.
So I guess people are reallycognizant now, probably more
than we were back in like 2009or 2010, about like how your
information is protected andlike what's private and what's
not, and like.
So it's interesting to me thatis just widely like available
information that you could justconglomerate all of these Google
searches.
And when you're doing that, arepeople Googling like symptoms
(07:25):
of flu or like symptoms of COVID, or is it like looking up news
articles that are related to itin the local community?
What were their kind of searchfeatures, I guess, to determine
yes, there's an outbreak.
Speaker 4 (07:35):
Regarding news, that's almost like a separate
area and I don't know that itwould necessarily fall so much
under infodemiology.
But there are communicationscholars who monitor media and
they use their own algorithms topull headlines and to analyze
data, and infodemiologists doincorporate some of that data,
but there's like a whole otherbranch of study that focuses on
(07:59):
mainstream media.
Yeah, patrick has some thingsto say.
I think about more like modernday concerns about privacy and
particularly like where AI isconcerned as well.
But at this time, the earlystudy, the first study that
Eisenbach did, trying to do thisflu infovalence thing, was
actually really genius, becausenowadays it's very easy I
(08:21):
shouldn't say that it is and itisn't easy to get this data,
because there are challengesthat I can get into a little bit
later, but at the time therewas no Google Trends or there
really weren't tools for doingdata scraping and then putting
it into a usable data set formatthat you could analyze.
So what Eisenbach did which Ithink this is just genius
(08:46):
Eisenbach did, which was I thinkthat this is just genius.
I really admire the way thatthis study was designed.
So what he did was he had apretty small budget.
I want to say it was somewherearound $500 or something, and he
bought a Google ad and when youbuy an ad a digital ad, and I'm
not sure what it's called now,but I think back then it was
called Google AdSense you couldtarget specific geographical
(09:06):
regions.
There were some rudimentarythings you could do to try and
to get to your audience, and sohe set those all very general
and then he used as his metricthe number of clicks.
So it was an advertisement thatsaid something about flu
resources or what to do if youhave the flu, and I think that
(09:26):
when people clicked on it itjust took them to like a WebMD
type site or something like that.
So it was harmless, basically,if people clicked on this ad.
But he used the number ofclicks in different areas and
then was able to again validatethat data against the CDC and
found that actually, yeah, likethe places where people were
(09:48):
clicking were the places whereflu was known to be in high
circulation.
Today it's a little bit, as Isaid, it's like easier to do
things but also more challengingin some ways as far as privacy
concerns.
A lot of infodemiology nowadaysuses social media, because this
is unfortunately or fortunately, I guess, depending on your
(10:10):
perspective, the number onesource of health information for
Americans.
And with social media, unlessyou have specific privacy
settings on your account so thatonly people that you've
approved can see your messages,that stuff's public like it's
out there.
There's nothing to stop anyonefrom collecting that data at any
(10:32):
time, which that might makesome people uncomfortable, to
which I would say check yourprivacy settings, but for the
most part, there aren't really aton of privacy concerns,
particularly with likeinfovalence, because anybody can
theoretically go and look atthat stuff.
The thing that's challenging,though, is that you need a tool
(10:52):
to collect that data, becauseyou can't just have a person sit
down and scroll through aTwitter feed and pull whatever
tweets they come across that youthink relate to the flu or
COVID or whatever.
I mean you could, but itwouldn't be very effective.
You need an algorithm to gothrough.
It's called data scraping.
You need something to scrapethat data from whatever part of
(11:15):
the internet you're interestedin, but the problem is that
tools for doing that tend tocome and go very quickly, and
that makes it difficult toreplicate or reproduce these
studies, because if you can'tuse the exact same tool for data
scraping that another scholarused when they published their
(11:36):
study.
You're already introducing newelements, and a lot of this
stuff is very black boxedbecause it's proprietary and
companies don't want people toknow how their algorithms work,
so that can make it verychallenging for the
infodemiologist.
Speaker 3 (11:52):
I think that's all really cool and not something
that I knew about, and I thinkit's particularly interesting
that certainly the apps that weuse for social media have
changed greatly and are peopleGoogling things anymore.
Are they just searching them onTikTok or something, and so I
think that the way that youcollect that information
probably has to change with that.
So I could see that would alsointroduce different apps, or
you're going to have differentinformation available on them,
(12:14):
based on what people are doingor what they're interested in
using it.
It is pretty scary to me tothink about that.
Socially, it's like the numberone source of health information
for Americans.
That's really, if you just takea second to think about it,
it's really wild.
And so then you're talkingabout, like, how this data is
collected and how it's scraped,and so that this is just public
information.
It's out there, people can useit, but what then are
(12:37):
infodemiologists looking for?
What are they interested in toscrape from these that to then
draw conclusions from?
Speaker 4 (12:44):
Yeah.
So, like I said before,infodemiology as a field is
pretty much as wide asepidemiology, although I will
say there are certain areas thatare overrepresented, not in a
bad way, just that that's wherea lot of these techniques have
been developed.
So I would say, as far as whatinfodemiologists are doing,
(13:04):
they're doing everything fromtrying to predict disease
outbreaks, like we've discussedbefore, but we're also
interested in things like howare people reacting to health
guidance, right?
Are they having positivereactions?
Are they having negativereactions?
Are they confused?
Trying to study if and thisactually, I think does get into
(13:25):
a slightly more controversialarea of this field but trying to
screen people on social mediafor things like suicide risk.
And then also there's a wholearea now that is interested in I
use the term mis-disinformationbecause it's just faster to say
, I think Looking at, like, howdoes information flow through
social media?
(13:45):
Who is primarily responsiblefor creating disinformation?
Who's magnifying it, right?
So again, we're interested inhuman behavior and human
reactions to this stuff.
And we're also interested incan we take people's behavior
and what they're saying aboutthemselves and what they're
searching for and predict whatsort of diseases or risks they
(14:07):
might be at and that sort ofthing.
So I've looked at somesystematic reviews that have
tried to divide up the field tolook at where who's doing what
research.
What's the bulk of it focusedon.
A lot of it still is focused onflu, which is kind of where it
started, and, of course, duringCOVID that expanded to cover
COVID as well, interested instudying health communications
(14:32):
and trying to use data fromsocial media to determine what's
most effective.
That's, I think, a very smallarea, but one that is definitely
growing and that has receivedmore attention because of the
COVID pandemic.
So, yeah, there's really allkinds of neat things, and
there's definitely a hugeoverlap in infodemiology with
social science, because a lot ofthese things are concerns of
(14:54):
social scientists as well.
Speaker 3 (14:56):
So can we talk a bit about accuracy for any of these
for predicting flu or forpredicting mental health states,
because I could see thatcertainly what a wonderful tool
if you can improve health withthis.
But I could also see this sortof having more of a negative
side if we don't get it rightand so like, then there's
complications if you don't getit right, and that's true for
(15:19):
any science, right?
Not?
Speaker 4 (15:20):
just this.
So could you talk about whatthat looks like?
Yeah, so as far as accuracygoes, I mean, there's a big
cautionary tale here as wellregarding Google flu trends,
which that was sort of anexample of where things fell
apart and didn't work the waythat they were supposed to, but
(15:42):
in terms of accuracy it's prettygood scary good, to be honest.
So this field has really beenin existence for depending on
when you want to say it startedanywhere from like almost 30 to
25 years, so it's stillrelatively new, but we have
enough of a body of literatureat this point to be able to say
that, yeah, we can actually getespecially with things like
respiratory diseases.
We can get it pretty close towhat our traditional
(16:05):
surveillance methods are tellingus.
Those have their problems too,but this stuff has been
validated again and again andit's pretty decent.
I think that where people startto get uncomfortable and the
mental health stuff is a bigarea where this starts to.
For me and a lot of people whowork in the area of, like
(16:25):
psychiatric epidemiology is thatit's not so much about being
accurate as it is.
What do we do once we'veidentified someone who's say a
suicide risk?
Because it's one thing to beable to say okay, this syntax or
these terms appearing in asocial media post, or maybe even
(16:46):
the frequency of posting or whothey're engaging with online.
We know that these might bemarkers for suicidal ideation,
but once you've done that, arethere then resources to connect
that person to?
Can we do something about it?
Because if we can't dosomething about it, it's almost
like an ethical breach.
(17:06):
What do you do with thatinformation?
Is it actually helpful, or isthis purely an academic exercise
and then we can't actually doanything to make the problem
better?
So that's where I think the bigquestions still are, and, of
course, we're now in this agewhere AI is being integrated
with everything and AI is makinginfodemiology even more
(17:27):
accurate, but there are alsosome things that are a little
bit uncomfortable about that aswell.
Speaker 3 (17:32):
So I'm definitely curious to learn more about how
AI is changing infodemiology asfar as accuracy or just like the
sheer amount of informationthat can be scraped, because
that's also something that it'sa whole other line of people
utilizing something and then usbeing able to look at
user-generated data.
But I'm curious with this ideaof connection to resources,
(17:53):
particularly what you're sayingwith mental health.
So I worked as a peer counselorat Cornell University where I
did my undergrad, because therewas just like huge lack of
access to care, and so a studentorganization formed and
actually trained for two years.
Then you were certified as apeer counselor because
undergraduate, there's certainlya large proportion of change
going on and things like that,and so being able to address
peers' mental health concernswas really important.
(18:14):
But I could definitely see whereso you're analyzing this you're
saying like this syntax or thisparticular pattern of
engagement might be moreindicative of something like
suicidal ideation.
I could see there being thisdisconnect of struggling to
connect someone to resources andthen being like how did you
find this out?
Your data was great, and Icould see someone feeling very
violated by that, and so that'ssuch a quandary that I hadn't
(18:36):
considered that like you mightbe able to say you're at risk
for this or that, whether it'sflu or whether it's suicide, but
then how do you connect peopleto resources?
And I think that's always thislike perennial, like wicked
problem in public health.
You can know something but like, how do you help to actually be
part of the solution andresolve?
Speaker 1 (18:53):
it.
Speaker 3 (18:53):
Or is it still just now?
This is coming up in the fieldwhich is relatively new.
Being 25, 30 years old, isthere now this kind of now?
We know this, but how do weactually make the connection?
Speaker 4 (19:03):
Yeah, and as far as I know, I don't think anyone
really has a perfect answer forthat.
To get really dystopian, I knowthat there have been attempts
at having rather than having ahuman person reach out to
someone who seems to be at riskto have an AI chatbot reach out
(19:24):
to them, and I know that therehave been studies looking at
whether therapy can be done.
I think there's a lot ofexcitement right now about
artificial intelligence andabout big data, and I think the
excitement is maybe a little bitpremature in a lot of ways.
I think that there's a lot ofthings that we can do my area of
(19:44):
expertise, although I do workwith some people that are in
that area but I think that theinfodemiology studies that have
(20:05):
been done with, specifically,mental health I believe a
majority of those have been donewith people who knew they were
in the study, so they basicallysigned up to have their content
monitored by a researcher.
Speaker 3 (20:20):
That makes me feel better.
Speaker 4 (20:21):
Yeah, and so I don't know.
I mean, there probably arestudies out there that are just
looking at whatever people puton Facebook or Twitter or
Instagram or whatever.
But yeah, and I think also thisis slightly off topic.
But another thing that getsmore complicated too is when a
platform like TikTok and YouTubeare also big sources of health
(20:43):
information and places wherepeople are connecting with each
other to talk about things likemental health.
But there's more nuance becauseyou've got a video element in
addition to a text element.
And that's again where I thinkAI comes in, because I think
we're going to begin relying onAI more and more to try and
interpret things like visualcues and body language and
(21:04):
things like that.
Because there's such a rich dataenvironment on those platforms,
a lot of things can get lost.
Like sarcasm is notoriouslydifficult for artificial
intelligence to process andunderstand, and the more times
that you scrape data, becausethen it has to be stored
somewhere people are going to beaccessing it.
(21:25):
Every time people access stuff,there's a risk of someone that
you don't want accessing thatstuff getting to it.
So I think that we are likelyto see some sort of like major
event where people's health dataunfortunately gets breached and
leaked and like theconsequences of that could be
pretty far reaching.
So it's something that I'm surewe will be for sure, using that
(21:49):
technology to do those things.
I think it'll be interesting tosee the directions that this
stuff goes in the next likedecade or so.
I've been in some.
Interesting to see thedirections that this stuff goes
in the next like decade or so.
Speaker 3 (21:57):
I've been in some classes where that's been
discussed as like a bioethicsconundrum of, yeah, people are
at risk for us, you can treatthem and prescreen them, but
also maybe their insurance nolonger wants to cover them
because they're at risk, right,I think sometimes actually a lot
of times technology movesfaster than legislation and if
you don't have laws around howsomething is being used, it's
essentially very unregulated andthat's certainly probably
(22:17):
riskier in the long run for usthan having some type of
regulation around it.
But there's also that aspect ofonce you start regulating
things, there's probably lessgrowth on like how far you can
go and do things.
I'm kind of curious to touchmore on so with what you're
talking about, infotemiology,when I think of something like
you're gathering user generateddata and even if you're using
something like AI, the sets it'strained on.
(22:39):
The data sets it's trained onreally are what's super
important for what it's going tocatch for nuance.
Is that being addressed inepidemiology, like the cultural
relevance of someone I grew upin New York?
Someone in New York probably issearching on what happens if I
have the flu and clicking on anad, but there's places where
that's not necessarily happeningright, where, like the aspect
of what you're looking for thecontent you are generating
(23:01):
online or the searches you'regenerating is different
depending on where you're fromand how you have been brought up
using it, or I can even thinkof a generational difference on
how technology is used.
So is that something that alsois being factored into
epidemiology?
Like this, like almost likecultural relevance of like how
good is the data we're getting?
Speaker 4 (23:21):
Yeah, it's tricky because it sort of depends on
your research question, right?
I mean, I think maybe a goodexample of this would be like
the way that people expresstheir symptoms.
That might vary according to,like, culture and age group and
maybe education level.
As far as I know, that is whereI think the human side of this
(23:42):
has not been replaced, right?
Yes, ais can be trained onmultiple different data sets.
They can learn over time, butas far as designing studies,
it's still human beings that aresitting down for the most part
and going.
Okay, what are the ways thatpeople talk about being sick?
Are young people using newslang words to talk about, like
(24:06):
sneezing or coughing, or youknow?
And I think that there's also avery strong awareness among
people in this field that if youare not someone who is, like,
chronically online I forget theexact term that people use for
that but if you're not someonewho's really submerged in
(24:26):
digital spaces, certain areas ofthis field are going to be very
challenging for you to breakinto and to do really good work
in, because you really do haveto know what's happening in
online spaces, and I think a lotof us too are probably like
what you would call lurkers.
Like we're people that sort ofsit in the background and watch
what other people do.
Speaker 3 (24:47):
The people watchers of the internet.
Speaker 4 (24:49):
Yeah, exactly, and I know that's my own position.
I have a lot of social mediaaccounts and I spend a lot of
time on social platforms, butnot necessarily engaging.
I really am just watching andreading and listening, trying to
get some of my own insights.
But, yeah, and I think alsoCamille.
Another thing that questionraises for me is the issue of
(25:10):
whether a sample taken online isever truly representative of
the population, and initially,especially in the early days,
that was a huge concern becausenot everyone had access to the
internet not in the UnitedStates, certainly not in the
world.
But I think that with timepassing and with the internet
becoming this obligatory part oflife that you need to be online
(25:34):
in some capacity, that concernhas decreased a bit because even
in parts of the world with lessinternet coverage, people are
finding ways to get online.
A lot of developing nations aretrying to get online.
People will primarily accessthrough their phones or through
a mobile device of some sort, soI think that aspect of it is
(25:55):
becoming less of a concern withtime.
But also, as with most research, english language content is
vastly overrepresented in thisfield as well, although again,
that's also changing, becauseone thing that AI is really good
at is learning differentlanguages and being able to
process content in differentlanguages.
So there are increasingly morestudies that are not only using
(26:20):
English language content, butalso using other languages as
well, and so I think that we areimproving our ability to
capture a true signal in thatregard, but it's certainly still
a concern signal in that regard, but it's certainly still.
It's still a concern, I wouldsay.
Speaker 3 (26:33):
And so, speaking of kind of the current context and
how that has changed and startedto shape infodemiology,
something that we think about alot in science and health spaces
is this sort of myths anddisinformation really campaigns
that are going on, right?
Some of this is intentional,and so what can infodemiology do
to teach us about what's goingon, right?
Some of this is intentional,and so what can infotemiology do
to teach us about what's goingon, but also to help intervene?
(26:56):
Right, because this is reallyimpacting health outcomes for a
lot of people, and your accessto information and whether or
not that information is valid orperceived as trustworthy to you
matters a lot when it comes toyour health.
Speaker 4 (27:09):
Yeah, this is one of my big interests as well is what
can we do as people or aspublic health professionals, as
epidemiologists, to counteractthis?
And I think one thing is thatwe really need to understand
what we're dealing with betterthan we currently do.
There's a study that came outduring the COVID-19 pandemic
(27:31):
that was really interesting andit was produced by Center for
Countering Digital Hate, that'swho produced it.
But it was this study wherethey basically tracked down who
was responsible for producing amajority of the disinformation
that was going around.
I think it was Twitter specificnot 100% sure, but they found
(27:55):
that there were essentially 12people behind 65% of the
disinformation that wascirculating on the internet.
Wow, yeah.
And also to go back to the fluinfovalence stuff, when
Eisenbach and I think the paperthat I'm thinking of is by
Eisenbach and Chu, but they didsome analysis of
(28:18):
misdisinformation then as well,because that was like where that
started and they found that,although there was a perception
among internet users that therewas like this vast amount of
misinformation going around,when they actually analyzed
content, only like 5% wasflagged as misdisinformation.
(28:40):
So that's not to say that theproblem today isn't much bigger,
because I think it is, and Ithink that, whereas back during
the swine flu H1N1 pandemic, Idon't think there was quite the
same degree of intent, right, Ithink nowadays there are bad
actors, so to speak, that arevery much intentionally cranking
(29:01):
out content that is false andthat is misleading and that
serves to divide people andcause people to turn against
each other.
But I think that we really needto do more research to try and,
first of all, just characterizethe problem, because if that
study that found that 12 peoplewere responsible for the
majority of the content goingaround hadn't been done, it
(29:24):
makes it seem like this problemis really huge and amorphous and
there's not much we can do toget our hands on it.
But actually, if you targetthose 12 people, maybe there is
something that you can do aboutit.
Yeah, I think, unfortunately,we're going in the wrong
direction right now, like withMeta recently announcing that
they're no longer going to dofact checking on their platforms
(29:45):
.
But I do think that we havesome models as far as this goes
for fighting back against it.
I can get into that a littlebit more if you guys are
interested, but it's a littlebit, I would say, beyond the
purview of just specificallyinfodemiology, which to me is
more about characterizing whatthe problem is and how big it is
and what types ofmis-disinformation there are,
(30:09):
and I think from there we canbegin to start developing
interventions and then testingthose interventions to see what
works and what doesn't.
Speaker 3 (30:17):
Yeah, no, I think that's really powerful.
This is really cool.
I knew nothing aboutepidemiology, so this has been
very cool to learn all of it,and so what in particular like
is your project withinepidemiology.
What are you working on?
Speaker 4 (30:30):
What I'm interested in doing and what I would really
like to focus on in the nextfew years is I think I mentioned
earlier that one of my bigconcerns is with everything
that's going on with thetransition to the new
administration and just thegeneral trend over the past five
to 10 years.
I think that public healthagencies as a whole need to
(30:53):
prepare for the fact that theremay be less and less government
support for what they're doing,and that a lot of these
traditional surveillanceactivities are very manpower and
time intensive and they take alot of resources, staff and even
if you're in this is like aproblem in New York State for
our local and city healthdepartments is that they get
(31:16):
tons of data.
There is no shortage of dataout there on every type of
health problem that there is,but there aren't people to
process and analyze that data.
There's a lack of people withthe time and a lack of skill as
well.
There's a lack of people withthe time and a lack of skill as
well, especially at the locallevel, but they're being asked
(31:37):
increasingly to do somethingwith it, and so I really want to
develop tools that areautomated and easy to use, that
would incorporate infodemiologyinto other forms of surveillance
that are already validated andestablished, like, specifically,
wastewater monitoring is one ofthem, because wastewater
monitoring also has its own setof challenges and problems and
(32:01):
quality issues and things likethat.
But I'm interested in findingthese sort of passive sources of
data.
Now, I know wastewatermonitoring isn't passive per se,
but infodemiology can be a verypassive activity if you have
the automation and thealgorithms and the tools set up,
and so I really want to developdata dashboards that these
(32:24):
smaller health departments andhealth agencies and it doesn't
have to only be public health,it could be nonprofits, even
private entities could use thesetools.
But I want to develop thingsthat basically take these
exciting new technologies thatwe have and make them accessible
and put them in the hands ofpeople who can use them for good
.
So that's our goal as anorganization and that's
(32:47):
hopefully what I will be workingon for the next couple of years
as I finish up my PhD.
Speaker 3 (32:53):
That is very cool.
We'll have to say weinterviewed you before you got
famous.
That is really cool.
Speaker 4 (33:00):
Oh gosh, I don't want to be famous, especially not on
the internet, please.
Speaker 3 (33:05):
That's why we don't have video recording yet for our
podcast.
We're not ready to be perceivedWell.
Thank you so much.
This was fantastic.
I feel like our listeners willget a lot out of this,
especially because I think thisis definitely as you're saying.
Things are expensive and wekind of are in this point where
we're able to basically automatemore with data collection and
we have more of an opportunityto do that, I think, with AI,
(33:27):
and the technology is alwaysshifting forward.
So I think that is potentiallythe way things might end up
going for how we surveil fordifferent health conditions, but
also how we decide whereinformation is coming from and
how trustworthy it is, andhelping us figure out what those
connections are.
So thank you so much.
This was an excellent overviewof it.
I feel like I learned so much.
Speaker 4 (33:46):
Yeah, and there was just one other thing that I
wanted to mention or add.
I can't remember how far backin the conversation it was, but
you mentioned something aboutregulation, I think earlier,
when we were talking aboutprivacy concerns and things like
that, and I just wanted tomention that another major
concern that is beginning to beraised by people who work in
(34:07):
this field is that it's not justabout protecting people.
People who work in this fieldis that it's not just about
protecting people.
It's also about ensuring thatwe can, as researchers, actually
do something with the data,because, I mentioned, there are
problems with how quickly toolsbecome defunct and things like
that, but there's also issueswith secrecy from the
(34:29):
organizations that produce thesetools as well and produce the
infrastructure for socialnetworks, because there's a lot
of things that they're notwilling to share that, if they
were willing to share,infodemiologists could really
improve their models and coulddesign tools that would produce
more accurate results and thingslike that.
So I think it's like we tend tothink about regulation in this
(34:52):
context as being reactive orprotective in some way for users
, which is absolutely necessary,don't get me wrong but also I
think we need to be looking atregulation, not just for that,
but also for can we force thesecompanies that use a vast amount
of energy resources in thiscountry?
(35:12):
Can we also require them topackage their data in ways that
would make it useful, becausethere's so much that we can
learn from all of the datathat's out there, but we need to
know more about its provenance,basically, and we need to have
things be more standardized andbe packaged better, and so I
think that's something that wemay start to see more calls from
(35:34):
researchers.
I know this specific issue wasbrought up recently, so it's
like the dialogue is out there,and I think it'll be interesting
to see whether that becomesmore a part of the public
discourse around thesetechnologies or if it's
something that kind of stays aniche concern technologies or if
it's something that kind ofstays a niche concern.
But yeah, I just wanted to saythat about regulation, because I
think that's something thatpeople don't really think about
(35:56):
in those terms.
Speaker 3 (35:57):
Infotemiology our next move as scientists to
collect data.
I love it.
Speaker 4 (36:01):
Yeah, I hope so honestly because I feel like AI,
especially these days, andinfotemiology and AI are
different things, but they'retools that are used together
frequently and I think thatthere's such a negative
perception out there, butthere's really so much good we
can do, especially when we addit to things that we already
know work well.
Speaker 3 (36:21):
All right.
Thank you so much for joiningus for this episode of
Infectious Science For everyonelistening.
Thank you so much.
We hope you enjoyed it.
Leave us a comment, as always.
Let us know if there's anytopics you want to hear about.
Thanks so much.
We hope you enjoyed it.
Leave us a comment, as always.
Let us know if there's anytopics you want to hear about.
Speaker 2 (36:34):
Thanks, so much.
Thanks for listening to theInfectious Science podcast.
Be sure to hit subscribe andvisit infectiousscienceorg to
join the conversation, accessthe show notes and to sign up
for our newsletter and receiveour free materials.
Speaker 1 (36:45):
If you enjoyed this new episode of Infectious
Science, please leave us areview on Apple Podcasts and
Spotify, and go ahead and sharethis episode with some of your
friends.
Speaker 2 (36:54):
Also, don't hesitate to ask questions and tell us
what topics you'd like us tocover for future episodes.
To get in touch, drop a line inthe comments section or send us
a message on social media.
Speaker 1 (37:03):
So we'll see you next time for a new episode, and in
the meantime, stay happy stayhealthy, stay, stay happy, stay
healthy Stay interested.
This is a podcast about One Health the idea that
the health of humans, animals,plants and the environment that
we all share are intrinsicallylinked.
Speaker 2 (00:17):
Coming to you from the University of Texas Medical
Branch and the GalvestonNational Laboratory.
Speaker 1 (00:21):
This is Infectious Science.
Where enthusiasm for science?
Speaker 2 (00:25):
is contagious.
Speaker 3 (00:30):
Everybody, welcome to this episode of Infectious
Science.
We are excited to be back forthis week's episode.
So we're going to be talkingabout infodemiology.
So we are joined by two gueststoday Dr Heather Duncan and Dr
Patrick Murphy.
Heather, would you introduceyourself and tell us how you got
into epidemiology?
Speaker 4 (00:49):
Sure.
So I originally started mycareer as a literary studies
scholar and decided a couple ofyears ago that I was really
interested in public health andwanted to be an epidemiologist.
So I went back to school andgot my master's in public health
and I am now finishing up myfirst year of my PhD in this
(01:10):
field, and I initially had noidea that infodemiology existed.
So I came to the field becauseof my former interest in
studying digital media as ahumanities scholar, and then,
when I realized that was alsohappening in epidemiology, I
became pretty interested in it.
So the work that I'm doing nowis more focused on infodemiology
(01:33):
.
Speaker 3 (01:33):
Very cool, Dr Murphy.
How did you come toinfodemiology?
Speaker 4 (01:37):
Surprisingly through Heather.
I had never heard of it untilshe brought it up and I think
it's fascinating.
It has just so manypossibilities and we're all
about linking things together.
This is a great way to do that,yeah, and Patrick and I co-own a
science and healthcommunications and medical
writing business, so we alsocame at it from that angle as
(01:59):
well, and one of the things thatwe are interested in working on
together is developing usingthe field of infodemiology that
can be hopefully one day in thefuture utilized by smaller
health departments to do butbasically to do more with
surveillance than they'recurrently capable of doing.
Gotcha.
Speaker 3 (02:20):
Very cool, I'd say.
When I first heard ofepidemiology, I immediately
thought of epidemiology and kindof the study of how diseases
are moving through populationsand case numbers, things like
that.
So I know, when I originallywas talking to you about the
episode, I was like oh, isinfodemiology just basically how
misinformation can spread inthe virtual world, just the
pathogens do in reality.
But it's actually broader thanthat and the definition that you
(02:42):
sent back to me was that andjust I want to just make it
clear to our listeners that way,as we go forward, when we're
talking about infodemiology,we're talking about the science
of distribution of determinantsof information, particularly via
the internet, in a populationwith the ultimate aim to inform
public health and public policy.
So can you get into?
Because there's different termswithin it.
(03:02):
Can you get into?
So that's like whatinfodemiology is, but can you
get into, because there'sdifferent terms within it?
Can you get into?
So that's like whatinfodemiology?
Speaker 4 (03:08):
is.
But can you get into what isinfovalence?
Or like, what's an infodemic?
Yeah, that definition that youjust gave.
I just want to give creditbecause he's known as, like the
father of this field, I guess,if you want to use parental
terms, but Gunther Eisenbach isthe one who came up with that
definition in the early 2000sand it's essentially the exact
classical definition ofepidemiology, which is the
(03:29):
distribution and determinants ofdisease with an eye towards
fixing them, basically.
But we're specificallyinterested in digital spaces and
you're right, there's a hugebreadth, I would say basically
everything that epidemiologydoes as well, and actually even
a little bit more.
I think it's just that we areinterested instead of human
(03:52):
beings.
Our unit of study is digitalcontent, basically, and that
could be anything from a tweetto a Google search or a unit of
information essentially, andwithin infodemiology, you
mentioned these two terms info,which is a play on the word
surveillance.
So I'll start with infovalence,because this is where the field
(04:16):
started out.
Well, to give you a little, totell this more as a story,
basically, the earliestinfodemiology studies were
people who were in the mid tolate 90s were starting to notice
that websites were beginning topop up that were providing
health information, and so someof the first questions people
(04:37):
had were like is this goodquality information?
How do we decide whether awebsite is trustworthy?
And nowadays, that's such a, Iguess, a simplistic question
compared to the way our digitallandscape looks.
But that's all it was initiallyso.
Gunther Eisenbach I don't knowexactly how he entered the field
(04:57):
, but he became interested inthese very early studies.
Except he wanted to ask adifferent question.
He wasn't just interested inevaluating the quality of health
websites.
He wanted to know can wesomehow use the internet to do
disease surveillance?
Because there's a lot ofdifferent things that
epidemiologists do, butsurveillance is one of the most
(05:19):
crucial because it's how we knowthat a disease outbreak is
occurring, and especially ifit's something infectious.
Obviously that's one of thecore functions of our public
health departments.
So he wanted to know can we usespecifically Google searches to
find out if there is anoutbreak occurring?
(05:40):
And he and his colleagues had achance to really try this out.
I don't know if you guysremember the swine flu outbreak
of 2009, 2010.
So that was the first time thathe really got a chance to apply
this idea that maybe we couldlook at people's search behavior
and use that to figure out,faster than our sentinel
(06:05):
surveillance networks, whether aflu outbreak is taking hold in
a specific community.
So that's where the idea ofinfovalence was born, and those
early studies by Eisenbach andhis colleagues were pretty
successful.
They validated their dataagainst the CDC's data and they
found that it very closelymatched the predictions that the
(06:28):
CDC was making with their data.
Except it was much more timelyBecause when we're talking about
the internet this stuff doesn'ttake particularly long to
process.
It doesn't have to go throughall of the bureaucracy and pass
from person to person the waythat traditional surveillance
does.
So that was a huge advantage,and at this time also this is
(06:51):
shortly after 9-11, that wasalso when, like the anthrax
attacks took place.
So the Department of Defensewas also really interested in
this technology and insurveillance specifically.
Speaker 3 (07:03):
Out of curiosity.
So I guess people are reallycognizant now, probably more
than we were back in like 2009or 2010, about like how your
information is protected andlike what's private and what's
not, and like.
So it's interesting to me thatis just widely like available
information that you could justconglomerate all of these Google
searches.
And when you're doing that, arepeople Googling like symptoms
(07:25):
of flu or like symptoms of COVID, or is it like looking up news
articles that are related to itin the local community?
What were their kind of searchfeatures, I guess, to determine
yes, there's an outbreak.
Speaker 4 (07:35):
Regarding news, that's almost like a separate
area and I don't know that itwould necessarily fall so much
under infodemiology.
But there are communicationscholars who monitor media and
they use their own algorithms topull headlines and to analyze
data, and infodemiologists doincorporate some of that data,
but there's like a whole otherbranch of study that focuses on
(07:59):
mainstream media.
Yeah, patrick has some thingsto say.
I think about more like modernday concerns about privacy and
particularly like where AI isconcerned as well.
But at this time, the earlystudy, the first study that
Eisenbach did, trying to do thisflu infovalence thing, was
actually really genius, becausenowadays it's very easy I
(08:21):
shouldn't say that it is and itisn't easy to get this data,
because there are challengesthat I can get into a little bit
later, but at the time therewas no Google Trends or there
really weren't tools for doingdata scraping and then putting
it into a usable data set formatthat you could analyze.
So what Eisenbach did which Ithink this is just genius
(08:46):
Eisenbach did, which was I thinkthat this is just genius.
I really admire the way thatthis study was designed.
So what he did was he had apretty small budget.
I want to say it was somewherearound $500 or something, and he
bought a Google ad and when youbuy an ad a digital ad, and I'm
not sure what it's called now,but I think back then it was
called Google AdSense you couldtarget specific geographical
(09:06):
regions.
There were some rudimentarythings you could do to try and
to get to your audience, and sohe set those all very general
and then he used as his metricthe number of clicks.
So it was an advertisement thatsaid something about flu
resources or what to do if youhave the flu, and I think that
(09:26):
when people clicked on it itjust took them to like a WebMD
type site or something like that.
So it was harmless, basically,if people clicked on this ad.
But he used the number ofclicks in different areas and
then was able to again validatethat data against the CDC and
found that actually, yeah, likethe places where people were
(09:48):
clicking were the places whereflu was known to be in high
circulation.
Today it's a little bit, as Isaid, it's like easier to do
things but also more challengingin some ways as far as privacy
concerns.
A lot of infodemiology nowadaysuses social media, because this
is unfortunately or fortunately, I guess, depending on your
(10:10):
perspective, the number onesource of health information for
Americans.
And with social media, unlessyou have specific privacy
settings on your account so thatonly people that you've
approved can see your messages,that stuff's public like it's
out there.
There's nothing to stop anyonefrom collecting that data at any
(10:32):
time, which that might makesome people uncomfortable, to
which I would say check yourprivacy settings, but for the
most part, there aren't really aton of privacy concerns,
particularly with likeinfovalence, because anybody can
theoretically go and look atthat stuff.
The thing that's challenging,though, is that you need a tool
(10:52):
to collect that data, becauseyou can't just have a person sit
down and scroll through aTwitter feed and pull whatever
tweets they come across that youthink relate to the flu or
COVID or whatever.
I mean you could, but itwouldn't be very effective.
You need an algorithm to gothrough.
It's called data scraping.
You need something to scrapethat data from whatever part of
(11:15):
the internet you're interestedin, but the problem is that
tools for doing that tend tocome and go very quickly, and
that makes it difficult toreplicate or reproduce these
studies, because if you can'tuse the exact same tool for data
scraping that another scholarused when they published their
(11:36):
study.
You're already introducing newelements, and a lot of this
stuff is very black boxedbecause it's proprietary and
companies don't want people toknow how their algorithms work,
so that can make it verychallenging for the
infodemiologist.
Speaker 3 (11:52):
I think that's all really cool and not something
that I knew about, and I thinkit's particularly interesting
that certainly the apps that weuse for social media have
changed greatly and are peopleGoogling things anymore.
Are they just searching them onTikTok or something, and so I
think that the way that youcollect that information
probably has to change with that.
So I could see that would alsointroduce different apps, or
you're going to have differentinformation available on them,
(12:14):
based on what people are doingor what they're interested in
using it.
It is pretty scary to me tothink about that.
Socially, it's like the numberone source of health information
for Americans.
That's really, if you just takea second to think about it,
it's really wild.
And so then you're talkingabout, like, how this data is
collected and how it's scraped,and so that this is just public
information.
It's out there, people can useit, but what then are
(12:37):
infodemiologists looking for?
What are they interested in toscrape from these that to then
draw conclusions from?
Speaker 4 (12:44):
Yeah.
So, like I said before,infodemiology as a field is
pretty much as wide asepidemiology, although I will
say there are certain areas thatare overrepresented, not in a
bad way, just that that's wherea lot of these techniques have
been developed.
So I would say, as far as whatinfodemiologists are doing,
(13:04):
they're doing everything fromtrying to predict disease
outbreaks, like we've discussedbefore, but we're also
interested in things like howare people reacting to health
guidance, right?
Are they having positivereactions?
Are they having negativereactions?
Are they confused?
Trying to study if and thisactually, I think does get into
(13:25):
a slightly more controversialarea of this field but trying to
screen people on social mediafor things like suicide risk.
And then also there's a wholearea now that is interested in I
use the term mis-disinformationbecause it's just faster to say
, I think Looking at, like, howdoes information flow through
social media?
(13:45):
Who is primarily responsiblefor creating disinformation?
Who's magnifying it, right?
So again, we're interested inhuman behavior and human
reactions to this stuff.
And we're also interested incan we take people's behavior
and what they're saying aboutthemselves and what they're
searching for and predict whatsort of diseases or risks they
(14:07):
might be at and that sort ofthing.
So I've looked at somesystematic reviews that have
tried to divide up the field tolook at where who's doing what
research.
What's the bulk of it focusedon.
A lot of it still is focused onflu, which is kind of where it
started, and, of course, duringCOVID that expanded to cover
COVID as well, interested instudying health communications
(14:32):
and trying to use data fromsocial media to determine what's
most effective.
That's, I think, a very smallarea, but one that is definitely
growing and that has receivedmore attention because of the
COVID pandemic.
So, yeah, there's really allkinds of neat things, and
there's definitely a hugeoverlap in infodemiology with
social science, because a lot ofthese things are concerns of
(14:54):
social scientists as well.
Speaker 3 (14:56):
So can we talk a bit about accuracy for any of these
for predicting flu or forpredicting mental health states,
because I could see thatcertainly what a wonderful tool
if you can improve health withthis.
But I could also see this sortof having more of a negative
side if we don't get it rightand so like, then there's
complications if you don't getit right, and that's true for
(15:19):
any science, right?
Not?
Speaker 4 (15:20):
just this.
So could you talk about whatthat looks like?
Yeah, so as far as accuracygoes, I mean, there's a big
cautionary tale here as wellregarding Google flu trends,
which that was sort of anexample of where things fell
apart and didn't work the waythat they were supposed to, but
(15:42):
in terms of accuracy it's prettygood scary good, to be honest.
So this field has really beenin existence for depending on
when you want to say it startedanywhere from like almost 30 to
25 years, so it's stillrelatively new, but we have
enough of a body of literatureat this point to be able to say
that, yeah, we can actually getespecially with things like
respiratory diseases.
We can get it pretty close towhat our traditional
(16:05):
surveillance methods are tellingus.
Those have their problems too,but this stuff has been
validated again and again andit's pretty decent.
I think that where people startto get uncomfortable and the
mental health stuff is a bigarea where this starts to.
For me and a lot of people whowork in the area of, like
(16:25):
psychiatric epidemiology is thatit's not so much about being
accurate as it is.
What do we do once we'veidentified someone who's say a
suicide risk?
Because it's one thing to beable to say okay, this syntax or
these terms appearing in asocial media post, or maybe even
(16:46):
the frequency of posting or whothey're engaging with online.
We know that these might bemarkers for suicidal ideation,
but once you've done that, arethere then resources to connect
that person to?
Can we do something about it?
Because if we can't dosomething about it, it's almost
like an ethical breach.
(17:06):
What do you do with thatinformation?
Is it actually helpful, or isthis purely an academic exercise
and then we can't actually doanything to make the problem
better?
So that's where I think the bigquestions still are, and, of
course, we're now in this agewhere AI is being integrated
with everything and AI is makinginfodemiology even more
(17:27):
accurate, but there are alsosome things that are a little
bit uncomfortable about that aswell.
Speaker 3 (17:32):
So I'm definitely curious to learn more about how
AI is changing infodemiology asfar as accuracy or just like the
sheer amount of informationthat can be scraped, because
that's also something that it'sa whole other line of people
utilizing something and then usbeing able to look at
user-generated data.
But I'm curious with this ideaof connection to resources,
(17:53):
particularly what you're sayingwith mental health.
So I worked as a peer counselorat Cornell University where I
did my undergrad, because therewas just like huge lack of
access to care, and so a studentorganization formed and
actually trained for two years.
Then you were certified as apeer counselor because
undergraduate, there's certainlya large proportion of change
going on and things like that,and so being able to address
peers' mental health concernswas really important.
(18:14):
But I could definitely see whereso you're analyzing this you're
saying like this syntax or thisparticular pattern of
engagement might be moreindicative of something like
suicidal ideation.
I could see there being thisdisconnect of struggling to
connect someone to resources andthen being like how did you
find this out?
Your data was great, and Icould see someone feeling very
violated by that, and so that'ssuch a quandary that I hadn't
(18:36):
considered that like you mightbe able to say you're at risk
for this or that, whether it'sflu or whether it's suicide, but
then how do you connect peopleto resources?
And I think that's always thislike perennial, like wicked
problem in public health.
You can know something but like, how do you help to actually be
part of the solution andresolve?
Speaker 1 (18:53):
it.
Speaker 3 (18:53):
Or is it still just now?
This is coming up in the fieldwhich is relatively new.
Being 25, 30 years old, isthere now this kind of now?
We know this, but how do weactually make the connection?
Speaker 4 (19:03):
Yeah, and as far as I know, I don't think anyone
really has a perfect answer forthat.
To get really dystopian, I knowthat there have been attempts
at having rather than having ahuman person reach out to
someone who seems to be at riskto have an AI chatbot reach out
(19:24):
to them, and I know that therehave been studies looking at
whether therapy can be done.
I think there's a lot ofexcitement right now about
artificial intelligence andabout big data, and I think the
excitement is maybe a little bitpremature in a lot of ways.
I think that there's a lot ofthings that we can do my area of
(19:44):
expertise, although I do workwith some people that are in
that area but I think that theinfodemiology studies that have
(20:05):
been done with, specifically,mental health I believe a
majority of those have been donewith people who knew they were
in the study, so they basicallysigned up to have their content
monitored by a researcher.
Speaker 3 (20:20):
That makes me feel better.
Speaker 4 (20:21):
Yeah, and so I don't know.
I mean, there probably arestudies out there that are just
looking at whatever people puton Facebook or Twitter or
Instagram or whatever.
But yeah, and I think also thisis slightly off topic.
But another thing that getsmore complicated too is when a
platform like TikTok and YouTubeare also big sources of health
(20:43):
information and places wherepeople are connecting with each
other to talk about things likemental health.
But there's more nuance becauseyou've got a video element in
addition to a text element.
And that's again where I thinkAI comes in, because I think
we're going to begin relying onAI more and more to try and
interpret things like visualcues and body language and
(21:04):
things like that.
Because there's such a rich dataenvironment on those platforms,
a lot of things can get lost.
Like sarcasm is notoriouslydifficult for artificial
intelligence to process andunderstand, and the more times
that you scrape data, becausethen it has to be stored
somewhere people are going to beaccessing it.
(21:25):
Every time people access stuff,there's a risk of someone that
you don't want accessing thatstuff getting to it.
So I think that we are likelyto see some sort of like major
event where people's health dataunfortunately gets breached and
leaked and like theconsequences of that could be
pretty far reaching.
So it's something that I'm surewe will be for sure, using that
(21:49):
technology to do those things.
I think it'll be interesting tosee the directions that this
stuff goes in the next likedecade or so.
I've been in some.
Interesting to see thedirections that this stuff goes
in the next like decade or so.
Speaker 3 (21:57):
I've been in some classes where that's been
discussed as like a bioethicsconundrum of, yeah, people are
at risk for us, you can treatthem and prescreen them, but
also maybe their insurance nolonger wants to cover them
because they're at risk, right,I think sometimes actually a lot
of times technology movesfaster than legislation and if
you don't have laws around howsomething is being used, it's
essentially very unregulated andthat's certainly probably
(22:17):
riskier in the long run for usthan having some type of
regulation around it.
But there's also that aspect ofonce you start regulating
things, there's probably lessgrowth on like how far you can
go and do things.
I'm kind of curious to touchmore on so with what you're
talking about, infotemiology,when I think of something like
you're gathering user generateddata and even if you're using
something like AI, the sets it'strained on.
(22:39):
The data sets it's trained onreally are what's super
important for what it's going tocatch for nuance.
Is that being addressed inepidemiology, like the cultural
relevance of someone I grew upin New York?
Someone in New York probably issearching on what happens if I
have the flu and clicking on anad, but there's places where
that's not necessarily happeningright, where, like the aspect
of what you're looking for thecontent you are generating
(23:01):
online or the searches you'regenerating is different
depending on where you're fromand how you have been brought up
using it, or I can even thinkof a generational difference on
how technology is used.
So is that something that alsois being factored into
epidemiology?
Like this, like almost likecultural relevance of like how
good is the data we're getting?
Speaker 4 (23:21):
Yeah, it's tricky because it sort of depends on
your research question, right?
I mean, I think maybe a goodexample of this would be like
the way that people expresstheir symptoms.
That might vary according to,like, culture and age group and
maybe education level.
As far as I know, that is whereI think the human side of this
(23:42):
has not been replaced, right?
Yes, ais can be trained onmultiple different data sets.
They can learn over time, butas far as designing studies,
it's still human beings that aresitting down for the most part
and going.
Okay, what are the ways thatpeople talk about being sick?
Are young people using newslang words to talk about, like
(24:06):
sneezing or coughing, or youknow?
And I think that there's also avery strong awareness among
people in this field that if youare not someone who is, like,
chronically online I forget theexact term that people use for
that but if you're not someonewho's really submerged in
(24:26):
digital spaces, certain areas ofthis field are going to be very
challenging for you to breakinto and to do really good work
in, because you really do haveto know what's happening in
online spaces, and I think a lotof us too are probably like
what you would call lurkers.
Like we're people that sort ofsit in the background and watch
what other people do.
Speaker 3 (24:47):
The people watchers of the internet.
Speaker 4 (24:49):
Yeah, exactly, and I know that's my own position.
I have a lot of social mediaaccounts and I spend a lot of
time on social platforms, butnot necessarily engaging.
I really am just watching andreading and listening, trying to
get some of my own insights.
But, yeah, and I think alsoCamille.
Another thing that questionraises for me is the issue of
(25:10):
whether a sample taken online isever truly representative of
the population, and initially,especially in the early days,
that was a huge concern becausenot everyone had access to the
internet not in the UnitedStates, certainly not in the
world.
But I think that with timepassing and with the internet
becoming this obligatory part oflife that you need to be online
(25:34):
in some capacity, that concernhas decreased a bit because even
in parts of the world with lessinternet coverage, people are
finding ways to get online.
A lot of developing nations aretrying to get online.
People will primarily accessthrough their phones or through
a mobile device of some sort, soI think that aspect of it is
(25:55):
becoming less of a concern withtime.
But also, as with most research, english language content is
vastly overrepresented in thisfield as well, although again,
that's also changing, becauseone thing that AI is really good
at is learning differentlanguages and being able to
process content in differentlanguages.
So there are increasingly morestudies that are not only using
(26:20):
English language content, butalso using other languages as
well, and so I think that we areimproving our ability to
capture a true signal in thatregard, but it's certainly still
a concern signal in that regard, but it's certainly still.
It's still a concern, I wouldsay.
Speaker 3 (26:33):
And so, speaking of kind of the current context and
how that has changed and startedto shape infodemiology,
something that we think about alot in science and health spaces
is this sort of myths anddisinformation really campaigns
that are going on, right?
Some of this is intentional,and so what can infodemiology do
to teach us about what's goingon, right?
Some of this is intentional,and so what can infotemiology do
to teach us about what's goingon, but also to help intervene?
(26:56):
Right, because this is reallyimpacting health outcomes for a
lot of people, and your accessto information and whether or
not that information is valid orperceived as trustworthy to you
matters a lot when it comes toyour health.
Speaker 4 (27:09):
Yeah, this is one of my big interests as well is what
can we do as people or aspublic health professionals, as
epidemiologists, to counteractthis?
And I think one thing is thatwe really need to understand
what we're dealing with betterthan we currently do.
There's a study that came outduring the COVID-19 pandemic
(27:31):
that was really interesting andit was produced by Center for
Countering Digital Hate, that'swho produced it.
But it was this study wherethey basically tracked down who
was responsible for producing amajority of the disinformation
that was going around.
I think it was Twitter specificnot 100% sure, but they found
(27:55):
that there were essentially 12people behind 65% of the
disinformation that wascirculating on the internet.
Wow, yeah.
And also to go back to the fluinfovalence stuff, when
Eisenbach and I think the paperthat I'm thinking of is by
Eisenbach and Chu, but they didsome analysis of
(28:18):
misdisinformation then as well,because that was like where that
started and they found that,although there was a perception
among internet users that therewas like this vast amount of
misinformation going around,when they actually analyzed
content, only like 5% wasflagged as misdisinformation.
(28:40):
So that's not to say that theproblem today isn't much bigger,
because I think it is, and Ithink that, whereas back during
the swine flu H1N1 pandemic, Idon't think there was quite the
same degree of intent, right, Ithink nowadays there are bad
actors, so to speak, that arevery much intentionally cranking
(29:01):
out content that is false andthat is misleading and that
serves to divide people andcause people to turn against
each other.
But I think that we really needto do more research to try and,
first of all, just characterizethe problem, because if that
study that found that 12 peoplewere responsible for the
majority of the content goingaround hadn't been done, it
(29:24):
makes it seem like this problemis really huge and amorphous and
there's not much we can do toget our hands on it.
But actually, if you targetthose 12 people, maybe there is
something that you can do aboutit.
Yeah, I think, unfortunately,we're going in the wrong
direction right now, like withMeta recently announcing that
they're no longer going to dofact checking on their platforms
(29:45):
.
But I do think that we havesome models as far as this goes
for fighting back against it.
I can get into that a littlebit more if you guys are
interested, but it's a littlebit, I would say, beyond the
purview of just specificallyinfodemiology, which to me is
more about characterizing whatthe problem is and how big it is
and what types ofmis-disinformation there are,
(30:09):
and I think from there we canbegin to start developing
interventions and then testingthose interventions to see what
works and what doesn't.
Speaker 3 (30:17):
Yeah, no, I think that's really powerful.
This is really cool.
I knew nothing aboutepidemiology, so this has been
very cool to learn all of it,and so what in particular like
is your project withinepidemiology.
What are you working on?
Speaker 4 (30:30):
What I'm interested in doing and what I would really
like to focus on in the nextfew years is I think I mentioned
earlier that one of my bigconcerns is with everything
that's going on with thetransition to the new
administration and just thegeneral trend over the past five
to 10 years.
I think that public healthagencies as a whole need to
(30:53):
prepare for the fact that theremay be less and less government
support for what they're doing,and that a lot of these
traditional surveillanceactivities are very manpower and
time intensive and they take alot of resources, staff and even
if you're in this is like aproblem in New York State for
our local and city healthdepartments is that they get
(31:16):
tons of data.
There is no shortage of dataout there on every type of
health problem that there is,but there aren't people to
process and analyze that data.
There's a lack of people withthe time and a lack of skill as
well.
There's a lack of people withthe time and a lack of skill as
well, especially at the locallevel, but they're being asked
(31:37):
increasingly to do somethingwith it, and so I really want to
develop tools that areautomated and easy to use, that
would incorporate infodemiologyinto other forms of surveillance
that are already validated andestablished, like, specifically,
wastewater monitoring is one ofthem, because wastewater
monitoring also has its own setof challenges and problems and
(32:01):
quality issues and things likethat.
But I'm interested in findingthese sort of passive sources of
data.
Now, I know wastewatermonitoring isn't passive per se,
but infodemiology can be a verypassive activity if you have
the automation and thealgorithms and the tools set up,
and so I really want to developdata dashboards that these
(32:24):
smaller health departments andhealth agencies and it doesn't
have to only be public health,it could be nonprofits, even
private entities could use thesetools.
But I want to develop thingsthat basically take these
exciting new technologies thatwe have and make them accessible
and put them in the hands ofpeople who can use them for good
.
So that's our goal as anorganization and that's
(32:47):
hopefully what I will be workingon for the next couple of years
as I finish up my PhD.
Speaker 3 (32:53):
That is very cool.
We'll have to say weinterviewed you before you got
famous.
That is really cool.
Speaker 4 (33:00):
Oh gosh, I don't want to be famous, especially not on
the internet, please.
Speaker 3 (33:05):
That's why we don't have video recording yet for our
podcast.
We're not ready to be perceivedWell.
Thank you so much.
This was fantastic.
I feel like our listeners willget a lot out of this,
especially because I think thisis definitely as you're saying.
Things are expensive and wekind of are in this point where
we're able to basically automatemore with data collection and
we have more of an opportunityto do that, I think, with AI,
(33:27):
and the technology is alwaysshifting forward.
So I think that is potentiallythe way things might end up
going for how we surveil fordifferent health conditions, but
also how we decide whereinformation is coming from and
how trustworthy it is, andhelping us figure out what those
connections are.
So thank you so much.
This was an excellent overviewof it.
I feel like I learned so much.
Speaker 4 (33:46):
Yeah, and there was just one other thing that I
wanted to mention or add.
I can't remember how far backin the conversation it was, but
you mentioned something aboutregulation, I think earlier,
when we were talking aboutprivacy concerns and things like
that, and I just wanted tomention that another major
concern that is beginning to beraised by people who work in
(34:07):
this field is that it's not justabout protecting people.
People who work in this fieldis that it's not just about
protecting people.
It's also about ensuring thatwe can, as researchers, actually
do something with the data,because, I mentioned, there are
problems with how quickly toolsbecome defunct and things like
that, but there's also issueswith secrecy from the
(34:29):
organizations that produce thesetools as well and produce the
infrastructure for socialnetworks, because there's a lot
of things that they're notwilling to share that, if they
were willing to share,infodemiologists could really
improve their models and coulddesign tools that would produce
more accurate results and thingslike that.
So I think it's like we tend tothink about regulation in this
(34:52):
context as being reactive orprotective in some way for users
, which is absolutely necessary,don't get me wrong but also I
think we need to be looking atregulation, not just for that,
but also for can we force thesecompanies that use a vast amount
of energy resources in thiscountry?
(35:12):
Can we also require them topackage their data in ways that
would make it useful, becausethere's so much that we can
learn from all of the datathat's out there, but we need to
know more about its provenance,basically, and we need to have
things be more standardized andbe packaged better, and so I
think that's something that wemay start to see more calls from
(35:34):
researchers.
I know this specific issue wasbrought up recently, so it's
like the dialogue is out there,and I think it'll be interesting
to see whether that becomesmore a part of the public
discourse around thesetechnologies or if it's
something that kind of stays aniche concern technologies or if
it's something that kind ofstays a niche concern.
But yeah, I just wanted to saythat about regulation, because I
think that's something thatpeople don't really think about
(35:56):
in those terms.
Speaker 3 (35:57):
Infotemiology our next move as scientists to
collect data.
I love it.
Speaker 4 (36:01):
Yeah, I hope so honestly because I feel like AI,
especially these days, andinfotemiology and AI are
different things, but they'retools that are used together
frequently and I think thatthere's such a negative
perception out there, butthere's really so much good we
can do, especially when we addit to things that we already
know work well.
Speaker 3 (36:21):
All right.
Thank you so much for joiningus for this episode of
Infectious Science For everyonelistening.
Thank you so much.
We hope you enjoyed it.
Leave us a comment, as always.
Let us know if there's anytopics you want to hear about.
Thanks so much.
We hope you enjoyed it.
Leave us a comment, as always.
Let us know if there's anytopics you want to hear about.
Speaker 2 (36:34):
Thanks, so much.
Thanks for listening to theInfectious Science podcast.
Be sure to hit subscribe andvisit infectiousscienceorg to
join the conversation, accessthe show notes and to sign up
for our newsletter and receiveour free materials.
Speaker 1 (36:45):
If you enjoyed this new episode of Infectious
Science, please leave us areview on Apple Podcasts and
Spotify, and go ahead and sharethis episode with some of your
friends.
Speaker 2 (36:54):
Also, don't hesitate to ask questions and tell us
what topics you'd like us tocover for future episodes.
To get in touch, drop a line inthe comments section or send us
a message on social media.
Speaker 1 (37:03):
So we'll see you next time for a new episode, and in
the meantime, stay happy stayhealthy, stay, stay happy, stay
healthy Stay interested.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.