The surprising seeds of a big-data revolution in healthcare | Joel Selanikio | TEDxAustin
-
0:18 - 0:20There's an old joke about a cop
-
0:20 - 0:22who's walking his beat
in the middle of the night, -
0:22 - 0:25and he comes across a guy
under a street lamp -
0:25 - 0:28who's looking at the ground
and moving from side to side, -
0:28 - 0:29and the cop asks him what he's doing.
-
0:29 - 0:31The guys says he's looking for his keys.
-
0:31 - 0:33So the cop takes his time
-
0:33 - 0:35and looks over and kind of
makes a little matrix -
0:35 - 0:37and looks for about two, three minutes.
-
0:37 - 0:38No keys.
-
0:38 - 0:40The cop says, "Are you sure?
-
0:40 - 0:43Hey buddy, are you sure
you lost your keys here?" -
0:43 - 0:45And the guy says,
"No, actually I lost them -
0:45 - 0:47down at the other end of the street,
-
0:47 - 0:48but the light is better here."
-
0:48 - 0:51(Laughter)
-
0:52 - 0:55There's a concept that people talk
about nowadays called "big data." -
0:55 - 0:58And what they're talking
about is all of the information -
0:58 - 1:00that we're generating
through our interaction -
1:00 - 1:01with and over the Internet,
-
1:01 - 1:03everything from Facebook and Twitter
-
1:03 - 1:07to music downloads, movies,
streaming, all this kind of stuff, -
1:07 - 1:08the live streaming of TED.
-
1:08 - 1:11And the folks who work
with big data, for them, -
1:11 - 1:13they talk about that their biggest problem
-
1:13 - 1:15is we have so much information.
-
1:15 - 1:18The biggest problem is: how do we
organize all that information? -
1:19 - 1:21I can tell you that,
working in global health, -
1:21 - 1:23that is not our biggest problem.
-
1:24 - 1:27Because for us, even though
the light is better on the Internet, -
1:28 - 1:32the data that would help us solve
the problems we're trying to solve -
1:32 - 1:34is not actually present on the Internet.
-
1:34 - 1:36So we don't know, for example,
-
1:36 - 1:39how many people right now
are being affected by disasters -
1:39 - 1:41or by conflict situations.
-
1:41 - 1:45We don't know for, really,
basically, any of the clinics -
1:45 - 1:46in the developing world,
-
1:46 - 1:48which ones have medicines
and which ones don't. -
1:48 - 1:52We have no idea of what
the supply chain is for those clinics. -
1:52 - 1:55We don't know -- and this is really
amazing to me -- we don't know -
1:55 - 1:59how many children were born --
or how many children there are -- -
1:59 - 2:02in Bolivia or Botswana or Bhutan.
-
2:04 - 2:06We don't know how many kids died last week
-
2:06 - 2:07in any of those countries.
-
2:07 - 2:10We don't know the needs
of the elderly, the mentally ill. -
2:10 - 2:13For all of these different
critically important problems -
2:13 - 2:16or critically important areas
that we want to solve problems in, -
2:16 - 2:19we basically know nothing at all.
-
2:21 - 2:24And part of the reason
why we don't know anything at all -
2:24 - 2:28is that the information technology systems
that we use in global health -
2:28 - 2:33to find the data to solve
these problems is what you see here. -
2:33 - 2:35This is about a 5,000-year-old technology.
-
2:35 - 2:37Some of you may have used it before.
-
2:37 - 2:38It's kind of on its way out now,
-
2:38 - 2:41but we still use it
for 99 percent of our stuff. -
2:41 - 2:43This is a paper form.
-
2:44 - 2:46And what you're looking at is a paper form
-
2:46 - 2:49in the hand of a Ministry of Health
nurse in Indonesia, -
2:49 - 2:51who is tramping out across the countryside
-
2:51 - 2:54in Indonesia on, I'm sure,
a very hot and humid day, -
2:54 - 2:58and she is going to be knocking
on thousands of doors -
2:58 - 2:59over a period of weeks or months,
-
2:59 - 3:01knocking on the doors and saying,
-
3:01 - 3:04"Excuse me, we'd like to ask
you some questions. -
3:04 - 3:08Do you have any children?
Were your children vaccinated?" -
3:08 - 3:10Because the only way
we can actually find out -
3:10 - 3:13how many children were vaccinated
in the country of Indonesia, -
3:13 - 3:14what percentage were vaccinated,
-
3:14 - 3:18is actually not on the Internet,
but by going out and knocking on doors, -
3:18 - 3:21sometimes tens of thousands of doors.
-
3:21 - 3:24Sometimes it takes months to even years
to do something like this. -
3:24 - 3:27You know, a census of Indonesia
-
3:27 - 3:29would probably take
two years to accomplish. -
3:29 - 3:31And the problem, of course,
with all of this -
3:31 - 3:33is that, with all those paper forms --
-
3:33 - 3:36and I'm telling you, we have
paper forms for every possible thing: -
3:36 - 3:38We have paper forms
for vaccination surveys. -
3:38 - 3:41We have paper forms to track
people who come into clinics. -
3:41 - 3:46We have paper forms to track
drug supplies, blood supplies -- -
3:46 - 3:49all these different paper forms
for many different topics, -
3:49 - 3:51they all have a single, common endpoint,
-
3:51 - 3:54and the common endpoint
looks something like this. -
3:54 - 3:57And what we're looking
at here is a truckful of data. -
3:59 - 4:02This is the data from a single
vaccination coverage survey -
4:02 - 4:05in a single district
in the country of Zambia -
4:05 - 4:07from a few years ago,
that I participated in. -
4:07 - 4:10The only thing anyone
was trying to find out -
4:10 - 4:13is what percentage of Zambian
children are vaccinated, -
4:13 - 4:16and this is the data,
collected on paper over weeks, -
4:16 - 4:17from a single district,
-
4:17 - 4:20which is something like a county
in the United States. -
4:20 - 4:23You can imagine that,
for the entire country of Zambia, -
4:23 - 4:25answering just that single question ...
-
4:26 - 4:28looks something like this.
-
4:28 - 4:31Truck after truck after truck,
-
4:31 - 4:34filled with stack after stack
after stack of data. -
4:34 - 4:37And what makes it even worse
is that's just the beginning. -
4:37 - 4:39Because once you've collected
all that data, -
4:39 - 4:42of course, someone --
some unfortunate person -- -
4:42 - 4:44is going to have to type that
into a computer. -
4:44 - 4:46When I was a graduate student,
-
4:46 - 4:48I actually was that unfortunate
person sometimes. -
4:48 - 4:51I can tell you, I often wasn't
really paying attention. -
4:51 - 4:53I probably made a lot
of mistakes when I did it -
4:53 - 4:56that no one ever discovered,
so data quality goes down. -
4:56 - 4:59But eventually that data, hopefully,
gets typed into a computer, -
4:59 - 5:01and someone can begin to analyze it,
-
5:01 - 5:03and once they have
an analysis and a report, -
5:03 - 5:06hopefully, then you can take
the results of that data collection -
5:06 - 5:09and use it to vaccinate children better.
-
5:09 - 5:14Because if there's anything worse
in the field of global public health -- -
5:14 - 5:17I don't know what's worse
than allowing children on this planet -
5:17 - 5:20to die of vaccine-preventable diseases --
-
5:20 - 5:23diseases for which
the vaccine costs a dollar. -
5:23 - 5:27And millions of children die
of these diseases every year. -
5:27 - 5:30And the fact is, millions
is a gross estimate, -
5:30 - 5:33because we don't really know
how many kids die each year of this. -
5:33 - 5:37What makes it even more frustrating
is that the data-entry part, -
5:37 - 5:39the part that I used to do
as a grad student, -
5:39 - 5:41can take sometimes six months.
-
5:41 - 5:44Sometimes it can take two years
to type that information into a computer, -
5:44 - 5:46And sometimes, actually not infrequently,
-
5:46 - 5:48it actually never happens.
-
5:48 - 5:50Now try and wrap your head
around that for a second. -
5:50 - 5:53You just had teams of hundreds of people.
-
5:53 - 5:56They went out into the field
to answer a particular question. -
5:56 - 5:58You probably spent hundreds
of thousands of dollars -
5:58 - 6:01on fuel and photocopying and per diem.
-
6:02 - 6:04And then for some reason, momentum is lost
-
6:04 - 6:05or there's no money left,
-
6:05 - 6:07and all of that comes to nothing,
-
6:07 - 6:10because no one actually types it
into the computer at all. -
6:10 - 6:12The process just stops.
-
6:12 - 6:13Happens all the time.
-
6:13 - 6:16This is what we base
our decisions on in global health: -
6:16 - 6:19little data, old data, no data.
-
6:21 - 6:23So back in 1995,
-
6:23 - 6:26I began to think about ways
in which we could improve this process. -
6:26 - 6:29Now 1995 -- obviously,
that was quite a long time ago. -
6:29 - 6:32It kind of frightens me to think
of how long ago that was. -
6:32 - 6:35The top movie of the year
was "Die Hard with a Vengeance." -
6:35 - 6:38As you can see, Bruce Willis
had a lot more hair back then. -
6:38 - 6:40I was working in the Centers
for Disease Control -
6:40 - 6:42and I had a lot more
hair back then as well. -
6:43 - 6:46But to me, the most significant
thing that I saw in 1995 -
6:46 - 6:47was this.
-
6:47 - 6:50Hard for us to imagine, but in 1995,
-
6:50 - 6:53this was the ultimate elite mobile device.
-
6:54 - 6:56It wasn't an iPhone.
It wasn't a Galaxy phone. -
6:56 - 6:58It was a PalmPilot.
-
6:58 - 7:01And when I saw the PalmPilot
for the first time, I thought, -
7:01 - 7:04"Why can't we put the forms
on these PalmPilots? -
7:04 - 7:07And go out into the field
just carrying one PalmPilot, -
7:07 - 7:11which can hold the capacity
of tens of thousands of paper forms? -
7:11 - 7:12Why don't we try to do that?
-
7:12 - 7:14Because if we can do that,
-
7:14 - 7:17if we can actually just collect
the data electronically, digitally, -
7:17 - 7:19from the very beginning,
-
7:19 - 7:22we can just put a shortcut right
through that whole process -
7:24 - 7:28of typing, of having somebody type
that stuff into the computer. -
7:28 - 7:29We can skip straight to the analysis
-
7:29 - 7:33and then straight to the use
of the data to actually save lives." -
7:33 - 7:35So that's what I began to do.
-
7:35 - 7:39Working at CDC, I began to travel
to different programs around the world -
7:39 - 7:43and to train them in using
PalmPilots to do data collection, -
7:43 - 7:45instead of using paper.
-
7:45 - 7:47And it actually worked great.
-
7:47 - 7:49It worked exactly as well
as anybody would have predicted. -
7:49 - 7:50What do you know?
-
7:50 - 7:54Digital data collection is actually
more efficient than collecting on paper. -
7:54 - 7:56While I was doing it, my business partner,
-
7:56 - 7:59Rose, who's here with her husband,
Matthew, here in the audience, -
7:59 - 8:02Rose was out doing similar stuff
for the American Red Cross. -
8:02 - 8:05The problem was,
after a few years of doing that, -
8:05 - 8:09I realized -- I had been to maybe
six or seven programs -- -
8:09 - 8:12and I thought, you know,
if I keep this up at this pace, -
8:12 - 8:13over my whole career,
-
8:13 - 8:16maybe I'm going to go
to maybe 20 or 30 programs. -
8:16 - 8:19But the problem is, 20 or 30 programs,
-
8:19 - 8:22like, training 20 or 30 programs
to use this technology, -
8:22 - 8:24that is a tiny drop in the bucket.
-
8:24 - 8:28The demand for this, the need
for data to run better programs -
8:28 - 8:31just within health -- not to mention
all of the other fields -
8:31 - 8:34in developing countries -- is enormous.
-
8:34 - 8:38There are millions and millions
and millions of programs, -
8:38 - 8:40millions of clinics
that need to track drugs, -
8:40 - 8:41millions of vaccine programs.
-
8:41 - 8:44There are schools
that need to track attendance. -
8:44 - 8:48There are all these different things
for us to get the data that we need to do. -
8:48 - 8:51And I realized if I kept up
the way that I was doing, -
8:52 - 8:55I was basically hardly
going to make any impact -
8:55 - 8:57by the end of my career.
-
8:57 - 8:59And so I began to rack my brain,
-
8:59 - 9:02trying to think about, what
was the process that I was doing? -
9:02 - 9:04How was I training folks,
-
9:04 - 9:06and what were the bottlenecks
and what were the obstacles -
9:06 - 9:09to doing it faster
and to doing it more efficiently? -
9:09 - 9:12And, unfortunately, after thinking
about this for some time, -
9:14 - 9:16I identified the main obstacle.
-
9:16 - 9:18And the main obstacle, it turned out --
-
9:18 - 9:20and this is a sad realization --
-
9:20 - 9:22the main obstacle was me.
-
9:22 - 9:24So what do I mean by that?
-
9:24 - 9:25I had developed a process
-
9:25 - 9:30whereby I was the center
of the universe of this technology. -
9:32 - 9:35If you wanted to use this technology,
you had to get in touch with me. -
9:35 - 9:37That means you had to know I existed.
-
9:37 - 9:40Then you had to find the money
to pay for me to fly out to your country -
9:40 - 9:43and the money to pay for my hotel
and my per diem and my daily rate. -
9:44 - 9:47So you could be talking
about 10- or 20- or 30,000 dollars, -
9:47 - 9:49if I actually had the time
or it fit my schedule -
9:49 - 9:51and I wasn't on vacation.
-
9:51 - 9:53The point is that anything,
-
9:53 - 9:56any system that depends
on a single human being -
9:56 - 9:57or two or three or five human beings --
-
9:57 - 9:59it just doesn't scale.
-
9:59 - 10:02And this is a problem for which
we need to scale this technology, -
10:02 - 10:04and we need to scale it now.
-
10:04 - 10:07And so I began to think of ways
in which I could basically -
10:07 - 10:08take myself out of the picture.
-
10:11 - 10:13And, you know, I was thinking,
-
10:13 - 10:15"How could I take myself
out of the picture?" -
10:15 - 10:17for quite some time.
-
10:17 - 10:21I'd been trained that the way
you distribute technology -
10:21 - 10:22within international development
-
10:22 - 10:24is always consultant-based.
-
10:24 - 10:26It's always guys
that look pretty much like me, -
10:26 - 10:29flying from countries
that look pretty much like this -
10:29 - 10:32to other countries
with people with darker skin. -
10:32 - 10:35And you go out there,
and you spend money on airfare -
10:35 - 10:38and you spend time and you spend per diem
-
10:38 - 10:40and you spend for a hotel
and all that stuff. -
10:40 - 10:44As far as I knew, that was the only way
you could distribute technology, -
10:44 - 10:46and I couldn't figure out a way around it.
-
10:46 - 10:48But the miracle that happened --
-
10:48 - 10:51I'm going to call it Hotmail for short.
-
10:51 - 10:53You may not think of Hotmail
as being miraculous, -
10:53 - 10:55but for me it was miraculous,
-
10:55 - 10:58because I noticed, just as I
was wrestling with this problem -- -
10:59 - 11:02I was working in sub-Saharan
Africa, mostly, at the time -- -
11:02 - 11:04I noticed that every sub-Saharan
African health worker -
11:04 - 11:08that I was working with
had a Hotmail account. -
11:10 - 11:12And it struck me, "Wait a minute --
-
11:12 - 11:16I know the Hotmail people surely didn't
fly to the Ministry of Health in Kenya -
11:16 - 11:18to train people in how to use Hotmail.
-
11:19 - 11:23So these guys are distributing technology,
getting software capacity out there, -
11:23 - 11:25but they're not actually
flying around the world. -
11:25 - 11:27I need to think about this more."
-
11:27 - 11:29While I was thinking about it,
-
11:29 - 11:32people started using even more
things like this, just as we were. -
11:32 - 11:35They started using LinkedIn and Flickr
and Gmail and Google Maps -- -
11:35 - 11:36all these things.
-
11:36 - 11:39Of course, all of these things
are cloud based -
11:39 - 11:41and don't require any training.
-
11:41 - 11:43They don't require any programmers.
-
11:43 - 11:44They don't require consultants.
-
11:44 - 11:47Because the business model
for all these businesses -
11:47 - 11:50requires that something be so simple
we can use it ourselves, -
11:50 - 11:51with little or no training.
-
11:51 - 11:54You just have to hear about it
and go to the website. -
11:54 - 11:57And so I thought, what would happen
if we built software -
11:58 - 12:00to do what I'd been consulting in?
-
12:00 - 12:04Instead of training people
how to put forms onto mobile devices, -
12:04 - 12:07let's create software that lets them
do it themselves with no training -
12:07 - 12:09and without me being involved.
-
12:09 - 12:10And that's exactly what we did.
-
12:10 - 12:16So we created software called Magpi,
which has an online form creator. -
12:16 - 12:17No one has to speak to me,
-
12:17 - 12:20you just have to hear about it
and go to the website. -
12:20 - 12:22You can create forms,
and once you've created the forms, -
12:22 - 12:25you push them to a variety
of common mobile phones. -
12:25 - 12:28Obviously, nowadays, we've moved
past PalmPilots to mobile phones. -
12:28 - 12:31And it doesn't have to be a smartphone,
it can be a basic phone, -
12:31 - 12:34like the phone on the right,
the basic Symbian phone -
12:34 - 12:36that's very common
in developing countries. -
12:36 - 12:40And the great part about this
is it's just like Hotmail. -
12:40 - 12:41It's cloud based,
-
12:41 - 12:44and it doesn't require any training,
programming, consultants. -
12:44 - 12:46But there are some
additional benefits as well. -
12:46 - 12:49Now we knew when we built this system,
-
12:49 - 12:51the whole point of it,
just like with the PalmPilots, -
12:51 - 12:54was that you'd be able to collect the data
-
12:54 - 12:56and immediately upload
the data and get your data set. -
12:56 - 12:59But what we found, of course,
since it's already on a computer, -
12:59 - 13:02we can deliver instant maps
and analysis and graphing. -
13:02 - 13:04We can take a process that took two years
-
13:04 - 13:08and compress that
down to the space of five minutes. -
13:08 - 13:10Unbelievable improvements in efficiency.
-
13:10 - 13:14Cloud based, no training,
no consultants, no me. -
13:15 - 13:17And I told you that in the first few years
-
13:17 - 13:19of trying to do this
the old-fashioned way, -
13:19 - 13:21going out to each country,
-
13:23 - 13:26we probably trained about 1,000 people.
-
13:26 - 13:27What happened after we did this?
-
13:27 - 13:29In the second three years,
-
13:29 - 13:31we had 14,000 people find the website,
-
13:31 - 13:33sign up and start using it
to collect data: -
13:33 - 13:35data for disaster response,
-
13:35 - 13:38Canadian pig farmers
tracking pig disease and pig herds, -
13:39 - 13:41people tracking drug supplies.
-
13:42 - 13:45One of my favorite examples, the IRC,
International Rescue Committee, -
13:45 - 13:48they have a program
where semi-literate midwives, -
13:48 - 13:51using $10 mobile phones,
-
13:51 - 13:54send a text message
using our software, once a week, -
13:54 - 13:57with the number of births
and the number of deaths, -
13:57 - 14:00which gives IRC something that no one
in global health has ever had: -
14:00 - 14:04a near-real-time system
of counting babies, -
14:04 - 14:05of knowing how many kids are born,
-
14:05 - 14:08of knowing how many children
there are in Sierra Leone, -
14:08 - 14:10which is the country
where this is happening, -
14:10 - 14:12and knowing how many children die.
-
14:13 - 14:15Physicians for Human Rights --
-
14:15 - 14:17this is moving a little bit
outside the health field -- -
14:17 - 14:22they're basically training people
to do rape exams in Congo, -
14:22 - 14:24where this is an epidemic,
-
14:24 - 14:25a horrible epidemic,
-
14:25 - 14:28and they're using our software
to document the evidence they find, -
14:28 - 14:31including photographically,
-
14:31 - 14:33so that they can bring
the perpetrators to justice. -
14:34 - 14:38Camfed, another charity
based out of the UK -- -
14:38 - 14:41Camfed pays girls' families
to keep them in school. -
14:41 - 14:45They understand this is the most
significant intervention they can make. -
14:45 - 14:49They used to track the disbursements,
the attendance, the grades, on paper. -
14:49 - 14:52The turnaround time between a teacher
writing down grades or attendance -
14:52 - 14:55and getting that into a report
was about two to three years. -
14:55 - 14:56Now it's real time.
-
14:56 - 14:59And because this is such a low-cost
system and based in the cloud, -
14:59 - 15:04it costs, for the entire five countries
that Camfed runs this in, -
15:04 - 15:05with tens of thousands of girls,
-
15:05 - 15:09the whole cost combined
is 10,000 dollars a year. -
15:09 - 15:11That's less than I used to get
-
15:11 - 15:14just traveling out for two weeks
to do a consultation. -
15:15 - 15:19So I told you before that when
we were doing it the old-fashioned way, -
15:19 - 15:23I realized all of our work was really
adding up to just a drop in the bucket -- -
15:23 - 15:2510, 20, 30 different programs.
-
15:25 - 15:27We've made a lot of progress,
-
15:27 - 15:29but I recognize that right now,
-
15:29 - 15:31even the work that we've done
with 14,000 people using this -
15:31 - 15:33is still a drop in the bucket.
-
15:33 - 15:36But something's changed,
and I think it should be obvious. -
15:36 - 15:38What's changed now is,
-
15:38 - 15:42instead of having a program
in which we're scaling at such a slow rate -
15:42 - 15:45that we can never reach
all the people who need us, -
15:45 - 15:48we've made it unnecessary
for people to get reached by us. -
15:48 - 15:53We've created a tool
that lets programs keep kids in school, -
15:53 - 15:57track the number of babies that are born
and the number of babies that die, -
15:57 - 16:01catch criminals and successfully
prosecute them -- -
16:01 - 16:05to do all these different things
to learn more about what's going on, -
16:06 - 16:07to understand more,
-
16:07 - 16:08to see more ...
-
16:10 - 16:12and to save lives and improve lives.
-
16:14 - 16:15Thank you.
-
16:15 - 16:17(Applause)
- Title:
- The surprising seeds of a big-data revolution in healthcare | Joel Selanikio | TEDxAustin
- Description:
-
Collecting global health data was an imperfect science: Workers tramped through villages to knock on doors and ask questions, wrote the answers on paper forms, then input the data -- and from this gappy information, countries would make huge decisions. Data geek Joel Selanikio talks through the sea change in collecting health data in the past decade -- starting with the Palm Pilot and Hotmail, and now moving into the cloud.
This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at http://ted.com/tedx
- Video Language:
- English
- Team:
- closed TED
- Project:
- TEDxTalks
- Duration:
- 16:22