The surprising seeds of a big-data revolution in healthcare
-
0:00 - 0:03There's an old joke about a cop who's walking his beat
-
0:03 - 0:04in the middle of the night,
-
0:04 - 0:06and he comes across a guy under a street lamp
-
0:06 - 0:09who's looking at the ground and moving from side to side,
-
0:09 - 0:11and the cop asks him what he's doing.
-
0:11 - 0:13The guys says he's looking for his keys.
-
0:13 - 0:15So the cop takes his time and looks over
-
0:15 - 0:17and kind of makes a little matrix and looks
-
0:17 - 0:20for about two, three minutes. No keys.
-
0:20 - 0:23The cop says, "Are you sure? Hey buddy,
-
0:23 - 0:25are you sure you lost your keys here?"
-
0:25 - 0:26And the guy says, "No, no, actually I lost them
-
0:26 - 0:28down at the other end of the street,
-
0:28 - 0:34but the light is better here."
-
0:34 - 0:35There's a concept that people talk about nowadays
-
0:35 - 0:38called big data, and what they're talking about
-
0:38 - 0:40is all of the information that we're generating
-
0:40 - 0:43through our interaction with and over the Internet,
-
0:43 - 0:44everything from Facebook and Twitter
-
0:44 - 0:49to music downloads, movies, streaming, all this kind of stuff,
-
0:49 - 0:50the live streaming of TED.
-
0:50 - 0:53And the folks who work with big data, for them,
-
0:53 - 0:55they talk about that their biggest problem is
-
0:55 - 0:57we have so much information,
-
0:57 - 1:00the biggest problem is, how do we organize all that information?
-
1:00 - 1:03I can tell you that working in global health,
-
1:03 - 1:06that is not our biggest problem.
-
1:06 - 1:07Because for us, even though the light
-
1:07 - 1:10is better on the Internet,
-
1:10 - 1:13the data that would help us solve the problems
-
1:13 - 1:16we're trying to solve is not actually present on the Internet.
-
1:16 - 1:18So we don't know, for example, how many people
-
1:18 - 1:20right now are being affected by disasters
-
1:20 - 1:23or by conflict situations.
-
1:23 - 1:27We don't know for really basically any of the clinics
-
1:27 - 1:29in the developing world, which ones have medicines
-
1:29 - 1:30and which ones don't.
-
1:30 - 1:33We have no idea of what the supply chain is for those clinics.
-
1:33 - 1:36We don't know -- and this is really amazing to me --
-
1:36 - 1:39we don't know how many children were born,
-
1:39 - 1:42or how many children there are in Bolivia
-
1:42 - 1:45or Botswana or Bhutan.
-
1:45 - 1:47We don't know how many kids died last week
-
1:47 - 1:49in any of those countries.
-
1:49 - 1:52We don't know the needs of the elderly, the mentally ill.
-
1:52 - 1:55For all of these different critically important problems
-
1:55 - 1:58or critically important areas that we want to solve problems in,
-
1:58 - 2:03we basically know nothing at all.
-
2:03 - 2:06And part of the reason why we don't know anything at all
-
2:06 - 2:08is that the information technology systems
-
2:08 - 2:12that we use in global health to find the data
-
2:12 - 2:15to solve these problems is what you see here.
-
2:15 - 2:17And this is about a 5,000-year-old technology.
-
2:17 - 2:18Some of you may have used it before.
-
2:18 - 2:20It's kind of on its way out now, but we still use it
-
2:20 - 2:22for 99 percent of our stuff.
-
2:22 - 2:26This is a paper form, and what you're looking at
-
2:26 - 2:30is a paper form in the hand of a Ministry of Health nurse
-
2:30 - 2:33in Indonesia who is tramping out across the countryside
-
2:33 - 2:37in Indonesia on, I'm sure, a very hot and humid day,
-
2:37 - 2:39and she is going to be knocking on thousands of doors
-
2:39 - 2:41over a period of weeks or months,
-
2:41 - 2:44knocking on the doors and saying, "Excuse me,
-
2:44 - 2:46we'd like to ask you some questions.
-
2:46 - 2:50Do you have any children? Were your children vaccinated?"
-
2:50 - 2:51Because the only way we can actually find out
-
2:51 - 2:54how many children were vaccinated in the country of Indonesia,
-
2:54 - 2:57what percentage were vaccinated, is actually not
-
2:57 - 3:00on the Internet but by going out and knocking on doors,
-
3:00 - 3:03sometimes tens of thousands of doors.
-
3:03 - 3:05Sometimes it takes months to even years
-
3:05 - 3:07to do something like this.
-
3:07 - 3:09You know, a census of Indonesia
-
3:09 - 3:11would probably take two years to accomplish.
-
3:11 - 3:13And the problem, of course, with all of this is that
-
3:13 - 3:15with all those paper forms — and I'm telling you
-
3:15 - 3:17we have paper forms for every possible thing.
-
3:17 - 3:20We have paper forms for vaccination surveys.
-
3:20 - 3:24We have paper forms to track people who come into clinics.
-
3:24 - 3:26We have paper forms to track drug supplies,
-
3:26 - 3:29blood supplies, all these different paper forms
-
3:29 - 3:31for many different topics,
-
3:31 - 3:33they all have a single common endpoint,
-
3:33 - 3:36and the common endpoint looks something like this.
-
3:36 - 3:40And what we're looking at here is a truckful o' data.
-
3:40 - 3:45This is the data from a single vaccination coverage survey
-
3:45 - 3:47in a single district in the country of Zambia
-
3:47 - 3:49from a few years ago that I participated in.
-
3:49 - 3:52The only thing anyone was trying to find out
-
3:52 - 3:55is what percentage of Zambian children are vaccinated,
-
3:55 - 3:58and this is the data, collected on paper over weeks
-
3:58 - 4:01from a single district, which is something like a county
-
4:01 - 4:02in the United States.
-
4:02 - 4:04You can imagine that, for the entire country of Zambia,
-
4:04 - 4:08answering just that single question
-
4:08 - 4:10looks something like this.
-
4:10 - 4:12Truck after truck after truck
-
4:12 - 4:16filled with stack after stack after stack of data.
-
4:16 - 4:17And what makes it even worse is that
-
4:17 - 4:19that's just the beginning,
-
4:19 - 4:21because once you've collected all that data,
-
4:21 - 4:23of course someone's going to have to --
-
4:23 - 4:26some unfortunate person is going to have to type that into a computer.
-
4:26 - 4:28When I was a graduate student, I actually was
-
4:28 - 4:30that unfortunate person sometimes.
-
4:30 - 4:33I can tell you, I often wasn't really paying attention.
-
4:33 - 4:35I probably made a lot of mistakes when I did it
-
4:35 - 4:38that no one ever discovered, so data quality goes down.
-
4:38 - 4:41But eventually that data hopefully gets typed into a computer,
-
4:41 - 4:43and someone can begin to analyze it,
-
4:43 - 4:45and once they have an analysis and a report,
-
4:45 - 4:49hopefully then you can take the results of that data collection
-
4:49 - 4:51and use it to vaccinate children better.
-
4:51 - 4:54Because if there's anything worse
-
4:54 - 4:56in the field of global public health,
-
4:56 - 4:59I don't know what's worse than allowing children on this planet
-
4:59 - 5:02to die of vaccine-preventable diseases,
-
5:02 - 5:05diseases for which the vaccine costs a dollar.
-
5:05 - 5:08And millions of children die of these diseases every year.
-
5:08 - 5:12And the fact is, millions is a gross estimate because
-
5:12 - 5:15we don't really know how many kids die each year of this.
-
5:15 - 5:17What makes it even more frustrating is that
-
5:17 - 5:20the data entry part, the part that I used to do as a grad student,
-
5:20 - 5:22can take sometimes six months.
-
5:22 - 5:25Sometimes it can take two years to type that information
-
5:25 - 5:28into a computer, and sometimes, actually not infrequently,
-
5:28 - 5:30it actually never happens.
-
5:30 - 5:32Now try and wrap your head around that for a second.
-
5:32 - 5:35You just had teams of hundreds of people.
-
5:35 - 5:37They went out into the field to answer a particular question.
-
5:37 - 5:39You probably spent hundreds of thousands of dollars
-
5:39 - 5:43on fuel and photocopying and per diem,
-
5:43 - 5:46and then for some reason, momentum is lost
-
5:46 - 5:47or there's no money left,
-
5:47 - 5:49and all of that comes to nothing
-
5:49 - 5:52because no one actually types it into the computer at all.
-
5:52 - 5:55The process just stops. Happens all the time.
-
5:55 - 5:58This is what we base our decisions on in global health:
-
5:58 - 6:03little data, old data, no data.
-
6:03 - 6:06So back in 1995, I began to think about ways
-
6:06 - 6:08in which we could improve this process.
-
6:08 - 6:11Now 1995, obviously that was quite a long time ago.
-
6:11 - 6:13It kind of frightens me to think of how long ago that was.
-
6:13 - 6:15The top movie of the year was
-
6:15 - 6:16"Die Hard with a Vengeance."
-
6:16 - 6:19As you can see, Bruce Willis had a lot more hair back then.
-
6:19 - 6:22I was working in the Centers for Disease Control,
-
6:22 - 6:25and I had a lot more hair back then as well.
-
6:25 - 6:28But to me, the most significant thing that I saw in 1995
-
6:28 - 6:29was this.
-
6:29 - 6:32Hard for us to imagine, but in 1995,
-
6:32 - 6:36this was the ultimate elite mobile device.
-
6:36 - 6:38Right? It wasn't an iPhone. It wasn't a Galaxy phone.
-
6:38 - 6:40It was a Palm Pilot.
-
6:40 - 6:43And when I saw the Palm Pilot for the first time, I thought,
-
6:43 - 6:46why can't we put the forms on these Palm Pilots
-
6:46 - 6:48and go out into the field just carrying one Palm Pilot,
-
6:48 - 6:52which can hold the capacity of tens of thousands
-
6:52 - 6:54of paper forms? Why don't we try to do that?
-
6:54 - 6:57Because if we can do that, if we can actually just
-
6:57 - 6:59collect the data electronically, digitally,
-
6:59 - 7:01from the very beginning,
-
7:01 - 7:04we can just put a shortcut right through that whole process
-
7:04 - 7:07of typing,
-
7:07 - 7:09of having somebody type that stuff into the computer.
-
7:09 - 7:11We can skip straight to the analysis
-
7:11 - 7:14and then straight to the use of the data to actually save lives.
-
7:14 - 7:17So that's actually what I began to do.
-
7:17 - 7:20Working at CDC, I began to travel to different programs
-
7:20 - 7:24around the world and to train them in using Palm Pilots
-
7:24 - 7:27to do data collection instead of using paper.
-
7:27 - 7:29And it actually worked great.
-
7:29 - 7:31It worked exactly as well as anybody would have predicted.
-
7:31 - 7:34What do you know? Digital data collection
-
7:34 - 7:36is actually more efficient than collecting on paper.
-
7:36 - 7:38While I was doing it, my business partner, Rose,
-
7:38 - 7:41who's here with her husband, Matthew, here in the audience,
-
7:41 - 7:44Rose was out doing similar stuff for the American Red Cross.
-
7:44 - 7:46The problem was, after a few years of doing that,
-
7:46 - 7:49I realized I had done -- I had been to maybe
-
7:49 - 7:52six or seven programs, and I thought,
-
7:52 - 7:54you know, if I keep this up at this pace,
-
7:54 - 7:56over my whole career, maybe I'm going to go
-
7:56 - 7:58to maybe 20 or 30 programs.
-
7:58 - 8:01But the problem is, 20 or 30 programs,
-
8:01 - 8:04like, training 20 or 30 programs to use this technology,
-
8:04 - 8:06that is a tiny drop in the bucket.
-
8:06 - 8:10The demand for this, the need for data to run better programs,
-
8:10 - 8:13just within health, not to mention all of the other fields
-
8:13 - 8:15in developing countries, is enormous.
-
8:15 - 8:19There are millions and millions and millions of programs,
-
8:19 - 8:22millions of clinics that need to track drugs,
-
8:22 - 8:23millions of vaccine programs.
-
8:23 - 8:25There are schools that need to track attendance.
-
8:25 - 8:27There are all these different things
-
8:27 - 8:29for us to get the data that we need to do.
-
8:29 - 8:34And I realized, if I kept up the way that I was doing,
-
8:34 - 8:37I was basically hardly going to make any impact
-
8:37 - 8:39by the end of my career.
-
8:39 - 8:41And so I began to wrack my brain
-
8:41 - 8:42trying to think about, you know,
-
8:42 - 8:44what was the process that I was doing,
-
8:44 - 8:47how was I training folks, and what were the bottlenecks
-
8:47 - 8:49and what were the obstacles to doing it faster
-
8:49 - 8:51and to doing it more efficiently?
-
8:51 - 8:54And unfortunately, after thinking about this for some time,
-
8:54 - 8:58I realized -- I identified the main obstacle.
-
8:58 - 9:00And the main obstacle, it turned out,
-
9:00 - 9:01and this is a sad realization,
-
9:01 - 9:04the main obstacle was me.
-
9:04 - 9:06So what do I mean by that?
-
9:06 - 9:08I had developed a process whereby
-
9:08 - 9:13I was the center of the universe of this technology.
-
9:13 - 9:16If you wanted to use this technology, you had to get in touch with me.
-
9:16 - 9:18That means you had to know I existed.
-
9:18 - 9:20Then you had to find the money to pay for me
-
9:20 - 9:21to fly out to your country
-
9:21 - 9:23and the money to pay for my hotel
-
9:23 - 9:26and my per diem and my daily rate.
-
9:26 - 9:29So you could be talking about 10,000 or 20,000 or 30,000 dollars
-
9:29 - 9:31if I actually had the time or it fit my schedule
-
9:31 - 9:33and I wasn't on vacation.
-
9:33 - 9:36The point is that anything, any system that depends
-
9:36 - 9:39on a single human being or two or three or five human beings,
-
9:39 - 9:41it just doesn't scale.
-
9:41 - 9:43And this is a problem for which we need to scale
-
9:43 - 9:46this technology and we need to scale it now.
-
9:46 - 9:48And so I began to think of ways in which I could basically
-
9:48 - 9:50take myself out of the picture.
-
9:50 - 9:55And, you know, I was thinking,
-
9:55 - 9:57how could I take myself out of the picture
-
9:57 - 9:59for quite some time.
-
9:59 - 10:01You know, I'd been trained that the way that
-
10:01 - 10:04you distribute technology within international development
-
10:04 - 10:06is always consultant-based.
-
10:06 - 10:09It's always guys that look pretty much like me
-
10:09 - 10:11flying from countries that look pretty much like this
-
10:11 - 10:14to other countries with people with darker skin.
-
10:14 - 10:17And you go out there, and you spend money on airfare
-
10:17 - 10:20and you spend time and you spend per diem
-
10:20 - 10:22and you spend [on a] hotel and you spend all that stuff.
-
10:22 - 10:24As far as I knew, that was the only way
-
10:24 - 10:27you could distribute technology, and I couldn't figure out a way around it.
-
10:27 - 10:30But the miracle that happened,
-
10:30 - 10:33I'm going to call it Hotmail for short.
-
10:33 - 10:35Now you may not think of Hotmail as being miraculous,
-
10:35 - 10:38but for me it was miraculous, because I noticed,
-
10:38 - 10:40just as I was wrestling with this problem,
-
10:40 - 10:44I was working in sub-Saharan Africa mostly at the time.
-
10:44 - 10:46I noticed that every sub-Saharan African health worker
-
10:46 - 10:50that I was working with had a Hotmail account.
-
10:50 - 10:53And I thought, it struck me,
-
10:53 - 10:55wait a minute, I know that the Hotmail people
-
10:55 - 10:58surely didn't fly to the Ministry of Health of Kenya
-
10:58 - 11:01to train people in how to use Hotmail.
-
11:01 - 11:03So these guys are distributing technology.
-
11:03 - 11:05They're getting software capacity out there
-
11:05 - 11:07but they're not actually flying around the world.
-
11:07 - 11:09I need to think about this some more.
-
11:09 - 11:11While I was thinking about it, people started using
-
11:11 - 11:14even more things just like this, just as we were.
-
11:14 - 11:15They started using LinkedIn and Flickr
-
11:15 - 11:18and Gmail and Google Maps, all these things.
-
11:18 - 11:21Of course, all of these things are cloud-based
-
11:21 - 11:23and don't require any training.
-
11:23 - 11:25They don't require any programmers.
-
11:25 - 11:26They don't require any consultants, because
-
11:26 - 11:29the business model for all these businesses
-
11:29 - 11:32requires that something be so simple we can use it ourselves
-
11:32 - 11:33with little or no training.
-
11:33 - 11:35You just have to hear about it and go to the website.
-
11:35 - 11:40And so I thought, what would happen if we built software
-
11:40 - 11:42to do what I'd been consulting in?
-
11:42 - 11:43Instead of training people how
-
11:43 - 11:46to put forms onto mobile devices,
-
11:46 - 11:48let's create software that lets them do it themselves
-
11:48 - 11:50with no training and without me being involved?
-
11:50 - 11:52And that's exactly what we did.
-
11:52 - 11:56So we created software called Magpi,
-
11:56 - 11:58which has an online form creator.
-
11:58 - 11:59No one has to speak to me.
-
11:59 - 12:02You just have to hear about it and go to the website.
-
12:02 - 12:04You can create forms, and once you've created the forms,
-
12:04 - 12:07you push them to a variety of common mobile phones.
-
12:07 - 12:09Obviously nowadays, we've moved past Palm Pilots
-
12:09 - 12:10to mobile phones.
-
12:10 - 12:12And it doesn't have to be a smartphone.
-
12:12 - 12:14It can be a basic phone like the phone on the right there,
-
12:14 - 12:16you know, the basic kind of Symbian phone
-
12:16 - 12:18that's very common in developing countries.
-
12:18 - 12:22And the great part about this is, it's just like Hotmail.
-
12:22 - 12:24It's cloud-based, and it doesn't require any training,
-
12:24 - 12:26programming, consultants.
-
12:26 - 12:28But there are some additional benefits as well.
-
12:28 - 12:30Now we knew, when we built this system,
-
12:30 - 12:33the whole point of it, just like with the Palm Pilots,
-
12:33 - 12:35was that you'd have to, you'd be able to
-
12:35 - 12:38collect the data and immediately upload the data and get your data set.
-
12:38 - 12:41But what we found, of course, since it's already on a computer,
-
12:41 - 12:44we can deliver instant maps and analysis and graphing.
-
12:44 - 12:46We can take a process that took two years
-
12:46 - 12:49and compress that down to the space of five minutes.
-
12:49 - 12:52Unbelievable improvements in efficiency.
-
12:52 - 12:57Cloud-based, no training, no consultants, no me.
-
12:57 - 12:59And I told you that in the first few years
-
12:59 - 13:01of trying to do this the old-fashioned way,
-
13:01 - 13:02going out to each country,
-
13:02 - 13:05we reached about, I don't know,
-
13:05 - 13:07probably trained about 1,000 people.
-
13:07 - 13:09What happened after we did this?
-
13:09 - 13:12In the second three years, we had 14,000 people
-
13:12 - 13:15find the website, sign up, and start using it to collect data,
-
13:15 - 13:16data for disaster response,
-
13:16 - 13:21Canadian pig farmers tracking pig disease and pig herds,
-
13:21 - 13:24people tracking drug supplies.
-
13:24 - 13:25One of my favorite examples, the IRC,
-
13:25 - 13:27International Rescue Committee,
-
13:27 - 13:30they have a program where semi-literate midwives
-
13:30 - 13:33using $10 mobile phones
-
13:33 - 13:35send a text message using our software
-
13:35 - 13:37once a week with the number of births
-
13:37 - 13:40and the number of deaths, which gives IRC
-
13:40 - 13:42something that no one in global health has ever had:
-
13:42 - 13:46a near real-time system of counting babies,
-
13:46 - 13:47of knowing how many kids are born,
-
13:47 - 13:49of knowing how many children there are
-
13:49 - 13:52in Sierra Leone, which is the country where this is happening,
-
13:52 - 13:55and knowing how many children die.
-
13:55 - 13:56Physicians for Human Rights --
-
13:56 - 13:59this is moving a little bit outside the health field —
-
13:59 - 14:02they are gathering, they're basically training people
-
14:02 - 14:05to do rape exams in Congo, where this is an epidemic,
-
14:05 - 14:07a horrible epidemic,
-
14:07 - 14:09and they're using our software to document
-
14:09 - 14:12the evidence they find, including photographically,
-
14:12 - 14:16so that they can bring the perpetrators to justice.
-
14:16 - 14:20Camfed, another charity based out of the U.K.,
-
14:20 - 14:24Camfed pays girls' families to keep them in school.
-
14:24 - 14:26They understand this is the most significant intervention
-
14:26 - 14:29they can make. They used to track the dispersements,
-
14:29 - 14:31the attendance, the grades, on paper.
-
14:31 - 14:32The turnaround time between a teacher
-
14:32 - 14:34writing down grades or attendance
-
14:34 - 14:37and getting that into a report was about two to three years.
-
14:37 - 14:39Now it's real time, and because this is such
-
14:39 - 14:42a low-cost system and based in the cloud, it costs,
-
14:42 - 14:45for the entire five countries that Camfed runs this in
-
14:45 - 14:47with tens of thousands of girls,
-
14:47 - 14:51the whole cost combined is 10,000 dollars a year.
-
14:51 - 14:52That's less than I used to get
-
14:52 - 14:58just traveling out for two weeks to do a consultation.
-
14:58 - 15:00So I told you before that
-
15:00 - 15:02when we were doing it the old-fashioned way, I realized
-
15:02 - 15:05all of our work was really adding up to just a drop in the bucket --
-
15:05 - 15:0710, 20, 30 different programs.
-
15:07 - 15:09We've made a lot of progress, but I recognize
-
15:09 - 15:11that right now, even the work that we've done
-
15:11 - 15:14with 14,000 people using this,
-
15:14 - 15:17is still a drop in the bucket. But something's changed.
-
15:17 - 15:18And I think it should be obvious.
-
15:18 - 15:20What's changed now is,
-
15:20 - 15:24instead of having a program in which we're scaling at such a slow rate
-
15:24 - 15:27that we can never reach all the people who need us,
-
15:27 - 15:31we've made it unnecessary for people to get reached by us.
-
15:31 - 15:34We've created a tool that lets programs
-
15:34 - 15:37keep kids in school, track the number of babies
-
15:37 - 15:40that are born and the number of babies that die,
-
15:40 - 15:43to catch criminals and successfully prosecute them,
-
15:43 - 15:46to do all these different things to learn more
-
15:46 - 15:51about what's going on, to understand more, to see more,
-
15:51 - 15:55and to save lives and improve lives.
-
15:55 - 15:57Thank you.
-
15:57 - 16:01(Applause)
- Title:
- The surprising seeds of a big-data revolution in healthcare
- Speaker:
- Joel Selanikio
- 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. (Filmed at TEDxAustin.)
- Video Language:
- English
- Team:
- closed TED
- Project:
- TEDTalks
- Duration:
- 16:18
Brian Greene edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Krystian Aparta commented on English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Krystian Aparta edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Krystian Aparta edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Krystian Aparta edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Krystian Aparta edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Thu-Huong Ha edited English subtitles for The surprising seeds of a big-data revolution in healthcare | ||
Thu-Huong Ha approved English subtitles for The surprising seeds of a big-data revolution in healthcare |
Krystian Aparta
The English transcript was updated on 1/11/2016.