TEDxBoston 2009
George Whitesides: A lab the size of a postage stamp
Filmed:
Readability: 3.2
544,696 views
Traditional lab tests for disease diagnosis can be too expensive and cumbersome for the regions most in need. George Whitesides' ingenious answer is a foolproof tool that can be manufactured at virtually zero cost.
George Whitesides - Chemist
In his legendary career in chemistry, George Whitesides has been a pioneer in microfabrication and nanoscale self-assembly. Now, he's fabbing a diagnostic lab on a chip. Full bio
In his legendary career in chemistry, George Whitesides has been a pioneer in microfabrication and nanoscale self-assembly. Now, he's fabbing a diagnostic lab on a chip. Full bio
Double-click the English transcript below to play the video.
00:15
The problem that I want to talk with you about
0
0
3000
00:18
is really the problem of:
1
3000
2000
00:20
How does one supply healthcare
2
5000
4000
00:24
in a world in which cost is everything?
3
9000
4000
00:28
How do you do that?
4
13000
2000
00:30
And the basic paradigm we want to suggest to you,
5
15000
2000
00:32
I want to suggest to you, is
6
17000
2000
00:34
one in which you say that in order to
7
19000
2000
00:36
treat disease you have to first know what you're treating --
8
21000
4000
00:40
that's diagnostics -- and then you have to do something.
9
25000
2000
00:42
So, the program that we're involved in is something which we call
10
27000
3000
00:45
Diagnostics for All, or zero-cost diagnostics.
11
30000
4000
00:49
How do you provide medically relevant information
12
34000
3000
00:52
at as close as possible to zero cost? How do you do it?
13
37000
4000
00:56
Let me just give you two examples.
14
41000
2000
00:58
The rigors of military medicine
15
43000
4000
01:02
are not so dissimilar from the third world --
16
47000
2000
01:04
poor resources, a rigorous environment,
17
49000
3000
01:07
a series of problems in lightweight, and things of this kind --
18
52000
4000
01:11
and also not so different from the home healthcare
19
56000
3000
01:14
and diagnostic system world.
20
59000
3000
01:17
So, the technology that I want to talk about
21
62000
3000
01:20
is for the third world, for the developing world,
22
65000
3000
01:23
but it has, I think, much broader application,
23
68000
2000
01:25
because information is so important in the healthcare system.
24
70000
5000
01:30
So, you see two examples here.
25
75000
2000
01:32
One is a lab that is actually a fairly high-end laboratory in Africa.
26
77000
5000
01:37
The second is basically an entrepreneur
27
82000
2000
01:39
who is set up and doing who-knows-what in a table in a market.
28
84000
4000
01:43
I don't know what kind of healthcare is delivered there.
29
88000
3000
01:46
But it's not really what is probably most efficient.
30
91000
5000
01:51
What is our approach?
31
96000
3000
01:54
And the way in which one typically approaches
32
99000
3000
01:57
a problem of lowering cost,
33
102000
3000
02:00
starting from the perspective of the United States,
34
105000
3000
02:03
is to take our solution,
35
108000
2000
02:05
and then to try to cut cost out of it.
36
110000
2000
02:07
No matter how you do that,
37
112000
2000
02:09
you're not going to start with a 100,000-dollar instrument
38
114000
2000
02:11
and bring it down to no-cost. It isn't going to work.
39
116000
3000
02:14
So, the approach that we took was the other way around.
40
119000
3000
02:17
To ask, "What is the cheapest possible stuff
41
122000
2000
02:19
that you could make a diagnostic system out of,
42
124000
3000
02:22
and get useful information,
43
127000
2000
02:24
add function?" And what we've chosen is paper.
44
129000
3000
02:27
What you see here is a prototypic device.
45
132000
3000
02:30
It's about a centimeter on the side.
46
135000
2000
02:32
It's about the size of a fingernail.
47
137000
2000
02:34
The lines around the edges are
48
139000
2000
02:36
a polymer.
49
141000
2000
02:38
It's made of paper and paper, of course, wicks fluid,
50
143000
4000
02:42
as you know, paper, cloth -- drop wine on the tablecloth,
51
147000
4000
02:46
and the wine wicks all over everything.
52
151000
3000
02:49
Put it on your shirt, it ruins the shirt.
53
154000
2000
02:51
That's what a hydrophilic surface does.
54
156000
3000
02:54
So, in this device the idea is that you drip
55
159000
2000
02:56
the bottom end of it in a drop of,
56
161000
2000
02:58
in this case, urine.
57
163000
2000
03:00
The fluid wicks its way into those chambers at the top.
58
165000
4000
03:04
The brown color indicates the amount of glucose in the urine,
59
169000
4000
03:08
the blue color indicates the amount of protein in the urine.
60
173000
3000
03:11
And the combination of those two
61
176000
2000
03:13
is a first order shot at a number of
62
178000
2000
03:15
useful things that you want.
63
180000
3000
03:18
So, this is an example of a device made from a simple piece of paper.
64
183000
3000
03:21
Now, how simple can you make the production?
65
186000
3000
03:24
Why do we choose paper?
66
189000
2000
03:26
There's an example of the same thing on a finger,
67
191000
3000
03:29
showing you basically what it looks like.
68
194000
2000
03:31
One reason for using paper is that it's everywhere.
69
196000
3000
03:34
We have made these kinds of devices using
70
199000
2000
03:36
napkins and toilet paper
71
201000
3000
03:39
and wraps, and all kinds of stuff.
72
204000
2000
03:41
So, the production capability is there.
73
206000
3000
03:44
The second is, you can put lots and lots
74
209000
2000
03:46
of tests in a very small place.
75
211000
2000
03:48
I'll show you in a moment that the stack of paper there
76
213000
2000
03:50
would probably hold something like
77
215000
2000
03:52
100,000 tests, something of that kind.
78
217000
3000
03:55
And then finally, a point that you don't think of so much
79
220000
3000
03:58
in developed world medicine:
80
223000
3000
04:01
it eliminates sharps.
81
226000
2000
04:03
And what sharps means is needles, things that stick.
82
228000
3000
04:06
If you've taken a sample of someone's blood
83
231000
2000
04:08
and the someone might have hepatitis C,
84
233000
3000
04:11
you don't want to make a mistake and stick it in you.
85
236000
2000
04:13
It just -- you don't want to do that.
86
238000
2000
04:15
So, how do you dispose of that? It's a problem everywhere.
87
240000
2000
04:17
And here you simply burn it.
88
242000
2000
04:19
So, it's a sort of a practical approach
89
244000
2000
04:21
to starting on things.
90
246000
3000
04:24
Now, you say, "If paper is a good idea,
91
249000
3000
04:27
other people have surely thought of it."
92
252000
2000
04:29
And the answer is, of course, yes.
93
254000
3000
04:32
Those half of you, roughly,
94
257000
2000
04:34
who are women,
95
259000
2000
04:36
at some point may have had a pregnancy test.
96
261000
2000
04:38
And the most common of these
97
263000
3000
04:41
is in a device that looks like the thing on the left.
98
266000
3000
04:44
It's something called a lateral flow immunoassay.
99
269000
2000
04:46
In that particular test,
100
271000
2000
04:48
urine either, containing
101
273000
2000
04:50
a hormone called HCG, does or does not
102
275000
3000
04:53
flow across a piece of paper.
103
278000
2000
04:55
And there are two bars. One bar indicates that the test is working,
104
280000
4000
04:59
and if the second bar shows up, you're pregnant.
105
284000
3000
05:02
This is a terrific kind of test in a binary world,
106
287000
3000
05:05
and the nice thing about pregnancy
107
290000
2000
05:07
is either you are pregnant or you're not pregnant.
108
292000
2000
05:09
You're not partially pregnant or thinking about being pregnant
109
294000
2000
05:11
or something of that sort.
110
296000
2000
05:13
So, it works very well there,
111
298000
2000
05:15
but it doesn't work very well when you need more quantitative information.
112
300000
3000
05:18
There are also dipsticks,
113
303000
2000
05:20
but if you look at the dipsticks, they're for
114
305000
2000
05:22
another kind of urine analysis.
115
307000
2000
05:24
There are an awful lot of colors and things like that.
116
309000
3000
05:27
What do you actually do about that in a difficult circumstance?
117
312000
3000
05:30
So, the approach that we started with is to ask:
118
315000
5000
05:35
Is it really practical to make things of this sort?
119
320000
4000
05:39
And that problem is now, in a purely engineering way, solved.
120
324000
4000
05:43
And the procedure that we have is simply to start with paper.
121
328000
4000
05:47
You run it through a new kind of printer called a wax printer.
122
332000
3000
05:50
The wax printer does what looks like printing.
123
335000
3000
05:53
It is printing. You put that on, you warm it a little bit,
124
338000
3000
05:56
the wax prints through so it absorbs into the paper,
125
341000
3000
05:59
and you end up with the device that you want.
126
344000
2000
06:01
The printers cost 800 bucks now.
127
346000
4000
06:05
They'll make, we estimate that if you were to run them 24 hours a day
128
350000
3000
06:08
they'd make about 10 million tests a year.
129
353000
3000
06:11
So, it's a solved problem, that particular problem is solved.
130
356000
3000
06:14
And there is an example of the kind of thing that you see.
131
359000
2000
06:16
That's on a piece of 8 by 12 paper.
132
361000
3000
06:19
That takes about two seconds to make.
133
364000
2000
06:21
And so I regard that as done.
134
366000
2000
06:23
There is a very important issue here,
135
368000
2000
06:25
which is that because it's a printer,
136
370000
3000
06:28
a color printer, it prints colors. That's what color printers do.
137
373000
3000
06:31
I'll show you in a moment, that's actually quite useful.
138
376000
4000
06:35
Now, the next question that you would like to ask
139
380000
3000
06:38
is: What would you like to measure? What would you like to analyze?
140
383000
3000
06:41
And the thing which you'd most like to analyze,
141
386000
3000
06:44
we're a fair distance from.
142
389000
2000
06:46
It's what's called "fever of undiagnosed origin."
143
391000
4000
06:50
Someone comes into the clinic,
144
395000
2000
06:52
they have a fever, they feel bad. What do they have?
145
397000
2000
06:54
Do they have T.B.? Do they have AIDS?
146
399000
2000
06:56
Do they have a common cold?
147
401000
2000
06:58
The triage problem. That's a hard problem
148
403000
2000
07:00
for reasons that I won't go through.
149
405000
2000
07:02
There are an awful lot of things that you'd like to distinguish among.
150
407000
3000
07:05
But then there are a series of things:
151
410000
2000
07:07
AIDS, hepatitis, malaria,
152
412000
2000
07:09
TB, others
153
414000
2000
07:11
and simpler ones, such as guidance of treatment.
154
416000
4000
07:15
Now even that's more complicated than you think.
155
420000
3000
07:18
A friend of mine works in transcultural psychiatry,
156
423000
4000
07:22
and he is interested in the question of
157
427000
2000
07:24
why people do and don't take their meds.
158
429000
3000
07:27
So, Dapsone, or something like that,
159
432000
2000
07:29
you have to take it for a while.
160
434000
2000
07:31
He has a wonderful story of talking to a villager in India
161
436000
3000
07:34
and saying, "Have you taken your Dapsone?" "Yes."
162
439000
2000
07:36
"Have you taken it every day?" "Yes."
163
441000
3000
07:39
"Have you taken if for a month?" "Yes."
164
444000
2000
07:41
What the guy actually meant
165
446000
2000
07:43
was that he'd fed a 30-day dose of Dapsone
166
448000
2000
07:45
to his dog, that morning.
167
450000
2000
07:47
(Laughter)
168
452000
1000
07:48
He was telling the truth. Because
169
453000
2000
07:50
in a different culture,
170
455000
2000
07:52
the dog is a surrogate for you,
171
457000
2000
07:54
you know, "today," "this month," "since the rainy season" --
172
459000
3000
07:57
there are lots of opportunities for misunderstanding,
173
462000
3000
08:00
and so an issue here is to,
174
465000
2000
08:02
in some cases, to figure out
175
467000
2000
08:04
how to deal with matters that seem uninteresting,
176
469000
3000
08:07
like compliance.
177
472000
3000
08:10
Now, take a look at what a typical test looks like.
178
475000
4000
08:14
Prick a finger, you get some blood,
179
479000
2000
08:16
about 50 microliters.
180
481000
2000
08:18
That's about all you're going to get,
181
483000
2000
08:20
because you can't use the usual sort of systems.
182
485000
4000
08:24
You can't manipulate it very well,
183
489000
2000
08:26
although I'll show something about that in a moment.
184
491000
2000
08:28
So, you take the drop of blood, no further manipulations,
185
493000
3000
08:31
you put it on a little device,
186
496000
2000
08:33
the device filters out the blood cells, lets the serum go through,
187
498000
4000
08:37
and you get a series of colors
188
502000
2000
08:39
down in the bottom there.
189
504000
2000
08:41
And the colors indicate "disease" or "normal."
190
506000
4000
08:45
But even that's complicated,
191
510000
2000
08:47
because to you, to me, colors might indicate "normal,"
192
512000
4000
08:51
but, after all, we're all suffering from
193
516000
2000
08:53
probably an excess of education.
194
518000
3000
08:56
What you do about something which requires
195
521000
2000
08:58
quantitative analysis?
196
523000
2000
09:00
And so the solution that we and many other people
197
525000
3000
09:03
are thinking about there,
198
528000
2000
09:05
and at this point there is a dramatic flourish,
199
530000
2000
09:07
and out comes the universal solution to everything these days,
200
532000
3000
09:10
which is a cell phone. In this particular case, a camera phone.
201
535000
3000
09:13
They're everywhere, six billion a month in India.
202
538000
5000
09:18
And the idea is that what one does,
203
543000
3000
09:21
is to take the device,
204
546000
2000
09:23
you dip it, you develop the color,
205
548000
3000
09:26
you take a picture, the picture goes to a central laboratory.
206
551000
3000
09:29
You don't have to send out a doctor,
207
554000
2000
09:31
you send out somebody who can just take the sample,
208
556000
3000
09:34
and in the clinic either a doctor, or ideally a computer
209
559000
3000
09:37
in this case, does the analysis.
210
562000
2000
09:39
Turns out to work actually quite well, particularly when your
211
564000
2000
09:41
color printer has printed the color bars
212
566000
2000
09:43
that indicate how things work.
213
568000
2000
09:45
So, my view of the health care worker of the future
214
570000
3000
09:48
is not a doctor,
215
573000
2000
09:50
but is an 18-year-old, otherwise unemployed,
216
575000
3000
09:53
who has two things: He has a backpack full of these tests,
217
578000
2000
09:55
and a lancet to occasionally take a blood sample,
218
580000
3000
09:58
and an AK-47.
219
583000
2000
10:00
And these are the things that get him through his day.
220
585000
5000
10:05
There's another very interesting connection here,
221
590000
2000
10:07
and that is that what one wants to do
222
592000
2000
10:09
is to pass through useful information
223
594000
3000
10:12
over what is generally a pretty awful telephone system.
224
597000
4000
10:16
It turns out there's an enormous amount of information
225
601000
3000
10:19
already available on that subject, which is the Mars rover problem.
226
604000
3000
10:22
How do you get back an accurate view of the color on Mars
227
607000
4000
10:26
if you have a really terrible bandwidth to do it with?
228
611000
4000
10:30
And the answer is not complicated
229
615000
2000
10:32
but it's one which I don't want to go through here,
230
617000
2000
10:34
other than to say that the communication systems
231
619000
3000
10:37
for doing this are really pretty well understood.
232
622000
2000
10:39
Also, a fact which you may not know
233
624000
3000
10:42
is that the compute capability of this thing
234
627000
3000
10:45
is not so different from the compute capability
235
630000
2000
10:47
of your desktop computer.
236
632000
2000
10:49
This is a fantastic device which is only beginning to be tapped.
237
634000
3000
10:52
I don't know whether the idea of one computer, one child
238
637000
4000
10:56
makes any sense. Here's the computer of the future,
239
641000
3000
10:59
because this screen is already there and they're ubiquitous.
240
644000
5000
11:04
All right now let me show you just a little bit about advanced devices.
241
649000
2000
11:06
And we'll start by posing a little problem.
242
651000
3000
11:09
What you see here is another centimeter-sized device,
243
654000
3000
11:12
and the different colors are different colors of dye.
244
657000
4000
11:16
And you notice something which might strike you as
245
661000
2000
11:18
a little bit interesting,
246
663000
2000
11:20
which is the yellow seems to disappear,
247
665000
3000
11:23
get through the blue, and then get through the red.
248
668000
3000
11:26
How does that happen? How do you make something flow through something?
249
671000
3000
11:29
And, of course the answer is, "You don't."
250
674000
2000
11:31
You make it flow under and over.
251
676000
2000
11:33
But now the question is: How do you make it flow
252
678000
2000
11:35
under and over in a piece of paper?
253
680000
3000
11:38
The answer is that what you do,
254
683000
3000
11:41
and the details are not terribly important here,
255
686000
3000
11:44
is to make something more elaborate:
256
689000
2000
11:46
You take several different layers of paper,
257
691000
2000
11:48
each one containing its own little fluid system,
258
693000
3000
11:51
and you separate them by pieces of,
259
696000
2000
11:53
literally, double-sided carpet tape,
260
698000
3000
11:56
the stuff you use to stick the carpets onto the floor.
261
701000
3000
11:59
And the fluid will flow from one layer into the next.
262
704000
3000
12:02
It distributes itself, flows through further holes,
263
707000
3000
12:05
distributes itself.
264
710000
2000
12:07
And what you see, at the lower right-hand side there,
265
712000
3000
12:10
is a sample in which a single sample
266
715000
2000
12:12
of blood has been put on the top,
267
717000
3000
12:15
and it has gone through and distributed itself
268
720000
3000
12:18
into these 16 holes on the bottom,
269
723000
3000
12:21
in a piece of paper -- basically it looks like a chip,
270
726000
2000
12:23
two pieces of paper thick.
271
728000
3000
12:26
And in this particular case we were just interested in
272
731000
2000
12:28
the replicability of that.
273
733000
2000
12:30
But that is, in principle, the way you solve
274
735000
2000
12:32
the "fever of unexplained origin" problem,
275
737000
2000
12:34
because each one of those spots then becomes
276
739000
2000
12:36
a test for a particular set of markers
277
741000
3000
12:39
of disease,
278
744000
2000
12:41
and this will work in due course.
279
746000
2000
12:43
Here is an example of a slightly more complicated device.
280
748000
3000
12:46
There's the chip.
281
751000
2000
12:48
You dip in a corner. The fluid goes into the center.
282
753000
2000
12:50
It distributes itself out into these various
283
755000
3000
12:53
wells or holes, and turns color,
284
758000
2000
12:55
and all done with paper and carpet tape.
285
760000
3000
12:58
So, I think it's as low-cost
286
763000
2000
13:00
as we're likely to be able to come up and make things.
287
765000
4000
13:04
Now, I have one last, two last little stories
288
769000
3000
13:07
to tell you, in finishing off this business.
289
772000
3000
13:10
This is one: One of the things that one does occasionally
290
775000
3000
13:13
need to do is to separate blood cells from serum.
291
778000
3000
13:16
And the question was,
292
781000
3000
13:19
here we do it by taking a sample,
293
784000
2000
13:21
we put it in a centrifuge,
294
786000
3000
13:24
we spin it, and you get blood cells out. Terrific.
295
789000
4000
13:28
What happens if you don't have an electricity,
296
793000
2000
13:30
and a centrifuge, and whatever?
297
795000
2000
13:32
And we thought for a while of how you might do this
298
797000
3000
13:35
and the way, in fact, you do it is what's shown here.
299
800000
2000
13:37
You get an eggbeater,
300
802000
2000
13:39
which is everywhere, and you saw off a blade,
301
804000
3000
13:42
and then you take tubing,
302
807000
2000
13:44
and you stick it on that. You put the blood in, you spin it --
303
809000
2000
13:46
somebody sits there and spins it.
304
811000
2000
13:48
It works really, really well.
305
813000
2000
13:50
And we sat down, we did the physics of eggbeaters
306
815000
2000
13:52
and self-aligning tubes and all the rest of that kind of thing,
307
817000
3000
13:55
sent it off to a journal.
308
820000
2000
13:57
We were very proud of this, particularly the title,
309
822000
2000
13:59
which was "Eggbeater as Centrifuge."
310
824000
2000
14:01
(Laughter)
311
826000
1000
14:02
And we sent it off, and by return mail it came back.
312
827000
3000
14:05
I called up the editor and I said,
313
830000
2000
14:07
"What's going on? How is this possible?"
314
832000
2000
14:09
The editor said, with enormous disdain,
315
834000
3000
14:12
"I read this.
316
837000
2000
14:14
And we're not going to publish it, because we only
317
839000
2000
14:16
publish science."
318
841000
2000
14:18
And it's an important issue
319
843000
2000
14:20
because it means that we have to,
320
845000
2000
14:22
as a society,
321
847000
2000
14:24
think about what we value.
322
849000
2000
14:26
And if it's just papers and phys. rev. letters,
323
851000
2000
14:28
we've got a problem.
324
853000
3000
14:31
Here is another example of something which is --
325
856000
3000
14:34
this is a little spectrophotometer.
326
859000
2000
14:36
It measures the absorption of light in a sample
327
861000
3000
14:39
The neat thing about this is, you have light source that flickers
328
864000
3000
14:42
on and off at about 1,000 hertz,
329
867000
2000
14:44
another light source that detects that light at 1,000 hertz,
330
869000
4000
14:48
and so you can run this system in broad daylight.
331
873000
3000
14:51
It performs about equivalently
332
876000
2000
14:53
to a system that's in the order of
333
878000
3000
14:56
100,000 dollars.
334
881000
2000
14:58
It costs 50 dollars. We can probably make it for 50 cents,
335
883000
3000
15:01
if we put our mind to it.
336
886000
2000
15:03
Why doesn't somebody do it? And the answer is,
337
888000
2000
15:05
"How do you make a profit in a capitalist system, doing that?"
338
890000
4000
15:09
Interesting problem.
339
894000
3000
15:12
So, let me finish by saying
340
897000
2000
15:14
that we've thought about this as a kind of engineering problem.
341
899000
4000
15:18
And we've asked: What is the scientific unifying idea here?
342
903000
6000
15:24
And we've decided that we should think about this
343
909000
1000
15:25
not so much in terms of cost,
344
910000
2000
15:27
but in terms of simplicity.
345
912000
2000
15:29
Simplicity is a neat word. And you've got to think about
346
914000
2000
15:31
what simplicity means.
347
916000
2000
15:33
I know what it is but I don't actually know what it means.
348
918000
4000
15:37
So, I actually was interested enough in this to put together
349
922000
2000
15:39
several groups of people.
350
924000
4000
15:43
And the most recent involved a couple of people at MIT,
351
928000
3000
15:46
one of them being an exceptionally bright kid
352
931000
2000
15:48
who is one of the very few people I would think of
353
933000
2000
15:50
who's an authentic genius.
354
935000
2000
15:52
We all struggled for an entire day to think about simplicity.
355
937000
4000
15:56
And I want to give you the answer of this
356
941000
2000
15:58
deep scientific thought.
357
943000
3000
16:01
(Laughter)
358
946000
3000
16:04
So, in a sense, you get what you pay for.
359
949000
3000
16:07
Thank you very much.
360
952000
2000
16:09
(Laughter)
361
954000
1000
ABOUT THE SPEAKER
George Whitesides - ChemistIn his legendary career in chemistry, George Whitesides has been a pioneer in microfabrication and nanoscale self-assembly. Now, he's fabbing a diagnostic lab on a chip.
Why you should listen
Someday Harvard chemistry professor George Whitesides will take the time to look back on the 950 scientific articles he's coauthored, the dozen companies he's co-founded or the 50-plus patents on which he's named. (He works in four main areas: biochemistry, materials science, catalysis and physical organic chemistry.) In the meantime, he's trying to invent a future where medical diagnosis can be done by anyone for little or no cost. He's co-founded a nonprofit called Diagnostics for All that aims to provide dirt-cheap diagnostic devices, to provide healthcare in a world where cost is everything.
Among his solutions is a low-cost "lab-on-a-chip," made of paper and carpet tape. The paper wicks bodily fluids -- urine, for example -- and turns color to provide diagnostic information, such as how much glucose or protein is present. His goal is to distribute these simple paper diagnostic systems to developing countries, where people with basic training can administer tests and send results to distant doctors via cameraphone.
George Whitesides | Speaker | TED.com