1
00:00:04,553 --> 00:00:06,553
I'm going to talk to that guy,

2
00:00:06,593 --> 00:00:09,616
I need one of those bottles
for my appartment in Park Slope.

3
00:00:10,347 --> 00:00:11,584
Thanks for having me.

4
00:00:11,609 --> 00:00:13,070
I'm going to talk to you today

5
00:00:13,095 --> 00:00:16,420
about the design of medical technology
for low-resource settings.

6
00:00:16,445 --> 00:00:18,801
Some of the countries
that Arun just highlighted

7
00:00:18,826 --> 00:00:20,468
that were dark in that map.

8
00:00:20,501 --> 00:00:22,854
I study health systems in these countries.

9
00:00:22,878 --> 00:00:24,793
And one of the major gaps in care,

10
00:00:24,817 --> 00:00:26,476
almost across the board,

11
00:00:26,500 --> 00:00:28,667
is access to safe surgery.

12
00:00:28,691 --> 00:00:31,350
Now one of the major
bottlenecks that we've found

13
00:00:31,374 --> 00:00:34,609
that's sort of preventing
both the access in the first place,

14
00:00:34,633 --> 00:00:38,069
and the safety of those surgeries
that do happen, is anesthesia.

15
00:00:38,546 --> 00:00:40,938
And actually, it's the model
that we expect to work

16
00:00:40,962 --> 00:00:43,661
for delivering anesthesia
in these environments.

17
00:00:44,320 --> 00:00:48,173
Here, we have a scene that you would find
in any operating room across the US,

18
00:00:48,593 --> 00:00:50,070
or any other developed country.

19
00:00:50,298 --> 00:00:51,485
In the background there

20
00:00:51,509 --> 00:00:54,080
is a very sophisticated
anesthesia machine.

21
00:00:54,104 --> 00:00:58,252
And this machine is able
to enable surgery and save lives

22
00:00:58,277 --> 00:01:01,576
because it was designed
with this environment in mind.

23
00:01:01,729 --> 00:01:04,661
In order to operate,
this machine needs a number of things

24
00:01:04,685 --> 00:01:06,585
that this hospital has to offer.

25
00:01:06,609 --> 00:01:09,977
It needs an extremely
well-trained anesthesiologist

26
00:01:10,001 --> 00:01:12,452
with years of training
with complex machines

27
00:01:12,476 --> 00:01:14,897
to help her monitor the flows of the gas

28
00:01:14,921 --> 00:01:17,484
and keep her patients
safe and anesthetized

29
00:01:17,508 --> 00:01:18,959
throughout the surgery.

30
00:01:18,983 --> 00:01:21,750
It's a delicate machine
running on computer algorithms,

31
00:01:21,774 --> 00:01:26,091
and it needs special care, TLC,
to keep it up and running,

32
00:01:26,115 --> 00:01:27,973
and it's going to break pretty easily.

33
00:01:27,998 --> 00:01:30,910
And when it does, it needs
a team of biomedical engineers

34
00:01:30,934 --> 00:01:34,243
who understand its complexities,
can fix it, can source the parts

35
00:01:34,267 --> 00:01:36,499
and keep it saving lives.

36
00:01:37,188 --> 00:01:38,722
It's a pretty expensive machine.

37
00:01:38,746 --> 00:01:41,387
It needs a hospital
whose budget can allow it

38
00:01:41,411 --> 00:01:46,305
to support one machine
costing upwards of 50 or $100,000.

39
00:01:46,796 --> 00:01:48,695
And perhaps most obviously,

40
00:01:48,719 --> 00:01:50,233
but also most importantly --

41
00:01:50,257 --> 00:01:53,093
and the path to concepts
that we've heard about

42
00:01:53,117 --> 00:01:54,403
kind of illustrates this --

43
00:01:54,427 --> 00:01:59,573
it needs infrastructure that can supply
an uninterrupted source of electricity,

44
00:01:59,597 --> 00:02:02,609
of compressed oxygen,
and other medical supplies

45
00:02:02,633 --> 00:02:06,911
that are so critical
to the functioning of this machine.

46
00:02:07,235 --> 00:02:10,533
In other words, this machine
requires a lot of stuff

47
00:02:10,558 --> 00:02:12,531
that this hospital cannot offer.

48
00:02:13,132 --> 00:02:16,152
This is the electrical supply
for a hospital in rural Malawi.

49
00:02:16,953 --> 00:02:18,114
In this hospital,

50
00:02:18,138 --> 00:02:20,809
there is one person qualified
to deliver anesthesia,

51
00:02:20,833 --> 00:02:21,985
and she's qualified

52
00:02:22,009 --> 00:02:26,585
because she has 12, maybe 18 months
of training in anesthesia.

53
00:02:26,609 --> 00:02:28,856
In the hospital and in the entire region

54
00:02:28,880 --> 00:02:30,887
there's not a single biomedical engineer.

55
00:02:31,211 --> 00:02:32,780
So when this machine breaks,

56
00:02:32,804 --> 00:02:35,013
the machines that they have
to work with break,

57
00:02:35,037 --> 00:02:36,806
they've got to try and figure it out,

58
00:02:36,830 --> 00:02:39,122
but most of the time,
that's the end of the road.

59
00:02:39,146 --> 00:02:41,363
Those machines go the proverbial junkyard.

60
00:02:42,287 --> 00:02:44,616
And the price tag
of the machine that I mentioned

61
00:02:44,640 --> 00:02:46,783
could represent maybe a quarter or a third

62
00:02:47,125 --> 00:02:49,937
of the annual operating budget
for this hospital.

63
00:02:50,980 --> 00:02:54,364
And finally, I think you can see
that infrastructure is not very strong.

64
00:02:54,388 --> 00:02:57,062
This hospital is connected
to a very weak power grid,

65
00:02:57,086 --> 00:02:58,868
one that goes down frequently.

66
00:02:58,892 --> 00:03:01,438
So it runs frequently,
the entire hospital,

67
00:03:01,462 --> 00:03:02,885
just on a generator.

68
00:03:02,909 --> 00:03:05,065
And you can imagine,
the generator breaks down

69
00:03:05,089 --> 00:03:06,401
or runs out of fuel.

70
00:03:06,761 --> 00:03:08,788
And the World Bank sees this

71
00:03:08,812 --> 00:03:12,343
and estimates that a hospital
in this setting in a low-income country

72
00:03:12,367 --> 00:03:15,443
can expect up to
18 power outages per month.

73
00:03:17,058 --> 00:03:20,251
Similarly, compressed oxygen
and other medical supplies

74
00:03:20,275 --> 00:03:21,457
are really a luxury,

75
00:03:21,481 --> 00:03:24,763
and can often be out of stock
for months or even a year.

76
00:03:24,787 --> 00:03:27,951
So it seems crazy, but the model
that we have right now

77
00:03:27,975 --> 00:03:30,180
is taking those machines
that were designed

78
00:03:30,204 --> 00:03:32,487
for that first environment
that I showed you

79
00:03:32,511 --> 00:03:36,375
and donating or selling them
to hospitals in this environment.

80
00:03:37,261 --> 00:03:38,837
It's not just inappropriate,

81
00:03:38,861 --> 00:03:41,399
it becomes really unsafe.

82
00:03:41,916 --> 00:03:43,838
One of our partners at Johns Hopkins

83
00:03:43,862 --> 00:03:48,583
was observing surgeries in Sierra Leone
about a year ago.

84
00:03:49,225 --> 00:03:52,558
And the first surgery of the day
happened to be an obstetrical case.

85
00:03:52,582 --> 00:03:55,565
A woman came in,
she needed an emergency C-section

86
00:03:55,589 --> 00:03:58,318
to save her life and the life of her baby.

87
00:03:58,805 --> 00:04:00,759
And everything began pretty auspiciously.

88
00:04:00,783 --> 00:04:03,036
The surgeon was on call and scrubbed in.

89
00:04:03,060 --> 00:04:04,348
The nurse was there.

90
00:04:04,372 --> 00:04:07,233
She was able to anesthetize her quickly,
and it was important

91
00:04:07,257 --> 00:04:09,593
because of the emergency
nature of the situation.

92
00:04:09,617 --> 00:04:11,334
And everything began well

93
00:04:13,750 --> 00:04:15,848
until the power went out.

94
00:04:16,873 --> 00:04:18,697
And now in the middle of this surgery,

95
00:04:18,721 --> 00:04:21,790
the surgeon is racing
against the clock to finish his case,

96
00:04:21,814 --> 00:04:24,166
which he can do -- he's got a headlamp.

97
00:04:24,190 --> 00:04:28,251
But the nurse is literally running
around a darkened operating theater

98
00:04:28,275 --> 00:04:31,230
trying to find anything
she can use to anesthetize her patient,

99
00:04:31,254 --> 00:04:32,943
to keep her patient asleep.

100
00:04:32,967 --> 00:04:35,850
Because her machine doesn't work
when there's no power.

101
00:04:36,537 --> 00:04:39,591
This routine surgery that many of you
have probably experienced,

102
00:04:39,615 --> 00:04:45,228
and others are probably the product of,
has now become a tragedy.

103
00:04:46,155 --> 00:04:48,891
And what's so frustrating
is this is not a singular event;

104
00:04:48,915 --> 00:04:51,414
this happens across the developing world.

105
00:04:51,438 --> 00:04:54,391
35 million surgeries
are attempted every year

106
00:04:54,415 --> 00:04:56,011
without safe anesthesia.

107
00:04:56,710 --> 00:04:59,622
My colleague, Dr. Paul Fenton,
was living this reality.

108
00:04:59,646 --> 00:05:01,476
He was the chief of anesthesiology

109
00:05:01,500 --> 00:05:04,000
in a hospital in Malawi,
a teaching hospital.

110
00:05:04,528 --> 00:05:06,022
He went to work every day

111
00:05:06,046 --> 00:05:08,014
in an operating theater like this one,

112
00:05:08,038 --> 00:05:11,177
trying to deliver anesthesia
and teach others how to do so

113
00:05:11,201 --> 00:05:12,476
using that same equipment

114
00:05:12,500 --> 00:05:16,811
that became so unreliable,
and frankly unsafe, in his hospital.

115
00:05:17,539 --> 00:05:19,319
And after umpteen surgeries

116
00:05:19,343 --> 00:05:21,911
and, you can imagine,
really unspeakable tragedy,

117
00:05:21,935 --> 00:05:24,277
he just said, "That's it.
I'm done. That's enough.

118
00:05:24,301 --> 00:05:26,371
There has to be something better."

119
00:05:26,729 --> 00:05:28,276
He took a walk down the hall

120
00:05:28,300 --> 00:05:31,731
to where they threw all those machines
that had just crapped out on them,

121
00:05:31,755 --> 00:05:33,432
I think that's the scientific term,

122
00:05:33,456 --> 00:05:34,798
and he started tinkering.

123
00:05:34,822 --> 00:05:37,175
He took one part from here
and another from there,

124
00:05:37,199 --> 00:05:39,781
and he tried to come up
with a machine that would work

125
00:05:39,805 --> 00:05:41,433
in the reality that he was facing.

126
00:05:41,957 --> 00:05:43,171
And what he came up with:

127
00:05:43,495 --> 00:05:44,818
was this guy.

128
00:05:44,842 --> 00:05:48,379
The prototype for the Universal
Anesthesia Machine --

129
00:05:48,403 --> 00:05:51,700
a machine that would work
and anesthetize his patients

130
00:05:51,724 --> 00:05:55,299
no matter the circumstances
that his hospital had to offer.

131
00:05:55,819 --> 00:05:57,300
Here it is, back at home

132
00:05:57,324 --> 00:06:00,753
at that same hospital, developed
a little further, 12 years later,

133
00:06:00,777 --> 00:06:03,957
working on patients
from pediatrics to geriatrics.

134
00:06:04,437 --> 00:06:07,207
Let me show you a little bit
about how this machine works.

135
00:06:07,231 --> 00:06:09,109
Voila!

136
00:06:10,052 --> 00:06:11,259
Here she is.

137
00:06:11,283 --> 00:06:12,527
When you have electricity,

138
00:06:12,551 --> 00:06:15,569
everything in this machine
begins in the base.

139
00:06:15,593 --> 00:06:18,190
There's a built-in
oxygen concentrator down there.

140
00:06:18,214 --> 00:06:21,343
Now you've heard me mention
oxygen a few times at this point.

141
00:06:21,367 --> 00:06:25,155
Essentially, to deliver anesthesia,
you want as pure oxygen as possible,

142
00:06:25,179 --> 00:06:28,602
because eventually you're going
to dilute it, essentially, with the gas.

143
00:06:28,626 --> 00:06:31,062
And the mixture that the patient inhales

144
00:06:31,086 --> 00:06:33,397
needs to be at least
a certain percentage oxygen

145
00:06:33,421 --> 00:06:35,107
or else it can become dangerous.

146
00:06:35,131 --> 00:06:37,049
But so in here when there's electricity,

147
00:06:37,073 --> 00:06:39,948
the oxygen concentrator takes in room air.

148
00:06:39,972 --> 00:06:43,251
Now we know room air is gloriously free,

149
00:06:43,275 --> 00:06:44,576
it is abundant,

150
00:06:44,600 --> 00:06:46,599
and it's already 21 percent oxygen.

151
00:06:46,936 --> 00:06:50,613
So all this concentrator does
is take that room air in, filter it

152
00:06:50,637 --> 00:06:54,457
and send 95 percent pure oxygen
up and across here,

153
00:06:54,481 --> 00:06:56,687
where it mixes with the anesthetic agent.

154
00:06:57,243 --> 00:07:00,867
Now before that mixture
hits the patient's lungs,

155
00:07:00,891 --> 00:07:03,105
it's going to pass by here --
you can't see it,

156
00:07:03,129 --> 00:07:04,911
but there's an oxygen sensor here --

157
00:07:04,935 --> 00:07:09,140
that's going to read out on this screen
the percentage of oxygen being delivered.

158
00:07:09,784 --> 00:07:11,758
Now if you don't have power,

159
00:07:11,782 --> 00:07:15,388
or, God forbid, the power cuts out
in the middle of a surgery,

160
00:07:15,412 --> 00:07:17,756
this machine transitions automatically,

161
00:07:17,780 --> 00:07:19,529
without even having to touch it,

162
00:07:19,553 --> 00:07:21,884
to drawing in room air from this inlet.

163
00:07:22,305 --> 00:07:23,641
Everything else is the same.

164
00:07:23,665 --> 00:07:25,261
The only difference is that now

165
00:07:25,285 --> 00:07:28,499
you're only working
with 21 percent oxygen.

166
00:07:28,906 --> 00:07:31,727
Now that used to be
a dangerous guessing game,

167
00:07:31,751 --> 00:07:34,137
because you only knew
if you gave too little oxygen

168
00:07:34,161 --> 00:07:35,537
once something bad happened.

169
00:07:35,561 --> 00:07:38,274
But we've put a long-life
battery backup on here.

170
00:07:38,298 --> 00:07:40,492
This is the only part
that's battery backed up.

171
00:07:40,516 --> 00:07:43,921
But this gives control to the provider,
whether there's power or not,

172
00:07:43,945 --> 00:07:45,538
because they can adjust the flows

173
00:07:45,562 --> 00:07:49,530
based on the percentage of oxygen
they see that they're giving the patient.

174
00:07:49,554 --> 00:07:52,899
In both cases,
whether you have power or not,

175
00:07:52,923 --> 00:07:54,963
sometimes the patient
needs help breathing.

176
00:07:54,987 --> 00:07:58,188
It's just a reality of anesthesia,
the lungs can be paralyzed.

177
00:07:58,212 --> 00:08:00,296
And so we've just added
this manual bellows.

178
00:08:00,320 --> 00:08:03,311
We've seen surgeries
for three or four hours

179
00:08:03,335 --> 00:08:05,193
to ventilate the patient on this.

180
00:08:05,924 --> 00:08:08,912
So it's a straightforward machine.

181
00:08:08,936 --> 00:08:11,841
I shudder to say simple;
it's straightforward.

182
00:08:12,436 --> 00:08:14,337
And it's by design.

183
00:08:14,361 --> 00:08:19,561
You do not need to be a highly trained,
specialized anesthesiologist

184
00:08:19,585 --> 00:08:20,756
to use this machine,

185
00:08:20,780 --> 00:08:23,532
which is good because,
in these rural district hospitals,

186
00:08:23,556 --> 00:08:26,151
you're not going to get
that level of training.

187
00:08:26,575 --> 00:08:29,703
It's also designed for the environment
that it will be used in.

188
00:08:29,727 --> 00:08:31,504
This is an incredibly rugged machine.

189
00:08:31,528 --> 00:08:34,982
It has to stand up to the heat
and the wear and tear

190
00:08:35,006 --> 00:08:38,125
that happens in hospitals
in these rural districts.

191
00:08:38,149 --> 00:08:40,544
And so it's not going
to break very easily,

192
00:08:40,568 --> 00:08:43,842
but if it does, virtually
every piece in this machine

193
00:08:43,866 --> 00:08:45,876
can be swapped out and replaced

194
00:08:45,900 --> 00:08:47,894
with a hex wrench and a screwdriver.

195
00:08:49,650 --> 00:08:51,910
And finally, it's affordable.

196
00:08:51,934 --> 00:08:55,587
This machine comes in
at an eighth of the cost

197
00:08:55,611 --> 00:08:58,689
of the conventional machine
that I showed you earlier.

198
00:08:58,713 --> 00:09:03,201
So in other words, what we have here
is a machine that can enable surgery

199
00:09:03,225 --> 00:09:04,397
and save lives,

200
00:09:04,421 --> 00:09:07,411
because it was designed
for its environment,

201
00:09:07,435 --> 00:09:09,853
just like the first machine I showed you.

202
00:09:09,877 --> 00:09:11,881
But we're not content to stop there.

203
00:09:11,905 --> 00:09:13,080
Is it working?

204
00:09:13,104 --> 00:09:15,686
Is this the design
that's going to work in place?

205
00:09:15,710 --> 00:09:17,477
Well, we've seen good results so far.

206
00:09:17,501 --> 00:09:20,886
This is in 13 hospitals in four countries,

207
00:09:20,910 --> 00:09:24,843
and since 2010, we've done
well over 2,000 surgeries

208
00:09:24,867 --> 00:09:27,100
with no clinically adverse events.

209
00:09:27,440 --> 00:09:28,607
So we're thrilled.

210
00:09:28,631 --> 00:09:33,053
This really seems like
a cost-effective, scalable solution

211
00:09:33,077 --> 00:09:35,110
to a problem that's really pervasive.

212
00:09:35,598 --> 00:09:37,198
But we still want to be sure

213
00:09:37,222 --> 00:09:39,639
that this is the most effective
and safe device

214
00:09:39,663 --> 00:09:41,720
that we can be putting into hospitals.

215
00:09:41,744 --> 00:09:44,268
So to do that, we've launched
a number of partnerships

216
00:09:44,292 --> 00:09:45,778
with NGOs and universities,

217
00:09:45,802 --> 00:09:48,451
to gather data on the user interface,

218
00:09:48,475 --> 00:09:50,667
on the types of surgeries
it's appropriate for,

219
00:09:50,691 --> 00:09:53,078
and ways we can enhance the device itself.

220
00:09:53,702 --> 00:09:56,356
One of those partnerships
is with Johns Hopkins

221
00:09:56,380 --> 00:09:57,753
just here in Baltimore.

222
00:09:57,777 --> 00:10:02,454
They have a really cool anesthesia
simulation lab out in Baltimore.

223
00:10:02,479 --> 00:10:04,360
So we're taking this machine

224
00:10:04,384 --> 00:10:08,257
and recreating some
of the operating theater crises

225
00:10:08,282 --> 00:10:09,789
that this machine might face

226
00:10:09,813 --> 00:10:12,176
in one of the hospitals
that it's intended for,

227
00:10:12,200 --> 00:10:14,324
and in a contained, safe environment,

228
00:10:14,348 --> 00:10:16,065
evaluating its effectiveness.

229
00:10:16,547 --> 00:10:20,179
We're then able to compare
the results from that study

230
00:10:20,203 --> 00:10:21,623
with real-world experience,

231
00:10:21,647 --> 00:10:23,844
because we're putting
two of these in hospitals

232
00:10:23,868 --> 00:10:26,019
that Johns Hopkins
works with in Sierra Leone,

233
00:10:26,043 --> 00:10:29,157
including the hospital
where that emergency C-section happened.

234
00:10:30,585 --> 00:10:33,713
So I've talked a lot about anesthesia,
and I tend to do that.

235
00:10:33,737 --> 00:10:37,829
I think it is incredibly fascinating
and an important component of health.

236
00:10:37,853 --> 00:10:40,484
And it really seems peripheral,
we never think about it,

237
00:10:41,279 --> 00:10:43,183
as he said when he was introducing me,

238
00:10:43,208 --> 00:10:45,412
until we don't have access to it,

239
00:10:45,736 --> 00:10:47,698
and then it becomes a gatekeeper.

240
00:10:47,722 --> 00:10:49,691
Who gets surgery and who doesn't?

241
00:10:50,024 --> 00:10:52,754
Who gets safe surgery and who doesn't?

242
00:10:53,416 --> 00:10:55,879
But you know,
it's just one of so many ways

243
00:10:55,903 --> 00:10:59,010
that design, appropriate design,

244
00:10:59,034 --> 00:11:01,010
can have an impact on health outcomes.

245
00:11:01,518 --> 00:11:03,637
If more people
in the health-delivery space

246
00:11:03,661 --> 00:11:07,230
really working on some of these
challenges in low-income countries

247
00:11:07,254 --> 00:11:10,907
could start their design process,
their solution search,

248
00:11:11,431 --> 00:11:13,369
from outside of that proverbial box

249
00:11:13,393 --> 00:11:15,331
and inside of the hospital --

250
00:11:15,355 --> 00:11:17,186
In other words, if we could design

251
00:11:17,210 --> 00:11:20,401
for the environment that exists
in so many parts of the world,

252
00:11:20,425 --> 00:11:22,834
rather than the one
that we wished existed --

253
00:11:24,458 --> 00:11:26,624
we might just save a lot of lives.

254
00:11:26,987 --> 00:11:28,455
Thank you very much.

255
00:11:28,479 --> 00:11:32,914
(Applause)