TEDxNijmegen
Peter van Manen: Better baby care -- thanks to Formula 1
Filmed:
Readability: 3.5
845,406 views
During a Formula 1 race, a car sends hundreds of millions of data points to its garage for real-time analysis and feedback. So why not use this detailed and rigorous data system elsewhere, like at children's hospitals? Peter van Manen tells us more.
Peter van Manen - Electronic systems expert
Peter van Manen is the Managing Director of McLaren Electronics, which provides data systems to major motorsports series. Full bio
Peter van Manen is the Managing Director of McLaren Electronics, which provides data systems to major motorsports series. Full bio
Double-click the English transcript below to play the video.
00:12
Motor racing is a funny old business.
0
336
2257
00:14
We make a new car every year,
1
2593
2317
00:16
and then we spend the rest of the season
2
4910
2188
00:19
trying to understand what it is we've built
3
7098
2776
00:21
to make it better, to make it faster.
4
9874
3221
00:25
And then the next year, we start again.
5
13095
3275
00:28
Now, the car you see in front of you is quite complicated.
6
16370
4238
00:32
The chassis is made up of about 11,000 components,
7
20608
3619
00:36
the engine another 6,000,
8
24227
2468
00:38
the electronics about eight and a half thousand.
9
26695
3093
00:41
So there's about 25,000 things there that can go wrong.
10
29788
4401
00:46
So motor racing is very much about attention to detail.
11
34189
4826
00:51
The other thing about Formula 1 in particular
12
39015
3263
00:54
is we're always changing the car.
13
42278
2124
00:56
We're always trying to make it faster.
14
44402
2280
00:58
So every two weeks, we will be making
15
46682
2984
01:01
about 5,000 new components to fit to the car.
16
49666
4200
01:05
Five to 10 percent of the race car
17
53866
2178
01:08
will be different every two weeks of the year.
18
56044
3752
01:11
So how do we do that?
19
59796
2309
01:14
Well, we start our life with the racing car.
20
62105
3744
01:17
We have a lot of sensors on the car to measure things.
21
65849
3991
01:21
On the race car in front of you here
22
69840
1882
01:23
there are about 120 sensors when it goes into a race.
23
71722
3159
01:26
It's measuring all sorts of things around the car.
24
74881
3652
01:30
That data is logged. We're logging about
25
78533
2052
01:32
500 different parameters within the data systems,
26
80585
3704
01:36
about 13,000 health parameters and events
27
84289
3665
01:39
to say when things are not working the way they should do,
28
87954
4565
01:44
and we're sending that data back to the garage
29
92519
2825
01:47
using telemetry at a rate of two to four megabits per second.
30
95344
4979
01:52
So during a two-hour race, each car will be sending
31
100323
3127
01:55
750 million numbers.
32
103450
2275
01:57
That's twice as many numbers as words that each of us
33
105725
3143
02:00
speaks in a lifetime.
34
108868
1631
02:02
It's a huge amount of data.
35
110499
2618
02:05
But it's not enough just to have data and measure it.
36
113117
2645
02:07
You need to be able to do something with it.
37
115762
2158
02:09
So we've spent a lot of time and effort
38
117920
2394
02:12
in turning the data into stories
39
120314
1869
02:14
to be able to tell, what's the state of the engine,
40
122183
3105
02:17
how are the tires degrading,
41
125288
2272
02:19
what's the situation with fuel consumption?
42
127560
3748
02:23
So all of this is taking data
43
131308
2788
02:26
and turning it into knowledge that we can act upon.
44
134096
3802
02:29
Okay, so let's have a look at a little bit of data.
45
137898
2638
02:32
Let's pick a bit of data from
46
140536
2030
02:34
another three-month-old patient.
47
142566
3079
02:37
This is a child, and what you're seeing here is real data,
48
145645
4171
02:41
and on the far right-hand side,
49
149816
1977
02:43
where everything starts getting a little bit catastrophic,
50
151793
2587
02:46
that is the patient going into cardiac arrest.
51
154380
3584
02:49
It was deemed to be an unpredictable event.
52
157964
3232
02:53
This was a heart attack that no one could see coming.
53
161196
3789
02:56
But when we look at the information there,
54
164985
2550
02:59
we can see that things are starting to become
55
167535
2349
03:01
a little fuzzy about five minutes or so before the cardiac arrest.
56
169884
4029
03:05
We can see small changes
57
173913
2037
03:07
in things like the heart rate moving.
58
175950
2383
03:10
These were all undetected by normal thresholds
59
178333
2486
03:12
which would be applied to data.
60
180819
2408
03:15
So the question is, why couldn't we see it?
61
183227
3143
03:18
Was this a predictable event?
62
186370
2581
03:20
Can we look more at the patterns in the data
63
188951
3010
03:23
to be able to do things better?
64
191961
3380
03:27
So this is a child,
65
195341
2650
03:29
about the same age as the racing car on stage,
66
197991
3232
03:33
three months old.
67
201223
1630
03:34
It's a patient with a heart problem.
68
202853
2605
03:37
Now, when you look at some of the data on the screen above,
69
205458
3468
03:40
things like heart rate, pulse, oxygen, respiration rates,
70
208926
4902
03:45
they're all unusual for a normal child,
71
213828
3076
03:48
but they're quite normal for the child there,
72
216904
2642
03:51
and so one of the challenges you have in health care is,
73
219546
4138
03:55
how can I look at the patient in front of me,
74
223684
2851
03:58
have something which is specific for her,
75
226535
3047
04:01
and be able to detect when things start to change,
76
229582
2788
04:04
when things start to deteriorate?
77
232370
2099
04:06
Because like a racing car, any patient,
78
234469
3050
04:09
when things start to go bad, you have a short time
79
237519
2976
04:12
to make a difference.
80
240495
1831
04:14
So what we did is we took a data system
81
242326
2754
04:17
which we run every two weeks of the year in Formula 1
82
245080
3131
04:20
and we installed it on the hospital computers
83
248211
3002
04:23
at Birmingham Children's Hospital.
84
251213
2290
04:25
We streamed data from the bedside instruments
85
253503
2439
04:27
in their pediatric intensive care
86
255942
2557
04:30
so that we could both look at the data in real time
87
258499
3456
04:33
and, more importantly, to store the data
88
261955
2871
04:36
so that we could start to learn from it.
89
264826
3057
04:39
And then, we applied an application on top
90
267883
4384
04:44
which would allow us to tease out the patterns in the data
91
272267
3270
04:47
in real time so we could see what was happening,
92
275537
2956
04:50
so we could determine when things started to change.
93
278493
3713
04:54
Now, in motor racing, we're all a little bit ambitious,
94
282206
3863
04:58
audacious, a little bit arrogant sometimes,
95
286069
2549
05:00
so we decided we would also look at the children
96
288618
3398
05:04
as they were being transported to intensive care.
97
292016
2957
05:06
Why should we wait until they arrived in the hospital
98
294973
2154
05:09
before we started to look?
99
297127
1994
05:11
And so we installed a real-time link
100
299121
2997
05:14
between the ambulance and the hospital,
101
302118
2836
05:16
just using normal 3G telephony to send that data
102
304954
3776
05:20
so that the ambulance became an extra bed
103
308730
2487
05:23
in intensive care.
104
311217
3136
05:26
And then we started looking at the data.
105
314353
3702
05:30
So the wiggly lines at the top, all the colors,
106
318055
2921
05:32
this is the normal sort of data you would see on a monitor --
107
320976
3194
05:36
heart rate, pulse, oxygen within the blood,
108
324170
3772
05:39
and respiration.
109
327942
2635
05:42
The lines on the bottom, the blue and the red,
110
330577
2753
05:45
these are the interesting ones.
111
333330
1360
05:46
The red line is showing an automated version
112
334690
3209
05:49
of the early warning score
113
337899
1597
05:51
that Birmingham Children's Hospital were already running.
114
339496
2487
05:53
They'd been running that since 2008,
115
341983
2338
05:56
and already have stopped cardiac arrests
116
344321
2256
05:58
and distress within the hospital.
117
346577
2757
06:01
The blue line is an indication
118
349334
2432
06:03
of when patterns start to change,
119
351766
2500
06:06
and immediately, before we even started
120
354266
2309
06:08
putting in clinical interpretation,
121
356575
1708
06:10
we can see that the data is speaking to us.
122
358283
2870
06:13
It's telling us that something is going wrong.
123
361153
3536
06:16
The plot with the red and the green blobs,
124
364689
3816
06:20
this is plotting different components
125
368505
2805
06:23
of the data against each other.
126
371310
2547
06:25
The green is us learning what is normal for that child.
127
373857
3840
06:29
We call it the cloud of normality.
128
377697
2610
06:32
And when things start to change,
129
380307
2241
06:34
when conditions start to deteriorate,
130
382548
2564
06:37
we move into the red line.
131
385112
2238
06:39
There's no rocket science here.
132
387350
1657
06:41
It is displaying data that exists already in a different way,
133
389007
4113
06:45
to amplify it, to provide cues to the doctors,
134
393120
3391
06:48
to the nurses, so they can see what's happening.
135
396511
2738
06:51
In the same way that a good racing driver
136
399249
3130
06:54
relies on cues to decide when to apply the brakes,
137
402379
4044
06:58
when to turn into a corner,
138
406423
1476
06:59
we need to help our physicians and our nurses
139
407899
2918
07:02
to see when things are starting to go wrong.
140
410817
3620
07:06
So we have a very ambitious program.
141
414437
2946
07:09
We think that the race is on to do something differently.
142
417383
4736
07:14
We are thinking big. It's the right thing to do.
143
422119
2904
07:17
We have an approach which, if it's successful,
144
425023
3412
07:20
there's no reason why it should stay within a hospital.
145
428435
2531
07:22
It can go beyond the walls.
146
430966
1841
07:24
With wireless connectivity these days,
147
432807
2071
07:26
there is no reason why patients, doctors and nurses
148
434878
3444
07:30
always have to be in the same place
149
438322
2171
07:32
at the same time.
150
440493
1993
07:34
And meanwhile, we'll take our little three-month-old baby,
151
442486
3995
07:38
keep taking it to the track, keeping it safe,
152
446481
3757
07:42
and making it faster and better.
153
450238
2333
07:44
Thank you very much.
154
452571
1405
07:45
(Applause)
155
453976
4954
ABOUT THE SPEAKER
Peter van Manen - Electronic systems expertPeter van Manen is the Managing Director of McLaren Electronics, which provides data systems to major motorsports series.
Why you should listen
To say that Peter van Manen has a high-speed job would be an understatement. As Managing Director of McLaren Electronics, which provides electronics and data collection software to motorsports events, he and his team work in real time during a race to improve cars on about 500 different parameters. That's about 750 million data points in two hours.
But recently van Manen and his team have been wondering: Why can't the extremely precise and subtle data-collection and analysis systems used in motorsports be applied elsewhere, for the benefit of all? They have applied their systems to ICU units at Birmingham Children's Hospital with real-time analysis that allows them to proactively prevent cardiac arrests. The unit has seen a 25 percent decrease in life-threatening events. And it's just the beginning.
Peter van Manen | Speaker | TED.com