TEDGlobal 2011
Ben Goldacre: Battling bad science
班．高達可 (Ben Goldacre)：擊敗壞科學
Filmed:
Readability: 4.3
2,713,579 views
每天都有新聞報告提出一些新的健康建議，但是我們要怎麼知道這些建議是否正確？身兼醫師與流行病學家，班．高達可滔滔不絕告訴我們證據是如何被扭曲的，從聲稱具有營養價值、讓人一目了然的噱頭手法，到醫藥工業細微的操弄手段。
Ben Goldacre - Debunker
Ben Goldacre unpicks dodgy scientific claims made by scaremongering journalists, dubious government reports, pharmaceutical corporations, PR companies and quacks. Full bio
Ben Goldacre unpicks dodgy scientific claims made by scaremongering journalists, dubious government reports, pharmaceutical corporations, PR companies and quacks. Full bio
Double-click the English transcript below to play the video.
00:15
So I'm a doctor, but I kind of slipped sideways into research,
0
0
3000
雖然我是醫生，但無心插柳做了點研究
00:18
and now I'm an epidemiologist.
1
3000
2000
所以現在我也是個流行病學家
00:20
And nobody really knows what epidemiology is.
2
5000
2000
沒有人真正知道流行病學是什麼
00:22
Epidemiology is the science of how we know in the real world
3
7000
3000
流行病學是一種科學，讓我們知道在生活中
00:25
if something is good for you or bad for you.
4
10000
2000
什麼東西對你好，或對你不好
00:27
And it's best understood through example
5
12000
2000
而有例可循最是簡單明瞭
00:29
as the science of those crazy, wacky newspaper headlines.
6
14000
5000
好比誇張、奇怪的新聞標題就是一例
00:34
And these are just some of the examples.
7
19000
2000
這裡有些例子
00:36
These are from the Daily Mail. Every country in the world has a newspaper like this.
8
21000
3000
它們來自《每日郵報》，全世界各國都有類似的報紙
00:39
It has this bizarre, ongoing philosophical project
9
24000
3000
這類報紙一直都有個古怪、看似有理的企畫
00:42
of dividing all the inanimate objects in the world
10
27000
2000
就是將世界上所有不會動的物件
00:44
into the ones that either cause or prevent cancer.
11
29000
3000
都分別歸類成致癌物或防癌物。
00:47
So here are some of the things they said cause cancer recently:
12
32000
2000
這些是一部分最近據說會致癌的玩意兒
00:49
divorce, Wi-Fi, toiletries and coffee.
13
34000
2000
離婚、無線網路、衛生用品和咖啡
00:51
Here are some of the things they say prevents cancer:
14
36000
2000
而這些則是據說能防癌的東西
00:53
crusts, red pepper, licorice and coffee.
15
38000
2000
麵包皮、紅椒、甘草和咖啡
00:55
So already you can see there are contradictions.
16
40000
2000
我想你看到了矛盾的地方
00:57
Coffee both causes and prevents cancer.
17
42000
2000
咖啡同時致癌也防癌
00:59
And as you start to read on, you can see
18
44000
2000
當你開始讀下去，你會發現
01:01
that maybe there's some kind of political valence behind some of this.
19
46000
3000
這些報導說不定別有用心。
01:04
So for women, housework prevents breast cancer,
20
49000
2000
拿女性來說，做家事能預防乳癌
01:06
but for men, shopping could make you impotent.
21
51000
3000
但對男性而言，購物會使他們性無能
01:09
So we know that we need to start
22
54000
3000
所以，我們要開始
01:12
unpicking the science behind this.
23
57000
3000
揭露這一切背後的科學真相
01:15
And what I hope to show
24
60000
2000
我想要做的
01:17
is that unpicking dodgy claims,
25
62000
2000
是揭穿那些騙人的說法
01:19
unpicking the evidence behind dodgy claims,
26
64000
2000
以及那些假象背後的證據，
01:21
isn't a kind of nasty carping activity;
27
66000
3000
這並非卑鄙、催毛求疵的事
01:24
it's socially useful,
28
69000
2000
這對社會有用，
01:26
but it's also an extremely valuable
29
71000
2000
同時，也是個極具價值的
01:28
explanatory tool.
30
73000
2000
解說工具。
01:30
Because real science is all about
31
75000
2000
因為真正的科學
01:32
critically appraising the evidence for somebody else's position.
32
77000
2000
是批判地評斷他人提出的證據
01:34
That's what happens in academic journals.
33
79000
2000
這就是在學術期刊，
01:36
That's what happens at academic conferences.
34
81000
2000
以及學術研討論上，上演的戲碼。
01:38
The Q&A session after a post-op presents data
35
83000
2000
有時在發言完後的問答時間
01:40
is often a blood bath.
36
85000
2000
就像場血戰。
01:42
And nobody minds that. We actively welcome it.
37
87000
2000
但沒人在乎，我們還熱烈歡迎。
01:44
It's like a consenting intellectual S&M activity.
38
89000
3000
就像是雙方同意，來場學術上的激烈性愛一樣。
01:47
So what I'm going to show you
39
92000
2000
所以，接下來要讓各位看的
01:49
is all of the main things,
40
94000
2000
是我的研究領域
01:51
all of the main features of my discipline --
41
96000
2000
也就是流行病學的主要特色－
01:53
evidence-based medicine.
42
98000
2000
循證醫學。
01:55
And I will talk you through all of these
43
100000
2000
我會一步步的講解，
01:57
and demonstrate how they work,
44
102000
2000
說明他們是如何運作的，
01:59
exclusively using examples of people getting stuff wrong.
45
104000
3000
但不會拿出錯的事情做為例子。
02:02
So we'll start with the absolute weakest form of evidence known to man,
46
107000
3000
一開始，我們先談最不堪一擊的證據來源
02:05
and that is authority.
47
110000
2000
就是權威。
02:07
In science, we don't care how many letters you have after your name.
48
112000
3000
在科學界，我們不在乎你名字後面有多少頭銜，
02:10
In science, we want to know what your reasons are for believing something.
49
115000
3000
我們想知道，你發表某樣事物的背後有什麼理由。
02:13
How do you know that something is good for us
50
118000
2000
你怎麼知道那對我們是好，
02:15
or bad for us?
51
120000
2000
還是不好？
02:17
But we're also unimpressed by authority,
52
122000
2000
但，因為權威容易偽造，
02:19
because it's so easy to contrive.
53
124000
2000
所以我們對權威並不有什麼感覺。
02:21
This is somebody called Dr. Gillian McKeith Ph.D,
54
126000
2000
這是麥基絲醫生博士
02:23
or, to give her full medical title, Gillian McKeith.
55
128000
3000
或者，給她一個完整的醫學頭銜，吉輪 ●麥基絲。
02:26
(Laughter)
56
131000
3000
(笑聲)
02:29
Again, every country has somebody like this.
57
134000
2000
每個國家都有像這個的一個人。
02:31
She is our TV diet guru.
58
136000
2000
她是電視上的健康飲食大師，
02:33
She has massive five series of prime-time television,
59
138000
3000
她有多達五個黃金時段的系列節目
02:36
giving out very lavish and exotic health advice.
60
141000
3000
分享又豐富，又充滿異國風情的保健建議
02:39
She, it turns out, has a non-accredited correspondence course Ph.D.
61
144000
3000
結果，她的學歷只是在美國某一大學的
02:42
from somewhere in America.
62
147000
2000
非公認的函授課程博士。
02:44
She also boasts that she's a certified professional member
63
149000
2000
她也自豪地說她是美國營養顧問協會
02:46
of the American Association of Nutritional Consultants,
64
151000
2000
的認證專業會員
02:48
which sounds very glamorous and exciting.
65
153000
2000
聽起來是多麼吸引人又讓人興奮
02:50
You get a certificate and everything.
66
155000
2000
你有張結業證書，就什麼都有了
02:52
This one belongs to my dead cat Hetti. She was a horrible cat.
67
157000
2000
這是我的貓，海地，她已經過世了，一隻很不乖的貓
02:54
You just go to the website, fill out the form,
68
159000
2000
你只要上網，填表格，
02:56
give them $60, and it arrives in the post.
69
161000
2000
交六十塊給他們，然後證書就郵寄給你
02:58
Now that's not the only reason that we think this person is an idiot.
70
163000
2000
但，這不是我們覺得這傢伙是笨蛋的唯一理由
03:00
She also goes and says things like,
71
165000
2000
她也會說些話像是
03:02
you should eat lots of dark green leaves,
72
167000
2000
你該吃很多的深綠色蔬菜，
03:04
because they contain lots of chlorophyll, and that will really oxygenate your blood.
73
169000
2000
因為深綠色蔬菜有大量葉綠素，可讓血液充滿氧氣
03:06
And anybody who's done school biology remembers
74
171000
2000
但，任何在學校做過生物實驗的人都記得
03:08
that chlorophyll and chloroplasts
75
173000
2000
葉綠素及葉綠粒
03:10
only make oxygen in sunlight,
76
175000
2000
只透過光合作用來製造氧氣
03:12
and it's quite dark in your bowels after you've eaten spinach.
77
177000
3000
而當你把菠菜吃掉時，身體裡其實是黑暗的
03:15
Next, we need proper science, proper evidence.
78
180000
3000
接著，我們需要正確的科學、有理的證據來證實
03:18
So, "Red wine can help prevent breast cancer."
79
183000
2000
「紅酒可預防乳癌。」
03:20
This is a headline from the Daily Telegraph in the U.K.
80
185000
2000
這是英國的《每日電訊》的頭條新聞
03:22
"A glass of red wine a day could help prevent breast cancer."
81
187000
3000
「一天一杯紅酒能幫助預防乳癌。」
03:25
So you go and find this paper, and what you find
82
190000
2000
接著你去找這則新聞，你會發現
03:27
is it is a real piece of science.
83
192000
2000
這確實有科學佐證
03:29
It is a description of the changes in one enzyme
84
194000
3000
這是描述某實驗室的工作臺上的培養皿，
03:32
when you drip a chemical extracted from some red grape skin
85
197000
3000
把紅葡萄皮上所萃取的化學物質，
03:35
onto some cancer cells
86
200000
2000
滴到一些癌細胞上，
03:37
in a dish on a bench in a laboratory somewhere.
87
202000
3000
酵素會產生變化
03:40
And that's a really useful thing to describe
88
205000
2000
就科學報告而言，
03:42
in a scientific paper,
89
207000
2000
這樣的實驗描述很有用，
03:44
but on the question of your own personal risk
90
209000
2000
但，一問到喝紅酒
03:46
of getting breast cancer if you drink red wine,
91
211000
2000
和得到乳癌的風險為何，
03:48
it tells you absolutely bugger all.
92
213000
2000
這樣的實驗描述什麼屁都沒告訴你。
03:50
Actually, it turns out that your risk of breast cancer
93
215000
2000
事實上，你得到乳癌的風險
03:52
actually increases slightly
94
217000
2000
會隨著你喝的含酒精飲料
03:54
with every amount of alcohol that you drink.
95
219000
2000
一點一點的增加
03:56
So what we want is studies in real human people.
96
221000
4000
因此，我們要看的，是拿真人做的研究
04:00
And here's another example.
97
225000
2000
這是另一個例子
04:02
This is from Britain's leading diet and nutritionist in the Daily Mirror,
98
227000
3000
來自英國的《每日鏡報》，一位頗具影響力的飲食營養學家
04:05
which is our second biggest selling newspaper.
99
230000
2000
《每日鏡報》是英國銷售量第二的報紙
04:07
"An Australian study in 2001
100
232000
2000
「2001年，澳洲當地的實驗發現
04:09
found that olive oil in combination with fruits, vegetables and pulses
101
234000
2000
若把橄欖油與水果、蔬菜，或豆類植物搭配食用，
04:11
offers measurable protection against skin wrinklings."
102
236000
2000
將對皮膚防皺有相當顯著效果。」
04:13
And then they give you advice:
103
238000
2000
然後，他們就給妳這樣的建議：
04:15
"If you eat olive oil and vegetables, you'll have fewer skin wrinkles."
104
240000
2000
「如果把橄欖油及蔬菜搭配著吃，你的皺紋會比較少。」
04:17
And they very helpfully tell you how to go and find the paper.
105
242000
2000
他們也很熱心地告訴你文獻出處
04:19
So you go and find the paper, and what you find is an observational study.
106
244000
3000
而在你看完報告後，你就會知道這是份觀測研究
04:22
Obviously nobody has been able
107
247000
2000
很顯然地，沒人有辦法
04:24
to go back to 1930,
108
249000
2000
回到1930年
04:26
get all the people born in one maternity unit,
109
251000
3000
把同個產科所有出生的小孩找來
04:29
and half of them eat lots of fruit and veg and olive oil,
110
254000
2000
其中一半給他們吃水果、蔬菜和橄欖油
04:31
and then half of them eat McDonald's,
111
256000
2000
另一半給他們吃麥當勞
04:33
and then we see how many wrinkles you've got later.
112
258000
2000
之後，看他們的皺紋誰多誰少
04:35
You have to take a snapshot of how people are now.
113
260000
2000
你得拍照記錄人們的樣子
04:37
And what you find is, of course,
114
262000
2000
當然，你會發現到
04:39
people who eat veg and olive oil have fewer skin wrinkles.
115
264000
3000
吃蔬菜和橄欖油的人，確實皺紋較少
04:42
But that's because people who eat fruit and veg and olive oil,
116
267000
3000
但，這是因為那些吃水果、蔬菜，橄欖油的人是怪胎
04:45
they're freaks, they're not normal, they're like you;
117
270000
3000
他們不正常，就像你們一樣
04:48
they come to events like this.
118
273000
2000
他們會出席TED這樣的活動
04:50
They are posh, they're wealthy, they're less likely to have outdoor jobs,
119
275000
3000
他們受歡迎、有錢、也不太可能做室外工作
04:53
they're less likely to do manual labor,
120
278000
2000
也不太從事勞力的工作
04:55
they have better social support, they're less likely to smoke --
121
280000
2000
他們在社會上獲得較多的支持，他們不太抽菸－
04:57
so for a whole host of fascinating, interlocking
122
282000
2000
因此，這麼多關於社會、政治，文化等等原因，
04:59
social, political and cultural reasons,
123
284000
2000
這些原因這麼棒，又如此環環相扣
05:01
they are less likely to have skin wrinkles.
124
286000
2000
而這些人，也就不太可能有皺紋
05:03
That doesn't mean that it's the vegetables or the olive oil.
125
288000
2000
這並不意味著蔬菜還是橄欖油的功勞
05:05
(Laughter)
126
290000
2000
(笑聲)
05:07
So ideally what you want to do is a trial.
127
292000
3000
所以，做實驗往往是最理想的選擇
05:10
And everybody thinks they're very familiar with the idea of a trial.
128
295000
2000
大家都認為，他們對於實驗都得很熟
05:12
Trials are very old. The first trial was in the Bible -- Daniel 1:12.
129
297000
3000
實驗是個老話題。第一個是出現在聖經－但以理書，第一章十二節
05:15
It's very straightforward -- you take a bunch of people, you split them in half,
130
300000
2000
這試驗很簡單，你找一群人，把他們分兩組
05:17
you treat one group one way, you treat the other group the other way,
131
302000
2000
這兩組人，你都用不同的方法去對待他們
05:19
and a little while later, you follow them up
132
304000
2000
過了一下子之後
05:21
and see what happened to each of them.
133
306000
2000
你就看會他們會發生什麼事
05:23
So I'm going to tell you about one trial,
134
308000
2000
現在我要和你說的這個試驗
05:25
which is probably the most well-reported trial
135
310000
2000
在過去十年的英國新聞媒體報導中
05:27
in the U.K. news media over the past decade.
136
312000
2000
可以算是名聲相當不錯的試驗
05:29
And this is the trial of fish oil pills.
137
314000
2000
這是個魚肝油的試驗
05:31
And the claim was fish oil pills improve school performance and behavior
138
316000
2000
魚肝油宣稱，可以改善大部分小孩的學業表現
05:33
in mainstream children.
139
318000
2000
及學校行為
05:35
And they said, "We've done a trial.
140
320000
2000
他們說：「我們做了個試驗
05:37
All the previous trials were positive, and we know this one's gonna be too."
141
322000
2000
而之前的試驗結果都是好的，我想我們的也會是好的。」
05:39
That should always ring alarm bells.
142
324000
2000
會這樣說，就是個警訊
05:41
Because if you already know the answer to your trial, you shouldn't be doing one.
143
326000
3000
當你已經知道實驗的結果，那還做它幹什麼
05:44
Either you've rigged it by design,
144
329000
2000
無論這試驗只是隨意起草
05:46
or you've got enough data so there's no need to randomize people anymore.
145
331000
3000
或是你已有充分的數據，而你已不需要隨機挑選更多的試驗者
05:49
So this is what they were going to do in their trial.
146
334000
3000
實驗就是這樣進行的
05:52
They were taking 3,000 children,
147
337000
2000
他們找了三千位孩童
05:54
they were going to give them all these huge fish oil pills,
148
339000
2000
並給孩童吃很多的魚肝油
05:56
six of them a day,
149
341000
2000
一天吃六顆
05:58
and then a year later, they were going to measure their school exam performance
150
343000
3000
一年後，他們來測量孩童的學業表現
06:01
and compare their school exam performance
151
346000
2000
並把得來的結果和
06:03
against what they predicted their exam performance would have been
152
348000
2000
這些孩童沒有吃魚肝油的預期結果
06:05
if they hadn't had the pills.
153
350000
3000
來兩相比較
06:08
Now can anybody spot a flaw in this design?
154
353000
3000
有人現在看出這實驗設計的缺失了嗎？
06:11
And no professors of clinical trial methodology
155
356000
3000
所有臨床試驗方法的教授
06:14
are allowed to answer this question.
156
359000
2000
都不準回答這問題
06:16
So there's no control; there's no control group.
157
361000
2000
這實驗沒有對照組
06:18
But that sounds really techie.
158
363000
2000
聽起來像電子迷才會做的事
06:20
That's a technical term.
159
365000
2000
那是一個專業術語
06:22
The kids got the pills, and then their performance improved.
160
367000
2000
孩童吃了魚肝油，他們的學業表現進步了
06:24
What else could it possibly be if it wasn't the pills?
161
369000
3000
除了魚肝油，還有什麼其它的原因嗎?
06:27
They got older. We all develop over time.
162
372000
3000
他們隨著時間長大了
06:30
And of course, also there's the placebo effect.
163
375000
2000
當然，還有安慰劑的影響
06:32
The placebo effect is one of the most fascinating things in the whole of medicine.
164
377000
2000
安慰劑的效果是所有藥品中，最令人神往的一個
06:34
It's not just about taking a pill, and your performance and your pain getting better.
165
379000
3000
不是吃藥而已，然後你的表現變好，疼痛舒緩如此而已
06:37
It's about our beliefs and expectations.
166
382000
2000
是我們的信念，以及期望
06:39
It's about the cultural meaning of a treatment.
167
384000
2000
這就是治療的文化定譯
06:41
And this has been demonstrated in a whole raft of fascinating studies
168
386000
3000
許多很好的研究也和安慰劑的效果
06:44
comparing one kind of placebo against another.
169
389000
3000
做了對照實驗
06:47
So we know, for example, that two sugar pills a day
170
392000
2000
舉例來說，一天吃兩顆糖的人
06:49
are a more effective treatment for getting rid of gastric ulcers
171
394000
2000
較不容易得到胃潰瘍
06:51
than one sugar pill.
172
396000
2000
和吃一顆糖的人相比
06:53
Two sugar pills a day beats one sugar pill a day.
173
398000
2000
一天吃兩顆糖比吃一顆糖更有用
06:55
And that's an outrageous and ridiculous finding, but it's true.
174
400000
3000
儘管這個發現讓人覺得奇怪又好笑，但這是事實
06:58
We know from three different studies on three different types of pain
175
403000
2000
從三種不同疼痛的研究當中，我們瞭解
07:00
that a saltwater injection is a more effective treatment for pain
176
405000
3000
注射生理食鹽水來治療疼痛，是個較有效的方法，
07:03
than taking a sugar pill, taking a dummy pill that has no medicine in it --
177
408000
4000
相較於吃糖果，安慰劑...等等沒有任何藥品成分的東西
07:07
not because the injection or the pills do anything physically to the body,
178
412000
3000
而這並非是注射或吃藥對生體產生了什麼影響
07:10
but because an injection feels like a much more dramatic intervention.
179
415000
3000
而是人們覺得注射是種更強力的手段
07:13
So we know that our beliefs and expectations
180
418000
2000
我們曉得自己的信念，期待
07:15
can be manipulated,
181
420000
2000
是可被操控的
07:17
which is why we do trials
182
422000
2000
這也為什麼我們做實驗時
07:19
where we control against a placebo --
183
424000
2000
需要拿安慰劑來對照
07:21
where one half of the people get the real treatment
184
426000
2000
一半的人接受真正的治療
07:23
and the other half get placebo.
185
428000
2000
另一半的人則使用安慰劑
07:25
But that's not enough.
186
430000
3000
但，這樣是不夠的
07:28
What I've just shown you are examples of the very simple and straightforward ways
187
433000
3000
目前為止，我給你們看的實驗例子的方法，都很簡單易懂
07:31
that journalists and food supplement pill peddlers
188
436000
2000
簡單到記者、食物供應商、糖果小販
07:33
and naturopaths
189
438000
2000
以及自然療法醫生
07:35
can distort evidence for their own purposes.
190
440000
3000
都可以根據他們自己的目的，來曲解證據
07:38
What I find really fascinating
191
443000
2000
令我相當驚訝的
07:40
is that the pharmaceutical industry
192
445000
2000
是那些醫藥工業
07:42
uses exactly the same kinds of tricks and devices,
193
447000
2000
他們使用相同的方法、設備
07:44
but slightly more sophisticated versions of them,
194
449000
3000
只不過把證據扭曲成較複雜的版本
07:47
in order to distort the evidence that they give to doctors and patients,
195
452000
3000
而醫生及病人，都根據那些扭曲的證據
07:50
and which we use to make vitally important decisions.
196
455000
3000
做出攸關性命的重要決定
07:53
So firstly, trials against placebo:
197
458000
2000
首先，大家都認為
07:55
everybody thinks they know that a trial should be
198
460000
2000
一個和安慰劑對照的實驗，應該是
07:57
a comparison of your new drug against placebo.
199
462000
2000
拿新藥品和安慰劑來相比較
07:59
But actually in a lot of situations that's wrong.
200
464000
2000
但，在很多情況下，這樣的方法是錯的
08:01
Because often we already have a very good treatment that is currently available,
201
466000
3000
因為，我們通常已經有了一個可行的良好療法
08:04
so we don't want to know that your alternative new treatment
202
469000
2000
我們就不需要知道那可有可無的
08:06
is better than nothing.
203
471000
2000
替代療法
08:08
We want to know that it's better than the best currently available treatment that we have.
204
473000
3000
我們想知道的是，替代療法是否比現行的最佳療法更好
08:11
And yet, repeatedly, you consistently see people doing trials
205
476000
3000
然而，你卻看到人們一再地
08:14
still against placebo.
206
479000
2000
拿安慰劑來對照做實驗
08:16
And you can get license to bring your drug to market
207
481000
2000
只要你得到許可證，你的藥品就可以上市
08:18
with only data showing that it's better than nothing,
208
483000
2000
藥品背後的數據，證明這種藥物可有可無
08:20
which is useless for a doctor like me trying to make a decision.
209
485000
3000
對一個像我這樣需要做決定的醫生來說是沒有任何用處的
08:23
But that's not the only way you can rig your data.
210
488000
2000
而這不是唯一可以扭曲數據的方法
08:25
You can also rig your data
211
490000
2000
你還可以把和新藥品要比較的物品
08:27
by making the thing you compare your new drug against
212
492000
2000
弄得像垃圾一樣
08:29
really rubbish.
213
494000
2000
毫無價值
08:31
You can give the competing drug in too low a dose,
214
496000
2000
在使用競爭者的藥品時，你也可以只給予很少的劑量，
08:33
so that people aren't properly treated.
215
498000
2000
而人們自然沒有辦法治療好
08:35
You can give the competing drug in too high a dose,
216
500000
2000
或是你把劑量提高
08:37
so that people get side effects.
217
502000
2000
這樣就會產生副作用
08:39
And this is exactly what happened
218
504000
2000
而這正是發生在治療人格分裂症的
08:41
which antipsychotic medication for schizophrenia.
219
506000
2000
抗精神病療程的事情
08:43
20 years ago, a new generation of antipsychotic drugs were brought in
220
508000
3000
二十年前，從國外引進了一批新的抗精神病藥物
08:46
and the promise was that they would have fewer side effects.
221
511000
3000
這些藥物都保證副作用比現有的藥物還少
08:49
So people set about doing trials of these new drugs
222
514000
2000
所以人們就開始拿這些新的藥物和舊的藥物
08:51
against the old drugs,
223
516000
2000
做實驗比較
08:53
but they gave the old drugs in ridiculously high doses --
224
518000
2000
但，他們卻把舊的藥物劑量提高很多－
08:55
20 milligrams a day of haloperidol.
225
520000
2000
一天要吃20毫克的施寧錠(抗精神病藥物)
08:57
And it's a foregone conclusion,
226
522000
2000
結論可想而知
08:59
if you give a drug at that high a dose,
227
524000
2000
要是你把一種藥物的劑量開這麼高
09:01
that it will have more side effects and that your new drug will look better.
228
526000
3000
它的副作用當然比新的藥物還要多
09:04
10 years ago, history repeated itself, interestingly,
229
529000
2000
過去十年，歷史不斷重演。有趣的是
09:06
when risperidone, which was the first of the new-generation antipscyhotic drugs,
230
531000
3000
當那一批抗精神病藥品，像是理思必妥
09:09
came off copyright, so anybody could make copies.
231
534000
3000
產權到期時，人人都可以自己做這些藥物
09:12
Everybody wanted to show that their drug was better than risperidone,
232
537000
2000
大家想證明，他們自己做的藥比理思必妥還好
09:14
so you see a bunch of trials comparing new antipsychotic drugs
233
539000
3000
所以你會看到，有一堆的抗精神病藥物
09:17
against risperidone at eight milligrams a day.
234
542000
2000
拿來和一天八毫克的理思必妥做比較
09:19
Again, not an insane dose, not an illegal dose,
235
544000
2000
八毫克不是多得很瘋狂，也沒有超出合法劑量
09:21
but very much at the high end of normal.
236
546000
2000
但幾乎游走在法律邊緣了
09:23
And so you're bound to make your new drug look better.
237
548000
3000
所以，你自己做的藥物當然比那些藥來的好
09:26
And so it's no surprise that overall,
238
551000
3000
果不其然
09:29
industry-funded trials
239
554000
2000
藥商贊助的實驗
09:31
are four times more likely to give a positive result
240
556000
2000
比那些獨立出資的實驗，得到陽性結果
09:33
than independently sponsored trials.
241
558000
3000
機率高出四倍
09:36
But -- and it's a big but --
242
561000
3000
但是，但但但但是
09:39
(Laughter)
243
564000
2000
(笑聲)
09:41
it turns out,
244
566000
2000
最後卻發現
09:43
when you look at the methods used by industry-funded trials,
245
568000
3000
那些藥商贊助實驗所使用的方法
09:46
that they're actually better
246
571000
2000
確實是比獨立出資的實驗
09:48
than independently sponsored trials.
247
573000
2000
要來的好
09:50
And yet, they always manage to to get the result that they want.
248
575000
3000
儘管如此，他們仍設法去得到他們想要的結果
09:53
So how does this work?
249
578000
2000
怎麼做？
09:55
How can we explain this strange phenomenon?
250
580000
3000
這樣如此奇怪的現象該如何解釋？
09:58
Well it turns out that what happens
251
583000
2000
原來，那些負面的數據
10:00
is the negative data goes missing in action;
252
585000
2000
在實驗進行中被隱匿起來。
10:02
it's withheld from doctors and patients.
253
587000
2000
醫生和病人都不知情
10:04
And this is the most important aspect of the whole story.
254
589000
2000
而這正是整件事情的重點所在
10:06
It's at the top of the pyramid of evidence.
255
591000
2000
證據都在這裡
10:08
We need to have all of the data on a particular treatment
256
593000
3000
任一特定的療法，我們需要所有的資訊
10:11
to know whether or not it really is effective.
257
596000
2000
來知道該療法是否有效
10:13
And there are two different ways that you can spot
258
598000
2000
而有兩種不同的方法可以讓你知道
10:15
whether some data has gone missing in action.
259
600000
2000
是否有數據在實驗中不見了
10:17
You can use statistics, or you can use stories.
260
602000
3000
妳可以用統計數字，或是實例
10:20
I personally prefer statistics, so that's what I'm going to do first.
261
605000
2000
我個人偏好統計數字，所以我先講這個
10:22
This is something called funnel plot.
262
607000
2000
這個圖表叫做漏斗圖
10:24
And a funnel plot is a very clever way of spotting
263
609000
2000
漏斗圖的設計非常聰明
10:26
if small negative trials have disappeared, have gone missing in action.
264
611000
3000
要是有少量的負面結果不見的話，你會看得出來
10:29
So this is a graph of all of the trials
265
614000
2000
這圖表上，就是針對某一種醫療法的
10:31
that have been done on a particular treatment.
266
616000
2000
所有實驗
10:33
And as you go up towards the top of the graph,
267
618000
2000
隨著你接近圖表的頂端
10:35
what you see is each dot is a trial.
268
620000
2000
你看到的每一個點，都是一個實驗
10:37
And as you go up, those are the bigger trials, so they've got less error in them.
269
622000
3000
越往上看，那些實驗規模越大，也就越少出錯
10:40
So they're less likely to be randomly false positives, randomly false negatives.
270
625000
3000
也就是越少隨便的假陽性效果、假的陰性效果
10:43
So they all cluster together.
271
628000
2000
他們全聚在一塊
10:45
The big trials are closer to the true answer.
272
630000
2000
規模較大的實驗，都較接近事實
10:47
Then as you go further down at the bottom,
273
632000
2000
而你越往底部看
10:49
what you can see is, over on this side, the spurious false negatives,
274
634000
3000
你就會看到捏造的假陰性結果
10:52
and over on this side, the spurious false positives.
275
637000
2000
而另一邊，你也是看到捏造的假陽性結果
10:54
If there is publication bias,
276
639000
2000
要是有發表篇倚(有統計同的結果更容易被發表)
10:56
if small negative trials have gone missing in action,
277
641000
3000
或是少量的負面結果在過程中消失
10:59
you can see it on one of these graphs.
278
644000
2000
你在這些圖表都會看到
11:01
So you can see here that the small negative trials
279
646000
2000
所以，那些本該在左下角的負面結果
11:03
that should be on the bottom left have disappeared.
280
648000
2000
全都消失了
11:05
This is a graph demonstrating the presence of publication bias
281
650000
3000
這圖表展示的，是關於發表篇倚的研究中
11:08
in studies of publication bias.
282
653000
2000
所出現的發表篇倚
11:10
And I think that's the funniest epidemiology joke
283
655000
2000
我認為那是個流行病學的笑話，而且，我想這是你聽過的笑話中
11:12
that you will ever hear.
284
657000
2000
最好笑的一個
11:14
That's how you can prove it statistically,
285
659000
2000
剛剛說的，就是說明該如何透過統記證明
11:16
but what about stories?
286
661000
2000
那麼實例呢？
11:18
Well they're heinous, they really are.
287
663000
2000
實例就比較可惡了，真的很可惡。
11:20
This is a drug called reboxetine.
288
665000
2000
這藥叫做瑞波西汀（抗憂鬱劑）
11:22
This is a drug that I myself have prescribed to patients.
289
667000
2000
我過去曾開過這種藥給病人
11:24
And I'm a very nerdy doctor.
290
669000
2000
而我是個蠻書呆子的醫生
11:26
I hope I try to go out of my way to try and read and understand all the literature.
291
671000
3000
我盡力去嘗試閱讀，並瞭解所有的文獻
11:29
I read the trials on this. They were all positive. They were all well-conducted.
292
674000
3000
我看了這些實驗，他們都是陽性的，也都進行得很好
11:32
I found no flaw.
293
677000
2000
我沒發現什麼缺失
11:34
Unfortunately, it turned out,
294
679000
2000
不幸地是
11:36
that many of these trials were withheld.
295
681000
2000
許多實驗並沒有公布出來
11:38
In fact, 76 percent
296
683000
2000
事實上，有76%關於這項藥品的實驗
11:40
of all of the trials that were done on this drug
297
685000
2000
並沒有讓醫生
11:42
were withheld from doctors and patients.
298
687000
2000
或是病人知道
11:44
Now if you think about it,
299
689000
2000
現在，請你想一想
11:46
if I tossed a coin a hundred times,
300
691000
2000
要是我擲一枚銅板一百次
11:48
and I'm allowed to withhold from you
301
693000
2000
而其中有一半的答案
11:50
the answers half the times,
302
695000
2000
我都不和你說
11:52
then I can convince you
303
697000
2000
那麼，我就有辦法說服你
11:54
that I have a coin with two heads.
304
699000
2000
這枚銅板兩面都是人頭
11:56
If we remove half of the data,
305
701000
2000
若我們刪除一半的數據
11:58
we can never know what the true effect size of these medicines is.
306
703000
3000
那我們永遠都不曉得，這些藥品實際的效果為何
12:01
And this is not an isolated story.
307
706000
2000
而這類的實例都是息息相關的
12:03
Around half of all of the trial data on antidepressants has been withheld,
308
708000
4000
大約有一半抗憂鬱藥物的實驗至今仍未公佈
12:07
but it goes way beyond that.
309
712000
2000
甚至還超出一半
12:09
The Nordic Cochrane Group were trying to get a hold of the data on that
310
714000
2000
而科克倫小組，正致力於獲得那些數據
12:11
to bring it all together.
311
716000
2000
來整合所有的事情
12:13
The Cochrane Groups are an international nonprofit collaboration
312
718000
3000
科克倫小組是個國際性，非營利的合作機構
12:16
that produce systematic reviews of all of the data that has ever been shown.
313
721000
3000
他們有系統地提供過去的數據回顧
12:19
And they need to have access to all of the trial data.
314
724000
3000
而他們需要管道，來獲得所有的數據
12:22
But the companies withheld that data from them,
315
727000
3000
但這些公司不讓他們知道那些數據
12:25
and so did the European Medicines Agency
316
730000
2000
歐洲藥物管理局也是如此
12:27
for three years.
317
732000
2000
已經達三年了
12:29
This is a problem that is currently lacking a solution.
318
734000
3000
這個問題現在仍少個解決方法
12:32
And to show how big it goes, this is a drug called Tamiflu,
319
737000
3000
要讓你們知道問題的嚴重性。來看看克流感
12:35
which governments around the world
320
740000
2000
世界上許多政府
12:37
have spent billions and billions of dollars on.
321
742000
2000
花了數十億的金錢在克流感上
12:39
And they spend that money on the promise
322
744000
2000
這些政府都承諾
12:41
that this is a drug which will reduce the rate
323
746000
2000
克流感可以降低
12:43
of complications with flu.
324
748000
2000
流感併發症的比率
12:45
We already have the data
325
750000
2000
我們既有的數據顯示
12:47
showing that it reduces the duration of your flu by a few hours.
326
752000
2000
克流感是降低流感的持續時間幾小時
12:49
But I don't really care about that. Governments don't care about that.
327
754000
2000
但我不在乎這個，政府也不在乎
12:51
I'm very sorry if you have the flu, I know it's horrible,
328
756000
3000
如果你得到流感，我很抱歉，我知道那很難受
12:54
but we're not going to spend billions of dollars
329
759000
2000
但，我們絕不會花了幾十億
12:56
trying to reduce the duration of your flu symptoms
330
761000
2000
只是要去縮短流感症狀的持續時間而已
12:58
by half a day.
331
763000
2000
還只減少個半天左右
13:00
We prescribe these drugs, we stockpile them for emergencies
332
765000
2000
我們會開這些藥，也會貯存他們為了緊急狀況
13:02
on the understanding that they will reduce the number of complications,
333
767000
2000
因為我們瞭解，他們可以降低併發症可能性
13:04
which means pneumonia and which means death.
334
769000
3000
併發症可能是肺炎，或是死亡
13:07
The infectious diseases Cochrane Group, which are based in Italy,
335
772000
3000
以義大利為根基的科克倫小組，正努力從藥品公司
13:10
has been trying to get
336
775000
2000
想辦法獲得感染疾病的完整數據
13:12
the full data in a usable form out of the drug companies
337
777000
3000
而這些數據都是正確可用的
13:15
so that they can make a full decision
338
780000
3000
如此一來，科克倫小組才能決定
13:18
about whether this drug is effective or not,
339
783000
2000
一個藥品有效與否
13:20
and they've not been able to get that information.
340
785000
3000
然而，他們卻無法獲得那樣的資訊
13:23
This is undoubtedly
341
788000
2000
無庸置疑
13:25
the single biggest ethical problem
342
790000
3000
照是現藥品面臨的
13:28
facing medicine today.
343
793000
2000
一個嚴重的道德問題
13:30
We cannot make decisions
344
795000
3000
若沒有這些資訊
13:33
in the absence of all of the information.
345
798000
4000
我們便沒有辦法下決定
13:37
So it's a little bit difficult from there
346
802000
3000
儘管，現在要下個有希望的結論
13:40
to spin in some kind of positive conclusion.
347
805000
4000
是有些困難
13:44
But I would say this:
348
809000
4000
但，我會說
13:48
I think that sunlight
349
813000
3000
我認為，陽光
13:51
is the best disinfectant.
350
816000
2000
是最好的消毒劑
13:53
All of these things are happening in plain sight,
351
818000
3000
我們眼前看到所發生的事物
13:56
and they're all protected
352
821000
2000
他們都被我們習以為常的陽光
13:58
by a force field of tediousness.
353
823000
3000
給保護著
14:01
And I think, with all of the problems in science,
354
826000
2000
儘管現今科學界存在著這些問題
14:03
one of the best things that we can do
355
828000
2000
我認為，我們最好能做的一件事
14:05
is to lift up the lid,
356
830000
2000
就是掀開事物的表層
14:07
finger around in the mechanics and peer in.
357
832000
2000
仔細探查，檢視一番
14:09
Thank you very much.
358
834000
2000
謝謝你們
14:11
(Applause)
359
836000
3000
(掌聲)
ABOUT THE SPEAKER
Ben Goldacre - DebunkerBen Goldacre unpicks dodgy scientific claims made by scaremongering journalists, dubious government reports, pharmaceutical corporations, PR companies and quacks.
Why you should listen
"It was the MMR story that finally made me crack," begins the Bad Science manifesto, referring to the sensationalized -- and now-refuted -- link between vaccines and autism. With that sentence Ben Goldacre fired the starting shot of a crusade waged from the pages of The Guardian from 2003 to 2011, on an addicitve Twitter feed, and in bestselling books, including Bad Science and his latest, Bad Pharma, which puts the $600 billion global pharmaceutical industry under the microscope. What he reveals is a fascinating, terrifying mess.
Goldacre was trained in medicine at Oxford and London, and works as an academic in epidemiology. Helped along by this inexhaustible supply of material, he also travels the speaking circuit, promoting skepticism and nerdish curiosity with fire, wit, fast delivery and a lovable kind of exasperation. (He might even convince you that real science, sober reporting and reason are going to win in the end.)
As he writes, "If you're a journalist who misrepresents science for the sake of a headline, a politician more interested in spin than evidence, or an advertiser who loves pictures of molecules in little white coats, then beware: your days are numbered."
Ben Goldacre | Speaker | TED.com