What really happens when you mix medications? | Russ Altman

翻译人员: Elvis Liu 校对人员: Tom Liu

So you go to the doctor and get some tests.

你去看医生的时候做了一些检查

The doctor determines that you have high cholesterol

医生说你的血脂高

and you would benefit from medication to treat it.

所以吃药会有帮助

So you get a pillbox.

那么你就买了一盒药

You have some confidence,

你有信心，

your physician has some confidence that this is going to work.

你的医生也有信心 觉得这会对你有帮助

The company that invented it did a lot of studies, submitted it to the FDA.

发明这个的药物公司做过很多研究 呈送到FDA

They studied it very carefully, skeptically, they approved it.

他们充满怀疑地研究，后来许可了。

They have a rough idea of how it works,

他们大概地知道这个药的机理

they have a rough idea of what the side effects are.

和药物的副作用

It should be OK.

它应该可以

You have a little more of a conversation with your physician

你跟你的医生又谈了一会儿

and the physician is a little worried because you've been blue,

医生对你有点儿担心 因为你有些抑郁

haven't felt like yourself,

感觉你不是自己

you haven't been able to enjoy things in life quite as much as you usually do.

你对生活不像以前那么充满兴趣

Your physician says, "You know, I think you have some depression.

你的医生说，“你知道吗， 我觉得你有些抑郁。

I'm going to have to give you another pill."

我会给你另一种药“

So now we're talking about two medications.

所以 现在我们所谈论的是两种药物

This pill also -- millions of people have taken it,

这种药－－上百万人服用过

the company did studies, the FDA looked at it -- all good.

公司做过很多研究，FDA许可的－－ 不会错

Think things should go OK.

你会想应该没问题

Think things should go OK.

应该没问题

Well, wait a minute.

等等

How much have we studied these two together?

我们知道两种药物一起服用的研究吗

Well, it's very hard to do that.

而这是很难做到的。

In fact, it's not traditionally done.

实际上 还从来没有

We totally depend on what we call "post-marketing surveillance,"

在药物上市之后

我们完全地依赖于我们叫做 “市场后监测”

after the drugs hit the market.

How can we figure out if bad things are happening

我们如何能够弄清楚

在两种，三种或五种药物

between two medications?

混合之后 会有哪些坏处呢？

Three? Five? Seven?

Ask your favorite person who has several diagnoses

问问你最喜欢的被诊断了 几个不同疾病的人

how many medications they're on.

他们在吃多少种药

Why do I care about this problem?

我为什么关心这个问题呢？

我对这非常在意

I care about it deeply.

我是信息和数据科学家 真的，以我的意见来说

I'm an informatics and data science guy and really, in my opinion,

唯一的能够理解这些药物相互作用 的希望

the only hope -- only hope -- to understand these interactions

is to leverage lots of different sources of data

就是平衡不同来源的数据

以便弄清楚药物在一起什么时候安全

in order to figure out when drugs can be used together safely

什么时候不安全

and when it's not so safe.

So let me tell you a data science story.

让我来告诉你们一些数据科学的故事

And it begins with my student Nick.

它来自于我的学生尼克

Let's call him "Nick," because that's his name.

我们叫他“尼克” 因为那是他的名字

笑声

(Laughter)

Nick was a young student.

尼克是一个年轻学生

我说，“你知道吗，尼克 我们需要理解药物的工作机理

I said, "You know, Nick, we have to understand how drugs work

不仅是它们单独作用 还有它们的协同机理

and how they work together and how they work separately,

and we don't have a great understanding.

目前我们还知道得不多

But the FDA has made available an amazing database.

但FDA已经提供了一个惊人的数据库

是关于反作用事件的数据库

It's a database of adverse events.

They literally put on the web --

他们发表在互联网上－－

publicly available, you could all download it right now --

可公开使用 你现在就可以下载

成千上万例的

hundreds of thousands of adverse event reports

从病人，医生，公司， 药房的反作用报告

from patients, doctors, companies, pharmacists.

And these reports are pretty simple:

这些报告很简单：

it has all the diseases that the patient has,

它具有病人所有的疾病

他们在用的所有药物

all the drugs that they're on,

and all the adverse events, or side effects, that they experience.

以及他们所经过的所有的 反作用、副作用

It is not all of the adverse events that are occurring in America today,

它还不是所有的在美國发生的 反作用事件

但它有成百上千种药物

but it's hundreds and hundreds of thousands of drugs.

So I said to Nick,

所以我对尼克说

“让我们考虑葡萄糖

"Let's think about glucose.

Glucose is very important, and we know it's involved with diabetes.

血糖很重要 我们知道它和糖尿病相关

让我们看看是否理解血糖的反应

Let's see if we can understand glucose response.

I sent Nick off. Nick came back.

我让尼克去做了。尼克又回来了。

”罗斯，“ 他说

"Russ," he said,

”我已经根据这个数据库 创建了一个分类

"I've created a classifier that can look at the side effects of a drug

可以查看一个药物的副作用

based on looking at this database,

并告诉你这个药物是否会改变血糖。“

and can tell you whether that drug is likely to change glucose or not."

He did it. It was very simple, in a way.

他这样做了，而且在他来说很简单

He took all the drugs that were known to change glucose

他把所有我们知道会改变血糖的药物

and a bunch of drugs that don't change glucose,

还有很多不会改变血糖的都分了类

and said, "What's the difference in their side effects?

他说，“他们的副作用有不同吗？”

Differences in fatigue? In appetite? In urination habits?"

在疲劳？胃口？ 以及排尿习惯方面有不同吗？

All those things conspired to give him a really good predictor.

所有这些指标都给予他 一个很好的预测

He said, "Russ, I can predict with 93 percent accuracy

他说，“罗斯， 我能以93%的精确度预测

一个药物会改变血糖”

when a drug will change glucose."

I said, "Nick, that's great."

我说，“尼克， 那很好。”

他是一个年轻的学生 你得帮他建立自信

He's a young student, you have to build his confidence.

“但是尼克， 有一个问题

"But Nick, there's a problem.

It's that every physician in the world knows all the drugs that change glucose,

要让这世界上的每一个医生都知道 所有改变血糖的药物

because it's core to our practice.

这是我们作业的核心

So it's great, good job, but not really that interesting,

很好 做的好 但并不是那么有趣

definitely not publishable."

绝对不能发表

（笑声）

(Laughter)

他说，“我知道，罗斯 我知道你会那样说。”

He said, "I know, Russ. I thought you might say that."

尼克很聪明

Nick is smart.

“我想到你会这么说， 所以我做了另一个试验。

"I thought you might say that, so I did one other experiment.

我查看了在这个数据中 服用两种药物的病人，

I looked at people in this database who were on two drugs,

and I looked for signals similar, glucose-changing signals,

我查看相似的，血糖改变信号，

for people taking two drugs,

对那些服用两种药物的人，

一种药物本身不改变血糖，

where each drug alone did not change glucose,

but together I saw a strong signal."

但一起的时候，我看到了很强的信号。“

And I said, "Oh! You're clever. Good idea. Show me the list."

我说，”喔！你真聪明。 好主意。给我看看列表。“

And there's a bunch of drugs, not very exciting.

有很多药物 但并不令人兴奋

But what caught my eye was, on the list there were two drugs:

但在列表上有两种药物 吸引了我的眼球：

paroxetine, or Paxil, an antidepressant;

paroxetine, 或 Paxil， 一种抗抑郁药物

and pravastatin, or Pravachol, a cholesterol medication.

和pravastatin, 或Pravachol， 抗胆固醇药物

And I said, "Huh. There are millions of Americans on those two drugs."

然后我说，“呵 上百万的美国人都在用这两种药物。”

In fact, we learned later,

事实上，我们后来知道，

15 million Americans on paroxetine at the time, 15 million on pravastatin,

一千五百万的美国人在用paroxetine 而同时 一千五百万人服用pravastatin

and a million, we estimated, on both.

我们估计 有一百万人 两者同时服用

So that's a million people

所以是一百万人

可能在血糖上会有问题

who might be having some problems with their glucose

if this machine-learning mumbo jumbo that he did in the FDA database

但会不会他在FDA数据库 的异想天开

actually holds up.

只是瞎猫碰上了死老鼠呢？

But I said, "It's still not publishable,

但我说，“还是不能发表。”

because I love what you did with the mumbo jumbo,

因为我喜欢你用搜索技术

所做出来的奇思妙想

with the machine learning,

but it's not really standard-of-proof evidence that we have."

但它不是我们真正的标准证据

So we have to do something else.

所以我们必须再做些其他的

让我们进入斯坦福的医疗记录电子库

Let's go into the Stanford electronic medical record.

我们有拷贝权，搜索时许可的

We have a copy of it that's OK for research,

我们挪开了个人信息

we removed identifying information.

And I said, "Let's see if people on these two drugs

然后我说， “让我们看看同时服用这两种药物的人

和他们的血糖问题。”

have problems with their glucose."

在斯坦福的医疗记录里 有成千上万的人

Now there are thousands and thousands of people

in the Stanford medical records that take paroxetine and pravastatin.

在服用paroxetine and pravastatin

But we needed special patients.

但我们需要很特别的病人

We needed patients who were on one of them and had a glucose measurement,

我们需要服用其中一种药物的病人 有血糖纪录

then got the second one and had another glucose measurement,

然后在服用第二种以后 有另一次血糖纪录

all within a reasonable period of time -- something like two months.

并且是在一个比较合理的阶段以内 比如像两个月

当我们这样做以后 我们发现了10个病人

And when we did that, we found 10 patients.

然而，10个中有8个在血糖上有变化

However, eight out of the 10 had a bump in their glucose

when they got the second P -- we call this P and P --

当他们服用第二个P药物的时候 我们把这个叫做P和P--

when they got the second P.

当他们服用第二个P时

可以是任意一个在先 第二个服用后

Either one could be first, the second one comes up,

glucose went up 20 milligrams per deciliter.

血糖升高了20mg／dl

Just as a reminder,

给一个小小的提示

you walk around normally, if you're not diabetic,

当你没有糖尿病 正常的四处活动时

你的血糖是90

with a glucose of around 90.

And if it gets up to 120, 125,

如果升高到120， 125，

your doctor begins to think about a potential diagnosis of diabetes.

你的医生会认为是潜在的糖尿病。

所以 一个20 的升高－－太明显了。

So a 20 bump -- pretty significant.

I said, "Nick, this is very cool.

我说，“尼克，这太好了。

但很抱歉，我们还是没有文章

But, I'm sorry, we still don't have a paper,

because this is 10 patients and -- give me a break --

因为这是10个病人，而且 让我想想

没有足够的病人。“

it's not enough patients."

So we said, what can we do?

所以我们说 我们还能怎么做呢？

And we said, let's call our friends at Harvard and Vanderbilt,

后来我们决定打电话 给我们在Harvard和Vanderbilt的朋友

who also -- Harvard in Boston, Vanderbilt in Nashville,

在波士顿的哈佛和纳什维尔的 范德比尔

who also have electronic medical records similar to ours.

也都有和我们相似的医疗电子记录

Let's see if they can find similar patients

我们想看看他们是否能够找到 相似的病人

with the one P, the other P, the glucose measurements

服用一种P， 然后另一种P

并在我们需要的那个范围内 做过血糖检测

in that range that we need.

God bless them, Vanderbilt in one week found 40 such patients,

上帝祝福他们。范德贝尔 在一周内发现40个这样的病人

same trend.

都有同样的血糖增长

Harvard found 100 patients, same trend.

哈佛发现100个同样的病人， 也有着一样的增长

So at the end, we had 150 patients from three diverse medical centers

所以，最后我们有150个病人 来自三个不同的的医学中心

that were telling us that patients getting these two drugs

这150个病人的记录告诉我们 这些使用这两种药物的病人

在某种程度上都有血糖的明显改变

were having their glucose bump somewhat significantly.

更令人感兴趣的是 我们没有算上糖尿病人

More interestingly, we had left out diabetics,

因为糖尿病人的血糖本身就是 一本糊涂账

because diabetics already have messed up glucose.

When we looked at the glucose of diabetics,

当我们查看糖尿病人的血糖

it was going up 60 milligrams per deciliter, not just 20.

它通常是升高60mg以上 而不是只有20

This was a big deal, and we said, "We've got to publish this."

这是一个了不起的结果。然后我们说， “我们一定要发表这个结果。”

我们呈送了文章

We submitted the paper.

It was all data evidence,

全是数据证据

data from the FDA, data from Stanford,

来自FDA，来自斯坦福

来自范德贝尔，来自哈佛

data from Vanderbilt, data from Harvard.

我们还没做一个实验

We had not done a single real experiment.

But we were nervous.

但我们很紧张

So Nick, while the paper was in review, went to the lab.

所以当文章在审查阶段 尼克去了实验室

We found somebody who knew about lab stuff.

我们找到了一些懂得实验的人

I don't do that.

我做不了那个活

I take care of patients, but I don't do pipettes.

我看病人 我不用移液器

They taught us how to feed mice drugs.

他们教我们怎样喂老鼠吃药

We took mice and we gave them one P, paroxetine.

我们拿过老鼠 给它们喂一种P paroxetine

We gave some other mice pravastatin.

我们又给某些老鼠pravastatin.

And we gave a third group of mice both of them.

我们给了第三组老鼠两种药

And lo and behold, glucose went up 20 to 60 milligrams per deciliter

老鼠的血糖

升高了20-60毫克／分升

in the mice.

So the paper was accepted based on the informatics evidence alone,

所以 基于尽有信息考据的文章 被接受了

but we added a little note at the end,

但是我门在文章的结尾 加上了一个小小的注解

顺便说一下 如果你给老鼠喂两种药 血糖会升高

saying, oh by the way, if you give these to mice, it goes up.

这太棒了 故事在此应该了结了

That was great, and the story could have ended there.

But I still have six and a half minutes.

但我还要讲六分半钟

(Laughter)

笑声

So we were sitting around thinking about all of this,

当我们坐在一起 想着这件事时

我记不得是谁说的了 但有人说：

and I don't remember who thought of it, but somebody said,

“我好奇那些服用 这两种药的病人

"I wonder if patients who are taking these two drugs

是否注意到自己有高血糖的症状

are noticing side effects of hyperglycemia.

They could and they should.

他们理应注意到的

How would we ever determine that?"

我们又怎样确定他们 是否真有呢

We said, well, what do you do?

我们说：那你怎么做呢？

You're taking a medication, one new medication or two,

“如果你在服用一种新药 或者是两种

and you get a funny feeling.

然后你有了一种奇怪的感觉

你会怎么做？

What do you do?

You go to Google

你会在谷歌上查找

and type in the two drugs you're taking or the one drug you're taking,

你会在搜索栏上打出 两种药物的名称

and you type in "side effects."

然后输入”副作用“

What are you experiencing?

你觉得这想法怎么样？”

So we said OK,

于是我们说还不错

let's ask Google if they will share their search logs with us,

我们可以试着问问谷歌 他们能不能与我们分享搜索记录

so that we can look at the search logs

然后我们可以通过这些搜索记录

and see if patients are doing these kinds of searches.

进而知道病人是否在做这种搜索

Google, I am sorry to say, denied our request.

很遗憾的是，谷歌拒绝了我们的请求

So I was bummed.

于是我有点闷闷不乐

I was at a dinner with a colleague who works at Microsoft Research

当时我在和一个微软公司的同事吃饭

我说：”我们想要做一个调查，

and I said, "We wanted to do this study,

但谷歌拒绝了，这真令人烦恼”

Google said no, it's kind of a bummer."

他说“哦，我们有必应bing搜索啊”

He said, "Well, we have the Bing searches."

(Laughter)

（笑声）

Yeah.

是的

That's great.

这太棒了

我感觉我就像要...了一样

Now I felt like I was --

(Laughter)

（笑声）

I felt like I was talking to Nick again.

我感觉我就像又在和尼克说话了

He works for one of the largest companies in the world,

他在全世界最大的公司工作

我不想伤害他的自信心

and I'm already trying to make him feel better.

但他说：“不，罗斯... 你可能不知道

But he said, "No, Russ -- you might not understand.

我们不只有必应bing

We not only have Bing searches,

但是如果你用IE浏览器 在谷歌上搜索词条

but if you use Internet Explorer to do searches at Google,

Yahoo, Bing, any ...

或是在雅虎，bing上

Then, for 18 months, we keep that data for research purposes only."

然后我们将这些搜索信息 为了学术目的自动保存18个月

I said, "Now you're talking!"

我于是说：”你真有两下子！“

他叫 Eric Horvitz，我在微软的朋友

This was Eric Horvitz, my friend at Microsoft.

So we did a study

因此我们就这样做了调查

where we defined 50 words that a regular person might type in

我们先确定了高血糖症患者

if they're having hyperglycemia,

可能会搜索的50个词条

like "fatigue," "loss of appetite," "urinating a lot," "peeing a lot" --

比如”疲劳“”食欲不振“”尿频“等

forgive me, but that's one of the things you might type in.

不好意思，但这些是你可能输入的词语

于是我们有了50个 叫做“肥胖词语”的词条

So we had 50 phrases that we called the "diabetes words."

And we did first a baseline.

我们先是确定了基线搜索率

And it turns out that about .5 to one percent

大概0.5-1%的网络搜索

of all searches on the Internet involve one of those words.

含有一个这些词语

So that's our baseline rate.

这就是我们的底线比率

If people type in "paroxetine" or "Paxil" -- those are synonyms --

如果人们输入paroxetine或Paxil —它们是同义词—

and one of those words,

它们其中的一个

the rate goes up to about two percent of diabetes-type words,

那么如果搜索者已经知道了 这个药物术语的话

if you already know that there's that "paroxetine" word.

则在肥胖类内容的搜索中 它们出现的概率升高到了大约2%

如果是pravastatin 概率则超过了基线3%

If it's "pravastatin," the rate goes up to about three percent from the baseline.

If both "paroxetine" and "pravastatin" are present in the query,

如果paroxetine和pravastatin同时出现

it goes up to 10 percent,

那么比例则到达了10%

a huge three- to four-fold increase

这是在那些肥胖类或高血糖类 搜索中

出现我们研究的两种药物的概率的

in those searches with the two drugs that we were interested in,

三至四倍的增长

and diabetes-type words or hyperglycemia-type words.

我们发表了这个结果

We published this,

and it got some attention.

获取了一些注意

The reason it deserves attention

这个研究值得注意的原因是

is that patients are telling us their side effects indirectly

病人在通过他们的网上搜索

向我们间接地传达他们的副作用

through their searches.

We brought this to the attention of the FDA.

我们吸引了FDA的注意

They were interested.

他们很感兴趣

They have set up social media surveillance programs

他们建立了社交网站监测项目

to collaborate with Microsoft,

和有着可以完成这些项目的设施的

which had a nice infrastructure for doing this, and others,

微软合作

to look at Twitter feeds,

在推特网页上

脸书上

to look at Facebook feeds,

观察人们的搜索内容

to look at search logs,

to try to see early signs that drugs, either individually or together,

以此来发现一种或多种药物 可能在产生问题的

are causing problems.

早期迹象

What do I take from this? Why tell this story?

那么我们由此学到了什么？ 为什么讲这个故事？

第一

Well, first of all,

我们现在有了大数据的支持

we have now the promise of big data and medium-sized data

to help us understand drug interactions

来帮助我们了解药物的相互作用

and really, fundamentally, drug actions.

更本上就是药物的机理

How do drugs work?

药物是怎样起效的？

这已经创造了一种新的系统

This will create and has created a new ecosystem

来了解药物的工作原理 以及优化它们的使用

for understanding how drugs work and to optimize their use.

Nick went on; he's a professor at Columbia now.

尼克继续从事着这事 他现在是哥伦比亚大学的教授

He did this in his PhD for hundreds of pairs of drugs.

他在他的PhD中研究了 成百对的药物

他发现了几种十分重要的药物反应

He found several very important interactions,

and so we replicated this

于是我们记录了这些结果

and we showed that this is a way that really works

而且我们展示了这种方法

for finding drug-drug interactions.

在发现药物相互作用上的可行性

However, there's a couple of things.

然而，这有几件事

我们不只是研究一对药物

We don't just use pairs of drugs at a time.

As I said before, there are patients on three, five, seven, nine drugs.

像我之前说的，有的人同时服用 3.5.7.9种药物

Have they been studied with respect to their nine-way interaction?

他们的九种药物反应有被研究过吗？

Yes, we can do pair-wise, A and B, A and C, A and D,

是的，我们确实可以用排列组合 a和b,a和c,a和d

but what about A, B, C, D, E, F, G all together,

但如果是a,b,c,d,e,f,g全部混在一起呢？

它们被同一个患者服用

being taken by the same patient,

perhaps interacting with each other

可能会和对方反应

有可能是让药效增强或是减弱

in ways that either makes them more effective or less effective

or causes side effects that are unexpected?

更甚是始料不及的副作用？

我们真不知道

We really have no idea.

It's a blue sky, open field for us to use data

我们可以很自由地使用数据

to try to understand the interaction of drugs.

来了解药物的协同机理

Two more lessons:

另外两个教训：

我想让你们想想 我们使用人们

I want you to think about the power that we were able to generate

通过他们的药师，医生或是自己 上传的药物反作用案例

with the data from people who had volunteered their adverse reactions

那些为斯坦福，哈佛和范德比尔数据库 提供了资料的案例

through their pharmacists, through themselves, through their doctors,

来用作研究

the people who allowed the databases at Stanford, Harvard, Vanderbilt,

能够产生的力量有多大

to be used for research.

People are worried about data.

人们担心自己的数据被泄露

They're worried about their privacy and security -- they should be.

他们害怕自己的隐私和信息安全被偷取 --他们理应这样想

因此我们需要安全的网络系统

We need secure systems.

But we can't have a system that closes that data off,

但是我们不应该容忍那些 垄断这些数据的网络系统

因为网络资源是在药理方面

because it is too rich of a source

of inspiration, innovation and discovery

创造灵感，创新和发现的

for new things in medicine.

强大资源

我最后想说的是

And the final thing I want to say is,

在这个案例中 我们发现了两种药物，十分遗憾

in this case we found two drugs and it was a little bit of a sad story.

The two drugs actually caused problems.

这两种药物实际上产生了麻烦

They increased glucose.

它们增加血糖含量

它们可能让 原本没有糖尿病的人

They could throw somebody into diabetes

who would otherwise not be in diabetes,

患上糖尿病

and so you would want to use the two drugs very carefully together,

所以当你同时使用这两种药时 会千万小心

分开用时也是

perhaps not together,

订购药物时做出其他选择

make different choices when you're prescribing.

But there was another possibility.

但也有另一种可能

We could have found two drugs or three drugs

我们可能可以发现 二至三种药物

that were interacting in a beneficial way.

能通过有益的方式相互反应

We could have found new effects of drugs

我们也可以发现药物的新作用

单独不具有的

that neither of them has alone,

but together, instead of causing a side effect,

但是在一起服用， 不是产生副作用

而是成为一种新型治疗手段

they could be a new and novel treatment

for diseases that don't have treatments

治疗那些无药可医的病症

或是旧的治疗方法效果不明显的疾病

or where the treatments are not effective.

如果我们今天纵观药物治疗

If we think about drug treatment today,

all the major breakthroughs --

所有的重大突破--

for HIV, for tuberculosis, for depression, for diabetes --

治疗艾滋病，肺结核，抑郁症 或是糖尿病的--

it's always a cocktail of drugs.

都是几种药物的混合疗法

And so the upside here,

所以我们目前所做的

and the subject for a different TED Talk on a different day,

也是TED大会今后探讨的话题

is how can we use the same data sources

就是我们怎样使用同样的数据资源

来寻找药物混合使用后的好处

to find good effects of drugs in combination

that will provide us new treatments,

这将会为我们提供新的疗法

new insights into how drugs work

药物工作原理的新视角

and enable us to take care of our patients even better?

使我们可以更好地治愈我们的病人

Thank you very much.

十分感谢

(Applause)

掌声

Subtitled by:治洋 Liu