What really happens when you mix medications? | Russ Altman

譯者: 易帆 余 審譯者: Adrienne Lin

So you go to the doctor and get some tests.

你去看醫生，接受了一些檢查。

The doctor determines that you have high cholesterol

醫生診斷出你的膽固醇過高，

and you would benefit from medication to treat it.

建議你服藥治療可能有幫助。

So you get a pillbox.

所以，你拿到了藥罐子。

You have some confidence,

你有點信心，

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

你的醫師也有信心，認為這藥會有效。

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

發明這個藥的公司做了很多的研究， 然後呈送給食品藥物管理局。

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

他們很仔細、審慎地研究， 並核准了這藥物上市。

They have a rough idea of how it works,

他們大概知道這藥物如何運作，

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

也大略知道會有什麼副作用，

It should be OK.

應該沒問題。

You have a little more of a conversation with your physician

你跟醫師又多聊了一會，

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

而醫師有點擔心，因為你很憂鬱，

haven't felt like yourself,

精神欠佳。

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

無法像平常一樣盡情享受生活點滴。

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

你的醫師說： 「我認為你有一點精神憂鬱，

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

我再開個藥給你。」

So now we're talking about two medications.

所以，我們現在有兩種藥了。

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

這個藥也有好幾百萬人服用過，

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

公司做了研究，食品藥物管理局 也檢查過，全部都沒問題。

Think things should go OK.

想一下，這東西沒問題，OK的。

Think things should go OK.

想一下，這東西沒問題，OK的。

Well, wait a minute.

但，請等一下。

How much have we studied these two together?

我們對這兩種藥混在一起吃 做了多少研究？

Well, it's very hard to do that.

其實，這很難評估。

In fact, it's not traditionally done.

事實上，傳統上都不會做。

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

在藥物上市後，我們完全倚賴一種

叫做「上市後監察系統」的機制，

after the drugs hit the market.

How can we figure out if bad things are happening

我們要如何確認，兩種藥之間

between two medications?

是否有什麼不好的事會發生？

Three? Five? Seven?

三種？五種？七種呢？

Ask your favorite person who has several diagnoses

問你身邊有各種疾病在身的人，

how many medications they're on.

他們正在吃多少藥。

Why do I care about this problem?

為什麼我在乎這個問題？

I care about it deeply.

我非常在乎。

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

我是念資訊和數據科學的人， 真的，在我看來，

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

了解藥物彼此間的交互影響 唯一的希望只有

is to leverage lots of different sources of data

運用不同來源的龐大資料，

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

才能找出這些藥 何時可以安全地一起服用，

and when it's not so safe.

以及何時不行。

So let me tell you a data science story.

所以，讓我來告訴各位 一個數據科學的故事。

And it begins with my student Nick.

這要從我的學生尼克開始講起。

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

我們就稱呼他為尼克吧， 因為那就是他的本名。

(Laughter)

（笑聲）

Nick was a young student.

尼克很年輕，

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

我說：「尼克， 我們必須了解藥物如何運作，

and how they work together and how they work separately,

以及藥物在一起會如何運作、 分開會如何運作，

and we don't have a great understanding.

而我們並沒有了解很深。」

But the FDA has made available an amazing database.

但食品藥物管理局已經 有一個很驚人的資料庫，

It's a database of adverse events.

是一個藥物不良反應通報資料庫。

They literally put on the web --

資料真的直接放在網路上

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

供大眾查詢，你現在就可以全部下載，

hundreds of thousands of adverse event reports

從病人、醫生、公司、藥劑師通報上來

from patients, doctors, companies, pharmacists.

好幾百萬個的藥物不良反應通報。

And these reports are pretty simple:

這些報告都相當簡單：

it has all the diseases that the patient has,

上面有病人所有疾病

all the drugs that they're on,

及所有藥物的使用狀況，

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

還有他們經歷過的 所有不良反應事件或副作用。

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

雖然沒有現今在美國 發生的所有不良反應事件，

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

但卻有上百萬種藥物資科。

So I said to Nick,

所以，我跟尼克說：

"Let's think about glucose.

「我們來想一想葡萄糖。

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

葡萄糖非常重要，而且 大家都知道它與糖尿病有關。

Let's see if we can understand glucose response.

讓我們來看看是否可以 了解葡萄糖的反應。」

I sent Nick off. Nick came back.

我請尼克去找資料，

"Russ," he said,

他回來後說：「洛斯，

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

我已經建造了一個分辨器， 可以透過這個資料庫

based on looking at this database,

來檢視一種藥物的副作用，

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

而且還可以告訴你，這個藥 會否改變病人血糖狀況。」

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

他用一個方法做到了，很簡單。

He took all the drugs that were known to change glucose

他把所有已知會改變葡萄糖的藥物

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

及所有不會改變的藥物拿出來做比較，

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

「它們之間的副作用有什麼分別？

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

疲勞狀況上的差異？食慾上的差異？ 排尿習慣上的差異？」

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

所有這些事情都可以協助他 做出一個很棒的預測器。

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

他說：「洛斯，我能預測 哪種藥可改變血糖，

when a drug will change glucose."

準確率可以高達93%。」

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

我說：「尼克，這太棒了！」 他是個年輕的學生，

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

你必須建立他的信心。

"But Nick, there's a problem.

「但，尼克，有一個問題。

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

就是全世界的醫師都知道 這些藥會改變葡萄糖，

because it's core to our practice.

因為這是我們實務上的核心。

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

所以，你很棒，幹得好， 但並沒有人對這有興趣，

definitely not publishable."

絕對還不適合公布你的研究結果。」

(Laughter)

（笑聲 ）

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

他說：「我知道，洛斯。 我知道你可能會這麼說。」

Nick is smart.

尼克很聰明。

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

「我知道你會這麼說， 所以我多做了另一項實驗。

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

我仔細觀察資料庫裡 同時服用兩種藥的人，

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

然後尋找他們之間

葡萄糖改變的相似訊號，

for people taking two drugs,

where each drug alone did not change glucose,

但前提是，這些藥單獨服用 不會改變葡萄糖，

but together I saw a strong signal."

一起服用時，會有強烈訊號的藥物。」

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

我說：「喔！你真聰明， 好主意，讓我看一下清單。」

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

有一大堆藥，並沒有令人非常興奮。

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

但引起我注意的是，清單上有兩種藥：

paroxetine, or Paxil, an antidepressant;

帕羅西汀或稱克憂果， 這是一種治療憂鬱症的藥，

and pravastatin, or Pravachol, a cholesterol medication.

還有普伐他汀或稱美百樂， 一種治療心臟疾病的藥。

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

然後我說：「哈！有上百萬 美國人正在服用這兩種藥」。

In fact, we learned later,

事實上，我們之後才知道，

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

當時有1500萬美國人正在服用帕羅西汀， 1500萬人正在服用普伐他汀，

and a million, we estimated, on both.

而我們預估有100萬人， 同時服用這兩個藥。

So that's a million people

所以，有100萬人

who might be having some problems with their glucose

可能有葡萄糖上的問題，

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

如果他用食品藥物管理局的資料庫

actually holds up.

做的機械學習判讀器真的有用的話。

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

但我說：「還是不能發表，

because I love what you did with the mumbo jumbo,

因為我雖然喜歡你做的

with the machine learning,

機械學習判讀器，

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

但我們沒有真正的證明標準 來證明我們是正確的。」

So we have to do something else.

所以，我們來必須做些其他事來驗證。

Let's go into the Stanford electronic medical record.

我們去找史丹佛的電子病例紀錄。

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

我們有一個副本，可以用來研究，

we removed identifying information.

我們移除了病人個資。

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

我說：「讓我們來看看， 服用這兩種藥的人

have problems with their glucose."

是否有葡萄糖上的疾病。」

Now there are thousands and thousands of people

在史丹佛病例紀錄中有成千上萬的人

in the Stanford medical records that take paroxetine and pravastatin.

同時服用這兩種藥。

But we needed special patients.

但我們需要特定病患。

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

我們需要已經做葡萄糖檢測 且服用其中一種藥的病人，

then got the second one and had another glucose measurement,

另外再找到另一個已經做過 另一個葡萄糖檢測的病人，

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

全部都在合理期間做的， 例如兩個月內。

And when we did that, we found 10 patients.

當我們開始著手進行時， 我們找到十個病人。

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

然而，十個人裡面 有八個葡萄糖異常增加現象，

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

在他們服用第二個P時 ─我們稱呼這個叫 P&P─

when they got the second P.

當他們服用了第二個 P。

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

哪一個先服用都行， 當第二個藥服用後，

glucose went up 20 milligrams per deciliter.

葡萄糖濃度每公升會增加20毫克。

Just as a reminder,

提醒各位一下，

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

如果你能正常走動，沒有糖尿病，

with a glucose of around 90.

你的葡萄糖濃度約90毫克／公升。

And if it gets up to 120, 125,

如果上升到120、125，

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

你的醫生會開始認為 你有潛在的糖尿病症狀。

So a 20 bump -- pretty significant.

所以，一下子增加20是相當明顯的。

I said, "Nick, this is very cool.

我說：「尼克，這很酷。

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

但，很抱歉，我們仍然沒辦法寫報告，

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

因為只有十個病人，饒了我吧，

it's not enough patients."

病人樣本數根本不夠。」

So we said, what can we do?

所以，那怎麼辦？

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

我們來打電話給哈佛 及范德堡大學的朋友，

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

就是波士頓的哈佛 及納許維爾的范德堡，

who also have electronic medical records similar to ours.

他們都有跟我們很像的 電子病歷紀錄。

Let's see if they can find similar patients

讓我們看看，他們是否 也可以找到相同的病人，

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

也有我們需要的已經服用這兩種藥，

in that range that we need.

並做過葡萄糖檢測的病人。

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

上天保佑，范德堡一個星期內找到40個

same trend.

有同樣趨勢的病人。

Harvard found 100 patients, same trend.

哈佛找到100個有同樣趨勢的病人。

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

所以，最後，我們從三個不同的 醫學中心找到150個病人

that were telling us that patients getting these two drugs

服用過這兩種藥，

然後有葡萄糖異常增加現象。

were having their glucose bump somewhat significantly.

More interestingly, we had left out diabetics,

有趣的是，我們沒有考慮糖尿病患者，

because diabetics already have messed up glucose.

因為糖尿病患者本身的 血糖濃度就已經很混亂。

When we looked at the glucose of diabetics,

當我們觀察糖尿病患者的血糖濃度時，

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

會上升到每公升60毫克， 不只20毫克。

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

這事情很重要，我們說： 「我們必須發佈這件事。」

We submitted the paper.

我們遞交報告，

It was all data evidence,

裡面全部都是資料證明，

data from the FDA, data from Stanford,

有來自食品藥物管理局、史丹佛的資料、

data from Vanderbilt, data from Harvard.

有來自范德堡、哈佛醫學院的資料，

We had not done a single real experiment.

我們完全沒有做任何實驗。

But we were nervous.

但我們很緊張。

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

所以，當報告送去審核時， 尼克就去了實驗室。

We found somebody who knew about lab stuff.

我們找到會做實驗的人。

I don't do that.

我不做實驗的。

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

我會看病人，但我不會做分量管。

They taught us how to feed mice drugs.

他們教我們如何餵老鼠吃藥。

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

我們給第一組老鼠餵食帕羅西汀，

We gave some other mice pravastatin.

給第二組老鼠餵食普伐他汀。

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

第三組的老鼠兩種藥都餵食。

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

驚奇的是，葡萄糖每公升上升20到60毫克，

老鼠也有相同的反應。

in the mice.

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

所以，只有資料證據的報告被接受了，

但我們在最後加了註記說，

but we added a little note at the end,

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

如果把藥物給老鼠，葡萄糖也會上升。

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

太棒了，故事其實就到這裡結束。

But I still have six and a half minutes.

但，我還有六分半鐘。

(Laughter)

（笑聲）

So we were sitting around thinking about all of this,

所以，我們坐下來想一下所有的事，

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

我忘記誰曾經說過，但有人說：

"I wonder if patients who are taking these two drugs

「不曉得同時服用這兩種藥的病人，

are noticing side effects of hyperglycemia.

是否有注意到高血糖症的副作用。

They could and they should.

他們可能知道，也必須知道。

How would we ever determine that?"

我們要如何確定？」

We said, well, what do you do?

我們說，好吧，你會怎麼做？

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

你服用了一種藥，一個或兩個新藥，

and you get a funny feeling.

然後你感覺怪怪的。

What do you do?

你會怎麼做？

You go to Google

你會去問 Google，

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

然後搜尋你在服用的一或兩個藥名，

and you type in "side effects."

然後加上「副作用」。

What are you experiencing?

你會找到什麼？

So we said OK,

所以，我們說，好，

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

我們來問 Google 能否 跟我們分享搜尋紀錄，

so that we can look at the search logs

讓我們可以觀察搜尋紀錄，

and see if patients are doing these kinds of searches.

看是否有病人也在做同樣的搜尋。

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

很抱歉我得這麼說， 但 Google 拒絕了我們的請求。

So I was bummed.

所以，我很煩惱。

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

我跟一個在微軟研究室的同事吃晚餐時，

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

我跟他說：「我們想做這個研究，

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

Google 說不行，我有點煩惱。」

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

他說：「我們有 Bing 搜尋引擎啊。」

(Laughter)

（笑聲）

Yeah.

是啊！

That's great.

太棒了。

Now I felt like I was --

現在，我感覺...

(Laughter)

（笑聲）

I felt like I was talking to Nick again.

我好像又在鼓勵尼克一樣。

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

他在全世界數一數二的公司上班，

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

我已經開始要安慰他了。

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

但他說：「不，洛斯，你可能沒搞懂。

We not only have Bing searches,

我們不只有 Bing 啊，

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

如果你用 IE 在 Google、

Yahoo, Bing, any ...

雅虎、Bing 等任何搜尋引擎，

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

之後18個月，我們保留這些數據 僅做研究目的使用。

I said, "Now you're talking!"

我說：「這才像話嘛！」

This was Eric Horvitz, my friend at Microsoft.

這就是我的微軟朋友艾瑞克．霍維茲。

So we did a study

我們做了一項研究，

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

我們定義出了50個

如果一般人有高血糖症時 會鍵入的關鍵字，

if they're having hyperglycemia,

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

像是疲勞、沒食慾、頻尿等。

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

請原諒我，但這些就是 你可能會鍵入的關鍵字。

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

所以，我們有了50個短語， 我們稱之為「糖尿病關鍵字」。

And we did first a baseline.

我們先設定了一條基準線。

And it turns out that about .5 to one percent

原來，網路上有包含這些關鍵字的搜尋

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

占了0.5~1%的比例。

So that's our baseline rate.

所以，這就是我們的基準線率，

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

如果大家鍵入「帕羅西汀」或「克憂果」 ──這些是同義字──

and one of those words,

以及剛剛其中一個關鍵字，

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

那糖尿病類型的基準線率會上升到2%，

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

如果你已經知道 「帕羅西汀」這個字的話。

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

如果是「普伐他汀」， 那比率會從基準線率上升到3％。

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

如果「帕羅西汀」 和「普伐他汀」同時出現，

it goes up to 10 percent,

那會上升到10%，

a huge three- to four-fold increase

有3到4倍的增加，

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

用這兩種藥搜尋，會出現 我們感興趣的字在裡面，

and diabetes-type words or hyperglycemia-type words.

像是糖尿病類的字 或高血糖症類的字。

我們發佈了這個研究，

We published this,

and it got some attention.

並得到一些關注。

The reason it deserves attention

它值得被關注的原因是，

is that patients are telling us their side effects indirectly

病人會透過搜尋，

直接告訴我們藥物的副作用。

through their searches.

We brought this to the attention of the FDA.

我們得到了食品藥物管理局的關注。

They were interested.

他們很感興趣。

They have set up social media surveillance programs

他們已經成立社會媒體監測計畫，

to collaborate with Microsoft,

與微軟展開合作，

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

他們有良好的設備來做這些事，

to look at Twitter feeds,

可以觀察推特的動態、

to look at Facebook feeds,

觀察臉書的動態、

to look at search logs,

觀察搜尋日誌、

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

嘗試觀察引發問題的

無論單一藥物或混合藥物的早期症狀。

are causing problems.

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

我從這件事學到什麼？ 為什麼要講這個故事？

Well, first of all,

首先，

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

我們現在有大數據及中型數據稱腰，

to help us understand drug interactions

來幫助我們了解藥物的相互作用，

and really, fundamentally, drug actions.

以及真實、基本的藥物作用。

How do drugs work?

藥物是如何作用？

This will create and has created a new ecosystem

這個將會創造一個新的生態系統，

for understanding how drugs work and to optimize their use.

來幫助我們了解藥物如何運作 以及有效使用它們。

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

尼克繼續往前走， 他現在是哥倫比亞的教授。

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

他用好幾百對藥物做為博士研究。

He found several very important interactions,

他找到一些非常重要的藥物交互作用，

and so we replicated this

所以，我們複製這個模式，

and we showed that this is a way that really works

展示出利用這樣做

for finding drug-drug interactions.

來尋找藥與藥之間的作用真的有效。

However, there's a couple of things.

然而，還有一些事。

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

我們不會同時一次只服用兩種藥。

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

就如我之前所說的， 有病人一次是服用三、五、七、九種藥。

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

他們有認真研究 這九種藥的相互作用嗎？

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

沒錯，我們可以做成對的藥， A+B、A+C、A+D，

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

但如果同一個病人 同時服用ABCDEFG，

being taken by the same patient,

perhaps interacting with each other

那可能會互相產生那些作用？

in ways that either makes them more effective or less effective

藥效更好或更不好？

or causes side effects that are unexpected?

或造成那些意想不到的副作用呢？

We really have no idea.

我們真的不知道。

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

它是個開放式的藍天領域， 讓我們可以使用數據，

to try to understand the interaction of drugs.

來嘗試了解藥物彼此間的作用。

Two more lessons:

另外兩件事：

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

我想要各位去想想 我們所創造出來的力量，

with the data from people who had volunteered their adverse reactions

就是我們已經可以透過藥劑師、 病人本身、病人的醫師，

through their pharmacists, through themselves, through their doctors,

來取得自願者身上 他們的藥物不良反應，

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

這些人同意他們的資料可以被 史丹佛、哈佛、范德堡醫學院

to be used for research.

來做研究使用。

People are worried about data.

大家都擔心個資問題。

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

他們擔心自己的隱私及安全 ──他們必須要擔心。

我們需要保全系統。

We need secure systems.

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

但我們不能有一個 把資料關起來的系統，

because it is too rich of a source

因為它的資源太豐盛了，

of inspiration, innovation and discovery

它對醫學界的鼓舞、 創新、發現新事物

for new things in medicine.

實在太重要了。

And the final thing I want to say is,

最後，我想說的是，

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

我們發現這兩個藥的案例， 的確是令人難過的故事。

The two drugs actually caused problems.

這兩個藥一起服用真的會有問題。

They increased glucose.

同時服用會增加葡萄糖，

They could throw somebody into diabetes

會造成一個原本沒糖尿病的人

who would otherwise not be in diabetes,

發生糖尿病情形，

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

所以，各位如果想一起使用 這兩種藥，一定要非常小心，

perhaps not together,

最好不要一起服用，

make different choices when you're prescribing.

當你要開處方簽時， 看看有沒有不同的選擇。

But there was another possibility.

但，也有其他的可能。

We could have found two drugs or three drugs

我們或許能找到兩或三種藥，

that were interacting in a beneficial way.

一起服用時也許可以更有效。

We could have found new effects of drugs

我們或許也可以找到

that neither of them has alone,

藥物本身沒有的作用，

but together, instead of causing a side effect,

但在一起服用時不但沒有產生副作用，

they could be a new and novel treatment

反而產生新作用，有可能變成最新的

for diseases that don't have treatments

絕症疾病治療方式，

or where the treatments are not effective.

或者原本的治療方式完全是無效的。

If we think about drug treatment today,

如果我們想想現今的藥物治療方式，

all the major breakthroughs --

所有的重大突破──

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

愛滋病、肺結核、 憂鬱症，糖尿病──

it's always a cocktail of drugs.

總像是藥物雞尾酒。

And so the upside here,

這件事的好處是，

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

也許哪一天不同的TED主題， 我們又會來到這裡分享，

is how can we use the same data sources

我們要如何用同樣的資料來源

to find good effects of drugs in combination

來找到藥物混用時產生的好效果，

that will provide us new treatments,

它將提供我們新的治療方式，

new insights into how drugs work

以及對藥物如何作用提供新的見解，

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

並且讓我們的病人得到更好的照顧。

Thank you very much.

非常謝謝各位。

(Applause)

（掌聲）