Thomas Goetz: 现在是重新设计医疗数据的时候

====================================
http://dotsub.com/view/6ae63f39-1370-4308-9155-eabf8f0c95bf
Thomas Goetz: 现在是重新设计医疗数据的时候
我准备和大家探讨 我们如何 开发一种没有得到有效利用的医疗资源， 那就是病人， 或者，用一个更专业的术语来说-- 人。 我们都是病人，我们也都是人。 就连医生有时候也可能成为病人。 我想借此机会 告诉大家 在这个国家，我们一直被蒙在鼓里， 事实上，全世界都是如此。 如果你从全局来看-- 也就是从公共卫生高度来看，这是我受训的领域-- 你会看到一些行为问题， 看到人们得到相关信息， 却没有完全按照这些信息行事。 问题是，有人得了糖尿病， 肥胖症，各种心脏疾病， 甚至某些癌症--如果人吸烟的话。 对于这些，人们都知道他们应该怎么做才是对的。 但是，尽管他们知道哪些事情该做， 他们就是不去做。
行为上的改变 是医学上的一个顽固问题。 这可以追溯到亚里士多德的时代。 医生们对此深恶痛绝。 他们总是为此而抱怨。 我们指的是人们并不配合治疗。 有些人不按要求吃药， 有些人不听从医嘱。 这些都是行为问题。 但是尽管临床医学 对病人的行为怨声载道， 他们并没有采取措施 来补救问题。 这样，问题的关键 在于如何帮助人们做出正确决定-- 给病人一张显示信息的表格， 这张表格不仅教育 或告知病人应该做什么， 能够引导如何他们为生活 做出更好的抉择。
然而，有一个医学的领域 在行为变化方面颇有建树。 那就是牙科。 牙科虽然看起来--个人意见-- 许多牙医可能会认为 牙科是医学中最没有发展前景的分支。 牙科领域没有新奇有趣的事情发生。 但是，牙科却很好地解决了 行为变化这个问题。 这是医疗系统中 我们所取得的重大成功。 人人都刷牙，使用牙线。 他们虽然做得还不够，但至少他们做了。
我下面会介绍一些牙科医生三十年前 在康涅狄格州 所做的一个试验。 这个试验有一定年头了，但却是一个非常棒的试验， 这个试验很简单， 所以故事很简短。 康涅狄格州的这些牙医决定 他们要让人们更加勤刷牙勤用牙线。 他们打算使用一个变量： 吓唬人们。 他们想告诉人们 如果不刷牙不用牙线，会有什么样的后果。 他们的病人很多。 他们这些病人分成两组。 一组是低恐惧人群， 他们给这些人做了一个十三分钟的演示， 完全依照科学事实进行， 但是，对他们说，如果你不刷牙，用牙线， 你就会得牙龈疾病。如果你的牙龈除了毛病，你的牙齿就会脱落， 但是，如果你安了假牙，问题就不大。 这是针对低恐惧群体的做法。 而对高恐惧群体，他们可下了狠药。 他们给这些人展示血淋淋的牙龈， 让他们看牙齿间淤出来的脓， 告诉他们，他们的牙齿很快就会脱落， 他们会得炎症 并且炎症会从空腔扩散到其它身体部位， 最后，是的，他们会失去自己的牙齿。 他们不得不使用假牙，如果你装了假牙， 你不能啃玉米棒， 不能咬苹果， 不能吃牛排； 在你的下半生里，你还有很多东西要吃呢， 还不赶快刷牙，用牙线。 这就是其中的信息；这个试验就是这样的。
接着他们测量了另一个变量。 他们想抓住这个变量， 那就是病人的自我效能感。 这是病人是否认为自己 会去刷牙和使用牙线的观念。 在一开始，他们问病人， “你认为你会坚持照做吗？” 那些说“是的，是的，我会好好做的，”的人 属于为高效能感人群。 而那些说 “呃，我平时刷牙或用牙线都不够勤，”的人 属于低效能感人群。 这是结果。 试验的结果是 恐惧并不是控制行为的 主要动力。 那些勤刷牙，用牙线的人， 并不一定是 那些害怕不良后果的人-- 他们是那些认为自己 有能力改变自己行为的人。 恐惧原来不是动因； 应该是自我感觉。
我想重点指出这个结果， 应为这是非常有用的发现-- 三十年前，对，三十前-- 这是当时尚未被研究的课题。 这实际上是从阿尔伯特. 班杜拉的研究中 引申出来的一个概念， 他研究了 人们是否会产生一种成就感。 效能感的信念最终归结到一点： 一个人是否相信他有能力改变自己的行为。 从医疗保健的角度看，可以这样描述这个信念： 一个人是否感觉 他看到更好的健康状况的前景， 看到自己的健康状况能够日渐转好。 这是个非常重要的信念。 非常奇妙的信念。 然而，我们还不知道怎样自如地控制这个信念。 或许，我们知道。
恐惧是不起作用的，对，恐惧不起作用。 这里有一个很好的例子， 证明了我们有多么不知悔改。 这是美国糖尿病协会的一次活动。 这仍然是我们传递保健信息的方式。 我昨晚给我三岁的儿子看了这张幻灯片， 他问：“爸爸，为什么这些人把救护车停在家里啊？” 我只能解释道：“他们想吓唬人。” 我不确定这么做是否起作用。
而这才是起作用的做法， 那就是个人化信息。 又是班杜拉， 他在几年前，甚至几十年前认识到了这一点。 当你给人们提供 关于他们健康状况的具体信息，他们的目前的病情， 以及他们预期的进展，他们可能的进展， 这个过程，对此过程的信念， 会改变病人的行为。 我来解释一下。 你得到一个病人的个人数据， 个人信息之后， 你必须把它与病人的生活联系起来。 你需要通过让他们了解这些信息 来把信息与他们的生活联系起来，而不是吓唬他们。 好，我知道我的病情。 那些抽象的数字， 还有摆在面前的 那一大堆健康信息让我摸不着头脑， 但这对我影响巨大。 这些信息不仅进入我们的头脑，还影响到我们的心理。 我们对这些信息有不解的情结， 因为这是关于我们自己的信息。 接着我们要把信息与人们的选择， 选择的范围， 以及方向联系起来-- 其中有哪些利弊。 最后，我们要清楚明了地告知人们他们应采取什么行动。 我们必须时刻把信息与行动 联系起来， 行动会产生反馈， 从而得到新的信息， 周而复始，形成一个反馈圈。
实践证实，这个信念对于改变行为 十分可靠。 然而，问题是，右上角的 个人数据，往往很难得到。 得到个人数据可谓困难重重，而且要花不少钱。 直到现在。 我接下来要给大家看一个例子，这个简单的例子展示我们如何得到个人数据。 我们都见过这些标示牌，这些是限速牌。 现在雷达都变便宜了， 所以我们随处都能看到这些牌子。 它们在反馈圈中的作用是这样的。 一开始是个性化数据， 此时你所在公路的限速 是25， 当然你有些超速了。 我们总是超速的。 在此情况下我们只有两种选择。 我们要么加速，要么减速。 我们或许应该减速， 这样行动点就是现在。 我们应该马上停止踩刹车。 我们一般都会这么做；这些限速牌非常有效地 提醒了人们要减速。 他们将速度降低了百分之十。 他们能坚持五英里路程， 接着我们才会再次踩油门。 但在还是起作用了，甚至还有益于身心健康。 你的血压可能没那么高了。 可能交通事故减少了，造福了公众健康。
但是，总的来看，这是个 难得的一流反馈圈。 因为在医疗保健方面，大多数情况下， 数据与行动相互分离。 很难把事情理得那么顺。 但是我们有一线希望。 我想将话题转移，来探讨 在我国，我们如何传递健康信息， 以及我们如何获取信息。 这是一个药品广告。 这实际上是在耍花腔；而不是真正的药品广告。 还没有人聪明到用”Havidol“ 来给药品命名的。 但是它看起来完全没问题。 这就是我们 得到健康信息和药品信息的方式， 看上去完美无缺。 当翻开杂志， 我们看到这样一页，对吧。 食品及药物管理局要求制药公司 把这一页放进他们广告中，或加在广告后面。 对于我而言，这是医药方面一大应受批判的做法。 因为我们知道， 我们中间有谁认为人们会读这一页？ 有谁认为 那些读了这一页的人 能够得到任何有用信息？ 这是医疗信息沟通的 昏庸之举。 没有人会买账。
而这是另一种方式。 这种方式是由 达特茅斯医学院的几位研究人员所开发。 他们是莉莎. 史沃兹和史蒂芬. 沃罗森。 他们的这个发明叫做药物成分说明表。 出乎人意料的是，他们的灵感来自 Cap‘n Crunch麦片。 他们看到上面的营养成分介绍， 并意识到麦片，食物的成分介绍， 实际上有助于人们了解他们食物中有哪些成分。 我们怎么就没想到， 可以把Cap'n Crunch麦片的成分介绍模式 推广到制药公司。 我快速给大家展示一下。 它清清楚楚地说明了该药的用途，使用对象， 这样你就可以对号入座， 看上面的信息是否与你的情况相符， 这药是否适合你服用。 你能够清楚地了解服药之后会有哪些好处。 它并没有含糊地承诺，无论怎样都能药到病除， 但是你可以通过数据来了解药的效果。 最后，你明确了自己有哪些选择。 你可以开始挑拣有关的选择， 因为你要考虑到药的副作用。 每一次吃药，你都有可能面临某种副作用。 它明明白白地列出了可能的副作用。 这很管用。
我特别喜欢这个药物说明表。 于是，我在思考， 我怎样才能 帮助人们理解信息呢？ 还有哪个隐藏的信息源 让人们无所适从呢？ 于是，我想到了这个：实验室测试结果。 这一大堆信息是血液的测试结果。 上面满满的都是信息。 这不是给我们看的；不是给人们看的；不是给病人看的。 是直接给医生看的。 上帝啊！我敢说，很多医生，如果你问他们， 他们也不完全明白这些东西。 这样的信息表述实在是糟透了。 如果你问塔夫特，他会说： “毫无疑问，这样的信息表述真是槽糕透顶了。”
在Wired杂志社， 我让我们的图表设计部 对这些测试报告进行重新设计。 我给大家简短地展示一下。 这是一项常规血液报告原来的版本， 这是修改后的版本，这就是我们的设计。 修改后的版本将原来的四页纸-- 上一张幻灯片实际上 四页纸的第一页， 这还只是一项常规的血液检查。 全篇洋洋洒洒，那么多的数值，那么多的数字，你根本看不懂。 这是我们一页纸的总结。 我们运用颜色作为信号。
（众人笑） 我们需要认识到这信息的力量， 以此拉近医患关系，帮助人们 改变他们生命的轨迹。
非常感谢
（鼓掌）

-----------------------
Thomas Goetz: It's time to redesign medical data
I'm going to be talking to you about how we can tap a really underutilized resource in health care, which is the patient, or -- as I like to use the scientific term -- people. Because we are all patients, we are all people. Even doctors are patients at some point. So I want to talk about that as an opportunity that we really have failed to engage with very well in this country and, in fact, worldwide. If you want to get at the big part -- I mean from a public health level, where my training is -- you're looking at behavioral issues, your looking at things where people are actually given information, and they're not following through with it. It's a problem that manifests itself in diabetes, obesity, many forms of heart disease, even some forms of cancer -- when you think of smoking. Those are all behaviors where people know what they're supposed to do. They know what they're supposed to be doing, but they're not doing it.

Now behavior change is something that is a long-standing problem in medicine. It goes all the way back to Aristotle. And doctors hate it, right. I mean, they complain about it all the time. We talk about it in terms of engagement, or non-compliance, when people don't take their pills, when people don't follow doctors' orders. These are behavior problems. But for as much as clinical medicine agonizes over behavior change, there's not a lot of work done in terms of trying to fix that problem. So the crux of it comes down to this notion of decision-making -- giving information to people in a form that doesn't just educate them or inform them, but actually leads them to make better decisions, better choices in their lives.

One part of medicine though has faced the problem of behavior change pretty well, and that's dentistry. Dentistry might seem -- and I think it is -- many dentists would have to acknowledge it's somewhat of a mundane backwater of medicine. Not a lot of cool, sexy stuff happening in dentistry. But they have really taken this problem of behavior change and solved it. It's the one great preventive health success we have in our health care system. People brush and floss their teeth. They don't do it as much as they should, but they do it.

So I'm going to talk about one experiment that a few dentists in Connecticut cooked up about 30 years ago. So this is an old experiment, but it's a really good one, because it was very simple, so it's an easy story to tell. So these Connecticut dentists decided that they wanted to get people to brush their teeth and floss their teeth more often. And they were going to use one variable: They wanted to scare them. They wanted to tell them how bad it would be if they didn't brush and floss their teeth. They had a big patient population. They divided them up into two groups. They had a low-fear population, where they basically gave them a 13-minute presentation, all based in science, but told them that, if you didn't brush and floss your teeth, you could get gum disease. If you get gum disease, you will lose your teeth, but you'll get dentures, and it won't be that bad. So that was the low-fear group. The high-fear group, they laid it on really thick. They showed bloody gums, they showed puss oozing out from between their teeth, they told them that their teeth were going to fall out, they said that they could have infections that would spread from their jaws to other parts of their bodies, and ultimately, yes, they would lose their teeth. They would get dentures, and if you got dentures, you weren't going to be able to eat corn on the cob, you weren't going to be able to eat apples, you weren't going to be able to eat steak; you'll eat mush for the rest of your life. So go brush and floss your teeth. That was the message; that was the experiment.

Now they measured one other variable. They wanted to capture one other variable, which was the patients' sense of efficacy. This was the notion of whether the patients felt that they actually would go ahead and brush and floss their teeth. So they asked them at the beginning, "Do you think you'll actually be able to stick with this program?" And people who said, "Yeah, yeah. I'm pretty good about that," they were characterized as high efficacy, and the people who said, "Eh, I never get around to brushing and flossing as much as I should," they characterized as low efficacy. So the upshot was this. The upshot of this experiment was that fear was not really a primary driver of the behavior at all. The people who brushed and flossed their teeth were not necessarily the people who were really scared about what would happen -- it's the people who simply felt that they had the capacity to change their behavior. So fear showed up as not really the driver; it was the sense of efficacy.

So I want to isolate this, because it was a great observation -- 30 years ago, right, 30 years ago -- and it's one that's laid fallow in research. It was a notion that really came out of Albert Bandura's work, who studied whether people could get a sense of empowerment. The notion of efficacy basically boils down to one that, if somebody believes that they have the capacity to change their behavior. In health care terms, you could characterize this as whether or not somebody feels that they see a path towards better health, that they can actually see their way towards getting better health. And that's a very important notion. It's an amazing notion. We don't really know how to manipulate it, though, that well. Except, maybe we do.

So fear doesn't work, right, fear doesn't work. And this is a great example of how we haven't learned that lesson at all. This is a campaign from the American Diabetes Association. This is still the way we're communicating messages about health. I mean, I showed my three year-old this slide last night, and he's like, "Papa, why is an ambulance in these people's homes?" And I had to explain, "They're trying to scare people." And I don't know if it works.

Now here's what does work, personalized information works. Again, Bandura recognized this years ago, decades ago. When you give people specific information about their health, where they stand, and where they want to get to, where they might get to, that path, that notion of a path, that tends to work for behavior change. So let me just spool it out a little bit. So you start with personalized data, personalized information, that comes from an individual, and then you need to connect it to their lives. You need to connect it to their lives, hopefully not in a fear-based way, but one that they understand. Okay, I know where I sit. I know where I'm situated. And that doesn't just work for me in terms of abstract numbers, this overload of health information that we're inundated with, but it actually hits home. It's not just hitting us in our heads, it's hitting us in our hearts. There's an emotional connection to information because it's from us. That information then needs to be connected to choices, needs to be connected to a range of options, directions that we might go to -- trade-offs, benefits. Finally, we need to be presented with a clear point of action. We need to connect the information always with the action, and then that action feeds back into different information, and it creates, of course, a feedback loop.

Now this is a very well-observed and well-established notion for behavior change. But the problem is that things in the upper-right corner there, personalized data, it's been pretty hard to come by. It's a difficult and expensive commodity, until now. So I'm going to give you an example, a very simple example of how this works. So we've all seen these. These are the "your speed limit" signs. You've seen them all around, especially these days as radars are cheaper. And here's how they work in the feedback loop. So you start with the personalized data where the speed limit on the road where you are at that point is 25, and, of course, you're going faster than that. We always are. We're always going above the speed limit. The choice in this case is pretty simple. We either keep going fast, or we slow down. We should probably slow down, and that point of action is probably now. We should take our foot off the peddle right now. And generally we do; these things are shown to be pretty effective in terms of getting people to slow down. They reduce speeds by about five to 10 percent. They last for about five miles, in which case we put our foot back on the peddle. But it works, and it even has some health repercussions. You blood pressure might drop a little bit. Maybe there's fewer accidents, so there's public health benefits.

But by-and-large, this is a feedback loop that's so nifty and too rare. Because in health care, most health care, the data is very removed from the action. It's very difficult to line things up so neatly. But we have an opportunity. So I want to talk about, I want to shift now to think about how we deliver health information in this country, how we actually get information. This is a pharmaceutical ad. Actually, it's a spoof; it's not a real pharmaceutical ad. Nobody's had the brilliant idea of calling their drug Havidol quite yet. But it looks completely right. So it's exactly the way we get health information and pharmaceutical information, and it just sounds perfect. And then we turn the page of the magazine, and we see this, right, we see this. Now this is the page the FDA requires pharmaceutical companies to put into their ads, or to follow their ads. And to me, this is one of the cynical exercises in medicine. Because we know. Who among us would actually say that people read this? And who among us would actually say that people who do try to read this actually get anything out of it? This is a bankrupt effort at communicating health information. There is no good faith in this.

So this is a different approach. This is an approach that has been developed by a couple researchers at Dartmouth Medical School, Lisa Schwartz and Steven Woloshin. And they created this thing called the drug facts box. They took inspiration from, of all things, Cap'n Crunch. They went to the nutritional information box and saw that what works for cereal, works for our food, actually helps people understand what's in their food. God forbid that we should use that same standard that we make Cap'n Crunch live by and bring it to drug companies. So let me just walk through this quickly. It says very clearly what the drug is for, specifically who is it good for, so you can start to personalize your understanding of whether the information is relevant to you or whether the drug is relevant to you. You can understand exactly what the benefits are. It isn't this kind of vague promise that it's going to work no matter what, but you get the statistics for how effective it is. And finally, you understand what those choices are. You can start to unpack the choices involved because of the side effects. Every time you take a drug, you're walking into a possible side effect. So it spells those out in very clean terms. And that works.

So I love this. I love that drug facts box. And so I was thinking about, what's an opportunity that I could have to help people understand information? What's another latent body of information that's out there that people are really not putting to use. And so I came up with this: lab test results. Blood test results are this great source of information. They're packed with information. They're just not for us; they're not for people; they're not for patients. They go right to doctors. And God forbid -- I think many doctors, if you really asked them, they don't really understand all this stuff either. This is the worst presented information. You ask Tufte, and he would say, "Yes, this is the worst presentation of information possible."

What we did at Wired was we went, and I got our graphic design department to re-imagine these lab reports. So that's what I want to walk you through. So this is the general blood work before, and this is the after, this is what we came up with. The after takes what was four pages -- that previous slide was actually the first of four pages of data that's just the general blood work. It goes on and on and on, all these values, all these numbers you don't know. This is our one-page summary. We use the notion of color. It's an amazing notion that color could be used. So on the top level you have your overall results, the things that might jump out at you from the fine print. Then you can drill down and understand how actually we put your level in context, and we use color to illustrate exactly where your value falls. In this case, this patient is slightly at risk of diabetes because of their glucose level.

Likewise, you can go over your lipids and, again, understand what your overall cholesterol level is and then break down into the HDL and the LDL if you so choose. But again, always using color and personalized proximity to that information. All those other values, all those pages and pages of values that are full of nothing, we summarize. We tell you that you're okay, you're normal. But you don't have to wade through it. You don't have to go through the junk. And then we do two other very important things that kind of help fill in this feedback loop. We help people understand in a little more detail what these values are and what they might indicate. And then we go a further step: We tell them what they can do. We give them some insight into what choices they can take, what actions they can take. So that's our general blood work test.

Then we went to CRP test. In this case, it's a sin of omission. They have this huge amount of space, and they don't use it for anything, so we do. Now the CRP test is often done following a cholesterol test, or in conjunction with a cholesterol test. So we take the bold step of putting the cholesterol information on the same page, which is the way the doctor is going to evaluate it. So we thought the patient might actually want to know the context as well. It's a protein that shows up when your blood vessels might be inflamed, which might be a risk for heart disease. What you're actually measuring is spelled out in clean language. Then we use the information that's already in the lab report. We use the person's age and their gender to start to fill in the personalized risks. So we start to use the data we have to run a very simple calculation that's on all sorts of online calculators to get a sense of what the actual risk is.

The last one I'll show you is a PSA test. Here's the before, and here's the after. Now a lot of our effort on this one -- as many of you probably know, a PSA test is a very controversial test. It's used to test for prostate cancer, but there are all sorts of reasons why your prostate might be enlarged. And so we spent a good deal of our time indicating that. We again personalized the risks. So this patient is in their 50s, so we can actually give them a very precise estimate of what their risk for prostate cancer is. In this case it's about 25 percent, based on that. And then again, the follow-up actions.

So our cost for this was less than $10,000, all right. That's what Wired magazine spent on this. Why is Wired magazine doing this? (Laughter) Quest Diagnostics and LabCorp, the two largest lab testing companies: Last year, they made profits of over 700 million dollars and over 500 million dollars respectively. Now this is not a problem of resources, this is a problem of incentives. We need to recognize that the target of this information should not be the doctor, should not be the insurance company; it should be the patient. It's the person who actually, in the end, is going to be having to change their lives and then start adopting new behaviors.

This is information that is incredibly powerful. It's an incredibly powerful catalyst to change. But we're not using it; it's just sitting there. It's being lost. So I want to just offer four questions that every patient should ask, because I don't actually expect people to start developing these lab test reports. But you can create your own feedback loop. Anybody can create their feedback loop by asking these simple questions: Can I have my results? And the only acceptable answer is -- (Audience: Yes.) -- yes. What does this mean? Help me understand what the data is. What are my options? What choices are now on the table? And then, what's next? How do I integrate this information into the longer course of my life?

So I want to wind up by just showing that people have the capacity to understand this information. This is not beyond the grasp of ordinary people. You do not need to have the education level of people in this room. Ordinary people are capable of understanding this information, if we only go to the effort of presenting it to them in a form that they can engage with. And engagement is essential here, because it's not just giving them information, it's giving them an opportunity to act. That's what engagement is; it's different than compliance. It works totally different than the way we talk about behavior in medicine today. And this information is out there.

I've been talking today about latent information, all this information that exists in the system that we're not putting to use. But there are all sorts of other bodies of information that are coming online. And we need to recognize the capacity of this information to engage people, to help people and to change the course of their lives.

Thank you very much.

(Applause)