美国国家公共电台 NPR Researchers Gather Health Data For 'All Of Us'

LAUREN FRAYER, HOST:

Now to a story about a big data approach to medicine with a big price tag - hundreds of millions of taxpayer 1 dollars. It's called the All of Us Precision Medicine Initiative. It's an effort to gather blood samples, medical information and fitness readouts from a million Americans. The goal - to create a huge pool of data for scientists to mine for clues about health and disease. Proponents 2 hope All of Us will be revolutionary. Critics aren't so sure, as NPR's Richard Harris reports.

RICHARD HARRIS, BYLINE 3: The plan is to recruit a million Americans to sign up for a program that will not only gather all sorts of medical data about them, but follow them for at least a decade, possibly much longer. Their electronic medical records will end up in huge databases. The physical samples of blood and urine will end up in an industrial park in Rochester, Minn.

MINE CICEK: This used to be an old warehouse 4, but when we moved in three, four years ago, we really built a laboratory, and that's in the space, so when you go in, you'll really see.

HARRIS: Mine Cicek is an assistant professor at the Mayo Clinic, which runs this cavernous lab. The warehouse roof covers an acre and a half of floor space. Power cords drop down from the ceiling to lab benches and robotic instruments lined up row after row - these will process blood and urine samples collected from around the country, which will ultimately end up in freezers. At the moment, the program is just running pilot studies.

CICEK: But when we really launch and up to speed, it's going to be close to, maybe even more than, a thousand participants.

HARRIS: Every day?

CICEK: Every day.

HARRIS: Cicek leads us into another part of the warehouse filled not only with ordinary-looking freezers, but one behemoth, 74 feet long and more than 15 feet wide.

CICEK: So this is the back of the freezer. So we'll go around so you'll have a appreciation 5 of how big this is.

HARRIS: It's bigger than a railroad boxcar. We loop around to the front, which has a glowing green slit 6.

CICEK: The green light that you see, it has a door. If you want to come this way...

HARRIS: Samples get fed through this door. Inside, robots pick up the samples, read bar codes stamped on them and store them in individual freezers hidden inside.

Do people ever go in there?

CICEK: Well, in an emergency, we have a red button in front. When you push that red button, everything stops, including the robotics because everything is in minus 80. And there's a section where it's only minus 20. But again, it's minus 20. You don't want to be in there unless you really, really want to go in there.

HARRIS: When the system is fully 7 up and running, she says, the incoming tubes of blood will be processed by a fully automated 8 system. So all she'll have to do is pick up boxes of samples and feed them into this freezer. And the scientific output for this enterprise - that's less tangible 9 at the moment.

ERIC DISHMAN: It's really a research resource that we're building for the whole country, and if not the whole world.

HARRIS: Eric Dishman runs the National Institutes of Health's All of Us Research Program from another anonymous 10 building, this one in the Washington, D.C., suburb of Rockville, Md. His job is to build it and then make the data widely available to top biomedical researchers on down.

DISHMAN: In fact, we want to work really hard to make that resource available to community colleges and even high school curriculums so sort of the brainpower per problem across the country is greater than what it is today.

HARRIS: It's kind of like the original effort to sequence the human genome. It's not an end in itself. It's a tool. And the hope is to bring the big data revolution that has reshaped online commerce and social media and apply it to science.

DISHMAN: We keep looking at diseases in isolation 11, and the All of Us Research Program is really trying to bring a diversity of people, a diversity of health conditions and a diversity of data sets to try to understand us more in our complexity 12 and not simplify everything as just one single disease.

HARRIS: Scientists might scour 13 this pile of data for unexpected patterns or dive into it looking for answers to specific questions. Dishman himself has lived through a 23-year odyssey 14 with kidney disease and kidney cancer and uses his experience as an example.

DISHMAN: I had $6 million of care. Now, looking back at it, scientists tell me that 90 percent of everything we ever did to me was destined 15 to be wrong.

HARRIS: Medicine is often a matter of trial and error, so that's not so unusual. But Dishman also had the DNA 16 in his tumor 17 deciphered, and that gave scientists yet one more clue about what to do and lead to an effective treatment, though even that process wasn't entirely 18 scientific.

DISHMAN: I'm nervous about sharing this story because it is both hopeful, but I also know there was a lot of science, but also a lot of luck that this worked for me.

HARRIS: Dishman regards his story as a lesson about what could be.

DISHMAN: We are in the early days of precision medicine. And this is exactly why we need to accelerate the science and the discovery so there's an evidence base for the decisions and choices that we're making for you as an individual, as well as the general population.

HARRIS: Among those skeptical 19 about the big talk and big investment behind precision medicine is Ken 20 Weiss, who recently retired 21 from his post as a genetics professor at Penn State.

KENNETH WEISS: I think there will be some progress, but I also think this is as much of a slogan to get funding as it is a serious promise.

HARRIS: Gathering 23 huge data sets may be useful for merchants trying to suss out your spending patterns, but he cautions that in biology, it may lead to more confusion rather than clarity. That's because many health conditions involve hundreds of genes 24, and the pattern is different in every individual. As it is, the more scientists look, the more variants 25 they find. So think about what that will look like when they have gathered a million samples.

WEISS: And bigger and bigger samples will just identify more and more very rare or very weak effects.

HARRIS: When the human genome was sequenced, many scientists hoped they would quickly be able to identify the common genes that are responsible for common diseases, like diabetes 26, heart disease, high blood pressure and so on. That simply didn't pan out. There were no such variants. Weiss says it's time to cut our losses pursuing that concept.

WEISS: I think we're already at the diminishing-returns point for many of the complex traits that are important to our society in terms of health.

HARRIS: The solutions to those common conditions lie largely in changing diets, exercise habits and tobacco addiction 27. Focusing genetic 22 resources on diseases that do have strong genetic components 28 does make a lot of sense, he says.

WEISS: But pouring more and more investment into these huge studies based on the idea that if you search enough computer data, you will get an answer, I think is a false promise.

HARRIS: Why not plunk down a couple hundred million dollars a year and see where this takes us?

WEISS: Why not plunk down that couple hundred million dollars a year to work on genetic therapies for known genetic traits? And once those are developed, which I think they will be because I think humans are very good at engineering, then we can extend to the less clear-cut genetic traits.

HARRIS: Given all the momentum 29 built up behind the Precision Medicine Initiative, Weiss is not voicing a popular point of view. But the retired geneticist says he has no axe 30 to grind and no brilliant insights about what would actually lead to medical breakthroughs. His concern is that biomedical research is all-in on this idea because they have a tool they are eager to use, not because they have a clear path ahead. Richard Harris, NPR News.