Larry Page: Healthcare Data Mining Could Save 100,000 Lives a Year

An anonymous reader writes Google often gets criticism for its seemingly boundless desire for data collection and analysis, but the company says it has higher ambitions than just figuring out how best to serve advertising. Speaking to the NY Times, Larry Page said, "We get so worried about these things that we don't get the benefits Right now we don't data-mine healthcare data. If we did we'd probably save 100,000 lives next year." By "these things," he means privacy concerns and fear that the data might be misused. But he also pointed to Street View as a case where privacy concerns mostly melted away after people used it and found it helpful. "In the early days of Street View, this was a huge issue, but it's not really a huge issue now. People understand it now and it's very useful. And it doesn't really change your privacy that much. A lot of these things are like that."

Comments

Hey Larry ...

[gstoddart]

How many fingers am I holding up?

Screw you Google. "Do no evil" my ass.

This is just another instance of him saying "trust us, we're google, give us all your private information, what could possibly go wrong".

Re:Hey Larry ...

[causality]

How many fingers am I holding up?

Screw you Google. "Do no evil" my ass.

This is just another instance of him saying "trust us, we're google, give us all your private information, what could possibly go wrong".

Yes, at some point it's quite rational to decide "this one entity has enough power". He's really very smooth, though. I'll hand him that:

By "these things," he means privacy concerns and fear that the data might be misused. But he also pointed to Street View as a case where privacy concerns mostly melted away after people used it and found it helpful. "In the early days of Street View, this was a huge issue, but it's not really a huge issue now. People understand it now and it's very useful. And it doesn't really change your privacy that much. A lot of these things are like that."

That's a very diplomatic way to go about it. People often mistake that for honesty and openness in fact. It's basically a highly polished way of saying, "if you were educated you would agree with me."

True in theory

[Thinking Tom]

It is true, healthcare data mining could save many lives. The problem is nobody trusts health insurance companies because most of them (a) deliberately make it hard to deal with them in order to get people to give up on collecting claims, (b) refuse to cover at least some of the people we know when they need medical treatment, and (c) limit the quality of care received from most doctors. So nobody trusts them to abstain from using the information to come up with some reason to exclude you either from coverage in the marketplace or for a particular condition.

Re:True in theory

[dnavid]

I would disagree. Having this data available to crunch the numbers would definitely benefit healthcare (in the saving lives aspect). Currently (perhaps forever) the dangers probably outweigh the benefits. There is far too much incentive to abuse this data and we require more than some guy saying "hey, it's totally going to be fine" to convince us.

It's like storing passwords as plaintext. It is super convenient for everyone involved as long as everyone involved is altruistic. But assuming everyone is altruistic is stupid so storing passwords in plaintext is generally regarded as stupid.

I don't think that's a reasonable analogy. Here's a better one. The only reason the government even possesses the capability to perform the sort of mass surveillance they currently conduct is because of massive improvements in computational, storage, and communications technology. They couldn't collect email metadata without the invention and widespread deployment of cheap or essentially free electronic mail. They couldn't keep the data meaningfully without modern storage, or use it without current computer technology. But suppose that right at the birth of the microprocessor someone said that the government couldn't be trusted with the kind of power that technology would eventually provide, and thus we should strive to ensure its never allowed to be developed. No personal computers, no internet, no cloud computing, no smartphones. We now know they would have been right. Even so how many of us would in retrospect eliminate these technologies or forego their benefits? I suspect extremely few. And not because they are stupid or ignorant, but because they genuinely feel those benefits outweigh the consequences of their abuse.

I would *presume* that any large-scale collection and analysis of medical information will eventually be abused by someone. That still leaves the question of whether its a reasonable tradeoff. Its easy to say no now, when the benefits are only theoretical and the deficits seem obvious. But I think the many people whose lives would be saved or radically improved due to advances in statistical medical analysis would disagree. If someone is going to stand on principle and say nothing is worth the potential abuses, they should at least be honest and accept that their stance isn't bloodless or abstract. The opposition always assumes that the proponents of an idea should be held responsible for its consequences, but they rarely accept responsibility for the consequences of inaction. They should be. Even if its the right thing to do. Perhaps especially so.

Think about it... Seriously...

[Snapple]

There is a HUGE pool of untapped resources. Insurance companies process claims for millions, and have all the data, what is being prescribed, what is not being prescribed.. how long the perscription is for..... Who is seeing a doctor on a regular basis, and who isn't.... Using this data you can find out what treatments are being effective, and which ones aren't. Or is it really worth going to the dentist every 6 months? Isn't that worth it's weight in gold?

Internally insurance companies can summarize data without compromising their client's as they have the data all ready. Moving it to an external company would involve generating an guid for each identifying piece of information before it leaves the company. Basically a complete scrubbing of the data, but it is not an impossible task.

Why won't this happen? It's not a privacy issue, it a $$$$ issue.... Drug companies wouldn't want you to find out that they are selling snake oil.. They could loose millions if a report showed that their drug is not as effective as a competitors....

Apples and Oranges

[Qzukk]

The number of people who don't get hired because the shrub in their front yard is trimmed crooked is considerably lower than the number of people who don't get hired because they have MS, cancer or some other chronic disease that will cost the company's insurer big bucks and drive up the cost of insurance and cost the company in lost productivity when they're incapacitated. Oh sorry, I meant, don't get hired because they "aren't a good fit with the company culture".

Re:Apples and Oranges

[ilsaloving]

As I don't have mod points, I'll just reply and say that you are correct, and it's not limited to just that.

There have already been documented incidents where people in Canada have been denied entry into the states just because they went into a hospital a decade ago for depression.

Unlike StreetView, it has *already* been demonstrated that easy access to health information will guarantee abuse.

It's already happening

[Tucan]

Larry Page is just complaining that Google doesn't have the data. These data already exist and are being extensively studied by researchers in academics, government, health insurers, employers, and pharmaceutical companies. The de-identified data can be licensed and analyzed by anyone. The fully identifiable information is routinely analyzed by the owners of the data.

The problem is not access to the data, the problem is that it is difficult to make valid inferences about causation from observational data.

Ridiculous fear factor

[dentin]

I'm actually shocked by the shortsightedness of the slashdot crowd on this one. I expected at least -some- positive responses to be moderated up. Instead, I see a lot of misconceptions and ignorance of the actual problem Page, like Aubrey Degray, is trying to address.

We have a hundred thousand people worldwide dying due to various medical problems and the diseases of old age. These medical problems and diseases are complicated. They consist of tens of thousands of interlocking subproblems, so many that we often take several thousand specific issues and lump them together to call them something like 'cancer'. Fixing these problems - all of them - isn't something that a single drug company, or a single nation is going to do.

It's going to take everybody, everywhere. And in order to fix all these things - cancers, diseases of old age, genetic problems, and more - is going to take research, time, and data. Lots of data.

Lots and lots of data.

People whine about privacy, oh no the bad guys are going to steal my information, ignoring the fact that a hundred thousand people a day die and that thier information could help. All of these medical problems are tractable, all of them are soluble, but they'd be a hell of a lot easier to solve if researchers weren't hamstrung by ridiculous information privacy restrictions.

I don't want immortality in good health just for me, I want it for everyone, and this idiotic fear of having information released is standing in the way of that. A hundred thousand people a day dead, because we fear someone might discover an abnormally BPH score, HIV, or a genetic propensity for Alzheimers. What a steaming load of shortsighted crap.