Health care data analytics for all, but at what cost? Teradata CTO Stephen Brobst shares his vision
Give up all your data, keep it all back, or find a compromise? The future's not yet written
"I know it's the buzzword right now, but sensor technology is very interesting to me, and in particular - and this is me personally, not Teradata generally - I believe that health care is a huge problem worldwide, and I believe that we're doing a pitiful job at making the right things happen," says Teradata CTO Stephen Brobst.
Brobst - in Anaheim, California at Teradata's Partners 2015 conference - believes new developments in data analytics and data warehousing could have a profound impact on the health care sector.
"If we can get the data integrated in such a way that we can start doing more advanced analytics, I believe we can significantly reduce the cost of health care while simultaneously increasing the quality," he says.
Brobst laments a world in which "health care costs are growing faster than GDP in basically every developed country". In the US, the Affordable Care Act - perhaps better known as "ObamaCare" - hasn't helped the situation, says Brobst, simply redistributing the cost of care while doing nothing to improve a health care sector that he considers is failing US citizens.
"I would say the UK is managing it more effectively than we are, although you're clearly not using the data well enough, either," he adds.
"The Affordable Care Act required all data to be in electronic form, and that's great, but we're still not doing anything with it. There's a lot more work to be done after that first step."
Brobst describes even the most up-to-date models of the FitBit health and fitness bracelet "primitive".
"[They measure] how many steps you take, and sleeping habits and so on, but they could be much more sophisticated," he says.
"Apple Watch could be much more than just taking your heart rate. I wanted to combine that with external data - what is the quality of the air you're breathing? That's absolutely possible today," he says, citing work by wireless communication tech firm Qualcomm to make this possible.
"There's all kinds of opportunities for sensors everywhere, and then feeding them into the health care of the individual. If you're in a hospital bed, it's obvious we're collecting all that data, but what about intervening before you end up in the hospital?" asks Brobst.
Creepy
"But it can also get a little bit creepy, right?" he continues.
"Because to do it right, I want your genetic encoding. I want not just your health care history, but also your parents' and grandparents'. Because if your grandparents had diabetes, your likelihood of diabetes is much higher. So I want all that data."
"But do you trust me?" asks Brobst.
"Health care insurance providers in the US we typically don't trust, as we give them all that data and they're going to raise my rates if they see I have diabetes. So there have to be regulatory things we do to prevent you from abusing that data."
Brobst says most people want increased quality of care and at a lower cost. "But I don't want to have to pay for something I can't control. And this is where it gets controversial," he adds
"One extreme is you can't use data at all. I think that's also wrong. If you smoke and I don't, I don't want to have to subsidise your health care, and I don't think I should. But what if you have diabetes and I don't?" asks Brobst.
"Is this Type 2 diabetes, because I have poor diet and brought it upon myself, or Type 1 because I'm genetically predisposed?" Computing fires back.
Brobst: "So now we're getting in the realm of what's controllable and what's not controllable, right? Now again this is my opinion, but if you have control over it, you should be encouraged - with financial incentives - to behave properly. So if you have high cholesterol and are eating two steaks a day, I'm not going to subsidise your trip to the emergency room, or your triple bypass surgery."
Computing: But what if I was ill from my poor diet, and I change and improve my life, but I have run-off conditions like, say, that Type 2 diabetes?
Brobst: "If you change your ways and you become a vegetarian, I'm okay subsidising you now because you're doing the right thing."
Computing: But how do we know all this? We'd have to measure every single one of these factors, and it feels like a rabbit hole...
Brobst: "I can measure all that now anyway - are you kidding me? Even your dentist can tell if you're a vegetarian, from the type of plaque in your teeth."
Computing: All right, but it still sounds like we're heading towards a bit of an elephant in the room here - namely embeddables.
Brobst: "Well,you can chip your pet already."
Computing: Pets can't complain about it, though.
Brobst: "Well, would you chip your children?"
Computing: I would not chip my children, no.
Brobst: "Well, in Hong Kong they're giving children bracelets. It depends on the kid and parent, but if they say wear the bracelet, [the kids] wear the bracelet. And I have a friend in the US who tracks his daughter through her iPhone - and that has value for him. He can keep her safe. She's of course fully aware of it, though.
Computing: Would you like to see a future where everyone is chipped legally, though? To help make all these health analysis ideas easier to carry out across the population?
Brobst: "No, no. But I guess it would be easier, though."
Computing: Well, yeah...
Brobst: "But [things like] physically where I am, I don't think [medical companies] need to know. For my child, maybe that's different, but does the government need to know? That would be a stretch for me. Although personally I don't care [if they know where I am]. But it's a need-to-know thing - if you don't need to know, I won't give you the data."
Computing: Right, but perhaps it's hard to keep control on specific things at some point. Once the horse has bolted, maybe it's too late?
Brobst: "Well, anyone who thinks we can't track you anywhere when we want to already... is confused."
Computing: Sure. Snowden, the NSA stuff, Prism back at home in the UK. The UK has more CCTV than anywhere else in the world.
Brobst: "Right. Especially at Manchester Airport, where they're actually matching criminal photos right in front of you to your passport. But that's fine, I have no problem with that."
Computing: But say in five years' time, where should we be in terms of how plugged in we are to analytics databases, and how much we freely share with both governments and private insurers?
Brobst: "I would like health care to be completely plugged in. That doesn't mean I think everyone should have access to it, and I as a consumer get to choose who uses it and for what purpose."
Computing: This is what Tim Berners-Lee has been saying, of course.
Brobst: "I get to choose which hospital has it."
Computing: And can you choose to sell it?
Brobst: "Right. I think we're moving towards being able to say ‘I own my data and how it will be used'. But to be economically efficient the selling of data needs to be large scale. Individual selling of data doesn't really work unless there's a standardised format. We do it in a different way - I give you a free t-shirt if you fill out my 15-question survey. That's essentially the same thing. We haven't really got good at doing that on aggregate."
Catch 22
At this point, the meeting was sadly cut short as Brobst had to run to another engagement. But it seems, while there are still some kinks to be ironed out, Brobst sees a largely positive future of connected health sensors, regulated industry practice and parents finding peace of mind in tracking their children.
Stopping short of Tim Berners-Lee's dream of data becoming a true ‘currency' for each individual - an admittedly difficult idea to carry out - Brobst seems nevertheless to be encouraging a greater level of trust in a big data future that has always seemed certain could revolutionise many aspects of life, but with it risk taking away a great many perceived freedoms.
Which choice would you make?