What is small data?

Our unique digital traces, or small data, are being collected and analyzed constantly by search engines and social media. What can we learn about our health through this data? Deborah Estrin, co-founder of the startup Open mHealth, joins Hari Sreenivasan to discuss this trend.


Our unique digital traces, or small data, are being collected and analyzed continuously by mobile networks, social media platforms, and search engines.

What can we learn about our health through this data?

Deborah Estrin, co-founder of the nonprofit start-up Open End Health joins us to discuss this trend.

So, I'm fairly positive that Google knows a tremendous amount about my life from obviously reading all my e-mails and figuring out where I am on my maps, et cetera, et cetera.

How does some of this information about me help me live a healthy, longer life?

Google does know a lot about you, which is how they've so improved search and maps and all those other tools they give you over time.

It's not really individuals in Google.

It's algorithms inside of Google.

We are starting to see many ways in which those same data streams can help you as an individual solve your problem, help understand how you're responding to a medication, help you adhere to some behavior that you would like to do because it's good for you.

How does the idea of what's in a smartphone now change our behaviors or at least help us learn from the behaviors that we're already performing?

Mobile health is this leveraging or pervasive technology to address pervasive problems, particularly when it comes to chronic disease.

Seven out of ten avoidable deaths are due to chronic disease.

86% of our health care costs are due to chronic disease.

Very interestingly, there are four behaviors that cause a lot of the suffering, reduced quality of life, and mortality associated with chronic disease.

Getting up and moving?

Getting up and moving.

What you eat.

How much alcohol you drink.

And still how much you smoke.

There are things that the public policy and that we do as a system, but then there's the individual.

And how do I become the best and healthiest, best feeling that I can be, and how do I take care of those chronic conditions that I have, because more than 50% of Americans have a chronic condition, at least one.

As we age we end up with more than one.

We end up taking more than one medication.

What mobile allows us to start to do is really have things be based on you.

How are you doing on and off this medication, with and without this dose, using this form of physical therapy?

Whatever is the behavior that you need to do that somebody prescribed to you to do or suggests for you to do, how can that mobile technology both really bring back the data of what's working and inform that feedback loop, and also help to prod you along and motivate you to continue to do it or to restart it again when we fall off the wagons, as routinely happens.

How do we take the data that we're generating and read it and interpret it well?

Let's say I put all that stuff into an app, the phone is with me, it knows I'm doing the exercises, I input exactly what pill I'm taking at what time of day.

Now, it goes somewhere, maybe to a doctor or a specialist.

How do they take all this and sift through it and say, this isn't working as well for you?

That's where the state of the science and technology is, this condition at a time, trying to bridge that gap.

You mentioned the word that is part of that solution.

In a sense, it's all about apps.

It's all about applications that take in data, a lot of which would be passively generated, so you don't have to enter it.

Maybe it's coming from the pharmacy or a smart pill box.

It's processing that data and making inferences and getting rid of the noise, yeah, this is vacation week, just because you're staying at home all the time doesn't mean you've entered a depressive episode or that your arthritis has flared.

In the same way that Google and our apps have become so much more smarter and adaptive, the same technology, the algorithms that become adaptive and allow things to personalize can be applied here.

What can an average consumer do today to figure out how to obtain the signals in the first place, gather it, and give it to the appropriate people to analyze it?

There are a growing number of applications and services that are already in the business of saying, give me access to your Fitbit data or give me access to your senior elder that you're taking care of, your father, mother, aunt, who you might have them wear a Fitbit, so that you can track how mobile they are from a day to day basis, because you're not always able to visit.

So the way that the average consumer needs to go about it for the most part is to look for those apps and services that are out there already, marketing to the particular demographic or condition.

And they are increasingly having you as part of your onboarding, letting you get access to some of your data.

I noticed in your TED med talk, you said something that was interesting.

It was about your father.

who had passed away, that there were signals that he was generating that now in hindsight you saw, oh, my gosh, certainly there was something coming, but we didn't really know about it, but now I think about it, maybe I should put a tracking watch on my elders in my family.

Explain that.

He was the primary caregiver for my mother who had slow onset dementia.

He had stopped preparing a meal or two at home during the week for her, they were eating out all the time.

Those trips and his patterns in retrospect changed.

Now, I have an older sister who is a real doctor.

And Margo noticed these changes.

But most people don't have that.

And truthfully, some of those things we had to catch up with.

When he went to see his cardiologist and the cardiologist asked him how are you doing, have things changed, his sense of self was that he was doing fine, and he loved his cardiologist, he wanted to be performing well for that doctor.

So I don't think it's so much about you putting a tracking device on your elder relatives, but rather people themselves, as they're living alone in context where they're separated from family members who in the old village context would be observing them all the time, might choose to be able to see for themselves whether their mobility is changed, whether they need a little more encouragement to get out and walk around the block.

Maybe sometime in the future we'll even have little agents that run on our behalf around our e-mail and our texting.

One of the things that's very difficult to observe is, as I mentioned about my mother, is cognitive impairment.

Advances in that field are very important, very slow to progress, because we only see patients when they're pretty far advanced into the disease.

What if we had agents that are looking at the anger we use in our e-mail or texts or the way in which we talk to our Amazon Echo, and start to detect changes in those language patterns.

Deborah Estrin, thanks so much for joining us.

Thank you.