Health 2.0 featured a panel devoted to the “new environment for better health care decisions.” I shared some new findings and I’d love to hear what you think:
[Update: the video of my talk is now online.]
Where I work, at the Pew Research Center, we use data to hold up a mirror to society so you can see yourselves clearly. We don’t tell you what to do about your reflection. We just want you to see yourself as you really are.
We also provide a window into other people’s lives, which may be very different from your own. We survey in 60 countries and on a wide range of topics so people can make informed decisions, based on data. We are huge fans of reality-based decision-making and I hope you are too.
In our health research, we visit other kinds of nations.
The land of the newly diagnosed cancer patient. The realm of rare disease. The mighty matriarchy of women whose kids are sick. You may not be a citizen of these nations, but you may want to understand them because they are your target audiences.
One of our surveys last year identified a previously unmeasured nation within our borders: Trackers.
This was the first national survey to measure the magnitude of health data tracking. We fielded it for 3 reasons:
1) The medical literature shows that tracking is a low-cost, effective health intervention for losing weight or managing chronic conditions.
2) My fieldwork in a variety of online patient communities turned up multiple examples of home-grown, kitchen-table tracking solutions.
3) Media stories about trackers often focused on fitness nuts and geeks who love gadgets and that didn’t fit the reality of what I was seeing in the field.
So the Pew Research Center and the California Healthcare Foundation went after the data.
[slideshare id=26508868&doc=foxhealth2conmondaypanel-130924140641-phpapp02]
7 in 10 US adults are tracking some aspect of health, their own or someone else’s.
They are tracking weight, diet, exercise routines, and health indicators like blood pressure, blood sugar, headaches, and sleep patterns. Most track for themselves, but some track on behalf of the loved ones they care for. Half track on a regular basis, half when something comes up, like a symptom flare or a new goal. Technology plays a minor role for most trackers. Pencil and paper is more common than apps or online tools. And fully half keep track just in their heads. Some share the data with their clinician. Many do not.
Let’s talk about one huge market within health care. We will release a report about this group later in October, but I wanted to share a sneak preview.
45% of US adults are living with a chronic disease.
About half of them are dealing with 2+ conditions. They are statistically less likely to have internet access or to own a cell phone, so let’s just keep things in perspective. 7 in 10 U.S. adults living with 1+ condition say they track some aspect of health; 8 in 10 U.S. adults living with 2+ conditions do so.
Tracking data is not a hobby for them. They may not have a choice, especially those with two or more conditions.
They are trying to use data as a mirror, to see themselves more clearly and do the best job they can, trying to stay well, given the tools and knowledge they have.
Trackers living with chronic conditions are more likely than other trackers to take formal notes and to use medical devices like glucometers, but less likely to use other types of technology. They are more likely to say that tracking has had an impact on a health care decision, led them to ask a doctor new questions or seek a second opinion, or changed their overall approach to health.
So yes, many people are collecting data. But what about connecting data? Connecting with other people, to share and learn from each other. That’s a big piece of the health care system and we need to talk about it when we talk about data.
Think about the last time you had a serious health issue or experienced any significant change in your physical health. That’s often a big decision-making moment for people. To whom did you turn for information, care or support?
Our national survey found:
- 70% of U.S. adults got information, care, or support from a doctor or other health care professional.
- 60% turned to friends and family.
- 24% turned to others who have the same health condition.
People living with chronic conditions are even more likely to say they connected with each of these sources, and are the most likely group to say they consulted a health professional.
Again, technology still plays a small role in those connections. Health care is still very much an offline activity for most people.
So picture this: There is a huge group of American adults who have cell phones – 91% – and those phones can serve as a tracking device.
There is another huge group – 7 in 10 adults – who are already tracking their health, but mostly offline, not using any technology. Yet they say that tracking is having a beneficial effect on their health and the literature shows it can.
45% of American adults are living with a chronic condition, doing a ton of tracking and a ton of interaction with the health care system, but again, mostly offline and not integrating the data within the system.
This is a huge opportunity. What are you going to do with it?
Hold data up as a mirror so people can see themselves more clearly.
Use data as a window, into other people’s lives, so you can help them reach their goals.
Seize this opportunity in the age of data and connection.
“This is a huge opportunity. What are you going to do with it?”
I’ve been meaning to get your thoughts on this for some time. You have this massive amount of useful information being generated for health. How do you distill it into the types of insights that can be universally available and that people can use to make good decisions?
I’m aware of large data sets being trimmed down into bite sized insights — we use them all the time in the form of Google, Pinterest, Quora, Amazon. But all of those require several rounds of human curation (linking, re-pinning, marking reviews helpful, etc.) to surface the most relevant insights. Is there a way to go from Big Data to relevant insights without human curation? If you or your readers know of successful examples I’d love to hear them.
To reappropriate one of your jokes, I’m asking, “How do you get the Big Data to fit into Skinny Jeans?” Ha ha ha.
I love it! Let’s spread this question far & wide — I’m sure someone in our community will have insights to share.
Thanks for your help Susannah! After falling down the rabbit hole that is Quora, I’ve emerged without an answer, so I posted the question, but it’s kind of hard to explain. 🙂 I do think it’s important for the self-tracking movement as it goes form DIY to DIT as you say.
http://www.quora.com/Quantified-Self/Is-it-possible-to-convert-the-insights-from-the-self-tracking-quantified-self-movement-into-relevant-insights-without-human-curation
I am not sure I completely understand the context of your question. I suspect that when you talk about getting relevant insight from Big Data, you have in mind an automatic, i.e., computer-based, means to do so. As you probably know, computer techniques that attempt to extract insight or knowledge from data come under various names such as machine learning, data mining, or data analysis. Whatever the name, all predictive techniques (i.e., algorithms that build models from data that can then be used for future prediction) rely on labeled data, hence human curation; but that is also what we humans require. I cannot guess that someone likes or dislikes something, finds a review interesting or not, without that individual telling me so. For the same reason, systems that purport to make recommendations for us, or predict our likes/dislikes, generally rely on some degree of feedback from us. And when they appear to be working without it (e.g., Amazon’s recommendation engine), they still like the users to answer questions like “was this review helpful?” so as to tune their recommendation engine. So, while they can do a lot with little input based on observed behavior (e.g., what else you have clicked on, what other movie you’ve seen, what other people who look like you are doing), they depend on feedback to improve their predictions. But I may be misunderstanding what you mean by “human curation.”
Hi Christophe, Yes, that is exactly what I mean. As Susannah commented, one of the driving goals today is “self-tracking => better decision making for myself” and in the future “self-tracking => better decision making for other people.” I agree with you that it will depend on “feedback to improve their predictions.”
The question received a couple of answers on Quora. I tend to agree with the CEO of Zenobase.
I can definitely see experts and machines curating the data generated by self tracking, but then they (the experts) would own the insights, and not the people doing the tracking. In any event, it will be fascinating to see how things shake out.
Thanks, Brett, for reporting back! I’ve been traveling (again) and am just now looking at the Quora thread.
I’ll quote your question and the first answer for posterity and convenience:
QUESTION: Quantified Self: Is it possible to convert the big data from the self-tracking (quantified self) movement into relevant insights without human curation?
For many of our decisions (where to eat, how to decorate) we use tools like Yelp and Pinterest that convert big data sets into relevant insights. But all of these tools rely on many levels of human curation (marking reviews helpful, repinning, tagging) to surface relevant insights. Is there a way of delivering self tracking derived insights to individuals without human curation?
ANSWER: Eric Jain, CEO & Founder, Zenobase.com
No, unless the target audience is expert users. Without human curation of the rules that are used to generate the insights, and without human review of the surfaced insights, much of the “insight” risks being either useless or misleading. There is less tolerance for misguided health recommendations than for poor book suggestions.
-end of quote-
My most recent trip was up to the Bronx to participate in a symposium for medical school faculty at Albert Einstein College of Medicine and affiliated NY-area schools (Columbia, NYU were two of the institutions who also sent clinicians and students).
Here’s a Storify that captures slides, images, and tweets from the two keynotes — mine and Kevin Pho (aka @KevinMD):
http://storify.com/SusannahFox/social-media-s-use-in-medicine
One of the points I don’t think I made clearly enough was something that E-patient Dave shared with me, along with slide #20 (which he, in turn, had borrowed from Lucien Engelen).
If you look at the slide, you’ll see it depicts the transition from Health 1.0 (institutions hold all knowledge) to 2.0 (exchange of information among networked patients and, separately, institutions) to 3.0 (true network of networks).
Dave says: “The connections among people are like capillaries, carrying nutrients and creating new processes. There is no substitute for the trained mind to put information in context but there is also no shame in a clinician not knowing a piece of information found by a layman.”
Eric Jain’s answer resonates with Dave’s insight, don’t you think? And it resonates with clinicians’ experience and Pew Research data that they are still a trusted resource — even THE trusted resource — because they are the trained minds which can help make sense of people’s collected, curated insights.
Thank you for your insights regarding this important topic, Susannah. Its clear that data can be leveraged to make decisions that improve healthcare outcomes. Increasingly, we hear about “Big Data” but I have been even more interested in what Francis Pedraza and others refer to as “Dark Data”. Dark data in healthcare includes all of the collected information that is A) not collected. B) Collected but not accessible for decision making or C) collected and accessible but not utilized. You’ve referenced all three forms of dark data in this article ( without explicitly mentioning the term).
The key to bringing dark data to light is digitization. Digital data is much easier, faster and efficient to work with. Think about the difference between physical books and ebooks.
Sadly, the majority of the really neat ways to digitally collect data right now is too labor intensive for all but the most dedicated data nerds.
But my experience with the quantified self movement has convinced me that we are moving towards harnessing such data in a much seamless way. Mobile phones will play a huge role.
I’m looking forward to your upcoming study on chronic conditions.
Thank you! The other term I have learned, from Gilles Frydman in particular, is the “Deep Web” — all the material that is not indexed by search engines, for example. How do we surface it and use it to improve health?
When people talk about Dark Data do they acknowledge all the paper note-taking that people do? It must fall under category B.
Exactly. Note taking is very helpful. Some of the best data I’ve seen from patients are simple symptom diaries and dietary log books. But so much of this is Category B Dark Data. Indeed. Even in a medical practice with a top notch EMR; most diabetic patients log books (even if scanned into the system) are still category B data. Very powerful but under utilized. The challenge is to make the act of digitally logging medical information as simple and effortless as scribbling in a log book. In this sense I think the interface is more important than the tool. Right now we are still focused on developing better tools.
Great topic Susannah. There are actually 2 issues here: 1) “mining” the wealth of offline data that already exists; and, 2) the information going forward. I will leave thoughts and suggestions on the former to the more technologically informed but I do have a thought or two on the latter.
In short, its a question of how you take a behavior pattern that is analog (being in person discussions whether on the phone or face to face) and making it digital (whereby records are created) without requiring a great change on the part of the individuals; otherwise frankly, they just won’t do it. What we have found is creating an digital dialogue between the “patient” and their care givers/friends family where, these individuals with chronic issues respond to inquiries (as created by them) about their condition(s), with an outreach to their care givers/friends/family when desired provides a huge step in the direction you are discussing. This Reflective Care approach opens up the ability to collect this mass of data and in a manner that provides immediate personal value to these individuals, requiring nominal change in how they interact now (if at all), and provides a wealth de-identified data for research.
The key to it all, to access this data, is to provide an almost invisible means of interaction, allowing the dialogue to continue as it does now, taking into account the practical realities of human behavior. The great news as this is not theory but something we are doing everyday.
This is a great discussion topic and I look forward to reading more insights. Thanks.
Howard
Thanks, Howard! For those who don’t know him, click on his name to check out LifeWIRE.
ALS has several good websites, particularly Patients Like Me, that helps track progression and techniques for coping with a fatal disease with no treatment or cure. They have offered to work with researchers and the FDA. This may be unique to the rare disease world. We often feel isolated as this devil of a disease progresses and these sites can make a remarkable difference.
Yes, it was field work in communities of people living with rare and life-changing conditions that clued me in to the importance of tracking.
What I tell doubters or newcomers: Tracking is a tool for solving personal mysteries. Rare disease is one of the biggest, most challenging mysteries you will encounter in life. Let’s learn from these pioneers. We are all going to face mysteries and puzzles, our own or those of people we love. Tracking is often DIY — what’s next is DIT (do it together).
Physicians prescribe tracking for a few conditions such as diabetes or CHF. Tracking is less essential to the management of most other conditions. I would be interested if patient tracking in these cases has been used to successfully initiate engagement.
Richard
Thanks, Susannah–for a great post–and a great talk at Health 2.0. I thought these new Pew findings were interesting. Several of us were pondering some of the implications of these tracking trends. In fact you inspired my post on THCB from last night–thank you! http://thehealthcareblog.com/blog/2013/10/11/confessions-of-a-self-tracker/comment-page-1/#comment-452836
Big data has been a huge topic at our recent conference of ACO and pharma industry leaders. Their big challenge is just what you have identified: how to connect (not just collect) data. ACOs want to know that their population health interventions are working to improve quality and reduce costs among the “sickest of the sick.” More of them are looking for meaningful outcomes data on the treatments they offer patients, including pharmaceuticals. I think we’re going to see tracking tools grow across the industry as various stakeholders try to help patients improve compliance, manage complications of treatment, and measure treatment progress.
Hi everyone, apologies for being late to the table!
We are currently developing a mobile app allowing people with health conditions to track their symptoms, medication and other variables they see as important. There’s a strong emphasis on making this as quick and easy as possible and providing users with a visual ‘report’ at the end of each week so as they and their doctors can gain insight into their condition. If anyone is interested to know more or get involved then do let me know. I’ve got an overview document that I’d be more than happy to share with you. More info on us available here: http://www.designandprosper.com
Thanks,
Stu (stu@designandprosper.com)
Thank you! The conversation is never over — and you’ll never be too late to share insights and tools like this.