Health Data: Let's Go Exploring
Think about it for a moment. When you consider what you "know" about health, where does that knowledge come from? While we all have our sources—doctors, friends, news articles—our knowledge at its core is derived from research. And that research is built on a foundation of data.
Data about health typically come from several types of sources: clinical data, gleaned from electronic health records or chart pulls, and billing and claims data, which are byproducts of the health care process; and public health surveillance data, which are specialized collections about particular topics or populations. All of these sources can then be supplemented, at a considerable cost, by original data collection efforts specific to a particular study.
These different types of data are like pieces of a jigsaw puzzle; when assembled, they create a more complete picture of health.
But a piece of the puzzle is missing. Or it has been up till now.
More and more of us are generating data about our day-to-day lives—either intentionally (we're talking about you, FitBit®, UP™ and Nike+ FuelBand® users!) or somewhat unwittingly, through the data-generating magic of our smartphones. These data could tell us a lot about our health.
In addition, there is an increasing recognition that health is not so much a factor of our engagement with the health care system (though that helps), but something we both experience and influence 24/7. We eat, we drink, we move around, we sleep. We experience pain and discomfort, elation and sadness.
These data — collected by and about people — could give researchers a new window into how people experience health in everyday life. They could help to fill out a new and potentially very important piece of the puzzle.
Tracking Everyday Health
Day-to-day variation in the experiences of daily living is important—as health parameters that may be important to people as they manage a chronic illness or as indicators to a healthy person about their wellbeing. But because tracking day-to-day variation requires day-to-day data collection, these experiences have typically been difficult to study.
Now, the barriers are starting to fall, and we find ourselves with a new opportunity. We have activity-tracking devices. We have smartphones—now used by half the U.S. population—that generate potentially valuable data on movement, location, social behavior, and even tone of voice. We have apps on those smartphones, prompting us to record data about our moods, pain scales, menstrual cycles, diet, sleep, exercise, and more. Add-on devices such as blood glucometers and asthma inhalers may soon be as common as the accessories that allow us to use our iPhones to process credit card purchases.
And it’s not just our phones and phone accessories that are tracking our data. Internet-connected scales and blood pressure cuffs, online patient communities, and provider affinity groups also generate data related to trends in health and health care. This is of course just a partial list -- please comment below with any sources we've left out.
We’re excited about the insights that are starting to emerge from analyzing these health data:
- Based on photographs of 500,000 meals, Massive Health posted a series of findings about how people eat, including the discovery that the healthiness of our food decreases with every passing hour of the day.
- CureTogether has shared results of user-reported treatments for arthritis and other conditions, showing which treatments are more common and which seem to more effective.
- PatientsLikeMe has published studies highlighting day-to-day experiences of those living with illness of which many providers were not aware.
- A team of researchers found unreported drug side effects in web search data before the FDA.
- Researchers have used Google search data to discover novel population-level seasonal trends in searching for mental health information.
These findings are a start, but what more is waiting to be discovered? We’d like to find out.
We’re not announcing a new research program—but we are starting an exploration. We’d love to bring companies that have these data together with researchers who can analyze them for new insights. And we’d love to have the people who create these data with their daily routines and observations engaged in generating new hypotheses about what could be found.
We're delighted to be working with Kevin Patrick and Jerry Sheehan at Calit2, who will be leading the Health Data Exploration project, and we've lined up some great advisors: Linda Avey, Hugo Campos, Sendhil Mullainathan, Tim O'Reilly, Larry Smarr, Martha Wofford and Gary Wolf.
We’d love for you to help us think it through as well. If you're interested in taking part in this exploration, let us know.