Ideally, we’re all supposed to see our doctor once a year for a checkup. It’s an opportunity to see how we’re doing from a health perspective, address any concerns or issues that we may have and catch any emerging issues before they become true problems.
But those visits are really only one-time, infrequent snapshots of health. They don’t give a full view of how we’re doing or feeling.
Now, think for a moment about how often you post something to Facebook or Twitter. Do you post anything about whether you’re feeling ill or down, or haven’t slept well? Ever share how far you ran, the route you biked or your number of steps for the day?
Every time you do, you’re creating a data point—another snapshot—about your health. Put those data points together, and what starts to emerge is a rich view of your health, much richer than one based on the records of your occasional medical visit.
As John Brownstein, PhD—director of the Computational Epidemiology Group (CEG) in Boston Children’s Hospital’s Computational Health Informatics Program and the hospital’s new Chief Innovation Officer—explains in this episode of the Harvard Medical School (HMS) Labcast (click the image above to hear it), this view has a name: your digital phenotype.
“People in their day-to-day lives are interacting with technology,” he told Labcast. “They’re interacting with their mobile phones, they’re on social media, they’re wearing their Fitbit device, they’re getting on their connected scale…. Basically, you’re quantifying your activity, quantifying your sleep.
“All of that data collectively can build a pretty rich characterization of an individual,” he continued. “What we’re suggesting is that that data is becoming increasingly important and ultimately should end up feeding back into clinical decision making…[and our] understanding of population health in very new ways—especially when it comes to health behaviors.”
It’s a concept Brownstein and several collaborators described in a recent paper in Nature Biotechology:
Through social media, forums and online communities, wearable technologies and mobile devices, there is a growing body of health-related data that can shape our assessment of human illness.… Through the lens of the digital phenotype, an individual’s interaction with digital technologies affects the full spectrum of human disease from diagnosis, to treatment, to chronic disease management.
Your sleep profile—on Twitter
As they reported in the Journal of Medical Internet Research, Brownstein and a team from Boston Children’s—including CEG researchers David McIver, PhD, and Jared Hawkins, MMSc, PhD—and Merck have put this concept to the test by building a digital phenotype of insomnia and other sleep disorders based on Twitter data.
“Sleep deprivation and chronic sleep disorders are not well understood,” Brownstein said in a statement about the JMIR study. “We wanted to see if we could use new forms of online data, such as Twitter, to characterize the sleep-disordered individual and possibly uncover new, previously undescribed populations of patients suffering sleep problems.”
The team collected a virtual cohort of 896 Twitter users whose tweets suggested sleep issues, like use of the hashtag #teamnosleep. They then compared information about these users (e.g., age, tweets per day, time zone, number of followers) with another group of randomly selected users who didn’t tweet sleep-related words.
The preliminary profile—and it is, as Brownstein points out, just preliminary—that emerged suggests that an insomniac tweeter is less active on Twitter on average but tweets more during traditional sleeping hours. The increase in negative sentiment in their tweets suggests that sleep-disordered users could be at an increased risk for psychosocial issues.
“But,” he added, “they suggest that social media can be a useful addition to our toolkit for studying the patient experience and behavioral epidemiology of sleep disorders.”
Bringing it back to one-on-one care
While Brownstein hopes this kind of research will create new opportunities to enhance public health, he also wants to see whether it can give clinicians a more direct, complete—and, ideally, actionable—window into a patient’s health. As he told HMS Labcast:
“When you interact with your physician, they say, ‘How many times a week are you going to the gym?’ We all overestimate, but the wristband doesn’t lie. And ultimately you want your physician to have the best possible view.”