In 2009, researchers at the Massachusetts Institute of Technology gave a dorm full of students smartphones and tracked where they went, who they called and texted, and at what times they communicated. The researchers found that the data pouring out of the phones could reliably tell when a student was ill: Those stricken with the flu moved around much less, and those who were depressed had fewer calls and interactions with others.
Anmol Madan, the PhD student who led the study, concluded that the findings might be useful outside of dorms. There are now more than 60 million smartphones in the U.S., and they’re “incredibly powerful diaries of a person’s life,” he says. So in November 2010, Madan and his classmate Karan Singh, both 29, started Ginger.io to mine those diaries and provide the kind of detailed, persistent health monitoring that doctors and researchers have only dreamed of. “There hasn’t been large-scale, real-world data about how people behave” before now, he says.
The seven-employee Cambridge (Mass.) company raised $1.7 million in venture capital in October from True Ventures and Kapor Capital. They’ll use the money to build a series of apps that health-care providers, drug companies, and insurers can offer to their patients. As in the MIT study, Ginger.io (the name is a nod to the health benefits associated with ginger) relies on a branch of computer science known as “machine learning” to sort through the tens of thousands of data points coming out of a smartphone each month, identifying a user’s typical pattern of behavior. When someone deviates from that pattern, Ginger.io can trigger a response that Singh likens to a car’s “check engine” light, alerting friends or doctors that they may need to intervene.
The Cincinnati Children’s Hospital Medical Center is currently conducting the first test of Ginger.io’s technology, a study of teens and young adults suffering from inflammatory bowel disease (IBD). The chronic condition causes occasional diarrhea, stomach pain, and fever, and is often treated with a two-week course of steroids, which can be harmful when used for too long. The hope is that Ginger.io can detect exactly when an attack ends by, say, noticing when someone leaves the house for the first time after several days at home, and reduce the time patients are on steroids. A handful of patients started using a Ginger.io app in the fall, and Michael Seid, the researcher who’s leading the study, expects to have 50 enrolled by early 2012. Ginger.io should be “an effortless way to get a much finer-grained continuous measure of health status,” he says.
Ginger.io also won $100,000 in a November contest sponsored by Sanofi-Aventis to come up with ways to help diabetics. Diabetes patients are more susceptible to depression, which in turn increases the chance they’ll stop taking their medication. Ginger.io’s prototype smartphone app raises an alert when a diabetic starts behaving in a way that signals depression, so that doctors or family can offer help. “People don’t self-report that isolation,” says Singh. “They don’t necessarily say, ‘Yes, I’m talking to less people this week,’ or they don’t necessarily realize that they’ve actually started to close themselves off.”
Other groups are working on similar, passive health monitoring: Researchers at the National Institutes of Health are using mobile phones to track recovering drug addicts in Baltimore to understand what triggers relapses, and companies including WellAware Systems place motion sensors in the homes of the elderly to detect if they’re in trouble. The challenge for tech startups in health care is that it’s “traditionally a conservative market, and it’s a regulated market,” says Jonathan Collins, a London-based analyst at tech researcher ABI Research. Ginger.io will also have to reassure users that their data won’t be misused. Madan and Singh say Ginger.io doesn’t read the content of conversations or text messages and limits the information that goes to organizations such as insurance companies, so they can’t use Ginger.io to tell how often someone goes out for a cigarette, for example.
The company expects its monitoring algorithms to get better as phones and accessories evolve to collect more data. One hint of the future: Some earbuds already measure heart rates for athletes, and Apple has a patent on earbuds that take body temperature readings. “There’s just so much information on [smartphones], and there’s so many new sensors,” says Singh, that the potential to improve people’s health with better data “is only going to continue to grow.”