An App that Looks for Signs of Sickness
Mobile-phone activity can provide a warning of disease flare-ups.
Ginger.io, a startup spun out of the MIT Media Lab, aims to use data collected automatically from mobile phones to warn users and their physicians that they may be on the verge of a manic episode or a bout of debilitating illness.
The company has developed a mobile-phone app called DailyData that analyzes information on the user’s location and the frequency of calls and text messages to determine whether that person is having health problems. Ginger.io will market the software to health insurers and others who could use the aggregate data to better understand the links between health and behavior.
“Changes in medication or mood are tied to communication and movement patterns,” says Karan Singh, one of the company’s founders. “Call diversity is a great example. When people fall into a cycle of depression, they tend to go into isolation and only call a couple of people.”
The start-up is part of a growing effort to use the sophistication of smart phones and other wireless devices to track behavior as it pertains to health. But Ginger.io’s app is unique in that it collects data automatically. Most health-tracking programs rely on information manually entered by the user, but many people eventually lose interest in using the program. Mood-tracking apps, for example, typically ask users to rate their moods, a task a depressed person is liable to neglect.
The DailyData app first creates a baseline model of a user’s mobile-phone activity and then searches for deviations from that pattern. For patients with bipolar disorder, a burst of text messaging or phone calls could signal a manic episode. “We can compare this to your past behavior, or to aggregate behavior of individuals of your approximate age and demographic,” says Anmol Madan, another of the company’s cofounders.
Users can supplement the automatically collected data with manually entered information on medication, symptoms, and social activity, and can look at visualizations of their data on their phone or on a website. When the app detects behavioral changes, it will send out alerts, such as “You’ve been working harder on the weekends” or “You seem really stressed, is everything OK?”
Accurately predicting the meaning of changes in communication patterns is likely to be challenge. For example, the cell-phone profile of a person who stays home and stops calling friends for several days in order to meet a work deadline would be similar to that of someone who stays in bed and stops answering the phone because his depression is getting worse. But Singh says the algorithms underlying the app are flexible and can be tuned to be more or less sensitive to behavior changes, and that user feedback will also improve them. “With more data and users, we expect to get better at predictions,” says Madan.
The initial release of DailyData gives only the user access to his or her results. But, Madan says, eventually the alerts could go to family members or caregivers, who could then intervene if the patient seems to be on a downward spiral.
Apps of this type may worry privacy advocates. Revelations about how Apple and Google use location data caused a public outcry last month. But in the case of DailyData, the benefit to the user may outweigh privacy concerns. Continuous monitoring via mobile phone could give both patients and their physicians better insight into health problems, says Deborah Estrin, founding director of the Center for Embedded Networked Sensing at the University of California, Los Angeles. A doctor treating a patient for depression, for example, typically only gets to see the patient for a few minutes once a week or once a month, and the patient’s mood that day could be influenced by something that happened that morning. “You have the opportunity to look more objectively at what the past two weeks or months have been like,” says Estrin. “It’s so powerful to leverage the technology that people carry around willingly. People are already harvesting this type of information to serve marketing,” she says. “Why not use it to help serve people themselves?”
Joseph Kvedar, director of the Center for Connected Health at Harvard Medical School, says that if such an app is to be adopted widely, it’s essential that it collect the data automatically. “You have to make it really easy,” he says. “Require even the least little step and people lose interest.”
On the flip side, a growing body of research suggests that giving people information on their behavior can benefit their health. “When we measure something and share it back with individual, it raises awareness in a special way,” says Kvedar. “It gives insight into how lifestyle is connected to health in a way you can’t get without quantification.”
Ginger.io plans to market its software to health-care providers, pharmaceutical companies, health insurers, large employers, and chronic-patient communities. These groups would offer the app to patients or employees, and would in turn get a set of aggregate statistics about and trends in the health and behavior of these groups. “For a [health-care] provider or academic researchers, this might help them understand how people behave when they’re symptomatic,” says Madan. A pharmaceutical company might gain insight into links between behavior and medication and health, such as whether physically active people get better faster. “These are all novel data that they never had access to before,” says Madan.
The company is working with Cincinnati Children’s Hospital on a pilot study of patients with inflammatory bowel disorder and Crohn’s disease, both painful intestinal conditions. Physicians will try to determine whether behavior changes prior to a flare-up.
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today