Apple made some grand claims for ResearchKit, a software platform it promises will turn the iPhone into a powerful tool for biomedical research. In a synchronized release, independent researchers also rolled out five apps to study an assortment of complicated diseases. It's an interesting shift in strategy for the world's most popular smartphone. For the first time, instead of you using the phone to do (nominally) useful stuff, other people will be using your phone, and the useful stuff will be you---or at least, data you generate.
The problem ResearchKit hopes to solve is the difficulty in getting big data about disease. Typically, you want as many people as possible involved in a study about, say, how heart disease develops. And indeed a few big, prospective, multi-year studies with thousands of subjects do exist, and have produced famously useful data. But they're rare, because they're hard to do, and thousands of subjects is nothing compared to, say, the 700 million or so people who own iPhones. “It really is a new way of doing business,” says Euan Ashley, a doctor at Stanford who helped create MyHeart Counter, an app designed to monitor physical activity and other risk factors for heart disease. "We’re going to start to see patterns that we weren’t able to see before. Hopefully we can represent all the different subgroups and have tremendous diversity across the population.” Other diseases under investigation---Parkinson's, asthma, diabetes, and breast-cancer treatment recovery---also have apps available.
Even though the data will, by definition, come only from people who can afford an iPhone, researchers building the apps say the relatively large number of iPhone users nationwide---estimated at 63.2 million in 2014---will help reduce what's called "selection bias," the problem of research studies looking only at one kind of person. “If you can do something at scale, then the challenges of selection bias becomes smaller,” Ashley says. Biases never go away completely, but current studies already suffer from plenty of problems. For example, studies may only include people who happen to live near a research center or have the time to participate. Because the apps are open-source, researchers hope to eventually expand to Android and other platforms. Still, regardless of platform, smartphone users skew toward the young, wealthy, and Asians and whites. Not exactly representative.
Also, data used in scientific research is supposed to be more accurate than what your typical wearable gives you. "From a research perspective, it's a little challenging," says Errol Ozdalga, a physician at Stanford who wrote a review article about smartphone apps for physicians and medical students. "You have to be aware that a smartphone is not going to give you all your tracking activity." In other words, any researcher hoping to use iPhone-gathered data will have to be cautious about the accuracy of what they're getting. "We don’t know how these are going to work and if they’re really going to advance research," says Eric Topol, director of the Scripps Translational Science Institute and author of The Creative Destruction of Medicine.
All of the ResearchKit apps use a combination of questionnaires and basic data your phone collects---for example, using your phone's accelerometer to count how many steps you've taken as a measure of physical activity. If you have a wearable like the Apple Watch, it can measure how your heart rate changes while exercising. The mPower app, which tracks Parkinson's disease, uses the iPhone touch screen to measure hand tremors and the microphone to gauge voice trembling.
But gathering data isn't the same as understanding it. "What needs to happen is more evaluation about how these apps are going to be used to achieve these meaningful changes in health," says Mitesh Patel, a health researcher at the University of Pennsylvania who studies wearables and health-related smartphone apps. ResearchKit will certainly give researchers more access, he says. And that's good---it at least means a start at finding ways to use it. "This allows us an easier way to access that data and to be able to critically evaluate what data is useful or what data is not," Patel says.
Yet Patel and Ozdalga both think that despite those caveats, apps represent the future of medicine. Apps that motivate people to change their behavior---to eat better and to exercise more, say---may still have a huge impact, Patel says. "The most important thing is how to use this data to create behavior," he says. "If you're not changing someone's behavior, what's the point?"
