Breast cancer is one of the diseases that researchers hope to study with data gathered from an iPhone app. Credit: Eric Risberg/AP/Press Association

Countless owners of smartphones and wearable devices are already using their devices to track their sleep, exercise, blood pressure and other measures of health. Now biomedical researchers could have an easier time recruiting these users to collect and share their data as part of large-scale clinical studies.

On 9 March, Apple launched ResearchKit, an iPhone-based platform that researchers can use to design and administer app-based studies. The company, headquartered in Cupertino, California, debuted the first five apps built with ResearchKit, which are designed to study asthma, breast cancer, cardiovascular health, diabetes and Parkinson's disease. ResearchKit will be made available as an open-source framework for other app developers in April.

Stephen Friend, president of Sage Bionetworks, a non-profit organization based in Seattle, Washington, spoke to Nature about the two apps that Sage has developed for ResearchKit. He was joined by John Wilbanks, Sage's chief commons officer, and neurologist Ray Dorsey, a collaborator with Sage at the University of Rochester in New York. An edited transcript of the interview follows.

What is the ResearchKit framework, and how can researchers use it to perform studies?

Wilbanks: ResearchKit is a set of tools and services that you can use to assemble a clinical research study. It doesn’t become a research study until a protocol gets filed and approved by an institutional review board. But what's different now is that you can actually build a lot of the patient-reported outcomes and the quantitative data-gathering into the phone itself, and you can enrol people natively through a phone.

Friend: The ResearchKit framework offers the ability to track disease signs by sensors that are on the phone or the Apple watch, or on third-party devices. Symptoms can also be tracked by people filling out survey information on themselves.

What is the Parkinson mPower app from Sage, and what data does it collect?

Friend: mPower is built to allow patients to track their signs and symptoms in Parkinson’s, and to determine if certain things they're doing are making them better or worse. It has ability to track the things that a physician would do in an office twice a year, but to do it continuously. It has a balance and walking assessment using the accelerometer in the phone. It has a way to do a test of speed and dexterity in your fingers — mimicking a classic test they use for Parkinson's. And if I say 'Ahhh…' into a phone, we can take the pitch, the tremor, all these different attributes, we can dissect those out, and they are powerful ways to assess problems in the muscles.

What advantages does mPower offer over conventional medical exams or clinical studies?

Friend: I think it's the real-time contour of the disease. Normally, you go to the doctor, you give a history. It's a snapshot. If you're trying to figure out what's the trajectory of the disease on a six-month by six-month basis, the doctor's visits probably do fine.

But on a day-by-day, week-by-week basis people are having better days and worse days. Having real-time data where you can look back and see gait and phonation — it's orders of magnitude more helpful in trying to find things that in real life make it better or worse.

What health-care question is Sage's Share the Journey app designed to address?

Friend: Breast cancer patients often get chemotherapy, radiation and hormonal therapy that leave them at the end with problems of cognition and mood. About 25% of women who have had these treatments have significant issues, and by the end of a year, about 5% still have those symptoms.

This app is designed to allow them, as they come off of therapy, to see what makes their symptoms get better or worse. It requests people who have been treated to track their signs and symptoms. The app makes it possible for women to track what it is that is happening to their mood, their energy, their cognition, their sleep, on a daily basis, and to weekly write diaries of what they're doing. And at the same time, it allows people to track their exercise, and their real-time activity.

How do ResearchKit apps handle information privacy?

Wilbanks: All data is de-identified. The data itself doesn't go to Apple. It lives on the phone. The app transmits them on a regular basis to servers, where we separate identity from study data.

Friend: The patient is given a choice: do I want to have this data seen by the researchers for this study at the university? Or do I want this data to be shared broadly worldwide for researchers?

At any time, participants can also choose to stop. The data that they have contributed stays in, because you don't know who they are. But more importantly, any day that they don't want to answer something or do some part of the study, they can pause, too.

How do those choices affect data collection for a scientific study?

Friend: I know what you mean, and we're hoping we have enough patients that it's not an issue.

How many participants do you expect to get in your studies?

Dorsey: The sheer volume of participants in the mPower study will be enormous. I was just told that in the first three hours the app became available, 680 people downloaded it. The largest clinical trial ever conducted in Parkinson's disease was 1,700, to give you a flavour of how large this study could be, and how valuable the data could be to the field. I think, with 680 people, we'll have informative results. We're hoping to get into the thousands or tens of thousands.

Wilbanks: Both for mPower and Share the Journey, we've been approved for 20,000 participants initially, and I think the goal is eventually to enrol 100,000 or more in each study.

What other types of sampling bias do you expect to encounter?

Dorsey: Clearly, one of the big limitations is that the study is only open to individuals who have an iPhone, specifically the more current editions of the iPhone. There's no great way to say how this will generalize to individuals who do not.

Wilbanks: That's totally an issue as we get started. One of the hardest pieces in this is not being able to get out beyond that bound to start with. But in many ways we have to prove that this works first before we can expand and move out to the rest of the platforms.

How do you see ResearchKit apps changing the way biomedical research is done in the future?

Wilbanks: I think it takes us from this geographic limitation where people have to go to clinical centres to get measured, or have to have their certified [licensed] professionals administer surveys.

It lets you generate valid, quantifiable information about individuals and it lets you survey those individuals with a device that they carry around with them. That means they do it a lot more often and in a lot more detail, and that lets us explode the sample size of clinical studies. It lets us explode the diversity: gender, ethnic, geographic, economic. And it lets us start to get a more stratified picture of individual variation. It's really hard to do that with the methods of traditional clinical studies.