Keeps the doctor away

The inside story of how Apple’s new medical research platform was born

Omar Bustamante/Fusion

On September 27, 2013, during a dimly-lit presentation at Stanford’s MedX conference, Dr. Stephen Friend told an audience about the future of medical research.

“Imagine ten trials, several thousand patients,” said Friend, the founder of Seattle-based Sage Bionetworks, a nonprofit that champions open science and data sharing. “Here you have genetic information, and you have what drugs they took, how they did. Put that up in the cloud, and you have a place where people can go and query it, [where] they can make discoveries.” In this scenario, Friend said, patients would be able to control who could access their information, and for which purposes. But their health data would be effectively open-sourced.

The crowd was receptive. Several people looking to share their data with scientists stood up to ask what options they had. There were a few open-source health data projects in the works, Friend replied, but nothing fully-formed. “We’re pretty close,” he reassured them.

He was closer than he thought. Sitting in the audience that day was Mike O’Reilly, a newly minted vice president for medical technologies at Apple. A few months earlier, Apple had poached O’Reilly from Masimo, a sensor company that developed portable iPhone-compatible health trackers. Now, he was interested in building something else, something that had the potential to implement Friend’s vision of a patient-centered, medical research utopia and radically change the way clinical studies were done.

After Friend’s talk, O’Reilly approached the doctor, and, in typical tight-lipped Apple fashion, said: “I can’t tell you where I work, and I can’t tell you what I do, but I need to talk to you,” Friend recalls. Friend was intrigued, and agreed to meet for coffee.

Stephen Friend. Photo: Mirella Boot/Flickr.

Stephen Friend. Photo: Mirella Boot/Flickr.

Fast-forward roughly 18 months. Last week, Apple unveiled ResearchKit, an open-source platform that will make it easier for scientists to build apps that collect health data for research studies from volunteers, along with five iPhone apps aimed at some of the most costly medical conditions in the world. A day later, thousands of people had already downloaded these apps. The sheer number of participants was so huge that many are already calling ResearchKit and its companion apps a revolution in how medical science will be done.

The idea behind ResearchKit was to use the iPhone’s ubiquity to give scientists unprecedented amounts of clinical data. By using the iPhone’s built-in accelerometer, microphone, camera, and pressure sensors—as well as a bevy of personal trackers that can be connected to the iPhone, like the FitBit, glucose monitors, or AliveCor’s portable electrocardiogram recorder—scientists would be able to gather activity and biometric data on people who opted in to be part of research studies. A typical clinical study might include hundreds or thousands of subjects; a ResearchKit study could easily include hundreds of thousands, perhaps even millions. Plus, ResearchKit studies would have another unparalleled advantage: the cost of recruiting subjects would be basically zero.

“Not to be part of it would have been crazy,” Euan Ashley, one of the Stanford University investigators behind the myHeart app, which tracks cardiovascular health, told me. He and his colleagues have been working on myHeart with Apple for more than a year.

Friend immediately grasped the potential benefits of collecting health data at Apple’s scale. In the wake of the MedX meeting with O’Reilly, he made frequent trips to Cupertino and other cities to meet with scientists, engineers and quantified-self geeks. During one such trip, Friend helped organize a DARPA-funded workshop on how biosensors might help scientists understand Parkinson’s disease, a condition that would turn out to be the focus of one of ResearchKit’s five debut apps.

researchkit

Given Apple’s historical approach to data-hoarding—and the way it has traditionally refused to open-source its code—its willingness to partner with an open-source advocate like Friend is a surprise. But Apple may not have had much of a choice. Given the upcoming release of the Watch, “there is probably some interest in Apple in leveraging that new [wearables] market…to attract more people to the iPhone platform,” said Bernard Munos, the founder of the Innothink Center for Research in Biomedical Innovation. “They could also invite entrepreneurs to make clever plug-in devices.” If people are hungry to track and share their symptoms, they’ll likely flock to the platform that gives them the best tools to do that. Other tech companies like Google and Microsoft are likely working on a similar projects, but now that Apple has beat them to the punch, they’ll have to play catch-up: Whatever they come up with will be measured against ResearchKit, especially when it comes to privacy.

“No one wants to entrust their health data to a company that’s going to sell them to the highest bidder, and the highest bidders usually include the worst privacy abusers. Apple has taken a very principled stance,” Munos added. “It’s the kind of reassurance people need.”

Apple’s privacy record is far from spotless, but its approach is different from that of Google or Facebook, both of which make money by selling access to their users to advertisers. Apple is a hardware manufacturer—it sells gadgets, not targeted ads—and that hardware focus is at the heart of why Friend, who’s been trying to build something like ResearchKit for years, chose to work with Apple rather than a competitor.

Companies like Google and Facebook “make their power by selling data…They get people information about other people,” Friend told me. “Apple has said, ‘We will not look at this data.’ Could you imagine Google saying that?”

Jeff William describes ResearchKit at Apple's Spring Forward event.

Jeff William describes ResearchKit at Apple’s Spring Forward event.

The first five ResearchKit apps will target some of the most pressing and costly medical conditions out there: breast cancer, heart disease, asthma, Parkinson’s disease and diabetes. As the apps were being built, “Apple was just there as a facilitator, building the framework in the background,” said Euan Ashley, the Stanford researcher who worked on the myHeart app. “We each designed our own app with no input from Apple.”

In terms of policy, though, Apple’s role was less passive. Just three months before the press event where ResearchKit and its new wearable, the Apple Watch were announced, Apple executives, including O’Reilly, met with the U.S. Food and Drug Administration, the government agency that oversees both medical devices and clinical trials. Apple told the FDA that it “sees mobile technology platforms as an opportunity for people to learn more about themselves. With the potential for more sensors on mobile devices, Apple believes there is the opportunity to do more with devices, and that there may be a moral obligation to do more,” according to a meeting memo obtained by Apple Toolbox through a Freedom of Information Act request. The memo didn’t mention ResearchKit explicitly, but we can reasonably infer that Apple sees clinical research, which falls under the purview of the FDA, as part of its “moral obligation.”

Beyond its genesis at Stanford, the other details of how ResearchKit came to be are still murky. It’s unclear, for instance, if Friend’s long-held vision for citizen-driven clinical research was the main driving force behind ResearchKit, or if Apple’s growing army of medical experts had already been cooking something up at One Infinite Loop before O’Reilly and Friend met. (Apple wouldn’t comment for this story and Friend, who’s now also a medical technology advisor for the company, was instructed by the company not to tell me more.)

However it was that ResearchKit evolved from an idea into a full-fledged platform, one thing is clear: it’s already instigating change. Whether it will do to medical data collection what iTunes did to the music industry, for better or worse, remains to be seen.

Story Tags