With Apple’s backing, can an app help heart attack patients navigate recovery?

In July 2016, two cardiologists and a handful of computer scientists and developers rolled up in a minivan to Apple’s special projects office in Cupertino, Calif., with a big idea to show a company with grand designs on transforming health care.

The team from Johns Hopkins University had received a rare invitation from Apple to workshop their mock-ups for Corrie, an app to guide heart attack patients through the maze of recovery. For a week, Apple and the Hopkins team labored on the design, carefully talking through the minutiae of each feature.

Corrie is designed to make everything that’s hard about managing recovery after a heart attack far easier for patients — and, in turn, keep them out of the hospital. Once home, the app helps track their vital signs and activity data with the help of an Apple Watch and a Bluetooth blood pressure cuff. It sends reminders when a patient needs to take a pill or head in for a follow-up appointment, and also serves as a hub of critical health information, including guidance on diet and exercise, that’s often lost in the chaos of a hospital discharge.


“Our health care system [is] built to treat and get patients out of the hospital and home,” said Francoise Marvel, the cardiologist who first came up with the idea. What it’s not designed to do, she said, is build understanding or “encourage a skill set” that can help patients prevent more health problems moving forward.

Marvel and her colleagues designed Corrie to fill that gap. And five years after their visit to Cupertino, the effort is finally starting to bear fruit. In July, the Corrie team published results from a study showing the app reduced risk of hospital readmissions by half in a pool of 200 patients at four hospitals.


The app and its development are exactly the kind of work Apple set out to make possible when it launched CareKit, the patient care framework Corrie is built on. The Corrie team’s work has kept the attention of Apple, which paraded the app at its Worldwide Developers Conference in June as a sterling example of how its tools could improve patient care.

“I think we’re great marketing material for them,” Matthias Lee, the technology lead on the project, told STAT. “And we’ve certainly benefited a lot from their marketing. I mean, being featured at WWDC was about as good as it can possibly get.”

Now, Lee and the rest of the team are hoping to ride the recent momentum through the thicket of challenges that remain as they aim to turn Corrie from a promising academic research project into a tool that’s widely used. Proving that Corrie works in the real world will require more evidence, and even then, there’s no clear path for changing the rigid systems of bustling hospitals.

“There are so many apps out there that promise the coolest things in the world and are going to change the world and all these sorts of things,” said Lee. “Even in health care, very few of them actually become commercial products.”

Marvel came up with the idea for the precursor to Corrie during her internship at Hopkins, where she saw heart attack patients receive lifesaving treatment, only to return to the hospital for preventable reasons. The problem, she concluded after conversations with patients, was that much of the most important messaging was being lost on overwhelmed patients who just wanted to go home.

With the help of a small grant and colleagues at the Johns Hopkins’ Technology Innovation Center, Marvel set about reimagining the communication that happens as heart attack patients prepare to leave the hospital — usually a rushed discussion and a handful of dense printouts.

By the end of 2015, she had developed a rudimentary app to guide patients through some of what they needed to do after being discharged, like taking medications.

Marvel quickly realized the team needed “a way to really grow this and really take this to the next level because it was just going to fizzle out and be about nothing.” So she brought on new collaborators, including undergraduate developers, Lee, and Seth Martin, a cardiologist who had a reputation for working in digital health. Marvel credits the platform’s success so far, in part, to the wide range of people involved in its development — “engineers and businesspeople and leaders that you may not encounter during grand rounds and during our usual academic conferences,” she said.

As they were working on the earliest iterations of Corrie, Apple was preparing a blitz in health care.

In 2015, the company had rolled out the Apple Watch, which would soon become a focus of Apple’s health pitch to consumers. Behind the scenes, Apple was crafting software frameworks for developers to build health care apps that could catapult Apple directly into hospitals and labs. That same year, the company introduced ResearchKit, a suite of tools to help medical investigators conduct studies, followed by the release of CareKit in March 2016.

Apple caught wind of the project through its connections to Hopkins, which led to a phone call among Marvel, Martin, and Divya Nag, the special projects lead on CareKit, who invited them to Cupertino.

The trip was a whirlwind. The newly formed team huddled in an Airbnb for a week, living and working together for long hours between several trips to Apple’s offices for meetings. Lee and another colleague devised the back-end data infrastructure for the app and Marvel and Martin drafted up the documents for the clinical trial. A late-night rainstorm produced the design of the heart-healthy “ABCs” that would become a central part of patient education on the platform. They also settled on the name Corrie, a play on the Latin word for heart, Cor, and the name of the system, COmbined-device Recovery Enhancement.

“It’s very rare in the world of medicine or research that you have a team staying in a house together,” said Martin. “There’s something about the focused energy and the camaraderie that’s, I think, very special to our story.”

The team eagerly presented their mock-ups on the first day at Apple — and immediately got a crash course in the company’s notoriously meticulous design methods. Apple’s designers and engineers studied and discussed the design with the team for hours, zeroing in on minute details like fonts and buttons, questioning whether the colors of the app felt too medical, and advising the team to streamline the menus and screens so that the app would be easier for older people to use.

“Of all the meetings that I’ve ever had in my career, it’s one of those meetings that really sticks with me just because of how much it kind of took me out of my usual ways of thinking,” said Martin.

Before leaving, the team had a golden opportunity to present to Jeff Williams, Apple’s chief operating officer. According to a report the team produced afterwards, Williams, impressed with their work, went so far as to comment that their app could save lives.

“We needed that vision and the open source software and that spark from such an important influencer,” said Marvel.

By the fall, Corrie was ready for testing with patients.

The non-randomized trial, which took over two years to complete, compared the outcomes of 200 patients at four sites who were given the Corrie app to those of a historical control of 800 patients who were treated at the four hospital sites over a one-year period from September 2015 to October 2016. The study found that the Corrie users were at 52% lower risk of 30-day readmission for any reason.

The Corrie team chose to focus on readmissions within 30 days of discharge, which is a critical period for recovery when adhering to medications and changing behaviors is essential. Patients who’ve had a stent inserted must take medication to keep platelets from sticking to it, or they’ll wind up back in the hospital. The duration also targets the window of time during which readmissions can result in financial penalties for hospitals.

People who did not have an iPhone were offered one through a novel loaner program, addressing the criticism that building interventions around Apple products can leave out low-income individuals.

Nursing Ph.D. candidate Erin Spaulding helped recruit patients recovering in the hospital and walked them through setting up the app. One of the early participants was Tammy Gannon, a retail worker who went from someone unfamiliar with smartphones to one of Corrie’s “super users,” according to Marvel.

Gannon told STAT that when she had a heart attack in 2016, she had already been diagnosed with COPD and was smoking two packs a day, eating poorly, and drinking “15 Cokes a day,” she said. When she woke up after surgery to clear an artery and insert stents, she was introduced to the app.

“I thought I was joining some support group,” she said. “But the support group [is] the devices.”

Gannon said that, in particular, she liked how the devices allowed her to track her activity and blood pressure. When she found out her insurance wouldn’t cover cardiac rehabilitation after her heart attack, she took it on herself to return to what she called the “Rocky Balboa stairs” at Johns Hopkins Bayview Medical Center for regular workouts, using her Apple Watch to track her activity. She said the app also helped her stay on top of medication and appointments.

“The Corrie program saved my life,” she said. “Had someone not come and talked to me, I wasn’t thinking about changing.”

While the results seen in Gannon and many of the study’s other participants are promising, outside experts said, they’re also preliminary and will need to be validated. David Cho, a cardiologist at UCLA and chair of the American College of Cardiology’s health care innovation council, said the fact that the study looked at the first 30 days after a heart attack is useful because it’s a “very vulnerable period for patients.”

Still, Cho said recovering from a heart attack is a lifelong process, and future work should track longer-term stickiness and outcomes. Martin said that he could foresee the app moving beyond the initial acute recovery stage to 90-day and long-term recovery plans.

Replicating the results in a randomized trial with a true control group would also help instill confidence that Martin and Marvel working at the study site did not influence the outcome, said John Spertus, an endowed chair and professor of medicine at University of Missouri-Kansas City and clinical director of outcomes research at Saint Luke’s Mid America Heart Institute.

“We’ve been humbled by good technology, good concepts, good ideas that, when they’re in the real world, end up not working.”

John Spertus, clinical director of outcomes research at Saint Luke’s Mid America Heart Institute

“We’ve been humbled by good technology, good concepts, good ideas that when they’re in the real world, end up not working,” he said, noting that when testing an intervention at their home institution, “developers can implement and explain and really motivate its initial site to to really use the technology correctly.”

Marvel said that Corrie has a number of ongoing and upcoming clinical trials, though some are covered by nondisclosure agreements.

She said Hopkins has plans to conduct two more studies on the Corrie platform: a randomized trial of virtual cardiac rehab and a study of the early detection and prevention of premature coronary artery disease and sudden cardiac death in young adults.

Even if future evidence supports the wider adoption of Corrie, getting health systems to buy in can be a huge challenge, said Spertus, pointing to his own experience as the founder of Health Outcomes Sciences. The company developed a tool to integrate predictive analytics and decision support into clinical workflows that was proven to improve care, but ultimately sold to a large medical technology company because it was too hard to generate interest from hospitals wary of spending money.

That underscores one of the challenges Corrie could face: getting the staff of any hospital that adopts it to actually use it consistently.

“This is an exciting innovation that in an ideal setting has initial promise,” Spertus said. “And as we continue to gain more data that it works, we then have to spend a lot of time thinking about how to reproducibly get this offered to every patient, every opportunity, every admission, and how we support it so that it’s used in a way that really improves the outcomes,” he added. That could prove especially challenging for under-resourced hospitals who don’t have the bandwidth to coach hospitalized patients on how to use an app, Cho added.

Spaulding, who helped conduct the study, agreed the logistics are an open question.

“Outside of the research setting, there can’t just necessarily be me running around helping patients get a set up with these devices,” she said. “So what’s feasible? Will it be videos like on hospital room iPads that help them get set up? Will it be really relying on maybe, like, volunteers in the hospital or family members?”

The Corrie team now faces an uncertain path as it tries to turn a promising academic research study into a business with a popular product.

Up until now, the researchers have relied on scattered grant funding.

“You get a grant, you do something and then the grant’s over and then no one sees it again,” said Martin. “And that’s absolutely not what we want, but I think part of the reason that that happens a lot is because it’s so hard to take it to the next level.”

Johns Hopkins has a small ownership stake in the company formed around Corrie, which has Marvel as CEO, Martin as chief medical officer, and Lee as chief technology officer. But despite a surge in investment in digital health, the team hasn’t yet raked in huge sums of outside money.

Garnering more investment could provide a possible path forward for the company, as could selling to a larger company better equipped to commercialize.

The next version of the platform, which the team broadly calls Corrie 2.0, expands beyond discharge guidance and includes a dashboard that allows clinicians to monitor the progress of patients. The team has plans to incorporate more interactive features that respond to data from patients with nudges and suggestions, and is launching a system to engage clinicians earlier with patients at risk of health problems or of winding up back in the hospital. The team also has a road map for building Corrie interventions for atrial fibrillation and hypertension. An Android version of the app has also been built.

With ambitious plans for the product and fresh promotion from Apple, Lee stresses the importance of finding a way to get out in front of patients as quickly as possible. That will make it possible to collect more feedback and continue refining the app so that it feels to patients like a “partner, not like someone screaming at them.”

It’s a design ethos that Lee knows sounds a bit like Apple’s: “It has to be dead simple to use, and people have to want to use it,” he said.