Making Digital Health Development More Accessible through an Unlikely Collaboration

What do researchers at Stanford Medicine have in common with a teacher at Walt Whitman High School in New York’s South Huntington Union Free School District?

Both were frustrated with the time and effort required to build digital apps to advance health-related research. And, now, both are using CardinalKit, developed at Stanford Biodesign, to more efficiently and effectively support digitally-enabled research.

The story begins in 2020, with Oliver Aalami, a vascular surgeon at Stanford Medicine and director for digital health at Stanford Biodesign. Aalami and his team had been conducting a large-scale study of peripheral artery disease (PAD) using iPhone data, including patient surveys and walk tests, to identify trends that could be leveraged to improve PAD care. In getting the study up and running, he was struck by how much time and money it took to build the supporting app, and how much deep programming expertise was needed to develop something that was not only effective, but compliant with patient consent and privacy requirements. As Aalami networked with other researchers at Stanford, he realized that they were wrestling with the same challenges, with everyone starting from scratch and having to go from “zero to one” each time they needed to develop an app to support their research.

Aalami envisioned an open-source suite of tools that he and his fellow researchers could leverage to minimize the time, cost, and effort of developing digital technology to support their work. With the help of Stanford medical students (Neil Gandhi, Raheel Ata) and a computer science student (Santiago Gutierrez), Aalami and the group developed the CardinalKit template application to simplify what’s required to build a digital experience from the ground up, from the app itself to storing and analyzing collected data in the cloud. The system includes detailed documentation, reusable code libraries, and templates for patient onboarding and consent, task scheduling, standardized surveys, and health data collection from wearables and connected medical devices. CardinalKit also provides a web dashboard, which allows researchers and clinicians to view and analyze data collected from study participants and patients.

Aalami piloted CardinalKit with a number of peers at Stanford Medicine. He also continued expanding the framework with input from students in the Stanford Biodesign course “Building for Digital Health” (CS342). The benefits realized by these CardinalKit users were immediate. Over time, he and the team have been able to demonstrate that the use of CardinalKit can reduce development costs by as much at 75% and accelerate development timelines by up to 12 months on projects focused on creating interoperable digital health solutions for research or clinical care. As a result, CardinalKit quickly garnered attention beyond Stanford, gaining adoption by clinical faculty and researchers at numerous institutions, including the University of Utah.

When something works well for world-class researchers at some of the United States’ most prestigious institutions, one doesn’t immediately think about trying it at the high school level. So, how did CardinalKit make this leap?

Enter Frederick Feraco, a computer science/research science teacher at Walt Whitman High School, who came across CardinalKit while searching for tools his students could use to build Apple iOS health applications. In his research science classes, students explore problems they want to solve and/or questions they would like to answer through research projects. In some cases, the projects are health-related, and Feraco was interested in finding a way to make it faster and easier for students to leverage the sensors in a typical smartphone to support their work. When he discovered CadrinalKit online, he recognized that it could make digital health development more efficient and approachable, especially for students with no prior programming experience.

“Starting a project from scratch can be discouraging and difficult for students, which makes it easy to give up,” he explained. “But with the CardinalKit package, students with only a basic understanding of computer science can get a digital app up and running in a few class sessions rather than spending multiple months on this work. It saves them a lot of technical time in building the app so they can spend more time being creative in how they plan and execute their research.”

As one example, Feraco introduced two 11^th grade students, Paul Chinsky and Jacob Albert, to Cardinal Kit at the beginning of the school year. Once they understood its capabilities, they started thinking about compelling research applications. One area of interest was stress in the school environment and, more specifically, whether binaural beats could be used to lower student anxiety. Chinsky and Albert spent weeks performing a literature review and developing a research proposal. Once their proposal was approved by Feraco, they built a digital app using CardinalKit to survey students on self-reported stress levels. The app pairs this input with a tremor test, to gauge physical anxiety, using sensors in the students’ smartphones. Then, based on that combined data, the app provides binaural beats and instructions for listening throughout the day until the students are prompted to repeat the survey and tremor test. Chinsky and Albert’s hypothesized that the binaural beats would contribute to decreased stress and anxiety compared to students who seek to manage their stress and anxiety without this type of intervention. They are still fine-tuning their study parameters, but were able to start gathering data in a matter of weeks, which they found encouraging.

“With no prior programming experience,” the boys reported, “CardinalKit was essential to the development of our research app.” Even more importantly, added Vishnu Ravi, lead architect for CardinalKit and the Building for Digital Health course, “There’s a dual benefit: CardinalKit is being used to teach students how to build apps; and the positive experience gets them excited about going into careers in digital health and health research.”

While this bi-coastal collaboration between university researchers and a high school teacher may seem a bit unlikely, Aalami and Ravi have learned a great deal from Feraco’s use of CardinalKit at Walt Whitman High. Feraco has made a series of suggestions for the enhancement of the tools and has become an active participant in the CardinalKit open-source community. “He has provided us with valuable input about features we want to develop in the future to support new use cases like this one,” Ravi said. “The whole goal with CardinalKit is to make digital health development more accessible,” he continued. “Our preliminary focus was on university faculty. But it’s gratifying to see that we’ve created something that’s so straight-forward and intuitive it can be successfully used to support health research at all levels – even high school.”

More examples of research projects being led by Feraco's students are available online.