News

Johns Hopkins Uses Augmented Reality to Deliver High-Quality, Low-Cost Pediatric CPR Feedback

To improve the quality of pediatric CPR, immersive technology specialists at the Johns Hopkins Applied Physics Laboratory (APL) in Laurel, Maryland, and pediatric emergency physicians from Johns Hopkins Children’s Center in Baltimore are developing an augmented-reality-based CPR coaching system to be used in community care settings.

The project began under a 2020 Johns Hopkins Digital Education and Learning Technology Acceleration (DELTA) grant, designed to encourage innovation in teaching and learning.

Kleinman using the CPR system
To develop a cost-effective CPR coaching system, Jeffers, Canares and Kleinman (pictured above) looked to Johns Hopkins APL’s XRCC to develop a proof-of-concept AR CPR coach — one that was dependable and modular and would reduce cognitive overload in a highly dynamic environment.

Credit: Johns Hopkins APL/Craig Weiman

Why Real-Time CPR Feedback Matters

There are many roadblocks to successfully resuscitating children with CPR. For one, it is difficult to retain CPR training without frequent refreshers and practice, said James Dean, an immersive technology applications engineer at APL. Pediatric CPR presents an additional challenge because guidelines change with age, added Principal Investigator Dr. Justin Jeffers, an emergency medicine physician at Johns Hopkins Children’s Center (JHCC) who leads the effort and specializes in CPR and simulation training: “Those changing guidelines make it more difficult to remember and retain appropriate pediatric CPR skills.”

To address this problem, researchers at Johns Hopkins Medicine developed a CPR coaching system designed to provide real-time, personalized feedback to the CPR performer. Composed of a human CPR performer and human coach, the system calls for the coach to monitor a device that displays metrics such as compression rate, compression depth and chest recoil. The coach then relays verbal feedback based on those metrics to the CPR performer, who adjusts performance accordingly.

A 2018 study found this coaching system had set a gold standard for CPR quality: Following the coaching system, a CPR performer who performed simulated CPR for 18 minutes stayed within the ideal range for compression depth and rate about 60% of the time, on average. This was a marked shift from the performance of individuals unassisted by the coaching system, who stayed in the ideal ranges only about 30% of the time, on average. The study provided promising news for the more than 20,000 children in the U.S. who undergo cardiac arrest each year.

“If you’re performing CPR and someone next to you is saying what you’re doing well and what you’re not doing well, you’re going to perform at a higher quality because what you have is essentially a personal CPR trainer,” said Dr. Keith Kleinman, a fellow at JHCC involved in the study.

When using the AR CPR coach, the CPR performer must keep two bars — one representing compression depth and the other representing rate — inside a green box at the same time for three consecutive seconds.

Credit: Johns Hopkins APL/James Dean

Augmented Reality as a Cost-Effective Solution

However, because the number of children who go into cardiac arrest is relatively small on the scale of the entire U.S. population, there’s often a lack of resources to effectively respond to this emergency. Additionally, most emergency pediatric care isn’t delivered at a specialized pediatric center.

“It’s delivered at community emergency rooms, at urgent cares or in the back of an ambulance,” Kleinman said. “If a community emergency room sees one pediatric arrest each year, it’s not going to be fully prepared for the event because of a lack of training, experience and resources.”

As a cost-effective option, Jeffers, Kleinman and Dr. Therese Canares, who is also a pediatric emergency medicine physician at JHCC, looked to APL’s Extended Reality Collaboration Center (XRCC) to develop a proof-of-concept augmented reality (AR) CPR coach — one that was dependable and modular and would reduce cognitive overload in a highly dynamic environment. The XRCC team supports and develops immersive technology to meet sponsor needs.

The AR CPR coaching system comprises an inertial measurement unit (IMU) sensor that sits under the performer’s palm and, powered by USB, measures changes in compression depth and rate; a lightweight and untethered headset that provides the performer with a heads-up display (HUD); and a laptop to process the IMU’s raw data and send it to the HUD, Dean explained​. All the system’s hardware and software communicate within a simple network environment.

“It works on a network like your home Wi-Fi, so doctors don’t need to do any custom network configuration,” said Blake Schreurs, a virtual and augmented reality subject-matter expert at APL. “We’re also using a lot of cross-platform tools and standard interfaces, so it’s pretty straightforward to swap out any one component.”

When using the AR CPR coach, the CPR performer must keep two bars — one representing compression depth and the other representing compression rate — inside a green box at the same time for three consecutive seconds. If one of the bars stays outside the green box for three consecutive seconds, a red screen with feedback text, like “go slower,” appears in the display.

“Our synergy is unlike any collaboration I’ve experienced,” Canares said. “This is a combination of brilliant engineers who can build solutions that solve critical health care problems, combined with our physician-researcher team who can run a research study and generate data. It’s a perfect pairing.”

The XRCC project team is part of APL’s Information Technology Services Department (ITSD) and includes Dean and Schreurs, as well as Brandon Scott, a software engineer, and Nick DeMatt, the project’s manager and ITSD’s chief engineer for engineering design and fabrication services.

“Collaborating with the JHCC team on the AR CPR project has given us an opportunity to demonstrate how AR can be valuable to the health care community,” said DeMatt, who founded the XRCC in 2020 with Dean and Schreurs. “The project’s mission is what truly drove our desire to partner.”

In the video above, Kleinman is helping the team conduct the system’s first feasibility test at JHCC. Results suggest that when it comes to affecting human CPR performance, feedback from the AR-based CPR coaching system is comparable to that of a human-based coaching system.

Credit: Johns Hopkins Children’s Center/Therese Canares

Feasibility Tests and Expanding Collaborations

The team conducted the system’s first feasibility test at JHCC and demonstrated that a performer who receives feedback from the AR CPR coach stayed within the green box about 60% of the time, on average. This suggests that when it comes to affecting human CPR performance, feedback from the AR-based coaching system is comparable to that of a human-based coaching system.

The next feasibility test will take place among non-pediatric trained emergency room providers in community care settings in Maryland: Johns Hopkins Bayview Medical Center in Baltimore and Howard County General Hospital in Columbia. As the team continues to assess the impact of the system’s feedback on CPR performers, it will also develop methods to validate the system’s benchmark metrics to determine whether the system improves the quality of CPR.

The team will continue to leverage multidisciplinary collaborations with APL researchers and engineers as the system undergoes design iterations and more testing. It is also targeting federal funding opportunities to further develop the system.

“APL is excited about the progress the project has achieved so far through its collaboration with Johns Hopkins Children’s Center,” said Alan Ravitz, chief engineer of APL’s National Health Mission Area. “We look forward to maturing beyond this prototype stage to eventually transition the concept to industry so that it may impact the lives of children worldwide.”