Smart CPR Device Aims to Restart More Hearts

By HospiMedica International staff writers
Posted on 29 Jul 2025

Each year in the United States, approximately 600,000 people experience sudden cardiac arrest (SCA), with over 350,000 occurring outside hospitals. Sadly, three out of four of these out-of-hospital cardiac arrests (OHCA) do not survive to hospital admission due to the inability of rescuers to restart the patients' hearts. A critical barrier to successful resuscitation is the lack of real-time, accurate assessment of blood flow to the heart during CPR. Diastolic blood pressure (DBP), which indicates the heart's blood supply during chest compressions, typically requires invasive arterial catheters, especially in prehospital settings. Even when catheters are in place, chest compressions can distort DBP readings, rendering bedside monitors unreliable. Now, researchers are developing a wearable device that measures DBP noninvasively in real time and provides feedback to rescuers, enabling them to personalize CPR for each patient and improve survival chances.

The University of Michigan Max Harry Weil Institute for Critical Care Research and Innovation (Ann Arbor, MI, USA) is leading the development of the device, called INSIGHT-CPR, with funding from the American Heart Association (Dallas, TX, USA). The technology integrates a wearable sensor adapted from earlier prototypes with a neural network trained on arterial waveform data to detect DBP in real time. The sensor, worn on the wrist or finger, transmits data wirelessly to a mobile device, offering live feedback on how well blood is flowing to the heart. This allows rescuers to optimize their resuscitation efforts, including adjusting hand placement or medication timing. The INSIGHT-CPR project is still in the development and preclinical testing phase. Data from adult and pediatric cardiac arrest cases—including in-hospital and out-of-hospital events—will be used to train and validate the AI model.


Image: INSIGHT-CPR captures and displays diastolic blood pressure information in real time (Photo courtesy of the University of Michigan)

The AI model is being trained on diverse waveform datasets, including those from prehospital catheterization cases, CHOP’s ICU-resus pediatric study, and Michigan Medicine’s in-hospital data. The device will be tested in large animal models to evaluate compatibility with defibrillation and accuracy in measuring DBP during CPR. Researchers aim to compare DBP-directed CPR using the device against standard life support protocols in terms of effectiveness. Once validated, the solution is expected to eliminate the need for invasive monitoring during resuscitation, filling a major gap in emergency medicine. The team envisions future deployment in prehospital, EMS, and combat casualty care settings, especially those lacking reliable connectivity.

“Our goal with INSIGHT-CPR is to fundamentally change how cardiac arrest is treated by removing the need for invasive monitoring, taking out the guesswork for rescuers and tailoring resuscitation strategies to the patient. Even if we can save an additional 10% of cardiac arrest patients, that’s over 60,000 more lives saved each year. But we’re hoping to achieve much more than that,” said Dr. Cindy Hsu, the project Principal Investigator.

Related Links:
Max Harry Weil Institute for Critical Care Research and Innovation
American Heart Association


Latest Critical Care News