PRECISION-CPR: AI-PoweRed fEedbaCk for Improving reScuer posture and compressION effectiveness in CardioPulmonary Resuscitation

Background and goals

Sudden cardiac arrest (SCA) remains a leading cause of mortality worldwide, with survival rates largely dependent on the quality of cardiopulmonary resuscitation (CPR) performed by first responders. Despite decades of CPR training programs, studies indicate that the quality of chest compressions (depth, rate, and recoil) during real-life emergencies often falls below recommended standards. Additionally, the physical effort required during CPR frequently leads to rescuer fatigue, compromising performance over time. These issues underscore the critical need for advanced training methodologies that not only teach CPR techniques but also provide precise feedback on performance and physical condition.

PRECISION-CPR (AI-PoweRed fEedbaCk for Improving reScuer posture and compressION effectiveness in CPR) aims to address these critical challenges by developing an AI-driven feedback system that leverages data from simulators and wearable sensors. This system will evaluate key metrics such as compression quality, rescuer posture, and fatigue levels during CPR training, providing corrective post-training feedback to learners. By enhancing the accuracy of CPR feedback and emphasizing ergonomic factors, the project seeks to improve both the quality of CPR and the well-being of rescuers.

Objectives and benefits

The primary objectives of PRECISION-CPR are:

Develop an AI-powered feedback system that includes:
Simulator data (e.g. Resusci Anne) to assess CPR quality (compression depth, rate, recoil).

Video analysis to track and evaluate rescuer posture and identify inefficiencies.

Fatigue analysis using EMG signals, on key muscle groups.

Develop a web-based platform that provide comprehensive post-training feedback through user-friendly reports highlighting CPR quality, posture, and fatigue. This will include ergonomic recommendations for optimal rescuer posture and positioning during CPR, to enhance performance and minimize fatigue.

Implement and test the system in diverse training sites across Europe (partner institutions), ensuring adaptability and reliability.

Promote adoption and dissemination through workshops, publications, and collaboration with CPR training institutions.

Results

Expected concrete results include:
1. An AI model capable of analyzing CPR performance by assessing rescuer posture and
fatigue indicators, and surrounding ergonomics.
2. A digital feedback platform that provides real-time AI-driven analysis and post-training
reports.
3. Open-access scientific publications and resources detailing project findings and
implementation guidelines.
4. Increased competence and confidence among CPR learners, demonstrated by
measurable improvements in performance metrics.

Societal impact

This initiative aligns with the Erasmus+ priority to "stimulate innovative learning and teaching practices" by fostering a personalised, technology-driven, and holistic approach to CPR training. Furthermore, the proposed system will be a scalable, adaptable, and cost-effective solution that can be implemented across diverse European HEIs (and other educational and training institutions) that deliver CPR training.