Back close

Architectures, Algorithms, and Models for Developing Remote Cardiac Rehabilitation

Thematic Area: Tele-cardiac Rehabilitation

Project Id: Tele-Cardiac Rehabilitation

Principal Investigator: Ms. Shereena Shaji

Amrita Team Members: Dr. Maneesha V. Ramesh, Dr. Rahul Krishnan, Prof. K. A. Unnikrishna Menon, Mr. Ramesh Guntha

Indian Collaborators: Dr. Ravi Sankaran, Rehabilitation Department, AIMS Hospital

International Collaborators: Dr. Prakash Ishwar, Electrical & Computer Engineering Boston University, United States

Architectures, Algorithms, and Models for Developing Remote Cardiac Rehabilitation

More than 60,000 cardiac surgeries are performed annually in India. Cardiac rehabilitation is a validated efficacious measure that is well-developed in Western countries. Participation in CR is associated with significant reductions in cardiovascular mortality, and rehospitalization. Currently, there is limited availability of outpatient CR programs in India. Even so, there is a consensus that CR is effective and essential in the CVD population. Cardiac rehab has four phases. Phase 1 occurs during the acute inpatient admission. Phase 2 consists of 16 sessions of guided exercise in a rehabilitation center. Phases 3 and 4 are maintenance and preventive measures. In India, less than 25% of patients receive phase 1, and fewer receive phase 2 or the latter stages. With the impact of the recent pandemic, segments of healthcare have migrated towards telehealth. It has been established that protocol-based Physical Therapy can be conducted on par with office visits over an online venue. Many of these are Orthopedic rehabilitation scenarios that are not hazardous to the health or in need of vital parameter monitoring. The goal of phase 2 cardiac rehabilitation is to take a person to their maximum heart rate over a fixed amount of time and develop their stamina. For this telemetry becomes critical to safety, operant feedback training, and patient education. Remote Cardiac Rehabilitation is a potential area of rehabilitation intervention that can significantly impact patient health and post-surgery

recovery. Ensuring CR compliance at home has proven to be challenging. Remote tele-CR is an emerging area, and technological drawbacks include but are not limited to: fail-safe, time-critical, remote cardiac monitoring, real-time exercise compliance assessment, and exercise routine risk stratification.

We collaborated with Physical Medicine Rehabilitation (PMR) clinicians at our University tertiary hospital to develop a reliable, real-time, tele-cardiac rehabilitation system enabled by the Internet of Medical Things (IoMTs), AI and fail-safe algorithms to ensure that crucial patient feedback is not compromised to address this unmet clinical need.

Our solution consists of:

  • Design and development of a reliable end-to-end tele-cardiac rehabilitation system that undergoes the tele-cardiac rehabilitation program.
  • Time criticality-based dynamic feedback system for decision-making.
  • Automatic detection of human activity to help physicians assist patients during rehabilitation programs.
  • Automatic detection of ECG abnormalities to ensure patient safety during rehabilitation programs.
  • To research into the design and development of a remote tele-cardiac rehabilitation system that could ensure safe remote exercise compliance in post-surgery patients
  • To explore the development of a self-feedback self-assessed AI remote CR routines which will enable minimum oversight of technician/physician taking into consideration patient safety and exercise compliance

Tele-rehabilitation has garnered significant interest among clinical researchers with its potential to transform cardiac rehabilitation. A critical requirement of tele-cardiac rehabilitation is timely feedback to the patient and the therapists/physicians. Current studies often assume ideal network conditions, neglecting the nuances of real-world deployment. We have designed, developed, and tested an end-to-end tele-cardiac rehabilitation system that seamlessly combines Internet of Medical Things (IoMT) devices and AI-powered abnormality and activity detection providing a fail-safe and real-time actionable closed-feedback loop system for the patient and the doctor. A pilot study evaluates system performance across diverse mobile networks in varying conditions (stable/ unstable). The RESNET-18 model for cardiac abnormality detection (71% accuracy, 0.71 F1 score) and the VGG-16 model for human activity classification (89.0% accuracy, 0.89 F1 score) demonstrate significant performance. Furthermore, we optimize these models for edge devices, demonstrating significant speed improvements compared to cloud servers (up to 33 times faster).

The proposed system offers time-critical feedback delivery to the patient and therapists/physicians, which is the fundamental necessity of tele-cardiac rehabilitation. With real-time audio, video, and vital signs, doctors can easily monitor the patient. The system helps therapists/physicians to perform early intervention, such as changing the type of exercise, which could reduce the need for unneeded visits to the emergency department, hospital stays, and related expenses. Our pilot deployment in the tertiary hospital setting demonstrated the utility and acceptability of the system among therapists and patients.

Proposed Future Work Details

  • Real-time Integration of VGG-16 transfer learning model for activity detection with webcam-based video capture code.
  • Develop a 3DCNN model for activity detection
  • Integration of Transfer Learning Models for Real-time ECG Data Capture from ASMD and Abnormality Detection
  • The entire system was integrated into our web application and deployed into our university, a large tertiary care hospital(AIMS, Kochi).

Publication Details

  • Shaji, Shereena, et al. “A Real-time IoMT Enabled Remote Cardiac Rehabilitation Framework.” 2023 15th International Conference on COMmunication Systems & NETworkS (COMSNETS). IEEE, 2023.
  • IoMT and AI Enabled Time Critical System for Tele-cardiac Rehabilitation – Submitted (International Journal of Telerehabilitation (IJT)- Q1-3.83, Impact factor)
  • Detection of Cardiovascular Disease with Minimal Leads using Efficient Machine Learning Techniques -Accepted (ICWITE-2024)

Patent Details

  • Systems and methods for time-critical remote tele-cardiac rehabilitation using IoMT and AI -Submitted (Prior Art Report)

Proposal

  • DST-Proposal -Architectures, Algorithms, and Models for Developing Remote Cardiac Rehabilitation -Submitted

Product Details

  • Amrita Tele-Cardiac Rehabilitation System (ReCardio)

Related Projects

Electrodeposition of Lead-Free Solder Materials for Microelectronic Packaging
Electrodeposition of Lead-Free Solder Materials for Microelectronic Packaging
Medical Imaging Studies Leading to Prediction of Neurological Disfunctions
Medical Imaging Studies Leading to Prediction of Neurological Disfunctions
Synthesis of Intermediate Analogues for Biologically Active Molecules
Synthesis of Intermediate Analogues for Biologically Active Molecules
A Robust Approach for Improving the Accuracy of IMU based Indoor Mobile Robot Localization
A Robust Approach for Improving the Accuracy of IMU based Indoor Mobile Robot Localization
A study on the utilization of student welfare schemes offered by Government of Tamilnadu in Coimbatore district
A study on the utilization of student welfare schemes offered by Government of Tamilnadu in Coimbatore district
Admissions Apply Now