AI-Driven Virtual Rehabilitation: Revolutionizing Home-Based Care

By Muhammad OsamaFeb 12 2024Reviewed by Susha Cheriyedath, M.Sc.

In an article published in the journal npj Digital Medicine, researchers from Canada explored the applications and effectiveness of artificial intelligence (AI) in home-based virtual rehabilitation (VRehab) programs for various patient populations. They highlighted that AI-driven VRehab can improve rehabilitation outcomes for various patient populations, such as stroke, cardiac, and orthopedic patients by providing personalized and real-time feedback, guidance, and monitoring.

Background

Rehabilitation is a process of providing interventions to patients to improve their recovery, reduce disability, and optimize functioning and health outcomes. However, traditional in-person rehabilitation faces several barriers, such as transportation needs, appointment scheduling conflicts, financial constraints, and staff shortages. Moreover, during the coronavirus disease 2019 (COVID-19) pandemic, millions of patients were affected worldwide.

VRehab is an alternative approach that virtually delivers rehabilitation programs to patient's homes using various technologies such as sensors, cameras, wearables, and smart devices. It can overcome many of the barriers of in-person rehabilitation and expand access to healthcare for diverse populations. AI is a field of computer science that aims to create systems that can perform tasks that require human intelligence, such as learning, reasoning, and decision-making. It can be applied to VRehab to automate different stages of rehabilitation, such as assessment, recognition, and prediction, and provide feedback to clinicians to improve the quality of care they provide to patients in their homes.

About the Research

In the present paper, the authors conducted a scoping review of the literature on the applications and effectiveness of AI in home-based VRehab programs for adult patients. They followed the framework proposed by Arksey and O’Malley and reported as per the preferred reporting items for systematic reviews and meta-analyses (PRISMA) extension for scoping reviews (PRISMA-ScR) checklist. The authors searched several electronic databases, such as PubMed/MEDLINE, Embase, Institute of Electrical and Electronics Engineers (IEEE) Xplore, and Web of Science, from starting to June 2023. They also searched Google Scholar for studies published between June 2023 and June 2024.

The researchers included studies that used any type of AI algorithm, such as machine learning, deep learning, fuzzy logic, or rule-based methods, for any type of rehabilitation, such as physical, cardiac, or cognitive rehabilitation. They only selected studies meeting the following criteria:

• Peer-reviewed journal or conference article written in English
• Presented the development or use of an AI-driven VRehab platform
• Evaluated the platform for adult patients undergoing any type of rehabilitation
• Assessed the platform on patients in their homes in a home-based or home-and hospital-based hybrid rehabilitation program

The paper screened 2172 unique titles and abstracts and retrieved 51 full-text studies for further review. After applying the inclusion and exclusion criteria, the study identified 13 studies that met the eligibility criteria. The authors extracted relevant information from the included studies, such as study characteristics, participants, settings, VRehab programs, AI algorithms, and effectiveness measures. They synthesized the results using descriptive analysis and narrative summary.

Research Findings

The outcomes revealed that 13 studies reported the utilization of AI-driven VRehab platforms in delivering rehabilitation services to patients, with the majority of them published between 2020 and 2024. The research was conducted in the United States, China, Spain, Greece, Italy, Tunisia, and Ukraine. They included various patient populations and types of rehabilitation, such as stroke, physical therapy, motor, cognitive, and cardiac rehabilitation.

The studies used different sensors to collect patient data at home, such as cameras, smartwatches, robots, and body sensor networks. They employed various AI algorithms, such as fuzzy logic, template matching, and regression, to analyze the data and make inferences about patients’ health outcomes. The research utilized the outcomes of AI algorithms in different ways, such as providing feedback, resources, and notifications to patients, as well as reporting progress and performance to clinicians. Additionally, some studies also employed deep learning methods, such as neural networks and convolutional neural networks.

The paper also found that the effectiveness of AI-driven VRehab platforms was evaluated using hospital readmission rate, patient satisfaction, perceived usefulness, perceived ease of use, and disease-specific assessment tool metrics. Additionally, it highlighted the effectiveness of AI-driven VRehab in improving health outcomes compared to non-AI-driven VRehab or in-person rehabilitation. Furthermore, the study demonstrated that patients were satisfied and engaged with the AI-driven VRehab platforms and appreciated the feedback and guidance they received.

Applications

AI-driven VRehab platforms could have various applications in different healthcare settings and scenarios, such as

• Providing accessible and affordable rehabilitation services to patients facing barriers to in-person rehabilitation, such as transportation, financial, or geographical constraints.
• Enhancing the quality and efficiency of rehabilitation care by automating repetitive tasks, reducing clinician workload, and optimizing resource allocation.
• Supporting patient empowerment and self-management by providing personalized feedback, guidance, and motivation.
• Improving patient outcomes and satisfaction by tailoring rehabilitation programs to individual needs, preferences, and goals.
• Enabling remote monitoring and assessing patient progress and recovery by collecting and analyzing data from various sensors.
• Facilitating communication and collaboration between patients, clinicians, and caregivers by providing reports and alerts.

Conclusion

The study summarized that AI-driven VRehab is a promising approach to improving the recovery, physical and mental functioning, and quality of life of patients living in the community. The AI algorithms can enable the measurement, detection, and prediction of various patients’ health outcomes based on data collected from various sensors at home. This can lead to improved feedback, guidance, and personalization of VRehab programs.

The researchers acknowledged challenges and limitations in the current research, such as the lack of standardized evaluation methods, co-design frameworks, and privacy-preserving techniques. They suggested that further research could fully assess the effectiveness and feasibility of various forms of AI-driven home-based VRehab, considering the unique needs and preferences of different patient populations and stakeholders.