AI to ease musculoskeletal pain in conditions such as osteoarthritis

Musculoskeletal injuries account for a significant proportion of appointments in primary healthcare. Degenerative pathologies are particularly common, including osteoarthritis, which affects older people and can cause chronic pain and limit mobility, as well as impairing the patient's quality of life.

Given the many appointments related to these pathologies in primary healthcare and the millions of euros per year that the healthcare system spends as a result, a master's degree final project by a researcher at the Universitat Oberta de Catalunya (UOC) has reviewed the studies published worldwide in the last decade concerning the implementation and impact of artificial intelligence (AI) as an autonomous tool for the prescription of musculoskeletal rehabilitation treatments.

This review of the scientific literature examined the impact of AI when applied to musculoskeletal rehabilitation and its effect on symptoms, functionality in the patient's everyday activities and quality of life, as well as its usefulness as a tool for educating and training patients in the self-management of their disease. "The objective is to describe the potential and limitations of AI when it's applied to musculoskeletal rehabilitation, and to determine its impact on the consumption of healthcare resources," explained Liubov Adrover Kirienko, student on the UOC's Master's Degree in E-Health.

Optimization of care and rehabilitation systems

Estimates suggest that around 40% of patients who attend primary healthcare appointments suffer from chronic pain. Furthermore, around 10 million Spaniards consult a doctor about a musculoskeletal problem every year. Meanwhile, the prevalence of osteoarthritis is 30% in the general population, and this figure rises to 80% in people over 65 years old.

In general, physical exercise, a therapy prescribed by specialized care rehabilitation services, is the main treatment other than surgery for these health problems. However, these services are currently overloaded, with long waiting lists and overcrowded gyms, which diminishes the quality of the care provided. "So, there's a need for training to enable primary care to deal with musculoskeletal pathologies of limited complexity, and to empower patients in the self-management of their musculoskeletal conditions. This would reduce the overload that care services experience, and improve the population's musculoskeletal health," said Adrover.

The results of the research show that adapted AI can contribute to musculoskeletal rehabilitation programmes being prescribed with no need for supervision by a healthcare professional. Applying AI during musculoskeletal rehabilitation could also reduce pain and the symptoms of pathologies such as osteoarthritis.

"By implementing AI, we could consider performing rehabilitation treatments in non-specialized healthcare environments, such as primary care. That would mean that if a primary care centre didn't have a rehabilitation service, the patient could be admitted to a rehabilitation programme after seeing their doctor without having to wait to be referred to the specialized rehabilitation service, and without being placed on a waiting list," she said.

However, the use of AI in musculoskeletal rehabilitation has greater potential in less severe cases, such as less complex musculoskeletal pathologies. The most important factors in these cases are educating the patient and the patient doing therapeutic exercises. The physiotherapist's physical presence is unnecessary. "This application is more viable in degenerative pathologies such as osteoarthritis or mild or moderate tendinopathy that doesn't require surgery," explained Adrover.

There are already various applications that focus on rehabilitation in Spain which use AI to improve the patients' quality of life and provide support for the healthcare system. However, they have some limitations. In fact, these initiatives generally focus on diagnosis or on helping the healthcare professional to prescribe exercise programmes based on a video library, or to monitor patients' evolution. "AI interventions for the prescription and creation of rehabilitation programmes autonomously and without the involvement of a human being haven't yet been developed," she said.

Options for improvement in the implementation of AI

An appropriate implementation of AI could therefore improve access to rehabilitation programmes, with the consequent savings on travel. This would enable patients to follow them outside working hours, and increase the range of rehabilitation available at centres that do not have a specialized service. "AI can also empower the patient in terms of their self-care through education, training, and instruction in therapeutic exercises," she said.

In the healthcare system, artificial intelligence could help ease the burden on overloaded specialized rehabilitation services, reduce waiting lists and overcrowding in therapeutic areas, and lower the ratio of patients to physiotherapists. It would thereby contribute to achieving better quality of care. In particular, with a specific design and programming, the health costs arising from musculoskeletal rehabilitation could be reduced, and rehabilitation programmes could be offered without increasing the costs of specialized rehabilitation.

"AI can be a helpful tool in situations where there's a lack of resources. However, the ideal situation is not having to depend on technology to provide good healthcare," warned Adrover.

Despite the benefits of AI, we must not neglect other aspects of the healthcare system in order to achieve greater effectiveness and efficiency. "It's very important that we focus on prevention, on raising awareness in the population from an early age, on education about this health problem and on providing the tools so that people can take care of themselves as well as possible from the first level of care, which is primary care," she concluded.

This project was carried out under the supervision and tutoring of Hilda Maria Rodrigues Moleda Constant.

Reference

Adrover Kirienko, Liubov. Inteligencia artificial para la prescripción de rehabilitación musculoesquelética. [Master's degree final project]. Universitat Oberta de Catalunya (UOC). http://hdl.handle.net/10609/146593.