Frailty prediction system in the elderly using artificial intelligence

Frailty is an age-related syndrome characterized by loss of strength and exhaustion, and is associated with multimorbidity. Machine learning techniques (a form of artificial intelligence) can help with early detection and prediction of its occurrence.

Some scientists from the Intelligent Systems Group (GSI) of the Polytechnic University of Madrid (UPM) in Spain have become interested in this topic and have developed a machine learning model for the prediction of frailty and prefragility, with special attention to the physical appearance. of the pathology. Given the increase in the average age of the population, the development of policies for the prevention and treatment of frailty is a topic of great interest to society, since the prevention of this condition can significantly improve the lives of our elderly and alleviate the burden of the health system. Machine learning techniques show promising results in creating a medical support tool for that task.

Frailty is a syndrome that affects the elderly population and is characterized by the decline in physiological reserve and physical and cognitive functions. It is correlated with muscle loss and weakness and is associated with an increased risk of falls, frequent hospitalization, or motor and cognitive dysfunction. In relation to this topic, the study carried out by researchers from the GSI group of the UPM set the objective of creating a data set for fragility studies based on machine learning. To do this, they have used the definition proposed by epidemiologist and geriatrician Linda P. Fried in 2001, which identifies a frailty phenotype through five criteria (involuntary weight loss, slowness, grip strength, level of physical activity and exhaustion). , thus dividing the population into three categories: fragile, pre-frail and robust.

To develop these types of models, you need a large amount of data from which the model can learn. For this, Matteo Leghissa, Álvaro Carrera and Carlos Á. Iglesias, the three from the UPM, used one of the most recognized studies on aging that exists, the ELSA (English Longitudinal Study of Aging), which has been collecting data from older people in the United Kingdom since 2001. After studying and processing With these data, they formulated a model that can give an estimate of the risk of fragility over a two-year time horizon. They have identified the most relevant variables and with them they have developed a questionnaire to ask older people, and thus obtain the input data of the model. The questions vary between medical, economic, social and cultural areas, and do not require testing or analysis of the patient.

The data obtained in the study can be used to find out the level of frailty of each elderly person, through previously trained machine learning architectures for frailty detection and prediction. The achievement of these models is part of the integration of data science with medicine and hospitals, a tool with great prospects for improving the health of the population.

“One of the achievements achieved as a result of the study is a smart mirror that is installed in the homes of older people with the aim of helping them in their daily lives to counteract the risk of frailty,” the researchers indicate.

Example of use of the smart mirror. (Photo: provided by the University of Castilla la Mancha, coordinator of the MIRATAR project)

The results obtained would not have been possible without the support and work of the other research groups participating in the project: the University of Castilla-La Mancha, the Center for Biomedical Research in Fragility and Healthy Aging Network (CIBERFES) and the University Carlos III, in Spain all these institutions.

The study is titled “FRELSA: A dataset for frailty in elderly people originated from ELSA and evaluated through machine learning models”. And it has been published in the academic journal International Journal of Medical Informatics. (Source: UPM)