A recent study has shown howthe degree of dependency among the elderly can automatically be measured in a non-intrusive way, using only a smart wristband, while they carry out their everyday activities
This solution can save time for health professionals and healthcare systems thanks to the early detection of dependency and other diseases or impairments
Researchers from the Departments of Software Engineering and Physiotherapy of the University of Granada (UGR) have developed a new method of assessing the state of dependency of people over the age of 65, based on artificial intelligence (AI).
The study, published in the International Journal of Medical Informatics,confirms that it is possible to automatically measure the degree of dependency among elderly people in a non-intrusive way, using only a smart wristband, while they perform their day-to-day activities. This solution can save time for health professionals and healthcare systems thanks to the early detection of dependency and other diseases or impairments.
Traditionally, the dependency status of people over 65 years of age is assessed with tests or questionnaires that are self-reported or administered by others-such as the Lawton & Brody questionnaire. In this approach, elderly people are asked to perform tasks relating to 8 'domains of function'-known as Instrumental Activities of Daily Living or IADLs-while a health professional observes whether they perform them adequately or if they have any difficulties.
These activities include the ability to use the telephone, do the shopping, prepare food, wash their clothes, etc. In general, these types of observation-based evaluations require considerable time and, in addition, present a subjective dimension. Therefore, they are often not carried out because they are resource-intensive.
In a bid to overcome these drawbacks, the authors of this UGR-based study used wearable devices such as smartwatches and wristbands to collect physiological data (objective data on vital signs) from the elderly subjects during a complex and highly comprehensive IADL, such as supermarket shopping.
After analysing the data, combined with machine learning techniques, the researchers successfully validated a model capable of distinguishing between dependent and independent people-accurately, non-intrusively, and inexpensively.
Bibliography:
M. Garcia-Moreno, M. Bermudez-Edo, E. Rodríguez-García, J. M. Pérez-Mármol, J. L.
Garrido, & M. J. Rodríguez-Fórtiz (2022), 'A machine learning approach for semi-automatic assessment of IADL dependence in older adults with wearable sensors', 157 International Journal of Medical Informatics, 104625. DOI: https://doi.org/10.1016/j.ijmedinf.2021.104625
