Home activity control system: artificial intelligence to empower patients with diabetes through daily activity monitoring

Jesus González-Lama, David Pérez-Cruzado, Nicomedes Rodríguez-Rodríguez, Matilde Romero-López, Gertrudis Roldán-Molina, David Requena-Polonio

Keywords: artificial intelligence, diabetes mellitus, e-health, self-care, telemedicine

Background:

Poorly controlled type 2 diabetes mellitus (T2DM) is associated with significant morbidity, mortality and healthcare costs. The effective behavioural change and extensive education and self-management is one of the key approaches to alleviate complications from diabetes. E-health solutions can help to patients and healthcare professionals to manage this challenge.

Research questions:

It would be feasible and acceptable a home activity control system (Beprevent) to help to patients with T2DM and their healthcare providers to manage their disease?

Method:

We have planned a pilot, controlled, not randomized feasibility study, to evaluate a home activity control system (Beprevent) in the management of patients with DMT2, through the labeling of tailored objects (selected by the patient) linked to behaviors related to the management of that disease. Beprevent is a multipurpose solution of home activity control that informs people about routines through a mobile application, which also allows sending messages. This study will provide information on the best way and sample size necessary for conducting a definitive trial using standardized continuation criteria. Twenty patients with T2DM (10 in the intervention group and 10 in the control group), living alone or with people who cannot move on their own, will be included. The primary outcomes will include measures of feasibility such as recruitment rate and proportion of patients who remain in the study at the end of the study (6 months), and also degree of satisfaction (acceptability) of patients and professionals. Secondary outcomes will include degree of adherence to each of the therapeutic objectives agreed with the patient (exercise, hygiene, food and medication), HbA1c and other intermediate variables.

Results:

Conclusions:

Points for discussion:

Improving the measurement of the adherence of T2DM patients to health behaviors with non-interventional artificial intelligence tools.

Difficulties in recruiting this type of patients.

#99