The EU funded project FrailSafe, focused on a personalized approach towards frailty using innovative technologies and running for the last three years, has reached its evaluation phase. In this phase, 75 older adults from Cyprus, France and Greece will test the final version of the FrailSafe system in its integrity and receive personalized health recommendations. The system and its various components, mostly high-tech devices, have been developed and adapted to user requirements according to the results of the pilot trials that took place from September 2016 to October 2018.

During the evaluation phase, older participants are using the FrailSafe system in their home setting, in the most unobtrusive and user-friendly manner, while performing their usual everyday activities. In the meantime, multiple health data are being recorded by the FrailSafe devices and managed through a secured online platform. The data are available to participants, as well as their authorized doctors and family members. Online data are then being manipulated by an automatic Decision Support System (DSS) which generates alerts when it detects measurements outside of the normal range. Deviations in health parameters are detected according to preset criteria extracted from well-documented health guidelines in recent literature, but can also be manipulated by the participant’s physician to meet a more personalized range. Based on these measurements, personalized recommendations are provided to the participants, in order to address possible health issues as early as possible.

The alerts and recommendations are categorized into five main health domains: 1) medical, 2) nutritional, 3) psychosocial, 4) cognitive, and 5) physical, including over 50 measured parameters, such as the presence of polypharmacy, Body Mass Index (BMI) and muscle mass, social life, performance on standardized cognitive tests, physical activity, indoor and outdoor movements . Researchers visit the FrailSafe platform and document all the alerts that the DSS has generated for each participant. Then, alerts and recommendations are combined into a final health report including an overview of the results by a medical professional from the FrailSafe team (see next image for an example).

Furthermore, according to which domain was identified as the most vulnerable by the alerts generated by the DSS, clinicians provide the participants with leaflets containing generalized recommendations and health management information based on recent literature.

Finally, using deep machine learning techniques, the FrailSafe team developed an algorithm to identify if a participant is at risk of developing frailty-related adverse health outcomes in the future according to his/her measurements from the FrailSafe devices. Hence, in this phase, researchers are also testing this feature, which is the system’s ability to identify risk of frailty evolution early on. Older adults who participate in the evaluation phase are informed if their profile on the FrailSafe system is indicative of some health issues which require their attention. However, they are prompted to consult their doctors in order to further assess if these indications constitute a medically significant finding.

Preliminary interviews with the participants show that they maintain a positive attitude towards the feedback received by the system and have increased feelings of self-confidence and satisfaction.