The first two Medical Objectives (MOs) of the project are to better understand frailty and its relation to co-morbidities (MO1) and develop quantitative and qualitative measures to define frailty (MO2).
To achieve MO1 and MO2 FrailSafe will extensively measure not only the physical domain of frailty, but also the cognitive, the psychological, the functional, and the social domains. Most of these measures have never been recorded simultaneously before, and consequently, have not been correlated with frailty; furthermore, many of the proposed measures are meant to be real life tools in a real life environment, and such tools are totally lacking in clinical practice. Eventually, FrailSafe will provide continuous physiological clinical state monitoring of older people based on both embedded and behavior monitoring sensors. Specifically, FrailSafe will measure:
- cardinal physiological parameters of the cardiovascular, nervous and respiratory systems as well as aspects of the motor and metabolic and arousal/sleep state of the user as her/his case may indicate
- indoor and outdoor activity (using various sensors), locomotion changes within the home to monitor movement frequency/patterns, use of major appliances/objects reflecting reduced movement/use, inefficient movement/object use, etc.
- social interaction (measured by the number of incoming/outgoing phone calls/sms, emails, use of social networks, etc.,) and other social and behavioural parameters (through linguistic analysis of text appearing in e-mails, chat sessions or other electronic message exchanges, monitoring their location throughout the day) while respecting privacy and without becoming invasive.
- physical and cognitive activity through the use of accelerometers and a augmented reality serious game designed specifically for the individual
- self-evaluation (using various tests taken by the individual on a PDA, memory recall tests, button to indicate important episodes (e.g., memory loss, confusion, sudden fatigue), questionnaires performed in an automated way, etc).
In particular for the social interaction particular attention will be paid on the development of a natural language analysis tool that will be able to detect signs of cognitive deficiencies in electronic written text.
The outcome of the analysis of this plethora of data will be a formal and quantitative definition of a frailty metric that will be based on the aforementioned sensing dimensions. For the analysis state-of-the-art data mining techniques will be extended to provide knowledge discovery. The quantitative frailty metric (QFM) will serve as a frailty biomarker.