ICT solution for balance assessment in elderly: the DOREMI system validation and applicability

Introduction: Aging process is characterized by decline of body functions: one of the major risks is represented by falls. Several screening tools/tests have been used to assess stability. We describe the integration between single-item Berg scale, Wii Balance Board and neural networks to create a new platform for balance assessment. Methods: Two cohorts of older people (age 65-80) were enrolled for DOREMI balance assessment. The BERG balance score and anthropometric parameters (weight, height, BMI) were assessed for each subject. Group 1 (25 participants) showed a BERG score between 41 and 56; Group 2 (50 participants), enrolled in collaboration with PERSSILAA project, had a wider score (20-56) allowing to test accuracy of DOREMI tool in people having severe instability. Item-10 of BERG scale has been performed on the Wii balance board, inviting participant to turn the head to look behind over toward the right shoulder and then over the left. Data signals have been collected by custom developed software and analysed by neural network system for human activity recognition. Results: In Group 1, total BERG score estimated using item-10 data presented a strong correlation with individual total BERG score test; in Group 2 data analysis is under investigation. Key conclusions: DOREMI balance assessment is simple, automated, cost-effective and time-saving ICT solution for prevention of frailty. Once validated in a broad range of BERG scale, this system could objectively predict subjects at higher risks of falls, as well as improvement in stability and/or positive evolution in physical health status after specific treatments.