Monitoring habits and physiological data in the frail elderly

Abstract

Introduction: There is a need for new models of care enabled by technology to support long-term and independent living of the elderly. Integrating telecare and telehealth technologies can be used to provide innovative support in an unobtrusive way. inCASA is a European Commission funded project that uses an integrated platform to monitor both health and habits data in the frail elderly and demonstrate the concept of integrated health and social services through pilot trials. This paper presents the joint analysis of habits and clinical data from the UK pilot. Aims and Objectives: The aim of the analysis is to determine the correlation between change in habits behaviour, change in physiological data and deterioration in the condition of the patient. Methods: 40 participants registered at Chorleywood Health Centre who are over the age of 65 and who have been assessed as frail have been recruited to participate in the project. Participants have been given a combination of physiological devices such as blood pressure, weight , SpO2 and blood glucose monitor according to their condition, together with telecare devices such as PIR, bed and chair sensors, and medication dispenser. All data is automatically transmitted wirelessly from the devices to the home gateway and then wirelessly to the remote server. The data is processed in order to detect deviations from the norm and notify clinicians by means of a clincial portal for possible intervention. Two types of thresholds were used for the clinical data: an absolute threshold taken from clinical assessment protocols (e.g. 140/85 mmHg for systolic/diastolic BP) and subject specific thresholds (mean+/-2SD). Habits data processing is divided into two main parts: training and detecting deviations. We divide a day into four periods: 00:00-06:00, 06:0012:00, 12:00-18:00, 18:00-24:00. We count the number of movements for each of these periods, and calculated their average and standard deviations (SD) over 15-20 days. Using the mean and SD of each period, period-specific thresholds were set to detect deviations from normal pattern: mean+/-2SD for major deviation alerts and mean+/-1.5SD for minor deviation alerts. Deviation from the norm is then detected by comparing the number of movements in a period with the threshold values for that period. We investigated whether activity is correlated with pysiological parameters.