A multisensor home monitoring system has been developed within a telecare project to help elderly people by observing mobility changes indicative of abnormal events. This paper assesses the motor activity and the relationships between activity measures using the system. The motor activity data (in bed, getting up, getting out, visiting the toilets) are analysed from a statistical perspective to assess changes in occurrence, time and duration. Changes in activity data and correlations between in-bed restlessness and getting up variables show interesting trends in the behavior of elderly people and can be used by the system as a predictive tool in abnormal situations. The objective of the telecare project is to build an abnormal event diagnosis system to help elderly people living alone.