With the implementation of the Advanced Metering Infrastructure (AMI), comes the opportunity to gain valuable insights into an individual’s daily habits, patterns and routines. A vital part of the AMI is the smart meter. It enables the monitoring of a consumer’s electricity usage with a high degree of accuracy. Each device reports and records a consumer’s energy usage readings at regular intervals. This facilitates the identification of emerging abnormal behaviours and trends, which can provide operative monitoring for people living alone with various health conditions. Through profiling, the detection of sudden changes in behaviour is made possible, based on the daily activities a patient is expected to undertake during a 24-h period. As such, this paper presents the development of a system which detects accurately the granular differences in energy usage which are the result of a change in an individual’s health state. Such a process provides accurate monitoring for people living with self-limiting conditions and enables an early intervention practice (EIP) when a patient’s condition is deteriorating. The results in this paper focus on one particular behavioural trend, the detection of sleep disturbances; which is related to various illnesses, such as depression and Alzheimer’s. The results demonstrate that it is possible to detect sleep pattern changes to an accuracy of 95.96% with 0.943 for sensitivity, 0.975 for specificity and an overall error of 0.040 when using the VPC Neural Network classifier. This type of behavioral detection can be used to provide a partial assessment of a patient’s wellbeing.