Mohiuddin, Syed Golam
Enabling health, independence and wellbeing for patients with bipolar disorder through Personalised Ambient Monitoring.
University of Southampton, School of Management,
This thesis describes the role of mathematical modelling in the evaluation of an innovative automated system of wearable and environmental sensors to monitor the activity patterns of patients with Bipolar Disorder (BD). BD is a chronic and recurrent mental disorder associated with severe episodes of mania and depression, interspersed with periods of remission. Early detection of transitions between the normal, manic and depressed stages is crucial for effective self-management and treatment. Personalised Ambient Monitoring (PAM) is an EPSRC-funded multidisciplinary project involving biomedical engineers, computer scientists and operational researchers. The broad aim of PAM is to build and test a network of sensors (chosen by the patient) to collect and analyse daily activity data in order to identify an ‘activity signature’ for that individual in various health states. The hypothesis is that small but potentially significant changes in this activity pattern can then be automatically detected and the patient alerted, enabling him/her to take appropriate action. The research presented in this thesis involves the development and use of a Monte Carlo simulation model to evaluate the potential of PAM without the need for a costly and time-consuming clinical trial. A unique and novel disease state transition model for bipolar disorder is developed, using data from the clinical literature. This model is then used stochastically to test many different scenarios, for example the removal or technical failure of a sensor, or the limited availability of various types of data, for various simulated patient types and a wide range of assumptions and conditions. The feasibility of obtaining sufficient information to derive clinically useful information from a limited set of sensors is analysed statistically. The minimum best set of sensors suitable to detect both aspects of the disorder is identified, and the performance of the PAM system evaluated for a range of personalised choices of sensors
Actions (login required)