Using remote monitoring to improve prediction and develop understanding of aggression in inpatient mental health services

Student thesis: Doctoral ThesisDoctor of Philosophy


Inpatient aggression in forensic mental health services has serious negative consequences for both staff and service users. Methods of monitoring and predicting risk of aggression are typically based on changeable dynamic risk factors. Previous research shows that risk assessments are typically inaccurate, particularly for events when aggression escalate rapidly. One reason for this is that these assessments are carried out infrequently and cannot capture short-term change in relevant risk factors. These assessments are also limited to staff-reported observable variables and cannot capture the psychophysiological deregulation which neurophysiological theories suggest are relevant to the risk of aggressive behaviour. Developments in experience sampling methods and passive remote monitoring technology now enable these dynamic variables to be measured over extended periods of time in participants’ typical daily environment with limited intrusion in their daily activities. Across five empirical studies the overall aim of this thesis was to investigate whether monitoring relevant dynamic risk factors frequently could predict inpatient aggression short time periods. A systematic review was conducted (Chapter 2) which identified 26 candidate dynamic risk factors in forensic mental health services for inclusion in subsequent studies in this thesis. A series of qualitative focus groups was then conducted with forensic mental health nursing staff (Chapter 3), which highlighted potential clinical benefits of remote monitoring in addition to challenges that would need to be overcome. Two pilot studies with service users and staff were then carried out (Chapters 4 and 5), which confirmed the acceptability of remote monitoring in forensic services and identified optimal timeframes for identifying change in dynamic risk factors. An empirical study was then conducted (Chapters 6, 7 and 8) with forensic mental health service users to investigate whether short term changes in dynamic risk factors were associated with future aggression. Findings from these studies identified a number of psychological and psychophysiological risk factors relating to Positive and Negative Affect, Aggression Ideation, Phasic Electrodermal Activity and reduced Heart Rate Variability which significantly predicted acts of physical and verbal aggression up to 40-minutes before the aggressive outcome was observed. These findings point to a window opportunity between the onset of a particular risk factors and occurrence of aggression in which effective and targeted intervention and support may avert the aggressive incident.
Date of Award1 Jun 2021
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorTil Wykes (Supervisor) & Matteo Cella (Supervisor)

Cite this