Untoward incidents (UIs) are unfortunately common within mental health inpatient services. Monitoring service user safety and wellbeing in inpatient settings can be highly challenging. Bedroom environments have changed in recent years, with fewer shared rooms and more ensuite bathrooms — improvements that have added to the challenge of carrying out observations. Sadly, suicides and self-harm still take place on wards, largely in non-public areas. Research on risks in bedrooms is sparse despite the large proportion of time spent there by service users.
Led by Professor Brimblecombe, researchers are analysing anonymised incident data from inpatient mental health settings to identify potential precursors to these incidents. The data consist of anonymised incident logs from NHS trusts and anonymised Oxevision log data (including, for example, the locations of movement within a room).
During the initial phase of the project, a tool to collect the data and to present it in suitable format for analysis will be developed, using a small sample of anonymised incidents combined with anonymised data from Oxevision. This work is taking place as part of a Health Research Authority Research Ethics Committee approved study.
In the second phase — after the methodology for identifying and recording possibly relevant behaviours has been defined and critiqued by a clinical group — large numbers of anonymised UIs and non-UI recordings will be reviewed under the appropriate governance.
Descriptive statistics will summarise characteristics of UIs and inferential statistics will explore correlations between variables. In this way, candidate precursors to harm can be identified and validated. The findings will be published with the aim of providing further insights to ward staff around identifying imminent incidents so that they can be averted. The feasibility of automated detection of precursors will also be explored, with appropriate medical device clearances.
Early findings suggest that this type of analysis can yield new insights with the ultimate aim of helping staff to avert certain categories of UIs. It is anticipated that results from the initial phase of the project will be published in late 2022.
