Today we sent a letter to the Health Research Authority (HRA) asking for the retraction of ethical approval of a study that involved the use of Oxevision. The study involved analysing blurred video of patients in their bedroom for up to two hours prior to a self-harm incident and matching that video with incident logs describing the event. The aim of the study was to improve safety in inpatient mental health settings by identifying precursors to harm. It was led by Professor Neil Brimblecombe of London Southbank University who describes the blurred Oxevision video as “a novel source of data”.
From our Freedom of Information (FOI) request we know the study was across three NHS Trusts:
- Essex Partnership University Trust (EPUT)
- Rotherham Doncaster and South Humber NHS Foundation Trust (RDaSH)
- Tees, Esk and Wear Valleys NHS Foundation Trust (TEWV)
We believe the study ran between October 2021 and April 2023 and is complete but not yet published. This means anyone who was in hospital during that time at one of the study sites, and had a self-harm incident, could have had their data used in this study. We believe they have a right to know.
Our concerns about the study are as follows:
Informed consent
The researchers say that because Oxevision is part of ‘routine care’ at these Trusts it is okay for them to use this video in the research.
It is clear to us that this is not the case. Indeed, NHS England wrote to all Trusts in September 2023 reminding them that Oxevision must not be used in a blanket way to be both ethical and lawful, stating:
“It is our view that Vision-based monitoring systems should never be implemented in a blanket way and that any decisions to use VBMS in patient bedrooms should be made in a person-centred way, with the patient themselves where they have capacity to make a decision or through a Best Interest process compliant with the Mental Capacity Act 2005 where they lack capacity to consent to the monitoring system. The use of such systems should be carefully considered on a case by case, patient by patient basis to ensure that any decision to use such systems has a legitimate aim and is both lawful and fair. Their use must also be proportionate to the aim.”
There is also nothing in the patient information we have received from any of these three Trusts that suggests patients are informed that research is a possible way Oxevision might be used. This can be seen in our analysis of the patient posters in which none mention research as a possibility and RDaSH don’t even use the word ‘camera’.
Lack of anonymity
The study authors argue that the “heavily blurred” footage makes it anonymous.
The above images depict the video footage before and after blurring. Although the study describes this as anonymised data, a number of details can be observed about the subjects, for example, their body shape and build. In particular, there are a range of potentially identifying characteristics which would not be obscured by blurring, for example, tattoos, limb difference, body posture or gait, body shape and build, skin tone, wearing of religious headwear, distinctive hairstyles, numerous physical disabilities, as well as a wide variety of features which could result in re-identification.
Physical and psychological safety of participants
Privacy in mental health settings extends beyond policy- it’s a fundamental psychological need. The study does not address the risk of profound psychological trauma of surveillance, particularly for the patients who have experienced childhood sexual abuse, domestic abuse, or politically motivated surveillance or the risk of exacerbating psychosis and paranoia related to surveillance. We know from lived experience the impact this can have on patients.
We are also aware that there were two deaths at EPUT during the study period in which Oxevision use was subsequently mentioned at inquest. We are horrified that these patients’ data could be part of this study.
If you believe your data could be in this study or have concerns you can contact the research department of each relevant Trust here:
Leave a Reply