Evaluating a pilot intervention development programme utilising respiratory monitoring to prevent drug-related deaths

Background:

RESCU is a mixed-methods observational cohort study addressing Scotland’s dramatic rise in drug-related deaths, most of which are caused by opioid-induced respiratory depression. RESCU investigates a chest-worn accelerometer sensor’s ability to accurately capture respiratory patterns of people who use drugs (PWUD) to determine overdose response trigger points. The study assesses device acceptability to PWUD, third sector and first responder groups to create an intervention pathway.

Methods:

Participants (n=70) were recruited from an injection equipment provision site in the centre of Dundee. Participants received a sensor and a gateway device to record their breathing for retrospective analysis. Participants recorded their substance use and respiratory data gathered over a four-week period, returning weekly for data download.

Semi-structured interviews and focus groups were conducted with participants who engaged in the study protocol (n=21) and stakeholder groups (n=8) about device acceptability. Transcripts were analysed using Reflexive Thematic Analysis. Factors influencing device acceptability to participants were interpreted using the COM-B Model of Behaviour. Normalisation Process Theory was used to assess device integration into existing services.

Recorded accelerometer data was analysed using developmental algorithms measuring respiratory rate and frequency and duration of apnoeic episodes.

Results:

During February – December 2022, 70 participants had been recruited in the study.

Experiences with overdose or drug-related death were identified in qualitative data as motivating factors for device wear. Patient choice and device accuracy were emphasised by first responders.

8,614 apnoeic episodes of >10s duration were detected at the highest probability level in 6,202.08 hours of respiratory data. Participants who engaged fully or partially in the study protocol (n=48) used the device for an average of 189.79 hours (SD: 132.70), over an average capture period of 673.38 hours (SD: 184.57). The mean coverage per participant was 28.06% (SD: 18.67). The average respiratory rate per participant was 11 breaths per minute (SD: 3.36).  

Data review identified the presence of Cheyne-Stokes respiration following consumption of opiates and benzodiazepines. No events leading to interventions occurred during monitoring.

Conclusions:

Current data analysis suggests the device successfully captures respiratory anomalies. The process evaluation revealed the importance of positive therapeutic relationships and participant choice. Data analysis is ongoing, a future study aim being identification of emergency response trigger points.