Healthcare utilization and mortality after overdose prevention site closure: A linked cohort analysis using segmented...

Evidence regarding the impacts of supervised drug consumption services (SDC) remains mixed, and few evaluations have used individual-level, linkable health data to examine service withdrawal. In September 2024, the Red Deer overdose prevention site (OPS) in Alberta, Canada, was scheduled for closure, with operations ceasing on 31 March 2025. This study examined: (1) changes in weekly opioid agonist therapy (OAT) dispensing among identifiable OPS clients associated with the closure announcement and subsequent service cessation; and (2) changes in acute healthcare utilization and mortality.

Retrospective cohort study using linked provincial administrative health data from 30 June 2024 to 6 months after the Red Deer site closure (27 September 2025). A segmented difference-in-differences interrupted time-series (DID-ITS) design compared Red Deer OPS clients with clients from a continuously operating OPS in Lethbridge. Within-site interrupted time-series (ITS) models were applied for Red Deer-only analyses.

Two provincially regulated OPS programs in Alberta, Canada. The Red Deer OPS closure was publicly announced on 24 September 2024 and implemented on 31 March 2025. The Lethbridge OPS remained operational throughout.

OPS clients with a linked personal health number (PHN) and at least one consumption event at the Red Deer (n = 381) and Lethbridge sites (n = 300).

Primary outcome: weekly proportion of clients receiving ≥1 OAT dispensing.

weekly emergency department (ED) visits, inpatient (IP) admissions, suspected opioid-related emergency medical services (EMS) events and mortality.

Prior to the closure announcement, baseline weekly OAT proportions were lower in Red Deer (9.9%) than in Lethbridge (12.0%), with broadly similar pre-announcement trends. After the closure, the proportion of clients on OAT in Red Deer exceeded those in Lethbridge (16.1% versus 14.4%, respectively; P < 0.021). After closure, there was no statistically significant difference in the number of ED visits or suspected opioid-related EMS events in both the Red Deer only ITS models or in the proportion of these outcomes in the DID-ITS model comparing Red Deer and Lethbridge. Mortality events were rare during the follow-up period, and no statistically detectable increase was observed over the available follow-up.

The announcement and subsequent closure of the Red Deer overdose prevention site in Alberta, Canada, were associated with increased opioid agonist therapy dispensing among identifiable site users. Short-term effects on acute healthcare utilization and mortality appear stable but remain inconclusive due to the limited 26-week follow-up period and low event counts.