How do naloxone-based interventions work to reduce overdose deaths: a realist review

Naloxone-based interventions as part of health systems can reverse an opioid overdose. Previous systematic reviews have identified the effectiveness of naloxone; however, the role of context and mechanisms for its use has not been explored. This realist systematic review aims to identify a theory of how naloxone works based on the contexts and mechanisms that contribute to the success of the intervention for improved outcomes.

Pre-registered at PROSPERO, this realist review followed RAMESES standards of reporting. Keywords included 'naloxone' and ' opioid overdose'. All study designs were included. Data extraction using 55 relevant outputs based on realist logic produced evidence of two middle-range theories: Naloxone Bystander Intervention Theory and Skills Transfer Theory.

Harm reduction and/or low threshold contexts provide a non-judgemental approach which support in-group norms of helping and empower the social identity of the trained and untrained bystander. This context also creates the conditions necessary for skills transfer and diffusion of the intervention into social networks. Stigma and negative attitudes held by first responders and stakeholders involved in the implementation process, such as police or GPs, can prohibit the bystander response by inducing fear in responding. This interferes with skills transfer, naloxone use and carriage of naloxone kits.

The findings provide theoretically informed guidance regarding the harm reduction contexts that are essential for the successful implementation of naloxone-based interventions. Peer-to-peer models of training are helpful as it reinforces social identity and successful skills transfer between bystanders. Health systems may want to assess the prevalence of, and take steps to reduce opioid-related stigma with key stakeholders in contexts using a low threshold training approach to build an environment  to support positive naloxone outcomes.

PROSPERO 2019 CRD42019141003.