Barriers impacting the POINT pragmatic trial: the unavoidable overlap between research and intervention procedures in “real-world” research

Abstract

Background: This manuscript provides a research update to the ongoing pragmatic trial of Project POINT (Planned Outreach, Intervention, Naloxone, and Treatment), an emergency department-based peer recovery coaching intervention for linking patients with opioid use disorder to evidence-based treatment. The research team has encountered a number of challenges related to the "real-world" study setting since the trial began. Using an implementation science lens, we sought to identify and describe barriers impacting both the intervention and research protocols of the POINT study, which are often intertwined in pragmatic trials due to the focus on external validity. Method: Qualitative data were collected from 3 peer recovery coaches, 2 peer recovery coach supervisors, and 3 members of the research team. Questions and deductive qualitative analysis were guided by the Consolidated Framework for Implementation Research (CFIR). Results: Nine unique barriers were noted, with 5 of these barriers impacting intervention and research protocol implementation simultaneously. These simultaneous barriers were timing of intervention delivery, ineffective communication with emergency department staff, lack of privacy in the emergency department, the fast-paced emergency department setting, and patient's limited resources. Together, these barriers represent the intervention characteristics, inner setting, and outer setting domains of the CFIR. Conclusion: Results highlight the utility of employing an implementation science framework to assess implementation issues in pragmatic trials and how this approach might be used as a quality assurance mechanism given the considerable overlap that exists between research and intervention protocols in real-world trial settings. Previously undocumented changes to the trial design that have been made as a result of the identified barriers are discussed.