Table 3 Secondary outcomes and analytic plans

Implementation outcomes

 Fidelity as measured by adherence to a critical action checklist regarding the provision of ED-initiated buprenorphine with referral for ongoing MOUD using standard models including mixed models, or other appropriate methodology.

 Rates of enrolled patients receiving an appointment for opioid treatment provider/program upon ED discharge using the model described for the primary analysis.

 Number of ED providers receiving DATA 2000 training using appropriate methodology. Permutation tests will be considered for this analysis.

 Number of clinicians providing ED-initiated buprenorphine with referral for ongoing MOUD using count models. Models considered will be Poisson, zero-inflated Poisson, negative binomial, and zero-inflated negative binomial. An offset will be considered for how long the provider has been trained.

 ED provider readiness and preparedness ruler scores to initiate buprenorphine and provide referral for ongoing MOUD using standard models including mixed models or other appropriate methodology.

 ED Organizational Readiness to Change Assessment (ORCA) scores relating to ED-initiated buprenorphine with referral for ongoing MOUD using standard models including mixed models or other appropriate methodology.

 Community opioid treatment provider/program readiness and preparedness ruler scores to continue MOUD for patients with OUD who have received ED-initiated buprenorphine using standard models including mixed models or other appropriate methodology.

 Community opioid treatment provider/program Organizational Readiness to Change Assessment (ORCA) scores relating to receiving patients with OUD who have received ED-initiated buprenorphine using standard models including mixed models or other appropriate methodology

Effectiveness outcomes

 Self-reported days of illicit opioid use (past 7 days) as measured by TLFB methods at 30 days using count models. Models considered will be Poisson, zero-inflated Poisson, negative binomial, and zero-inflated negative binomial.

 Overdose events (past 30 days) captured by participant self-report, state medical examiner records, National Death Index, and review of medical records will be compared using count models. Models considered will be Poisson, zero-inflated Poisson, negative binomial, and zero-inflated negative binomial. An offset will be considered for how long a participant was in a given period if less than 30 days.

 HIV risk-taking behaviors (past 30 days) as measured by HIV Risk Taking Behavior Score using standard models such as mixed models or other appropriate methodology.

 Healthcare service utilization (past 30 days) measured by Health Services Utilization Form will be compared using count models. Models considered will be Poisson, zero-inflated Poisson, negative binomial, and zero-inflated negative binomial. An offset will be considered for how long a participant was in a given period if less than 30 days.

 Rates of illicit opioid negative urines at 30 days using the model described for the primary analysis in section.