Predicting Intent to Refer Buprenorphine: The Role of Familiarity and Beliefs

Background: Despite higher rates of individuals with opioid use disorder, correctional facilities underutilize medication-assisted treatment. The mechanisms which motivate correctional employees to refer medications, such as buprenorphine, for treatment are currently unclear as is the relationship between previously established constructs, such as awareness of buprenorphine and harm reduction ideologies. This study aims to determine presence of a reciprocal relationship between familiarity with buprenorphine and harm reduction beliefs among community corrections employees. Furthermore, this study will help identify how this relationship influences future intent to refer among this population.

Methods:Data were used from the Criminal Justice Drug Abuse Treatment Studies 2 (CJ-DATS 2) to determine the presence of reciprocal causality between familiarity with buprenorphine and harm reduction beliefs among the correctional employees (n=873) taken at baseline and 12 months later. Structural equation modeling was also performed to test whether the training provided in the CJ-DATs influenced intent to refer buprenorphine in the future through its effects on the hypothesized mediators of familiarity and harm reduction beliefs.

Four models were used with the full sample of respondents: (1) baseline model with autoregressive paths; (2) autoregressive effects with familiarity predicting harm reduction beliefs later time point; (3) autoregressive effects with harm reduction beliefs predicting familiarity at a later time point; (4) a fully cross lagged model with autoregressive effects and both variables predicting the other at later time points.

Then, the best fitting model was used to test the impact of receipt of training on familiarity and harm reduction beliefs, and the effect on intent of future referral at time 2. The fit of the models was assessed by several fit indices including Chi-Square (X2), the Comparative Fit Index (CFI), the Normed Fit Index (NFI) and the Root Mean Square Error of Approximation (RMSEA).

Results: Results indicate that Model 4, the fully cross-lagged model, provided a significantly better fit to the data than Model 3, X²_diff (1)= 5.62, p <0.05. In addition, the autoregressive paths for the variables were strong in the final model, indicating that the constructs were stable over time. Lastly, the paths between variables at baseline and later time points were also statistically significant, indicating that reciprocal causality existed between familiarity and harm reduction beliefs.

The analysis of reciprocal causality was used to inform the structural equation model to evaluate the effect of training on intent to refer through the hypothesized mediators, familiarity with buprenorphine and harm reduction beliefs. The results show that training had positive, indirect effects on intent to refer in the future through the reciprocal effects on familiarity and harm reduction beliefs. Training also had a positive, indirect effect on intent to refer through the autoregressive effect of familiarity between the two time points

Conclusions: Findings show that training may influence correctional employees’ intent to refer individuals to proven effective treatment options, such as buprenorphine. The mechanism that supports the intent to refer involves familiarizing employees with medication options for opioids and harm reduction education.