Technical and relational process in MI sessions with a sample of Hispanic/Latinx adults who engage in heavy drinking: A latent growth mediation model

J Ethn Subst Abuse. 2024 Jan 9:1-20. doi: 10.1080/15332640.2024.2301718. Online ahead of print.


The current study tests the Motivational Interviewing (MI) technical and relational hypotheses in a sample of Hispanic/Latinx adults (N = 276) who engage in heavy alcohol consumption. MI causal theory hypothesizes that therapist use of MI consistent skills (i.e., technical hypothesis) and embodiment of the MI Spirit (i.e., relational hypothesis) will elicit client change talk, which is a putative mechanism of positive client outcome after the session. We tested these associations in a rigorous parallel process latent growth curve mediation modeling framework. The data are from a completed randomized clinical trial of a culturally-adapted (CAMI) versus un-adapted MI targeting hazardous alcohol use and consequences. Results. The unconditional growth models for the mediator (i.e., proportion of change talk relative to sustain talk) and two study outcomes (i.e., percent of heavy drinking days; alcohol-related consequences) showed a linear effect over a 12-month period with a slower rate of growth at later timepoints. Contrary to expectations, the latent growth mediation models did not show relationships between MI-consistent skills (i.e., technical predictor) or latent MI Spirit (i.e., relational indicator) and the slope factor for proportion change talk. The slope factor for proportion change talk was also not associated with the slope factors for percent heavy drinking and consequences over follow-up. Conclusions. In this novel population for MI process analysis, the technical and relational hypotheses were not supported. Studies that are exploratory may be needed to further investigate the causal model in populations that are not often represented in MI process research.

Keywords: Dyadic interactions; Latinx; health disparities; process research.