Is the quality of brief motivational interventions for drug use in primary care associated with subsequent drug use?

Background Although a number of brief intervention approaches for drug use are based on motivational interviewing (MI), relatively little is known about whether the quality of motivational interviewing skills is associated with intervention outcomes. Method The current study examined whether indices of motivational interviewing skill were associated with subsequent drug use outcomes following two different MI-based brief interventions delivered in primary care; a 15 min Brief Negotiated Interview (BNI) and a 45 min adaptation of motivational interviewing (MOTIV). Audio recordings from 351 participants in a randomized controlled trial for drug use in primary care were coded using the Motivational Interviewing Treatment Integrity Scale, (MITI Version 3.1.1). Separate negative binomial regression analyses, stratified by intervention condition, were used to examine the associations between six MITI skill variables and the number of days that the participant used his/her main drug 6 weeks after study entry. Results Only one of the MITI variables (% reflections to questions) was significantly associated with the frequency of drug use in the MOTIV condition and this was opposite to the hypothesized direction (global p = 0.01, adjusted IRR 1.50, 95%CI: 1.03–2.20 for middle vs. lowest tertile [higher skill, more drug use]. None were significantly associated with drug use in the BNI condition. Secondary analyses similarly failed to find consistent predictors of better drug outcomes. Conclusion Overall, this study provides little evidence to suggest that the level of MI intervention skills are linked with better drug use outcomes among people who use drugs and receive brief interventions in primary care. Findings should be considered in light of the fact that data from the study are from negative trial of SBI and was limited to primary care patients. Future work should consider alternative ways of examining these process variables (i.e., comparing thresholds of proficient versus non-proficient skills) or considering alternative methods of coding intervention skills.