Sensitivity Analysis of the No-Omitted Confounder Assumption in Latent Growth Curve Mediation Models.

Latent growth curve mediation models are increasingly used to assess mechanisms of behavior change. For latent growth mediation model, like any another mediation model, even with random treatment assignment, a critical but untestable assumption for valid and unbiased estimates of the indirect effects is that there should be no omitted variable that confounds indirect effects. One way to address this untestable assumption is to conduct sensitivity analysis to assess whether the inference about an indirect effect would change under varying degrees of confounding bias. We developed a sensitivity analysis technique for a latent growth curve mediation model. We compute the biasing effect of confounding on point and confidence interval estimates of the indirect effects in a structural equation modeling framework. We illustrate sensitivity plots to visualize the effects of confounding on each indirect effect and present an empirical example to illustrate the application of the sensitivity analysis.

Race/ethnicity and racial group composition moderate the effectiveness of mindfulness-based relapse prevention for substance use disorder.

INTRODUCTION: Mindfulness-based relapse prevention has shown promise as a treatment for substance use disorder but its efficacy according to racial/ethnic minority status and group composition is unknown. METHOD: This is a secondary analysis of existing data (Bowen et al., 2014) testing individual race/ethnicity and racial/ethnic group composition as moderators of mindfulness-based relapse prevention (MBRP). Participants (N = 191; 29% female; 47% racial/ethnic minority; mean age = 39) with substance use disorder were randomized to MBRP or relapse prevention (RP). Outcomes were heavy drinking days (HDD) and drug use days (DUD) 12 months after treatment completion. Negative binominal regression models were conducted. RESULTS: Analyses accounted for drug of choice. Individual race/ethnicity was a significant moderator of substance use outcomes. White participants had lower HDD in MBRP than RP (IRR = 0, 95% CI: 0,0), whereas for minority participants, there was no treatment difference in HDD. Conversely, minorities had lower DUD in MBRP than RP (IRR = 0.03, 95% CI: 0.01, 0.10), whereas for whites there was no treatment difference in DUD. Group racial/ethnic composition was a significant moderator. Participants in groups with more than half whites had lower HDD in MBRP than RP (IRR = 0.01, 95% CI: 0, 0.09), whereas for participants in groups with more than half minorities there was no treatment difference in HDD. Exploratory analyses suggested MBRP resulted in better outcomes than RP when individual race/ethnic status was reflected in the group race/ethnicity (i.e., whites in groups with more than half whites or minorities in groups with more than half minorities). CONCLUSIONS: Among whites, MBRP appears to be more effective than RP in preventing heavy drinking relapse. However, among racial/ethnic minorities, MBRP appears to more effective than RP in preventing drug use relapse. This suggests that the interaction between individual race/ethnicity and group composition may influence primary outcomes.