| Background: In the context of substance abuse research, self-efficacy refers to one's confidence to abstain or resist from drug use in high risk situations. Self-efficacy has been found to be an important determinant of behavior change and highly correlated with substance use. Although there is a relatively large body of research on self-efficacy in alcohol dependence, there is a dearth of research in cocaine dependence.;Objectives: The current study was conducted to provide a better understanding of the role of self-efficacy in cocaine addiction and recovery. The specific aims of the study were as follows: 1) evaluate whether self-efficacy is heterogeneous across different high risk situations, 2) understand how self-efficacy changes over time following treatment, and 3) determine whether self-efficacy following treatment can predict long-term abstinence from cocaine use.;Methods: The sample consisted of 216 male veteran's seeking treatment for cocaine dependence and admitted to the West Los Angles Veterans Affairs Medical Center. Self-efficacy data were prospectively collected at the time of admission, 1-year, and 2-years following treatment. Furthermore, long-term abstinence from cocaine use (defined as 5 or more years since last use) was collected at approximately 12-years following treatment.;Results: Cross-sectional data analysis at each time point was evaluated using the latent class analysis (LCA) and revealed there to be three distinct subpopulations; those with low, medium and high self-efficacy. Longitudinal data analysis across the three time points was evaluated using latent transition analysis (LTA) and suggested that self-efficacy was relatively stable over time. An LTA with time varying covariates, time invariant covariates, and distal outcome, suggested that self-efficacy at intake, 1-year, and 2-years did not significantly predict long-term abstinence after controlling for other variables in the model. Finally, a latent transition mixture modeling (LTMM) was utilized to examine the long-term predictive validity of self-efficacy but convergence of the model was not achieved and resulted in an inappropriate solution.;Discussion: The current study found that: 1) self-efficacy was not appreciably heterogeneous across different high risk situations, 2) self-efficacy was stable over time, and 3) self-efficacy was not a robust predictor of long-term abstinence. The findings can be used to: 1) establish cut-off scores for self-efficacy measures in substance abuse research, 2) aid in the development of custom tailored interventions, 3) understand the temporal stability of self-efficacy, 4) provide prevalence rates for low, medium, and high self-efficacy groups prior to and following treatment, 5) understand the relevance of self-efficacy in long-term recovery, and 6) further demonstrate how advanced latent variable models can contribute to substance abuse research. |
