Repeated measures latent class analysis identifies unique daily cocaine-use patterns and associated clinical characteristics: A timeline followback-based analytic approach.

Background: Cocaine use poses significant health risks and treatment challenges. Individual patterns of multi-daily use and their association with clinical characteristics remain poorly understood. Novel approaches that differentiate cocaine-use patterns may improve clinical understanding and guide treatments. This study aimed to identify distinct patterns of cocaine use with a timeline follow-back (TLFB) approach and to evaluate relationships with demographics, cocaine-use severity, and negative emotionality. Methods: Data were collected from 109 (88 male) non-treatment seeking people who use cocaine (mean + SD age 45.6+7.1 years) through past 15-day TLFB interviews tracking daily cocaine use and amounts. Bias-adjusted 3-step repeated-measures latent class analysis identified distinct patterns of daily binary (yes/no) cocaine use. Predictors of class membership including age, income, negative emotionality (depression and anxiety symptoms), and cocaine-use severity were evaluated via regression models as part of the last step of the 3-step RM-LCA approach. Results: Three cocaine-use patterns emerged: high variability across days (Class 1, Discussion: This study identifies distinct patterns of cocaine use associated with different behavioral and demographic characteristics, providing a foundation for tailored treatment development and stratification. Future work should expand these methods and findings to explore the broader context of polysubstance use.