Experimental and Clinical Psychopharmacology, Vol 33(6), Dec 2025, 531-575; doi:10.1037/pha0000791
Despite awareness of polysubstance use—the co-use of multiple drugs—and its associated risks, there is a lack of research consensus on how to identify and classify individuals engaging in polysubstance use. Latent class analysis (LCA) and latent profile analysis (LPA) are data-driven approaches that may improve the identification and classification of polysubstance use. By clustering data using different indicators, LCA/LPA can extract subgroups of common drug use patterns within a sample. Variability in how LCA/LPA are conducted, however, can substantially impact how subgroups are extracted and have not been thoroughly reviewed. The present review was one of a two-part series preregistered on PROSPERO entitled, “A systematic review of studies using latent class analysis to examine patterns of polysubstance use in adults (Part 1) and adolescents (Part 2)” (CRD42022352293). The present review sourced relevant studies using LCA/LPA in the context of characterizing adult polysubstance use and identified factors influencing the number of latent classes extracted. Across several articles using LCA/LPA (n = 136), the current review found differences, for example, in the number of extracted classes based on study sample sizes (ρ = .275, p r = .244, p = .004). The present review noted substantial heterogeneity in study methodologies, statistical analyses, and latent class solutions. For future research, the review suggests some methodological considerations including attention to sample sizes, study locations, and the number of indicators included in LCA/LPAs. (PsycInfo Database Record (c) 2025 APA, all rights reserved)