Characterizing the predictive validity of measures of susceptibility to future use of combustible, vaporized and edible cannabis products in adolescent never-usersedit
Background and aims: The construct of susceptibility to substance use initiation (i.e. cognitive proclivity to future use) is critical for prevention efforts in adolescent populations. This study aimed to provide empirical evidence for the validity of the susceptibility construct for different cannabis products (i.e. combustible, edible or vaporized cannabis), and evaluate whether susceptibility measures are predictive of subsequent initiation. Design: Prospective cohort study including baseline data (Spring 2015) and four follow-up surveys administered every 6 months through Spring 2017. Setting: Ten schools in the Los Angeles, California metropolitan area. Participants: Adolescents [n = 2100; mean age = 16.1; standard deviation (SD) = 0.41; 54% female] who reported never having used any cannabis product at baseline.Measurements: We assessed five indices of a susceptibility to use cannabis composite index at baseline, adapted from a validated tobacco use index (intention to use, willingness, curiosity and positive/negative cannabis use outcome expectancies, with four response categories, definitely not  to definitely yes ), by cannabis product (combustible, edible or vaporized). A composite index was created for each product by averaging responses across the five susceptibility items. Subsequent initiation of use of each cannabis product was assessed at each follow-up wave. Findings: Factorial validity for unidimensionality for each five-item index (by product) was confirmed. The composite index for susceptibility to cannabis use was greatest for combustible (mean = 1.44; SD = 0.58), moderate for edible (mean = 1.37; SD = 0.53) and lowest for vaporized cannabis (mean = 1.30; SD = 0.44). The associations of each composite susceptibility index with subsequent initiation of that product and each of the other cannabis products over follow-up (i.e. cross-product associations) were statistically significant, with hazard ratios ranging from 2.30 to 2.80 across 24 months of follow-up (all Ps < 0.05). Conclusions: A five-item susceptibility to cannabis use composite index (by product) appears to be useful for characterizing and predicting youth at risk for cannabis use initiation across a spectrum of cannabis products.