Concurrent Validity of Caffeine Problems and Diagnostic Criteria for Substance Use Disorders.

Background: The DSM-5 proposes caffeine use disorder (CUD) as a condition for further study. The objective of this study was to report on the prevalence of CUD and rates of endorsement for each substance use disorder (SUD) criterion in relation to caffeine compared to alcohol and marijuana in a sample of adolescents presenting for medical care in the primary, adolescent, and substance use clinics at an academic medical center. Methods: A convenience sample of patients ( N  = 213; 66.7% female) aged 12-17 presenting for medical care completed the Composite Diagnostic Interview-Substance Abuse Module questionnaire, with questions regarding use of caffeine, alcohol, and marijuana. Descriptive analyses were used to determine prevalence of CUD and frequency of each endorsed SUD criterion as applied to caffeine versus alcohol or marijuana. Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) were used to determine psychometric properties for CUD. Results: The majority of subjects ( N  = 153) reported past 30-day caffeine use and of these, six (4%) met criteria for CUD. All six also met criteria for either alcohol and/or marijuana use disorders. Of the three essential CUD criteria (failure to quit, use despite harm, and withdrawal), both harm and withdrawal were endorsed significantly more often in relation to caffeine versus alcohol. Descriptive fit indices for the CUD model were excellent (CFI = 0.994, TLI = 0.991). Conclusions: In our sample, the proportion of adolescents that met proposed CUD criteria was low, suggesting that the proposed criteria would not lead to overdiagnosis of CUD. CUD was highly correlated with other SUDs.