Do DSM-5 substance use disorder criteria differ by user care settings? An item response theory analysis approach. (May 2021)

• Record Type: Journal Article
• Authors: Kervran, C.; Shmulewitz, D.; Serre, F.; Denis, C.; Roux, P.; Jauffret-Roustide, M.; Lalanne, L.; Hasin, D.; Auriacombe, M.
• Is Part Of: Addictive behaviors. Volume 116(2021)
• Publication Date: 2021-05
• Subjects: Differential item functioning -- Substance use disorder -- DSM-5 -- Harm reduction setting -- Patient seeking treatment; Substance abuse -- Periodicals; Alcoholism -- Periodicals; Drug addiction -- Periodicals; Nicotine addiction -- Periodicals; Smoking -- Periodicals; Gambling -- Psychological aspects -- Periodicals; Electronic journals; 362.29
• DOI: 10.1016/j.addbeh.2020.106797
• Languages: English
• ISSNs: 0306-4603

Abstract

Highlights: The overall consistency of the SUD criteria was confirmed in a large French clinical sample. No systematic differential functioning of SUD criteria across substances and by care settings. Diagnostic criteria of loss of control of use, especially craving function well in different care settings. Some diagnostic criteria of consequence of use function differently by care settings. Abstract: Aim: To examine differences in the psychometric characteristics of diagnostic criteria for Substance Use Disorders (SUD) between substance users in harm reduction settings (HR) and substance users seeking treatment (Tx). Methods: Differential Item and Test Functioning (DIF & DTF) analysis were performed to examine differences in the difficulty of endorsement and in discrimination of the 11 diagnostic criteria and to test if the criteria set as a whole (the "test") functioned differently by care settings (Tx vs. HR) for alcohol, cocaine, cannabis, opiates and tobacco. To test uniform and nonuniform DIF, multiple indicator multiple cause (MIMIC) structural equation models were used. Results: Regardless of the substance, the DSM-5 criteria "craving", "large amount", "time spent", "tolerance" and "activities given up" had similar functioning by care settings. Little evidence for DIF was found for other criteria. The criteria set as a whole did not function differently by care settings for alcohol, cocaine and tobacco. At the same trait severity, compared to HR, the Tx subgroup had a greater number of endorsed criteria for cannabis and a smaller number of endorsed criteria for opioids. Conclusion: The unidimensionality of the 11 DSM-5 criteria and applicability of all criteria and diagnosis was confirmed in this large sample of problematic substance users. While the majority of the criteria related to loss of control of substance use, functioned well in both care settings, the criteria related to consequences of substance use had several differential functioning.