Background and objectives: The goal of this study was to determine the diagnostic accuracy of brief screens for Gambling Disorder within a sample of people receiving outpatient treatment for substance use disorders.
Methods: Individuals (n = 300) recruited from intensive outpatient substance use treatment (23.67%) or methadone maintenance programs (76.34%) participated in the study. Four brief screens for Gambling Disorder were administered and compared to DSM-5 criteria. Receiver operator curves were created and an Area Under the Curve (AUC) analysis (an overall summary of the utility of the scale to correctly identify Gambling Disorder) was assessed for each.
Results: On average participants were aged 46.4 years (SD = 10.2), African American/Black (70.7%), with an income less than $20,000/year (89.5%). Half the participants were female. Approximately 40% of participants (40.5%; n = 121) met DSM-5 criteria for Gambling Disorder. Accuracy of the brief screens as measured by hit rate were .88 for the BBGS, .77 for the Lie/Bet, .75 for NODS-PERC, and .73 for the NODS-CLiP. AUC analysis revealed that the NODS-PERC (AUC: .93 (95% CI: .91-.96)) and NODS-CLiP (AUC: .90 (95% CI: .86-.93)) had excellent accuracy.
Discussion and conclusions: The NODS-PERC and NODS-CLiP had excellent accuracy at all cut-off points. However, the BBGS appeared to have the best accuracy at its specified cut-off point.
Scientific significance: Commonly used brief screens for Gambling Disorder appear to be associated with good diagnostic accuracy when used in substance use treatment settings. The choice of which brief screen to use may best be decided by the needs of the clinical setting.
© American Academy of Addiction Psychiatry.