The Structured Clinical Interview for Gambling Disorder (SCI-GD) has the potential to bridge a diagnostic clinical gap, but psychometric evaluations have been scarce, in particular in relation to self-reported diagnostic criteria. This study analyzed existing data, including Swedish gamblers (N = 204) from treatment- and help-seeking contexts, self-help groups, and the general population, who were interviewed with the SCI-GD and completed self-report measures. The results indicated that fewer individuals fulfilled the diagnostic criteria for gambling disorder (GD) with the SCI-GD (n = 110, 54%), compared to a self-report Diagnostic and Statistical Manual of Mental Disorders:5th Edition (DSM-5) questionnaire on GD (n = 145, 71%; p < .001). Agreement between interviews and self-reported criteria was generally low (Fleiss kappa range: 0.31-0.52; r range: 0.35-0.55). A Rasch analysis showed that specific diagnostic criteria varied in difficulty, indicating a general pattern of higher item difficulty for the SCI-GD compared to self-reported DSM-5 criteria. Both the SCI-GD and the self-reported DSM-5 criteria performed well in terms of internal consistency, convergent, and discriminant validity. We conclude that the SCI-GD is a reliable and valid diagnostic tool to assess GD among individuals with various gambling behavior patterns. Further research-related and clinical implications are discussed.
Keywords: DSM-5; SCI-GD; SCI-PG; SCID; diagnosis; gambling disorder.