An empirical evaluation of proposed changes for gambling diagnosis in the DSM-5

Aims: Recommendations related to pathological gambling for the fifth edition of the Diagnostic and Statistic Manual for Mental Disorders (DSM) are to eliminate the criterion related to committing illegal acts and reduce the threshold for diagnosis from five to four criteria. This study evaluated the impact of these changes on prevalence rates and classification accuracy. Design: Data were analyzed from five samples, varying in severity of gambling problems. Settings and participants: Surveys of randomly selected household residents in the United States (US) (n=2417), gambling patrons (n=450), individuals in brief intervention studies (n=375), patients in community-based gambling treatment programs (n=149) and participants in randomized intervention studies (n=319). Measurements: The National Opinion Research Center DSM-IV Screen for Gambling Problems (NODS) was administered to all participants. Internal consistency and factor structure were evaluated using both 10 and nine criteria. Base rates, hit rates, sensitivity, specificity and overall agreement were compared across classification systems, using DSM-IV classification as the standard. Findings: Eliminating the illegal acts criterion did not impact internal consistency and modestly improved variance accounted for in the factor structure. In comparing a classification system using four of 10 criteria versus one using four of nine, the four of nine system yielded equal or slightly better classification accuracy in all comparisons and across all samples. Conclusions: The inclusion of the illegal acts criterion in the proposed DSM-V pathological gambling diagnosis does not appear necessary for diagnosis of pathological gambling and, if it is eliminated, reducing the cut-point to four results in more consistent diagnoses relative to the current classification system.