A psychometric evaluation of the DSM-IV pathological gambling diagnostic criteria.

Specific diagnostic criteria for pathological gambling (PG) have been available for 25 years, since the publication of DSM-III. Little research has examined the psychometric performance of the diagnostic criteria. The goal of the present report from the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project was to examine the sensitivity, specificity and predictive values of the DSM-IV PG criteria for psychiatric outpatients who screened positive for a gambling problem. A total of 1709 psychiatric outpatients were evaluated with a semistructured diagnostic interview for PG. Of all patients 88 screened positive for PG, 40 of whom met DSM-IV diagnostic criteria for a lifetime history of PG. All ten DSM-IV criteria were significantly more frequent in the PG group. The sensitivity of the criteria ranged from 25.0% to 90.0% (mean = 67.8%), whereas specificity ranged from 62.5% to 100% (mean = 81.9%). Positive predictive values ranged from 64.1% to 100% (mean = 78.9%), and negative predictive values ranged from 61.5% to 90.7% (mean = 77.1%). Guidelines are recommended for determining whether a diagnostic criterion should be retained as part of the set of diagnostic criteria, and our results suggested that two of the DSM-IV PG criteria are candidates for elimination (criterion 8--commitment of illegal acts; criterion 10--reliance on others for financial assistance to relieve a desperate financial problem).