Efficiency and validity of commonly used substance abuse screening instruments in public psychiatric patients

Background: The more commonly used screening instruments for substance abuse were largely developed for addictive populations. We compared several alcohol and drug abuse scales to determine their efficiency and validity for psychiatric patients. Method: The subjects were 100 consecutively admitted patients to a public psychiatric facility. DSM-III-R diagnoses, obtained from the alcohol and drug scales of the SCID-P, were the criterion measure. Methods of reliability included inter-rater agreement, estimates of internal consistency, and repeat test administration. Sensitivity, specificity and more infrequently used accuracy indices, such as likelihood ratios and Receiver Operating Characteristic (ROC) analysis, were utilized to assess scale validity. Results: First, the reliability of all scales was high. Second, the instruments generally demonstrated highly acceptable levels of screening accuracy. Third, the intake evaluation was as reliable and valid as screening after admission on the unit. Finally, instruments were least discriminating for current problems (past 30 day). Conclusions: Lifetime measures were found to be reliable and valid for public psychiatric patients but further research is needed on increasing the accuracy of screening for current substance abuse problems and the effectiveness of multiple screening approaches.