Comparison of trauma center patient self-reports and proxy reports on the alcohol use identification test (AUDIT)

Background: There has been increased attention on identifying alcohol problems among individuals admitted to trauma centers. Reports about patients' drinking made by significant others represent one potential method to address both concerns about the validity of patients' self-reports and the need to have a proxy measure for trauma patients who are unable to provide information about their drinking behavior. The present study evaluated the level of agreement between trauma patients' self-report and proxy report on the Alcohol Use Disorders Identification Test (AUDIT). Methods: The study consisted of patients admitted to a Level I trauma center who screened positive for potential alcohol problems, who completed the AUDIT and National Institute of Mental Health Diagnostic Interview Schedule and who had a collateral who provided proxy information about the patient. Proxy informants completed the AUDIT on the basis of their knowledge and perception of patients' drinking behavior, as well as a checklist of alcohol-related problems experienced by the patients in the 30 days before admission. Results: Patient- and proxy-completed AUDIT total and subscale scores were highly correlated and did not differ from one another except for the dependence subscale, with proxies indicating a higher number of symptoms than did patients. A relatively high percentage of agreement (77.5%) was found overall in the classification of patients as having or not having problem drinking on the basis of self-report and proxy AUDITs; 56.4% of both sources indicated a drinking problem and 21.1% of both agreed that there was not a problem. Patients classified as having problem drinking on self-report AUDITs were rated by proxies as having significantly higher levels of hazardous drinking, alcohol-related problems, and dependence symptoms than patients whose AUDIT classified their drinking as nonproblematic. Conclusion: The relative comparability of proxy- and patient-completed AUDIT scores and classifications suggests that patient and proxy reports corroborate and, in cases of trauma patients' inability to provide self-reports, may serve as proxies for patients' reports of drinking.