The psychometric properties of the AUDIT: a survey from a random sample of elderly Swedish adults

BACKGROUND Increasing alcohol consumption and related harms have been reported among the elderly population of Europe. Consequently, it is important to monitor patterns of alcohol use, and to use a valid and reliable tool when screening for risky consumption in this age group. The aim was to evaluate the internal consistency reliability and construct validity of the Alcohol Use Disorders Identification Test (AUDIT) in elderly Swedish adults, and to compare the results with the general Swedish population. Another aim was to calculate the level of alcohol consumption (AUDIT-C) to be used for comparison in future studies. METHODS The questionnaire was sent to 1459 Swedish adults aged 79-80 years with a response rate of 73.3%. Internal consistency reliability, were assessed using Cronbach alpha, and confirmatory factor analysis assessed construct validity of the Alcohol Use Disorders Identification Test (AUDIT) in elderly population as compared to a Swedish general population sample. RESULTS The results showed that AUDIT was more reliable and valid among the Swedish general population sample than among the elderly and that Item 1 and 4 in AUDIT was less reliable and valid among the elderly. CONCLUSIONS While the AUDIT showed acceptable psychometric properties in the general population sample, it's performance was of less quality among the elderly respondents. Further psychometric assessments of the AUDIT in elderly populations are required before it is implemented more widely.