[Construction and validation of an indicator for alcohol problems in the Swiss Health Survey].

A proposed scale for alcohol problems which consisted of five items, was psychometrically examined and validated. Latent class analysis was used in addition to classic psychometric measures. The use of this scale in health surveys of the general population seems to be justified, as it offers a reliable and very economic method to measure alcohol induced problems. It is therefore proposed that this scale should continue to be used in future rounds of the Swiss Health Survey. The context of the questioning has to be taken into account when interpreting the summary score, as items showed different patterns of response in personal vs interview questionnaires.