Measuring harm from drinking in Sweden : Self-reports from drinkers in the general population

There are several traditions of conceptualizing and measuring harm from drinking. Two main approaches are considered in the introduction – the psychiatric epidemiology and the social survey research traditions. The present thesis adopts the latter, although, as discussed, on the empirical level there is no sharp borderline between the two perspectives, as there is not between personal (i.e. physical and psychological) and social harm from drinking as such. But while methods for studying personal harm are fairly well developed, social harm, i.e. adverse consequences of alcohol that involve social interaction, has received less attention. One of the aims of the thesis has been to explore different dimensions of harm from drinking, identify different harm areas and develop and apply area-specific measures using general population survey data. Two papers examine psychometric properties of a widely used screening instrument, the AUDIT (Alcohol Use Disorder Identification Test). While the first paper is concerned with the validity of the AUDIT, testing it against different criteria, the second paper focuses on the test-retest reliability of the instrument. In the third paper, a number of summary measures of different areas of alcohol-related harm are constructed using factor analysis. These measures are then, in the last paper, applied to estimate prevalence and risk of alcohol-related harm in the Swedish general population. The analyses are based on data from a national survey on drinking problems in Sweden collected in 2001-2002. It is concluded that the AUDIT screens well for both impaired self-control and social harm from drinking (as well as for high volume drinking), but performs less well when screening for health problems. The test-retest reliability of the AUDIT is relatively high. In the other two papers, constructing new summary measures and applying them to estimate prevalence of harm, it is concluded that, except for being young, no particular sociodemographic risk groups can be identified for different areas of harm from the same level and pattern of drinking. To what extent this reflects reality or is an effect of the methods we use and kind of population we reach in surveys is discussed