Palm eBook

Alcohol & Alcoholism, 32, 51-64, 1997.

Fellow, The Lindesmith Center, New York

Abstract

Cultural differences in alcohol consumption are inescapable, but have been difficult to establish as predictor variables in epidemiological models. With respect to dependent variables, the behavioural outcomes of alcohol use have not been operationalized as successfully as the health outcomes. This study examined cultural differences in drinking by employing Levine's distinction between Temperance and non-Temperance cultures, along with other cultural, consumption, and policy predictor variables, among 21 Western countries. Dependent variables included the prevalence of Alcoholics Anonymous (AA) groups (as a measure of behavioural and social problems) and a range of alcohol consumption and health measures. Level of consumption was an important determinant of the health consequences of drinking among Western nations, but not so important in determining behavioural outcomes. Culture, on the other hand, is largely determinative of behavioral outcomes and also quite critical for some health outcomes. An inverse relationship between alcohol consumption and AA membership strongly indicated that consumption is modified by cultural styles in producing drinking behaviours. Temperance cultures, which are largely Protestant, have far more AA groups and higher rates of coronary heart disease mortality, but lower cirrhosis mortality. Overall mortality does not vary according to national alcohol consumption or cultural distinctions. The percentage of alcohol consumed as wine is a strong inverse predictor of mortality in the 55-64 age group, but the change in absolute national wine consumption is directly associated with overall all-age mortality. In conclusion, religious and cultural distinctions among Western nations strongly predict behavioural drinking problems and also enhance the prediction of death rates from diseases related to alcohol consumption. Social engineering techniques which attempt to modify well-established cultural drinking practices can have counterproductive results.

Introduction

This study represents an attempt to operationalize cultural factors in the epidemiology of alcohol use and to measure the impact of such cultural variations on an expanded range of behavioral as well as health outcomes. The need for an epidemiological model that (a) gives proper weight to culture as a predictor variable, and (b) considers behavioural as well as health consequences, arises from both the successes and the limitations of recent research which relates alcohol consumption to disease incidence and mortality.

Both case comparison and cohort epidemiologic research, as well as cross-cultural analysis, have now firmly established that alcohol reduces coronary heart disease (CHD) incidence and mortality (Criqui and Ringel 1994; Gaziano et al. 1993; Klatsky et al. 1992; Rimm et al. 1991; Stampfer et al. 1988; Suh et al. 1992). Prospective epidemiological studies also find that overall mortality is reduced by moderate alcohol consumption (Boffetta and Garfinkel 1990; Doll et al. 1994; Fuchs et al. 1995; Grnbæk 1994; Klatsky 1992). These benefits occur primarily for middle-aged men and women, for whom heart disease is the primary cause of death. However, they also apply to all adults at risk for heart disease, a substantial majority of both female and male adults (Fuchs et al. 1995).

Impressive as these findings are, to have real policy implications they must be considered in the light of an older body of research on cross-cultural differences in drinking styles. Some cultures, notably Mediterranean and other wine-drinking societies, unquestionably socialize the use of alcohol more effectively and display fewer behavioural drinking problems than other cultures (Blum and Blum 1969; Lolli et al. 1958; Maloff et al. 1982). These differences are reflected in epidemiological and community studies through lower rates of drinking problems and alcoholism (Cahalan and Room 1974; Glassner and Berg 1980; Greeley et al. 1980). Vaillant (1983), for example, found that Irish-Americans in an urban Boston setting were seven times as likely to become alcohol dependent over their lifetimes as were Mediterranean-Americans (Italian, Greek, and Jewish). Yet, quantitative cross-cultural research has rarely found systematic differences in drinking behavior (see Whitehead and Harvey 1974).

The question that remains is how culture, as a predictor variable, fits into an overall epidemiological model. To define the relevant cultural differences solely by the consumption patterns (i.e., amount and style) that different cultures have developed historically is tautological. It begs the question of whether the health and behavioural consequences of alcohol use in a given culture are attributable primarily to such consumption patterns or to some larger cultural gestalt. This is a crucial determination to make at the policy level, since it tells us whether it is possible to replace one culture's consumption patterns with patterns that have been more successful in other cultures.

Predictor Variables

The current study employs the cross-cultural epidemiologic model which Criqui and Ringel (1994) used to analyze the impact of alcohol consumption on heart disease and overall mortality independent of diet. The primary predictors examined are consumption and culture. Two variables are utilized to capture the overall character of a culture. One is religion. The other is Levine's (1992) concept of Temperance versus non-Temperance cultures. In regression analyses, 1990 consumption and Temperance were analyzed first; other consumption data and/or policies such as taxation were then considered to see if they provided a better fit.

Temperance cultures are strongly concerned with alcohol abuse and maintain activist approaches to combating drinking problems, because ostentatious behavioural drinking problems are more apparent in these societies than in non-Temperance cultures (Levine 1992; Peele 1993). Temperance countries consume less alcohol, and a smaller proportion of their beverage alcohol as wine, than non-Temperance cultures (Levine 1992). Temperance cultures also have significantly higher rates of CHD (Peele 1993), due to a strong inverse association between societal alcohol consumption and CHD (LaPorte et al. 1980). However, the increased incidence of accidents, cirrhosis, and cancer resulting from higher levels of alcohol consumption counterbalance the cardiovascular benefits of drinking in terms of overall mortality (Boffetta and Garfinkel 1990; Klatsky et al. 1992), which may also neutralize the benefits of higher alcohol consumption cross-nationally (Criqui and Ringel 1994).