Alcohol and suicide—the Portuguese experience
OLE-J0RGEN SKOG,' ZELIA TEIXEIRA/ JOSE BARRIAS' & RUI MOREIRA'
^National Institute for Alcohol and Drug Research, Dannevigsveien W, 0463 Oslo, Norway & ^Centro Regional de Akoologia do Porto, Hospital de Magalhaes Lemos, Circunvalagao, 4100 Porto, Portugal
An analysis of regional (18regions) and temporal (1931-89) covariation of suicide rates and indicators of alcohol use and abuse in Portugal is reported. In the time series analysis, a positive relationship was found. An increase in per capita alcohol consumption of one litre is accompanied by a simultaneous increase in the male suicide rate of 1.9 per cent. This is comparable to what has been reported for France andDenmark, but considerably less than thai found in Nonvay, Sweden and Hungary. In the regional data, there was a substantial negative correlation between the variables. However, after controlling for religious and family integration, the latter correlation became small and insignificant but still negative. A possible explanation is suggested for the different outcomes of the two analyses.
IntroductionAt the itidividual level, a link between alcohol abuse and suicide has, for some time, been fairly well established, and is not seriously disputed today (Murphy, 1992), However, it has not always been widely recognized that this link also carries over to the aggregate level. For instance, Durkheim (1951(1897]) claimed that alcohol is not a relevant causal factor in relation to the social suiciderate, while recognizing that it was a factor of some importance at the individual level. For many years Durkheim's position was taken for granted, and alcohol was more or less ignored in aggregate level studies of suicide. There are several theoretical reasons for expecting an aggregate level relationship between alcohol and suicide (cf. Skog, 1991). First, acute
Correspondence to: Ole-JtargenSkog.
intoxication may interact with suicidal impulses caused by other factors. Ceteris paribus, the higher the aggregate consumption in a culture, the more frequently people are under the influence of alcohol, and hence the higher the risk that a latent suicidal impulse may be triggered. Secondly, it is well established that the higher the overall level of consumption in a culture, the larger thenumber of chronic abusers (Popham, 1970; Bruun era/., 1975; Skog, 1985). As chronic heavy drinking is known to increase the risk for depression and weakened social relations (Cahalan & Room, 1974; Murphy, 1992; Phillips, 1976; Vaillant, 1983), and since depression and weak social relations increase the risk for suicide, a high per capita consumption level ought to produce more suicides.Consequently, there are at least two causal mechanisms that might produce a positive relationship be© 1995, Society for the Study of Addiction to Alcohol and Other Drugs
Ole-j0rgen Skog et al. shall examine the negative regional correlation to see if this may be an anefact caused by other causal factors, or if it still holds after controlling for these factors.Secondly, we shall analyse the temporal covariation between aicohol consumption and suicide rates in Portugal, in order to see if the hypothesis of a positive relationship is also violated by temporal data.
tween per capita alcohol consumption and the social suicide rate. In recent years an increasing number of studies has confirmed that the overall level of alcohol consumption in a culture has animpact on the social suicide rate: the higher the per capita consumption level, the higher the suicide rate. Such results, based on time series analyses, have been reported from the United States (Wasserman, 1989), Sweden (Norstrom, 1988), Norway (Rossow, 1993), Hungary (Skog & Hlekes, 1993), Denmark (Skog, 1993), the former USSR (Wasserman, Vamik & Eklund, 1994) and France (Norstrom, forthcoming)....