|
Article Excerpt
HEAVY DRINKERS AND ALCOHOLICS are more likely to die of accidental and violent causes than the general population (Mann et al., 1993; Pell and D'Alonzo, 1973; Schmidt and deLint, 1972). Many studies have demonstrated an important relationship between alcohol and suicide. At the individual level, suicide has been associated with alcohol misuse and dependence. Death rates from suicide and rates of suicide attempts have been shown to be substantially elevated in samples of heavy drinkers and alcoholics (e.g., Schmidt and deLint, 1972). Examinations of aggregate or population rates of suicide mortality show a similar relationship with aggregate or per capita alcohol consumption (e.g., Mann et al., 2001; Norstrom, 1988a; Ramstedt, 2001). This observation is consistent with a large body of evidence on the link between population alcohol consumption levels and mortality rates from a large number of causes (Babor et al., 2003; Chaloupka et al., 2002; Mann et al., 2003; Rush et al., 1986).
Other studies suggest that the relationship between alcohol consumption and mortality rates may be affected by beverage type and gender. In examining suicide rates in the United States, Gruenewald et al. (1995) found that these rates were most strongly related to distilled spirits consumption. Similarly, Kerr et al. (2000) reported that cirrhosis mortality rates were most strongly related to distilled spirits consumption in five Western countries. Other research suggests additional factors may be important influences on the alcohol-suicide relationship at the aggregate level.
Previous research has demonstrated that suicide mortality rates are also linked with unemployment (e.g., Dooley et al., 1996; Voss et al., 2004). Increased risk of suicide is found among the unemployed at all social levels, and risk increases with increasing length of unemployment (Voss et al., 2004). Caces and Harford (1998) examined both unemployment and alcohol consumption in relation to suicide mortality in the United States for the period between 1934 and 1987; both measures were significantly associated with suicide mortality rates. In addition, unemployment influenced the magnitude of the relationship between alcohol consumption and suicide.
The possibility that population-level alcohol problem rates have been influenced by treatment for alcohol misuse and related activities such as Alcoholics Anonymous (AA) membership has been suggested (Holder and Parker, 1992; Mann et al., 1988; Romelsjo, 1987; Smart et al., 1996). Individual-level studies have found that treatment for alcohol problems may reduce mortality rates (Cuijpers et al., 2004; Kristensen et al., 1983; Mann et al., 1994). Several studies have observed that increases in AA membership and alcoholism treatment rates are negatively associated with cirrhosis mortality rates (Mann et al., 1988, 1991; Romelsjo, 1987; Smart et al., 1993). Recently, Mann et al. (2005) utilized time series analyses to assess the effects of per capita alcohol consumption and AA membership rates on mortality rates from liver cirrhosis in Ontario between 1968 and 1989. They found that as per capita consumption increased, cirrhosis mortality increased, and as AA membership increased, cirrhosis mortality decreased.
It has been proposed that these negative associations of cirrhosis mortality rates with alcoholism treatment and AA membership are due to the beneficial effects of these activities on the individual, reflected at the aggregate or population level (Mann et al., 1988, 1991, 2003; Smart and Mann, 2000). Although this is a particularly interesting hypothesis, the inference of individual-level causes for aggregate-level effects must be made with caution (Schwartz, 1994). In the case of cirrhosis mortality rates, one possibility is that the negative associations with treatment and AA may reflect a biological cohort effect (Gustaven, 1996; Halliday et al., 1991; Kerr et al., 2000), rather than the cumulative beneficial effects of treatment and AA. If a particular birth cohort is more or less likely to develop cirrhosis for some reason, as this group reaches the age when cirrhosis occurs they may have a disproportionate effect on cirrhosis mortality rates, increasing or decreasing those rates accordingly. These increases or decreases in mortality rates could then erroneously be attributed to other factors. For example, Halliday et al. (1991) observed that succeeding generations in Ontario, Canada, were developing cirrhosis at successively younger ages; they speculated that this trend may have been related to a reduction in competing risks of dying resulting from the decline in mortality from childhood infectious diseases, as well as to an increase over time in the hepatitis B infection rate. One way to test this possibility would be to examine the association of AA membership rates with deaths from other alcohol-related causes--and, in particular, causes such as suicide that are more strongly related to acute alcohol effects. It is difficult to imagine a specific biological influence on a birth cohort that would increase susceptibility to both cirrhosis of the liver and suicide. A significant negative relationship between suicide mortality rates and AA membership rates would support the argument that the cumulative beneficial effects of AA on individuals are being translated into aggregate-level trends.
In the present study, we use time series methodology to assess the association of total alcohol consumption, as well as the consumption of distilled spirits, wine, and beer, with suicide mortality rates in Ontario between 1968 and 1991. Based on previous research we expect that suicide mortality rates will be significantly associated with total alcohol consumption, and with distilled spirits consumption specifically. We also assess the impact of AA membership rates, to test the prediction that these rates will be negatively related to suicide mortality rates. In addition, we examine the impact of unemployment rate, which we predict will be positively related...
|
