7.2.1 Age and gender differentials
Two main indings of this study were that rates of alcohol-related mortality and morbidity increased considerably on the whole after the reduction in alcohol prices, and the effects of the reduction on these rates varied widely among the population subgroups.
Time-series analysis revealed an increase in alcohol-related mortality among both men and women aged 40 years or more, and virtually no change among younger people, whereas Study I showed a declining trend in alcohol-related mortality in the 1990s and early 2000s among the under-45s. On the other hand, alcohol-related hospitalisations increased by 16 per cent after the price reduction among men aged under 40. On the basis of both the mortality and hospitalisation studies it could be said that the mortality impact was highest, in both relative and absolute terms, among persons aged 50–69 years.
The inding that the level of alcohol-related mortality did not increase after the price reduction but, if anything, decreased among younger people is of great interest because alcohol consumption is generally more responsive to increases in price among youths and young adults than among older people (Chaloupka et al. 2002). The implication is that this is also true when the price goes down. The observed increase in alcohol-related hospitalisation among men aged under 40 does not change the fact that the price reduction mainly affected the over-50s.
The different developments according to age group would appear to be accounted for by tha varying changes in alcohol consumption: survey-based studies have shown an increase among the over-50s after the mid 1990s, and also after 2004, whereas it was stable or decreased among younger Finns (Helakorpi et al. 2007, 2009; Mäkelä et al. 2009). The increase was particularly large and harmful in terms of drinking patterns among the 50–69-year-olds between 2000 and 2008 (Mäkelä et al. 2009). These results suggest that younger generations may be adopting less damaging alcohol-consumption patterns, even if recent surveys do not fully conirm this (Mäkelä et al. 2009; Mustonen et al. 2009). Overall, the phenomenon (i.e. a decrease in alcohol consumption and alcohol-related harm among younger generations) rather seems to be indicative of a longer-term change than a consequence of the price reduction.
With regard to gender differences in the effects, the increase in alcohol-related mortality and hospitalisation, on the whole, was larger among men than among women, consistently in absolute terms, and in most cases also in relative terms.
The larger relative increase among women in some cases could be attributable in part to the consistently much lower rates before the price reduction. Study I conirms this pre-price-reduction trend. Gender differences in alcohol consump-tion and the related problems are universal and well-documented (see e.g., Plant 1990; Wilsnack et al. 2000). However, the large increases in both alcohol-related mortality and hospitalisation among women aged 50–69 years is indicative of an increase in alcohol abuse. Two recent survey-based studies conirm this: the proportion of heavy drinkers increased from 0.3 to 1.5 per cent (Mäkelä et al.
2009), and the proportion of binge-drinking occasions increased from three to ive per cent between 2000 and 2008 among women aged 50–69 years (Mustonen et al. 2009). Gender convergence in this respect has also been reported elsewhere (e.g., McPherson et al. 2004).
7.2.2 Socioeconomic differentials
There is ample evidence of socioeconomic differences in health (Mackenbach et al. 1997, 1999, 2008; Martikainen et al. 2007; Leinsalu et al. 2009; Stirbu et al.
2009). For example, a study based on data from 22 European countries reported substantially higher mortality rates in groups of a lower socioeconomic status, and it was only the magnitude of the inequalities between the groups that varied across countries (Mackenbach et al. 2008).
The effects of the price reduction on alcohol-related mortality were assessed in the present study according to four socioeconomic factors, which implies a broad deinition of socioeconomic status: on top of the three conventional factors of education, household income and social class, economic activity was also used. In addition, Study I examined educational differentials in trends of alcohol-related mortality prior to the price reduction.
The most salient inding was a huge gap between the employed and those without work among the 30–59-year-olds. There was a particularly strong increase among the long-term unemployed and early-age pensioners, which was mostly attribut-able to chronic causes, but deaths from acute causes also increased to some extent.
At least two relevant but interlaced explanations for this divergence between the employed and others could be offered. First, for reasons to do with selection and causal processes (Martikainen and Valkonen 1996, 1998), those drinking heavily before the price reduction are overrepresented among the unemployed and
early-age pensioners, and chronic conditions may more quickly respond to an increase in consumption in the form of an increase in deaths. Secondly, the unemployed and early-age pensioners include more poor individuals whose drinking may have previously been restricted by the higher price of alcohol. With regard to acute causes, the mortality rates of the long-term unemployed increased but remained almost unchanged among the employed. This could indicate that hazardous alcohol consumption has not increased substantially among employed individuals.
The inding that the increase in alcohol-related mortality after the price reduc-tion was, in absolute terms, higher in the lower educareduc-tional and socioeconomic groups is in accordance with the long-term trends observed in Study I and in two earlier studies conducted by Mäkelä and colleagues in the Finnish context (1997, 2000): the pre-existing differential trends were clearly more detrimental to people of a lower socioeconomic and educational status than in 1987–2003.
Moreover, as reported in Study I, in the past 20 years economic luctuations have not been associated with alcohol-related mortality among better educated men and women. In the lowest educational group, alcohol consumption and related mortality appeared to follow economic cycles, increasing during upturns and de-creasing during downturns. Differences in harmful alcohol consumption mainly explain these results. According to one survey both men and women with the lowest level of education increased their annual binge drinking occasions in the economic upturn of the 2000s, whereas the more highly educated kept it stable or even reduced it (Mäkelä et al. 2009). Furthermore, changes in alcohol-related mortality by socioeconomic indicators could be partly attributable to beverage preferences: the beverages consumed by the lower educated were affected by the price changes to a greater extent than those consumed by the more highly educated (Metso et al. 2002). Those with a lower education are more likely to consume spirits than the more highly educated, and spirits were most affected by the price reduction (Metso et al. 2002; Mäkelä and Österberg 2009).
Although household income was inversely associated with mortality before the price reduction, the effect was not the largest among the poorest. One might assume that if, on the whole, alcohol-related mortality duly increased, it would happen among poor people whose consumption may have been restricted the most by the higher price. The lack of a perfect gradient in change according to household income could relect the possibility that alcohol is still too expensive for the poorest. In any case, this peculiarity in the results diminished when eco-nomic activity was controlled for.