Per capita alcohol consumption and all-cause mortality in 14 European countries.

AIMS: (1) To estimate the relationship between per capita alcohol consumption and male all-cause mortality in 14 European countries. (2) To compare the estimates with predictions from the U-shaped curve at the aggregate level. DATA AND METHOD: The outcome measures comprised annual data, after 1950, on male mortality (all-cause mortality and mortality from diseases) for the following age groups: 15+, 15-29, 30-49, 50-69 and 70+ years. Female mortality was included as a control variable. Alcohol sales were used as proxy for per capita consumption. The data were analysed using the Box-Jenkins technique. The estimated alcohol effects were pooled within low-, medium- and high-consumption countries. RESULTS: For all-cause mortality (15+), the effect estimates were significantly positive in eight of the 14 countries. The effect on mortality of a 1-litre increase in consumption tended to be stronger in low-consumption countries (3% per litre) than in medium- and high-consumption countries (1%). This pattern deviates from that predicted from the U-shaped curve. No significant impact of alcohol was found in the youngest age group when mortality from diseases was used as the outcome. CONCLUSIONS: Increases in overall consumption seem to be associated with increases in total mortality. Differences in drinking patterns are discussed as a possible explanation for the variation between country groups in alcohol effect.

Alcohol and mortality: methodological and analytical issues in aggregate analyses.

This supplement includes a collection of papers that aim at estimating the relationship between per capita alcohol consumption and various forms of mortality, including mortality from liver cirrhosis, accidents, suicide, homicide, ischaemic heart disease, and total mortality. The papers apply a uniform methodological protocol, and they are all based on time series data covering the post-war period in the present EU countries and Norway. In this paper we discuss various methodological and analytical issues that are common to these papers. We argue that analysis of time series data is the most feasible approach for assessing the aggregate health consequences of changes in population drinking. We further discuss how aggregate data may also be useful for judging the plausibility of individual-level relationships, particularly those prone to be confounded by selection effects. The aggregation of linear and curvilinear risk curves is treated as well as various methods for dealing with the time-lag problem. With regard to estimation techniques we find country specific analyses preferable to pooled cross-sectional/time series models since the latter incorporate the dubious element of geographical co-variation, and conceal potentially interesting variations in alcohol effects. The approach taken in the papers at hand is instead to pool the country specific results into three groups of countries that represent different drinking cultures; traditional wine countries of southern Europe, beer countries of central Europe and the British Isles and spirits countries of northern Europe. The findings of the papers reinforce the central tenet of the public health perspective that overall consumption is an important determinant of alcohol-related harm rates. However, there is a variation across country groups in alcohol effects, particularly those on violent deaths, that indicates the potential importance of drinking patterns. There is no support for the notion that increases in per capita consumption have any cardioprotective effects at the population level.

Outlet density and criminal violence in Norway, 1960-1995.

OBJECTIVE: To analyze the relationship between criminal violence and number of outlets for on-premise sales of alcohol. METHOD: The data comprise aggregate time series for Norway for the period 1960-95. Two crime indicators were used: (1) crimes of violence investigated by the police per 100,000 inhabitants aged 15 years and above (15+) and (2) convictions for criminal violence per 100,000 inhabitants (15+). Outlet density was measured as number of public drinking places per 10,000 inhabitants (15+). RESULTS: On the basis of autoregressive integrated moving average (ARIMA) analyses of differenced data, a positive and statistically significant relationship (p = .03) was found between outlet density and crimes of violence investigated by the police. This replicates the findings that are reported from cross-sectional studies. The relationship was also positive, but of borderline significance (p = .06), when convictions for criminal violence were used as outcome. CONCLUSIONS: The findings suggest that increases in the number of public drinking places are associated with increases in criminal violence. Although other kinds of data support the findings, these are the first to be based on longitudinal data. To test for robustness and cultural specificity it, therefore, seems warranted to replicate the study on data from other drinking cultures.

Effects on criminal violence of different beverage types and private and public drinking.

AIM: The purpose of the study is to analyse the relationship between homicide and assault rates on one hand, and various indicators of alcohol consumption on the other. The latter include private and public drinking with a further disaggregation into beverage-specific drinking. MEASUREMENTS: The data comprise aggregate time series for Sweden during the period 1956-94. The assault rate is measured as the number of police reported assaults (at all degrees of aggravation) per 100,000 inhabitants (15+). The homicide rate is measured as the number of homicides (where the victim was at least one year old) per 100,000 inhabitants (15+). Private consumption is gauged as retail sales of alcohol, and public consumption as on -premise sales (litres 100% per inhabitant, 15+). These two measures are disaggregated further into beverage specific sales (beer, spirits and wine). FINDINGS: According to the findings, there is a statistically significant relationship between the assault rate and a combined measure of on-premise sales of beer and spirits. The estimated relationship corresponds to an attributable fraction of about 40%. The homicide rate is significantly associated with retail sales of spirits; the attributable fraction is estimated at about 50%. Wine sales are not related to any of the two violence indicators. CONCLUSIONS: The findings suggest that the assault rate is related to consumption of beer and spirits in bars and restaurants, while the homicide rate is linked to consumption of spirits in private contexts. The findings, notably specific to Sweden during a certain time period, can be interpreted as the outcome of the interplay of a number of factors, including opportunity structure, social control and context of drinking, drinking patterns associated with the different beverage types and characteristics of the drinkers.

Estimating changes in unrecorded alcohol consumption in Norway using indicators of harm.

AIM: To assess the value of using indicators of alcohol-related harm to estimate changes in unrecorded per capita consumption of alcohol. DESIGN: Unrecorded consumption was estimated from the discrepancy between the observed changes in a number of alcohol-related harm indicators and the changes that would be expected from changes in recorded consumption. The results were compared with estimates of unrecorded consumption from survey data. MEASUREMENTS: Four indicators of alcohol-related harm were used: alcohol-related mortality, assaults, drunken driving, and suicide. Estimates of unrecorded consumption from survey data for five different years were used as benchmarks. FINDINGS: The best performing indicators were alcohol-related mortality, suicide and assaults, in that order. Combining these indicators yielded a prediction error averaging 12% in comparison with the benchmarks. CONCLUSIONS: The method seems worthy of further applications, but it should be regarded as a supplement rather than as a substitute for other approaches.

Social class, drinking and alcohol-related mortality.

The main purpose of this study is to investigate the magnitude of class differences in the distribution of alcohol consumption in Sweden, and to assess whether this is compatible with the observed pattern in class-specific mortality. The emphasis is put on analyses of survey data from 1990, but we also provide an overview of earlier published data. The findings suggest that, over the past 20-30 years, there has been an equalization, or even reversal, between social strata in the average consumption of alcohol, such that the consumption level for manual laborers is now at parity with, or possibly exceeds, the level for non-manual employees. This shift may also have been accompanied by a shift in the dispersion: the 1990 survey data indicate that the consumption distribution for the manual laborers is more skewed than that for the non-manual employees, while no such tendency was detected in the early data. Calculations demonstrated that this difference in dispersion may well account for the elevated rate of alcohol-related mortality among manual laborers. The results are finally discussed as a case of deviation from the general rule of collective change in consumption.

Per capita alcohol consumption and total mortality: an analysis of historical data.

Since total mortality is a classical proxy for the overall health status of the population, its degree of association with per capita alcohol consumption is of great interest. Existing evidence, based on historical data from the turn of the century, is mostly in graphical form. These data are analysed using modern statistical tools. The results suggest a significant alcohol effect; a 1-l increase in consumption is expected to increase mortality (middle-aged men) by about 1%. This effect might seem fairly modest but it is noted that it may well be locally substantial because of its concentration to specific categories of the population. The alcohol effect is also compared with the impact of a factor that is a surrogate for a large number of etiological agents, namely real wages. Although the latter factor seems to be the more important one the difference is not overwhelming. The shift in the cause of the death panorama during this century, with an increasing share of CHD-mortality, may well have attenuated the aggregate relationship between alcohol and mortality.

Potential consequences from possible changes to Nordic retail alcohol monopolies resulting from European Union membership.

This paper projects the consequences of modifying or eliminating the current national alcohol retail monopolies in Sweden, Norway and Finland as a possible result of those countries' membership in the European Union (EU). First, the authors project absolute alcohol consumption in each country based on different possible changes in alcohol price and availability. Then they predict the future levels of alcohol-related problems likely to result from increased per capita alcohol consumption (Sweden and Norway only). All of the scenarios examined in this paper are expected to lead to increases in per capita alcohol consumption. The smallest increase in consumption would result from a partial elimination of the current monopoly and a modest reduction in alcohol prices. In that case, projected per capita consumption in Sweden for inhabitants 15 years and older would rise from 6.3 to 9.3 litres; in Norway, from 4.7 to 6.7 litres; and in Finland, from 8.4 to 11.1 litres. The greatest projected increase in consumption would result from a complete elimination of the state monopolies such that all beer, wine and spirits were sold in food shops, grocery stores and gasoline stations, along with a substantial drop in alcohol prices as a result of private competition within each country and increased cross-border alcohol purchases. That scenario would result in projected per capita consumption of 12.7 litres in Sweden, 11.1 litres in Norway and 13.7 litres in Finland. The authors project that a 1-litre increase in consumption would result in a 9.5% increase in total alcohol-related mortality in Sweden and a 9.7% increase in Norway. Further, alcohol-related assaults would increase by 9% in Sweden and 9.6% in Norway. A 5-litre increase in consumption would result in a 62% increase in alcohol-related mortality in Sweden and a 60% increase in Norway, and a 57% increase in alcohol-involved assaults in both countries.

Alcohol and suicide: a comparative analysis of France and Sweden.

The aim of this paper is to test the hypothesis that the effect of alcohol consumption on suicide risk is stronger in Sweden than in France. The rationale of the hypothesis is that we should expect a difference between the two countries with respect to: (i) the composition of alcohol abusers; and (ii) the degree to which heavy drinking is accepted. Analyses of time-series data provide support for the hypothesis: the effect of per capita consumption of alcohol is significantly stronger in Sweden than in France.

Prevention strategies and alcohol policy.

This paper addresses the potentials of high risk strategy and population strategy for preventing mortality from liver cirrhosis, accidents and suicide. The methodological approach is based on an integration of an empirical distribution of consumption and the risk functions of the damages at issue. According to the findings, the rate of alcohol-induced mortality from the three causes combined would be halved either by a 25 per cent decrease in overall consumption or a 36 per cent decrease in the alcohol consumption of the heavy drinkers (the top 5 per cent). The high risk strategy is most efficient in preventing cirrhosis; however the effect of the population strategy is also quite substantial here. The comparative advantage of the population strategy is most marked in connection with accidents and suicide, but the high risk approach yields an appreciable impact in this context as well. That is, neither of the two strategies appears as clearly superior to the other in terms of efficiency. The high risk strategy thus seems to be a sensible complement to the population strategy, and should have the potential of yielding effects on the population level if implemented on a large scale.

The utilization of home help in Sweden: a multivariate analysis.

We estimated the effect of a variety of factors on the utilization of home help among the elderly in Sweden using a sample (N = 421) of people 75 years and older. A regression analysis singled out two factors as the most important: physical functioning (instrumental activities of daily living) and living alone. Calculations of attributable fractions substantiated this result. Implications for planning and allocating home help are discussed.

The structure of IADL and ADL measures: some findings from a Swedish study.

The purpose of this paper is to uncover the dimensional structure of a set of disability items. The analysis is based on a sample of 421 persons aged 75 years and older. Firstly, a factor analysis of the items reveals two distinct dimensions: IADL and ADL. The items with large loadings on the IADL-factor are next submitted to a Guttman analysis. An index is derived which strictly meets the requirements of a Guttman scale, but it is doubtful whether there is a true hierarchical structure among the items. The criterion validity of the index (in different versions) appears to be satisfactory when assessed in terms of the correlation with an outcome variable (home help).

The use of aggregate data in alcohol epidemiology.

Analyses of aggregate data have the potential of elucidating pivotal issues in alcohol epidemiology, in particular those singled out by the public health perspective, such as the impact of per capita alcohol consumption on various damage rates. This article discusses the possibilities and limitations of the two main approaches in the analysis of aggregate data, focusing upon ecological and time series data, respectively. In addition, a synthetic approach is outlined, which aims at an integration of macro and micro findings. In that context it is shown how effect measures which are normally based on micro data (relative risk and attributable fraction) can be derived from macro data.

Deriving relative risks from aggregate data. 1. Theory.

Sociological macro analyses of the association between risk factors and mortality can be seen as a valuable supplement to epidemiological micro studies. However, sociologists and epidemiologists typically employ different measures of association and this hampers strict comparisons of findings. This study presents a synthetic approach relying on both micro and macro data. In Part 1, the mathematical relations between the relative risk and the attributable fraction on the one hand, and the regression coefficient on the other are derived in order to make cross level comparisons possible. Part 2 provides an empirical illustration of the approach.

Deriving relative risks from aggregate data. 2. An application to the relationship between unemployment and suicide.

In Part 1 of this study, the mathematical relations between micro and macro measures of effect were derived. The formulas thus obtained can be used for cross level comparisons of findings concerning the effect of some specific risk factor on, eg, mortality. The approach is illustrated by means of an empirical example relating to the association between unemployment and suicide. This relationship is estimated on micro data as well as on aggregate time series data. The findings from the two levels are fairly consistent.

Real wages, alcohol consumption and mortality in Sweden, 1861-1913.

Factors affecting the decline in mortality among Swedish men that occurred during the latter part of the nineteenth century are analyzed. In addition to the impact on mortality of changes in real wages, the author develops the hypothesis that a decline in alcohol consumption was a contributing factor. These relationships are explored using time-series analyses on official data for the period 1861-1913. The results indicate that "real wages as well as alcohol consumption per capita had a statistically significant effect on male mortality. When the alcohol predictor was omitted in the model, the wage effect did not reach statistical significance. The decline in male mortality during the study period was, on the average, about one per cent per year. According to the results, a good half of this decrease is attributable to the trends in real wages and alcohol consumption." (SUMMARY IN FRE)