Are the 1976-1985 birth cohorts heavier drinkers? Age-period-cohort analyses of the National Alcohol Surveys 1979-2010

Abstract

Aims: To estimate age-period-cohort models predicting alcohol volume, heavy drinking and beverage-specific alcohol volume in order to evaluate whether the 1976-1985 birth cohorts drink relatively heavily.

Design: Data from seven cross-sectional surveys of the USA conducted between 1979 and 2010 were utilized in negative binomial generalized linear models of age, period and cohort effects predicting alcohol measures.

Setting: General population surveys of the USA.

Participants: Thirty-six thousand four hundred and thirty-two US adults (aged 18 years or older).

Measurements: Monthly number of alcohol drinks, beer, wine and spirits drinks, and days drinking five or more drinks in the past year derived from beverage-specific graduated frequency questions.

Findings: Relative to the reference 1956-60 birth cohort, men in the 1976-1980 cohort for were found to consume more alcohol [incidence rate ratio (IRR) = 1.222: confidence interval (CI) 1.07-1.39) and to have more 5+ days (the number of days having five or more drinks) (IRR = 1.365: CI 1.09-1.71) as were men in the 1980-85 cohort for volume (IRR = 1.284: CI 1.10-1.50) and 5+ days (IRR = 1.437: CI 1.09-1.89). For women, those in the 1980-85 cohort were found to have higher alcohol volume (IRR = 1.299: CI 1.07-1.58) and more 5+ days (IRR = 1.547: CI 1.01-2.36). Beverage-specific models found different age patterns of volume by beverage with a flat age pattern for both genders' spirits and women's wine, an increasing age pattern for men's wine and a declining age pattern from those in their early 20s for beer.

Conclusions: In the USA, men born between 1976 and 1985, and women born between 1981 and 1985 have higher alcohol consumption than in earlier or later years.

Figure 1

Estimated age, period and cohort effects on the volume of beer displayed as incidence rate ratios (IRR) from negative binomial generalized linear models controlling for individual socio-demographic characteristics. Reference groups are the 41–50 age group, the 2010 survey and the 1956–1960 birth cohort. Dotted lines represent 95% confidence intervals and large diamonds indicate a significant difference from the reference group.

Figure 2

Estimated age, period and cohort effects on the volume of spirits displayed as incidence rate ratios (IRR) from negative binomial generalized linear models controlling for individual socio-demographic characteristics. Reference groups are the 41–50 age group, the 2010 survey and the 1956–1960 birth cohort. Dotted lines represent 95% confidence intervals and large diamonds indicate a significant difference from the reference group.

Figure 3

Estimated age, period and cohort effects on the volume of wine displayed as incidence rate ratios (IRR) from negative binomial generalized linear models controlling for individual socio-demographic characteristics. Reference groups are the 41–50 age group, the 2010 survey and the 1956–1960 birth cohort. Dotted lines represent 95% confidence intervals and large diamonds indicate a significant difference from the reference group.