Deconstructing the age-prevalence curve of alcohol dependence: why "maturing out" is only a small piece of the puzzle

Abstract

Epidemiological studies have consistently demonstrated that heavy alcohol use and alcohol dependence (AD) tend to increase in adolescence and emerging adulthood and then show a large decline in the late 20s, a phenomenon called maturing out. This decline has been explained as an effect of "role incompatibility" in which involvement in new roles and activities interferes with a heavy drinking lifestyle. However, maturing out has been conceived mostly as a decrease in offset, with little attention paid to reductions in new onset or recurrence across decades of life. Moreover, although role incompatibility processes have been studied with young samples, little is known about the effect of life transitions (e.g., marriage, parenthood, changes in employment status) on AD later in life and whether similar effects are observed. Using longitudinal data from the National Epidemiologic Survey on Alcohol and Related Conditions, a nationally representative epidemiologic survey, we examined the patterns of stability and change in AD across the life span and the differential effect of life transitions on AD across different age strata. Results showed that persistence of AD tended to increase with age, although not dramatically, and that onset and recurrence tended to decrease with age. Moreover, the effects of life transitions on the course of AD varied across the life span and were different for men and women. These results indicate that life transitions differentially affect the patterns of stability and change in younger versus older people, have a different impact for men and women, and highlight the need to consider the unique aspects of each stage of adult development on the course of AD.

Figure 1

Relative contribution of the three AD groups to the overall AD prevalence across age groups. The Y-axis represents the percentage of participants diagnosed with past-12-month AD at Wave 2. Age groups were created based on age reported at Wave 1, so that the X-axis shows the approximated age at Wave 2. n = total number of participants within each age group.

Figure 2

Prevalences of the three AD groups among participants at risk. Brackets represent 95% confidence intervals around the mean. The Y-axis represents the percentage of participants diagnosed with past-12-month AD at Wave 2. Age groups were created based on age reported at Wave 1, so that the X-axis shows the approximated age at Wave 2. In the upper panel (new onset) participants at risk are those with no lifetime history of AD at Wave 1. In the middle panel (persistence) participants at risk are those with past-12-month AD diagnosis at Wave 1. In the bottom panel (recurrence) participants at risk are those with prior-to-past-12-month, but no diagnosis in the past 12 months at Wave 1. n = number of participants at risk within each age group.

Figure 3

Three-way interactions among age, sex, and life transitions. Brackets represent standard errors. Age groups were created based on age reported at Wave 1, so that the X-axis shows the approximated age at Wave 2. Parenthood = Had any children or any adopted/foster/step children from Wave 1 to Wave 2; No Parenthood = no change in parenthood from Wave 1 to Wave 2; Lose Job = Became unemployed or retired from Wave 1 to Wave 2; Stable = No change in employment status from Wave 1 to Wave 2; Get Job = Became employed from Wave 1 to Wave 2.