Paternal and Maternal Problem Drinking and Lifetime Problem Drinking of Their Adult Children

Parents’ alcohol use is associated with alcohol use of their adolescent offspring, but does this association extend to the adulthood of the offspring? We examined associations of paternal and maternal problem drinking with lifetime problem drinking of their adult offspring prospectively assessed in a population-based Finnish twin-family cohort (FinnTwin16). Problem drinking (Malmö-modified Michigan Alcoholism Screening Test) was self-reported separately by mothers and fathers when their children were 16. The children reported on an extended lifetime version of the same measure during their mid-twenties (21–28 years) and mid-thirties (31–37 years). 1235 sons and 1461 daughters in mid-twenties and 991 sons and 1278 daughters in mid-thirties had complete data. Correlations between fathers’ and their adult children’s problem drinking ranged from .12 to .18. For mothers and their adult children, these correlations ranged from .09 to .14. In multivariate models, adjustment for potential confounders had little effect on the observed associations. In this study, parental problem drinking was modestly associated with lifetime problem drinking of their adult children. This association could be detected even when the children had reached the fourth decade of life.

Supplementary Methods.Supplementary information on measures of problem drinking, heavy drinking occasions and covariates.
Problem drinking  For both measures of problem drinking, Mm-MAST and Mm-MAST-11, we included all responses with no more than two missing items, and substituted for the missing items the mean score of the available items of each included respondent.Substitutions were done for 6% of fathers and mothers and 1% of their adult children.Further, in wave 1, respondent parents were instructed to skip the entire scale if they did not drink at all.Therefore, fathers and mothers received an Mm-MAST score of zero if all items were missing and the respondent did not drink alcohol during the past year, but did not selfattribute lifetime abstinence.Fathers and mothers, as well as their sons and daughters, were excluded from the analyses if they were lifetime abstainers.Among all who returned questionnaires, 79% of fathers had answered all 9 items, 85% had at most two missing items, 5% did not drink during the past year and 7% were lifetime abstainers; similarly, 73% of mothers had answered all 9 items, 79% had at most two missing items, 7% did not drink during the past year and 13% were lifetime abstainers.At mid-twenties and mid-thirties, 98% and 94% of sons had answered all 11 items, 99% and 95% had at most two missing items and 4% and 2% were lifetime abstainers, respectively.For daughters at mid-twenties and mid-thirties, corresponding percentages were 98% and 95%, 99% and 95%, and 4% and 3%, respectively.

Fathers' and mothers' religiosity
Among those who returned the questionnaire, all 12 items of the Religious fundamentalism content scale of the Minnesota Multiphasic Personality Inventory (MMPI) were answered by 74% of fathers and 73% of mothers, and 91% of fathers and 93% of mothers had answered at least 9 items.We included those answering at least 9 items, and mean score of the available items of each particular individual was substituted for missing items.

Fathers' and mothers' personality
Among those who returned the questionnaire, all 50 items of the social deviance scale (Pd or "Psychopathic deviate" scale of the MMPI) were answered by 74% of fathers and 73% of mothers, and 96% of fathers and 97% of mothers had answered at least 40 items.Therefore, we included those answering at least 40 items, and mean score of the available items of each particular individual was substituted for missing items.

Table 1 .
Background characteristics of the participants by inclusion to versus exclusion from the mid-twenties main analysis Data are mean (standard deviation) or N (%).

Table 2 .
Background characteristics of the participants by inclusion to versus exclusion from the mid-thirties main analysis Data are mean (standard deviation) or N (%).

Table 5 .
Correlations between measures of problem drinking and heavy drinking occasions when the analysis was restricted to those who were living with both biological parents at age 16. *p<0.01,**p<0.001.aPearson'schi-squared test indicates violation of the underlying bivariate normality assumption.Spearman's rank order correlation coefficients (rho) given for comparison in parentheses for those polychoric correlations for which there was evidence for violation of the underlying bivariate normality assumption.Lifetime abstainers were excluded from the analysis.Abbreviations: Mm-MAST, Malmö-modified Michigan Alcoholism Screening Test (original 9-item version); Mm-.