Alcohol and Alcoholism Advance Access published online on October 12, 2007
Alcohol and Alcoholism, doi:10.1093/alcalc/agm146
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Typologies of alcohol consumption in adolescence: predictors and adult outcomes
Department of Epidemiology and Public Health, University College London, UK
* Author to whom correspondence should be addressed at: Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 6BT, UK. Tel: +44(0) 207 6791709; Fax: +44(0) 207 8130280; E-mail: n.cable{at}ucl.ac.uk
Received 17 April 2007;
first review notified 30 July 2007; in revised form 10 September 2007;
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionMethodsFindingsDiscussionReferences |
|---|
Aims: Data from the 1970 British Cohort Study were used to examine the effects of alcohol expectancies, norms, and openness of communication with parents on typologies of adolescent alcohol use and the subsequent risk of adult alcohol misuse from adolescent use. Methods: Of a population originally defined as all children born in the UK in 1 week of April 1970, 69.4% were interviewed at age 16 and 70.1% at age 30. Missing information was imputed using the multivariate imputation by chained equation (MICE) method, yielding a sample size of 7023 for men and 6896 for women. Four adolescent drinking typologies were defined by frequency and quantity of alcohol consumption at age 16. Results: Positive alcohol expectancies predicted all types of adolescent alcohol use in young men and women. Norms affected frequency of alcohol use over quantity, while openness of communication with parents affected quantity of alcohol use. All men who drank alcohol in adolescence were at risk of alcohol misuse (defined by the CAGE questionnaire) in adulthood, whereas the risk for women was limited to frequent drinkers. Conclusions: Drinking typologies were useful for understanding the mechanisms of adolescent alcohol use. Early prevention may be required to reduce alcohol related problems in later life.
|
Introduction |
|---|
|
TOPABSTRACTIntroductionMethodsFindingsDiscussionReferences |
|---|
More youth in England are exposed to alcohol and are consuming greater quantities of alcohol (Hewton and Moody, 2004
), causing social problems from binge drinking (Prime Minister's Strategy Unit, 2004). A greater risk for developing adult alcohol misuse among today's youth is envisaged (Pitkänen et al., 2005; Hingson et al., 2006). Alcohol misuse is a financial burden to the health care system in England, costing approximately £ 1 billion (Prime Minister's Strategy Unit, 2003). Thus, studies to identify factors associated with youth drinking, including binge drinking, and to estimate the anticipated risk of adult alcohol misuse from adolescent drinking are needed to provide evidence for effective preventive measures. The aims of this study are to identify precursors of various adolescent drinking patterns and to estimtate the longitudinal risks from these adolescent drinking patterns for adult alcohol misuse.
It has been suggested that two dimensions of alcohol use, frequency and quantity, be studied jointly (Oei and Morawska, 2004;Vogel-Sprott, 1974). Drinking patterns that combine quantity and frequency are better predictors of the social consequences of alcohol related problems (Rehm and Gmel, 1999) and have strong associations with risk factors (San José et al., 2000; Rehm et al., 2003). However, previous studies of adolescent alcohol use have tended to place greater emphasis on frequency (Callas et al., 2004; Helstrom et al., 2004; Li et al., 2002; Williams et al., 2003) over quantity of alcohol consumption. On the other hand, measures of binge drinking are largely quantity based (Chassin et al., 2002; Griffin et al., 2004; Guilamo-Ramos et al., 2005).
Measurements defined by different dimensions of alcohol use can compromise consolidation of the findings from studies of adolescent drinking because frequency of alcohol use is thought to be related to socially related factors, whereas quantity of alcohol use is believed to be associated with individual control (Vogel-Sprott, 1974). Previous research findings indicate differences between studies in the relationship between risk factors and alcohol use or binge drinking. Factors associated with frequency of alcohol use appear to reflect the social context of alcohol use, such as adolescents' externalizing problems (Helstrom et al., 2004), somatic symptoms and parental alcohol use (Williams et al., 2003), peer use (Li et al., 2002) and parental norms (Callas et al., 2004). In contrast, binge drinking seems to be related to intra- and extra-individual controls such as parental communication, control and supervision (Guilamo-Ramos et al., 2005), and low aspirations (Griffin et al., 2004).
Moreover, the ‘prevention paradox’, (that the larger number of problems from drinking in a society arises in the population of drinkers whose weekly total is ‘moderate’ than in the population of very heavy drinkers, because ‘moderate’ drinkers are very much more numerous), suggests the value of applying drinking typologies rather than the simultaneous use of frequency and quantity of alcohol consumption (Gmel et al., 2001; Rossow and Romelsjo, 2006; Weitzman and Nelson, 2004). However, there has been little consistency in the definition of typologies of adolescent drinking in the few studies that have employed them.
In addition to creating adolescent drinking typologies from frequency and quantity of alcohol usage, some studies have added other factors such as the number of heavy drinking episodes and the number of alcohol related problems in adolescence (Windle, 1996); or the largest amount of alcohol consumption on a single occasion and the number of alcohol related problems (Wells et al., 2004). Due to various individual physical responses to alcohol usage (Murgraff et al., 1999), inclusion of variables derived from subjective experiences can underestimate actual alcohol usage. To better understand adolescent alcohol use, these findings from previous studies clearly indicate the need for a straightforward classification of drinking typologies based on frequency and quantity of alcohol in preference to simultaneous use of these two dimensions of alcohol use.
Application of Kuther's (2002) theory of the rational and normative predictors of alcohol consumption to study adolescent drinking typologies guided us in the development of a study model. Kuther explained adolescent alcohol use in relation to positive and negative alcohol expectancies, parental and peer norms and behaviour, and perceived control. Moreover, there are suggested links between factors in his theory and certain types of alcohol use. Positive alcohol expectancy was related to binge drinking (Morawska and Oei, 2005), and peer drinking behaviour to heavy and frequent alcohol use (Beck and Treiman, 1996) and alcohol intoxication (Ariza Cardenal and Nebot Adell, 2000). In contrast, negative alcohol expectancies were related to non-beer usage (Callas et al., 2004).
However, some modifications to Kuther's theory are required in the light of recent research findings. Parents' drinking behaviours were treated separately because of a suggestion of an independent effect of maternal and paternal drinking (Zhang et al., 1999). In addition, adolescents' own norms are included in the model, indicated by the age at first exposure to alcohol and the timing of onset of regular alcohol use. Applying social learning theory (Bandura, 1977), learning to use alcohol is considered to be based on individual learning experiences of initial exposure and regular use as well as observation of others.
Protective effects of parental control (Van Der Vorst et al., 2006) and quality of communication with parents and parental monitoring (Guilamo-Ramos et al., 2005) have been found, suggesting openness of communication with parents may predict adolescent drinking typologies better than perceived control over alcohol use. Psychological well-being was included in the model to detect problematic usage of alcohol: depressed mood was associated with binge drinking (Kuntsche and Gmel, 2004) and with alcohol dependence (Clark, 2004) among adolescents.
Finally, alcohol usage in adolescence has been linked with later usage (Jefferis et al., 2005; McCarty et al., 2004). We anticipate that specific adolescent drinking typologies carry risks for adult alcohol misuse over and above the predictors for adolescent alcohol use. Drinking related factors were differently related to frequency based alcohol use, and to drinking behaviours by gender (Yeh et al., 2006). This underlies our decision to stratify our analyses by gender.
|
Methods |
|---|
|
TOPABSTRACTIntroductionMethodsFindingsDiscussionReferences |
|---|
Population
The 1970 British Cohort Study (BCS70) is an ongoing longitudinal study targeting all children born in 1 week of April 1970 (Bynner et al., 1996
; Ekinsmyth et al., 1992). Initially, 17298 children were included in the study (Elliott and Shepherd, 2006; Plewis et al., 2004). Deaths, emigration, loss of contacts, refusals, and non-responses, diminished the sample size, but 69.4% of the estimated target sample was observed at age 16 and 70.1% at age 30 (Plewis et al., 2004). BCS70 data at age 16 (N = 11206) and 30 (N = 10833) are used in this study. The achieved sample remains representative of the target population despite a slight under-representation of males and the most disadvantaged (Elliott and Shepherd, 2006; Plewis et al., 2004). Robson (2003) found that BCS70 cohort members were more likely to be unmarried, better educated and employed at age 30, compared with the same age group in the UK Labour Force Study. However, demographic differences between the two populations were marginal.
Adolescent drinking typologies
Cohort members' drinking typologies are identified based on information about frequency of drinking alcohol in the previous week and the average amount of alcohol per occasion. We adopted the clinical cut-off criteria specifically set for adolescent alcohol use and for the cohort members of the BCS70 by Blake et al. (2001). Frequent drinkers were those who drank on two or more days per week, while heavy drinkers are defined as those who drank four or more units per day. Four typologies are derived by combining these categories: none/infrequent (<4 units/day and <2 days/week), frequent light (<4 units/day, but 
2 days/week), infrequent heavy (4 units/day, but <2 days/week) and frequent heavy (4 units/day and 2 days/week).
Alcohol expectancies
Positive and negative alcohol expectancies are used for this study. Seven questions about positive experiences brought about by drinking alcohol were addressed to cohort members of age 16 asking, ‘What pleasant effects does drinking alcohol have on you?’ Guided by exploratory factor analysis, the items ‘less shy or more chatty’, ‘happy’, ‘more friendly’ and ‘more relaxed and confident’ were used to measure positive expectancies.
Negative alcohol expectancies are assessed through responses to eight items about unpleasant physical, behavioural, or psychological experiences from drinking alcohol: depressed mood, sleepiness, feeling ill, sickness, experiencing a blackout, having a headache, becoming abusive, and becoming violent. Adolescents who identified themselves as infrequent or non-drinkers also responded to all questions on alcohol expectancies, allowing us to sum their responses into the measures of positive and negative alcohol expectancies.
Norms
Adolescents' own norms are assessed by age at first alcohol use and onset of regular alcohol use. These were based on questions addressing when cohort members had their first taste of an alcoholic drink and when they started drinking with their friends. Responses ranged from 6 years old or under, to never having had alcohol, or initiated regular drinking habits by age 16. The order of the responses is reversed to identify the magnitude of the risk from early exposure to alcohol.
Peer norms are assessed using information about peer drinking behaviour gathered from two items on cohort members' best friend's and next best friend's drinking behaviour. The responses on these items were (i) ‘never’, (ii) ‘occasionally’, (iii) ‘some days’ or (iv) ‘most days’. Peer norms are measured by the mean of the scores on the two items, rounded up to give a 4-point scale equivalent in interpretation to the original scale.
Parental norms are assessed using information about perceived paternal and maternal drinking behaviour. Paternal drinking behaviour is assessed by cohort members’ perception of the frequency of their father's alcohol use with responses (i) ‘never’, (ii) ‘occasionally’, (iii) ‘some days’ or (iv) ‘most days’. Information about maternal drinking behaviour is collected in the same manner as the paternal drinking behaviour.
Openness of communication with parents
This factor is measured by two dimensions: perceived parental control and adolescent communication. Perceived parental control is measured through responses to items asking cohort members how often their parents questioned them when they went out: with whom they were going out, where they were going, and what they were going to do. Responses ranged from one (almost always) to four (hardly ever). A summed score of responses, with a range of 3–12, is used for the analysis.
Adolescent communication is measured by responses to questions on information freely given to parents on going out: with whom they were going, where they were going and what they were going to do. Each item ranged from one (almost always) to four (hardly ever) and a summed score (range 3–12) used for the analysis.
Psychological well-being
Psychological well-being is assessed through the responses on the 12-item General Health Questionnaire (GHQ–12) developed by Goldberg (1992). This questionnaire measures the self-assessed general psychological well-being of respondents relative to usual. Items are measured on a 4-point Likert scale corresponding to zero (better than usual) to three (much less than usual). A higher score indicates greater psychological distress.
Alcohol misuse at age 30
Alcohol misuse at age 30 is determined using the CAGE questionnaire (Ewing, 1984) administered at age 30. Alcohol misuse is viewed as alcohol use with the potential to harm individual health. The CAGE is a brief 4-item survey instrument designed to detect problems of alcohol abuse and dependence CAGE is an acronym for key words from the 4 questions (cut down, annoyed, guilty, eye-opener). A 2-point indicator, 0 being negative for alcohol misuse and 1 being positive, was created following the conventional cut-off point of endorsing two or more items on the questionnaire. Soderstrom et al. (1997) reported a sensitivity of 84% and specificity of 90% among trauma patients; Malet et al. (2005) reported a sensitivity of 77% and a specificity of 94% in a French clinical population.
Imputation of missing data
Multiple imputation is recommended for epidemiological studies with extensive missing values as long as these amount to less than 60% of the total (Barzi and Woodward, 2004; Greenland and Finkle, 1995). Missing data values were imputed using the MICE method implemented in Stata by Royston (2004). Using all variables for our models plus supplementary variables that predicted model variables or non-response, we imputed data for cases that had partial information at either age 16 or age 30, producing five replicates of the dataset. Each filled-in dataset comprised 7023 men and 6896 women. Estimation of model coefficients was produced using the MICOMBINE program in Stata, which averages all estimates and adjusts standard errors according to Rubin's rule (Rubin, 1987).
Statistical analysis
A multinomial logistic regression model is applied to identify the effects of predictors on the four typologies of adolescent drinking with infrequent drinkers set as the reference group. The contribution of each predictor to the model is assessed using the Wald chi-square statistic. The hypothesis that socially or control related factors are associated with certain typologies is provided by post-hoc contrasts of overall effects.
A binary logistic regression model is used to estimate specific risk of adult alcohol misuse manifested by each adolescent drinking typology over and above adolescent alcohol related factors. All analyses are stratified by gender and carried out using Stata v8.2 (Stata Corp, 2003).
|
Findings |
|---|
|
TOPABSTRACTIntroductionMethodsFindingsDiscussionReferences |
|---|
Descriptive findings
The distributions of the variables in the model by drinking typology are presented in Table 1. More men than women were drinkers at age 16; they consumed more alcohol (M = 6.03 units, SD = 8.27) than women (M = 4.78 units, SD = 6.52).
|
Mean values of positive and negative expectancies were the lowest for none/infrequent drinkers and highest in the frequent heavy drinking group. All drinkers, except none/infrequent drinkers appeared to be exposed to alcohol between the ages of 10 and 15 years and started using it regularly between the ages of 13 and 15 years.
Close friends of young men in the cohort appeared to drink slightly more frequently at age 16 than did the women's close friends, while many of the cohort members reported that their close friends used alcohol occasionally or on some days. They also reported that their father drank alcohol occasionally or on some days, whereas they tended to describe their mother as an occasional drinker.
Young women's mean levels of perceived parental control and adolescent communication with parents were lower than young men's, indicating that young women had more open communication with their parents than young men. Young women reported more psychological distress than men; young women's mean GHQ-12 (M = 10.13, SD = 5.69) was higher than young men's (M = 9.37, SD = 5.31).
At age 30, the majority of cohort members did not misuse alcohol. The proportion who misused alcohol was similar between the groups of frequent light and infrequent heavy that is higher than for the none/infrequent drinkers, but is lower than frequent heavy drinkers.
Adolescent drinking typologies
The results of the multinomial logistic regression model of adolescent drinking typologies are shown in Table 2. The findings for each determinant of drinking behaviour are discussed in turn.
|
Alcohol expectancies Among young men and women, expectations about the positive effects of alcohol consumption significantly increased the likelihood of being a drinker compared with being a none/infrequent drinker (See Table 2). Positive alcohol expectancies were associated most strongly with the frequent heavy drinking typology. The odds of being a frequent light or an infrequent heavy drinker relative to a none/infrequent drinker were of similar magnitude. Negative alcohol expectancies failed to predict membership of any of the drinking typologies in comparison with the none/infrequent drinking group. Positive alcohol expectancies dominated the independent contribution of the construct of alcohol expectancies to the models. Nevertheless, the Wald statistic showed that negative expectancies did contribute to the model for young men.
Post-hoc tests showed that alcohol expectancies distinguished frequency and quantity related drinking typologies slightly differently for young men and women (See Table 3). In young men, the three contrasts between drinking typologies were statistically significant, showing that expectancies differentiated between both quantity and frequency related typologies. Moreover, expectancies seemed to operate in an additive fashion on both quantity and frequency of alcohol consumption. Among women, there was a non-significant contrast between the infrequent heavy and frequent heavy drinking typologies (differing in the number of occasions of drinking), and significant contrasts between groups that differed as to the quantity consumed per occasion. This indicates a close association of alcohol expectancies with quantity related typologies, but not with frequency related typologies.
|
2 values (df) comparing the effect of each construct on adolescent drinking typologies by men (N = 7023) and women(N = 6896)Norms In young men and women, the odds of being a frequent light or a frequent heavy drinker associated with age at first alcohol use, and with the onset of regular alcohol use, were slightly greater than that of being an infrequent heavy drinker. Similarly, peer alcohol use was a significant predictor for all drinking typologies with reference to the none/infrequent drinking typology, especially so for frequent heavy drinking. The odds of being a frequent heavy drinker increased markedly as the frequency of consumption of alcohol by the adolescent's peer group increased. While the father's drinking did not predict adolescent drinking typologies, more frequent maternal alcohol use increased the likelihood of being a frequent light drinker in young men and women. Frequent heavy drinking was also predicted by the mother's alcohol use although comparisons with the none/infrequent drinkers were weaker.
Among all norm related factors, peer norms were the strongest independent predictors of drinking behaviour for both young men and women. In men, early exposure to alcohol had the second largest impact on alcohol use at age 16; however, the same factor had the least effect on adolescent alcohol use among young women. The effect of onset of regular alcohol use was the smallest of all norm related factors in young men.
When the contribution of norms to each drinking typology was compared, norms differentiated between frequent light and frequent heavy drinking typologies the least in young men and non-significantly differentiated between them in young women. By contrast, norms among infrequent heavy drinkers were different from both frequent light and frequent heavy drinkers. These findings indicate that norms contributed more to differences in the frequency than the quantity of adolescent alcohol consumption, especially among young women.
Openness of communication with parents The effect of openness of communication was dominated by perceived parental control (see Table 2). The independent contribution of adolescent communication just failed to reach statistical significance in both the male and female models. Weaker perceived parental control increased the likelihood of becoming either an infrequent heavy or a frequent heavy drinker in both young men and young women.
Post-hoc tests confirmed non-significant differences between the quantity related typologies (infrequent heavy vs frequent heavy) in men and women by openness of communication with parents, whereas this construct showed significant differences between the frequency related typologies (frequent light vs frequent heavy) (see Table 3). This indicates that openness of communication with parents contributed to quantity of alcohol use, but had no effect on frequency of alcohol use in youth.
Psychological well-being Contrary to expectations, in both bivariate analyses and multivariate analyses, psychological well-being was not significantly associated with particular drinking typologies in young men and women.
Adult alcohol misuse
The risk from the different typologies of adolescent drinking for adult alcohol misuse was assessed using logistic regression. All adolescent drinking typologies increased the odds of misusing alcohol at age 30 before adjusting for the effects of factors related to the adolescent drinking typologies (Table 4). The odds of misusing alcohol at age 30 from infrequent heavy drinking at age 16 were non-significantly lower than from frequent heavy drinking at age 16 in men and women. Adolescent frequent heavy drinking behaviour most increased the odds of misusing alcohol at age 30; the odds were over 2 times greater than that for the none/infrequent drinkers.
|
After adjusting for the factors associated with adolescent drinking typologies, the odds of adult alcohol misuse were considerably reduced to approximately 1.4–1.6 times the odds for male none/infrequent drinkers. Women showed a risk of alcohol misuse from frequent light and frequent heavy drinking in adolescence. However, women's infrequent heavy drinking at age 16 failed to show a significantly raised odds ratio for alcohol misuse at age 30. Thus, the risk of misusing alcohol in adulthood was related to any types of adolescent alcohol use in men, and to frequent alcohol use in women.
|
Discussion |
|---|
|
TOPABSTRACTIntroductionMethodsFindingsDiscussionReferences |
|---|
Four adolescent drinking typologies using frequency and quantity of alcohol use at age 16 were identified, with the expectation that a unique set of factors would predict either frequency or quantity related typologies. Alcohol expectancies and openness of communication with parents predicted quantity of adolescent alcohol use, while norms better predicted frequency. The pattern of influences that emerged provide support for Vogel-Sprott's (1974
) notion of an association between socially related factors and frequency of alcohol use and an association between individual control related factors and quantity of alcohol use. However, there was no support for the hypothesized relationship between psychological well-being and the adolescent drinking typologies. Moreover, adolescent drinking typologies did not specifically identify which young men were at risk of becoming problem drinkers in adulthood: any type of drinking behaviour in adolescence increased the odds of alcohol misuse in adult life in men. By contrast, frequent alcohol use by young women predicted alcohol misuse later in life. One might argue with the treatment of missing cases, the method to assess adolescent alcohol use, and the classification method for adolescent drinking typologies. Although available data on individual variables ranged from 4902 to 10968, using only complete cases reduced the sample size to 686 for men and 957 for women. The effects of some predictors of drinking typologies and the risk of drinking typologies for adult alcohol misuse were underestimated in a complete case analysis (results not shown). Estimating the models using multiple imputation under the missing-at-random assumption is likely to be less biased than using complete cases only.
Use of self reports is liable to underrate actual alcohol use among adolescents (Del Boca and Darkes, 2003; Del Boca and Noll, 2000; Wilson et al., 2004); thus the cohort member's actual alcohol use at age 16 could have been underestimated because of reliance on self-reported adolescent alcohol use. Inconsistent responses to questions on age at first taste of alcohol, onset of regular use, and alcohol consumption in the previous week were deliberately preserved by the BCS70 study team for the interesting contradictions they expose (Goodman and Butler, 1986). As for their actual alcohol use, young people in BCS70 were asked to recall immediate frequency and quantity of alcohol use rather than their distal use, with a guided reference. This should have prompted young respondents to recall their alcohol use more accurately.
One might also argue that the cut-points used for the drinking typologies are not gender specific. We used the cut-points developed by Blake et al. (2001) because of a lack of consensus among studies regarding the definition of binge drinking, even for adults (Gill, 2002). The increased risk of alcohol misuse from all drinking typologies compared with none/infrequent use in adolescence provides a clear warning that any use of alcohol carries a risk for adult alcohol misuse regardless of gender. Further research is awaited before the establishment of gender specific cut-points to identify adolescent drinking typologies, including binge drinking.
The prediction of the quantity related typologies (heavy infrequent and heavy frequent) by alcohol experiences is in line with previous studies (Lee et al., 1999; Morawska and Oei, 2005). Community mastery protected against frequent alcohol use among young men only (Piko, 2006). It is possible that the items in the positive expectancies scale tap into the social context of alcohol use, thus explaining the relationship with frequency related typologies as well as quantity related typologies.
Negative alcohol expectancies did not predict any type of drinking behaviour over none/infrequent drinking though they predicted overall alcohol use among young men. In another study, negative expectancies only produced the desired effect of reducing consumption after the age of 35 years (Leigh and Stacy, 2004) and this may be the reason for the failure of negative expectancies to protect young people from infrequent heavy drinking in this study.
Norms were the most important predictors of adolescent alcohol use and contributed more to frequency related drinking typologies than quantity related drinking typologies. Unexpectedly, paternal drinking did not significantly contribute to the models of alcohol use. Malone et al. (2002) found that paternal maximum use of alcohol predicted adolescents’ substance misuse and dependence. Our findings suggest that future studies may need to consider employing two dimensions of alcohol use for parents as we did for adolescent alcohol use.
A distinctive association between frequency related drinking typologies and norms was found among young women. In this study, peer drinking was associated with all types of young men's alcohol use as was found in the study by Beck and Treiman (1996). Increased frequency of peer drinking will tend to be correlated with quantity of alcohol use, thus acting as a marker for quantity of alcohol consumption. Again, future studies may need to include quantity as well as frequency of peer alcohol use.
Openness of communication with parents was identified as a predictor of quantity related drinking typologies. Quantity of alcohol consumption can be predicted by individual control (Vogel-Sprott, 1974). However, in this study, it was perceived parental control that mainly predicted heavy alcohol use among adolescents, a finding similar to that of Van Der Vorst et al. (2006). Further work is needed to explore the relationship between parental and individual control to assess the protective effect of each factor on adolescent alcohol use.
Unexpectedly, psychological well-being did not predict membership of any drinking typology groups. Cable and Sacker (2007) found that adult psychological well-being showed a weak relationship with adult men's heavy drinking, and a stronger association with adult alcohol misuse in both men and women. It is possible that the link between psychological well-being and problematic alcohol consumption is forged later in life.
Adolescent drinking was hypothesized to manifest a risk for misusing alcohol in adulthood, with the expectation of greater risks from more frequent and heavy alcohol use. Although adolescent drinking did constitute a risk for later alcohol misuse, the latter prediction was not upheld. In men, the likelihood of misusing alcohol at age 30 was predicted to be similar between infrequent heavy and frequent heavy drinking at age 16 in comparison to none/infrequent drinking behaviour. Either, quantity or frequency of adolescent alcohol use, increased the risk for adult alcohol misuse in men over and above the factors associating with adolescent alcohol use, with no evidence of any additive effect.
On the other hand, among women, the risk of misusing alcohol at age 30 was limited to more frequent alcohol use. It is possible that the statistical power from the smaller sample of female adolescent infrequent heavy drinkers is related to the finding that female infrequent heavy drinkers appeared to be immune to adult alcohol misuse. Alternatively, men might be more likely to maintain or increase their alcohol usage over time, whereas women may have more variable patterns of alcohol consumption.
The typologies adopted in this study have been a useful way of examining the specificity of predictors of alcohol use and whether they relate to the quantity or frequency of alcohol consumption. The effectiveness of using frequency and quantity of alcohol use to indicate adolescent drinking typologies was demonstrated by the findings that norms were associated more with frequency than with quantity of alcohol consumption, while the opposite was found for alcohol expectancies and openness of communication. Theories of adolescent alcohol consumption need different accounts for these two dimensions of alcohol use.
We found that the frequent light drinking pattern carries nearly a 2-fold increased risk of alcohol misuse at age 30. This suggests that interventions to reduce adult alcohol misuse should be started in adolescence or even earlier. Building an alcohol-free culture for youth is part of the Alcohol Harm Reduction Strategy for England (Department of Health et al., 2007). Our findings support this current political trend discouraging alcohol use among young people but go further in suggesting that different approaches are needed for reducing both the quantity and frequncy of adolescent alcohol consumption. To reduce frequency of alcohol use, delaying first exposure to alcohol could be helpful, as would the provision of alcohol-free environments for adolescent social interactions. Enhancing young peoples’ socialization skills could be a useful means of alternating their positive alcohol expectancies, in turn leading to a reduction in the quantity of alcohol consumed. Involving parents could be an additional focus for the prevention of heavy drinking. Effective early interventions for adolescent alcohol use might reduce later problems.
|
ACKNOWLEDGEMENTS |
|---|
This research was funded by the Economic Social Research Council, RES-337-25-0001. The 1970 British Cohort Study data used in this study were deposited by the Centre for Longitudinal Study, Institute of Education. The data were accessed via UK Data Archive. Neither Centre for Longitudinal Studies nor the UK Data Archive bear any responsibilities for their further analysis or interpretation. These secondary data are publicly held and ethical approval for the use of these data is not required. There are no conflicts of interest involved in this study.
|
References |
|---|
|
TOPABSTRACTIntroductionMethodsFindingsDiscussionReferences |
|---|
Ariza Cardenal C., Nebot Adell M. Factors associated with problematic alcohol consumption in schoolchildren. Journal of Adolescent Health (2000) 27:425–433.[CrossRef][Web of Science][Medline]
Bandura A. Social Learning Theory (1977) London: Prentice Hall. 1–247.
Barzi F., Woodward M. Imputations of missing values in practice: Results from imputations of serum cholesterol in 28 cohort studies. American Journal of Epidemiology (2004) 160:34–45.
Beck K. H., Treiman K. A. The relationship of social context of drinking, perceived social norms, and parental influence to various drinking patterns of adolescents. Addictive Behaviors (1996) 21:633–644.[CrossRef][Web of Science][Medline]
Blake H., Collier J., Polnay L., et al. Drinking habits of 16 year olds: secondary analysis of the 1970 British Cohort Study. Short Accessible Report-May 2001 (2001) London: AERC.
Bynner J., Wiggins R. D., Parsons S. An Exploratory Comparative Analysis of Data Collected in the 1958 and 1970 British Cohort Studies: the Role of Personal Agency in Entry to Employment under Different Labour Market Conditions (1996) London: SSRU, City University.
Cable N., Sacker A. The role of adolescent social disinhibition expectancies in moderating the relationship between psychological distress and alcohol use and misuse. Addictive Behaviors (2007) 32:282–295.[CrossRef][Web of Science][Medline]
Callas P. W., Flynn B. S., Worden J. K. Potentially modifiable psychosocial factors associated with alcohol use during early adolescence. Addictive Behaviors (2004) 29:1503–1515.[CrossRef][Web of Science][Medline]
Chassin L., Pitts S. C., Prost J. Binge drinking trajectories from adolescence to emerging adulthood in a high-risk sample: predictors and substance abuse outcomes. Journal of Consulting Clinical Psychology (2002) 70:67–78.[CrossRef][Web of Science][Medline]
Clark D. B. The natural history of adolescent alcohol use disorders. Addiction (2004) 99(Suppl. 2):5–22.[Web of Science][Medline]
Del Boca F. K., Darkes J. The validity of self-reports of alcohol consumption: state of the science and challenges for research. Addiction (2003) 98:S1–S12.
Del Boca F. K., Noll J. A. Truth of consequences: The validity of self-report data in health services research on addiction. Addiction (2000) 95:S347–S360.[Web of Science][Medline]
Department of Health, Home Office, Department of Education and Skills, and Department for Culture, Media and Sport. Safe. Sensible. Social. The Next Steps in the National Alcohol Strategy. (2007) accessed 6-6-2007. Available at: http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_075218.
Ekinsmyth C., Bynner J., Montgomery S., et al. An Integrated Approach to the Design and Analysis of the 1970 British Cohort Study (BCS 70) and the National Child Development Study (NCDS) (1992) London: SSRU, City University.
Elliott J., Shepherd P. Cohort profile: 1970 British Birth Cohort (BCS70). International Journal of Epidemiology (2006) 35:836–843.
Ewing J. A. Detecting alcoholism: The CAGE questionnaire. JAMA (1984) 252:1905–1907.
Gill J. S. Reported levels of alcohol consumption and binge drinking within the UK undergraduate student population over the last 25 years. Alcohol and Alcohoholism (2002) 37:109–120.[CrossRef]
Gmel G., Klingemann S., Muller R., et al. Revising the preventive paradox: The Swiss case. Addiction (2001) 96:273–284.[CrossRef][Web of Science][Medline]
Goldberg D. General Health Questionnaire (GHQ12) (1992) Windsor: NEFR-Nelson.
Goodman A., Butler N. BCS 1986- Guide to the BCS70 16-year Data Set (1986) London: Social Statistics Reserch Unit, City University.
Greenland S., Finkle W. D. A critical look at methods for handling missing covariates in epidemiologic regression analyses. American Journal of Epidemiology (1995) 142:1255–1264.
Griffin K. W., Botvin G. J., Nichols T. R., et al. Low perceived chances for success in life and binge drinking among inner-city minority youth. Journal of Adolescent Health (2004) 34:501–507.[Web of Science][Medline]
Guilamo-Ramos V., Jaccard J., Turrisi R., et al. Parental and school correlates of binge drinking among middle school students. American Journal of Public Health (2005) 95:894–899.
Helstrom A., Bryan A., Hutchison K. E., et al. Tobacco and alcohol use as an explanation for the association between externalizing behaviour and illicit drug use among delinquent adolescents. Preventive Science (2004) 5:267–277.[CrossRef]
Hewton J., Moody A. Drug use, Smoking and Drinking among Young People in England in 2003 (2004) accessed 13-5-2005. Available at: http://www.dh.gov.uk/assetRoot/04/09/89/17/04098917.pdf.
Hingson R. W., Heeren T., Winter M. R. Age at drinking onset and alcohol dependence: Age at onset, duration, and severity. Archives of Pediatric Adolescent Medicine (2006) 160:739–746.
Jefferis B. J. M. H., Power C., Manor O. Adolescent drinking level and adult binge drinking in a national birth cohort. Addiction (2005) 100:543–549.[CrossRef][Web of Science][Medline]
Kuntsche E. N., Gmel G. Emotional wellbeing and violence among social and solitary risky single occasion drinkers in adolescence. Addiction (2004) 98:331–339.
Lee N. K., Greely J., Oei T. P. S. The relationship of positive and negative alcohol expectancies to patterns of consumption of alcohol in social drinkers. Addictive Behaviors (1999) 24:359–369.[CrossRef][Web of Science][Medline]
Leigh B. C., Stacy A. W. Alcohol expectancies and drinking in different age groups. Addiction (2004) 99:215–227.[CrossRef][Web of Science][Medline]
Li F., Barrera M Jr, Hops H, et al. The longitudinal influence of peers on the development of alcohol use in late adolescence: a growth mixture analysis. Journal of Behavioral Medicine (2002) 25:293–315.[CrossRef][Web of Science][Medline]
Malone S. M., Iacono W. G., McGue M. Drinks of father: Father's maximum number of drinks consumed predicts externalizing disorders, substance use, and substance use disorders in preadolescent and adolescent offspring. Alcoholism: Clinical and Experimental Research (2002) 26:1823–1832.[Web of Science][Medline]
Malet L., Schwan R., Boussiron D., et al. Validity of the CAGE questionnaire in hospital. European Psychiatry (2005) 20:484–489.[CrossRef][Web of Science][Medline]
McCarty C. A., Ebel B. E., Garrison M. M., et al. Continuity of binge and harmful drinking from late adolescence to early adulthood. Pediatrics (2004) 114:714–719.
Morawska A., Oei T. P. S. Binge drinking in university students: a test of the cognitive model. Addictive Behaviors (2005) 30:203–218.[CrossRef][Web of Science][Medline]
Murgraff V., Parrott A., Bennett P. Risky single-occasion drinking amongst young people-definition, correlates, policy, and intervention: A broad overview of research findings. Alcohol and Alcoholism (1999) 34:3–14.
Oei T. P. S., Morawska A. A cognitive model of binge drinking: The influence of alcohol expectancies and drinking refusal self-efficacy. Addictive Behaviors (2004) 29:159–174.[CrossRef][Web of Science][Medline]
Piko B. F. Adolescent smoking and drinking: The role of communal mastery and other social influences. Addictive Behaviors (2006) 31:102–114.[CrossRef][Web of Science][Medline]
Pitkänen T., Lyyra A. L., Pulkkinen L. Age of onset of drinking and the use of alcohol in adulthood: A follow-up study from age 8–42 for females and males. Addiction (2005) 100:652–661.[CrossRef][Web of Science][Medline]
Plewis I., Claderwood L., Hawkes D., et al. National Child Development Study and 1970 British Cohort Study technical report: Changes in the NCDS and BCS70 Populations and Samples Over Time (2004) 1st edn. London: Centre for Longitudinal Studies.
Prime Minister's Strategy Unit. Alcohol Misuse: How much Does it Cost? (2003) accessed 4-7-2007. Available at: http://www.cabinetoffice.gov.uk/strategy.
Prime Minister's Strategy Unit. Alcohol Harm Reduction Strategy for England (2004) accessed 1-3-0005. Available at: http://www.cabinetoffice.gov.uk/strategy/work_areas/alcohol_misuse/index.asp.
Rehm J., Gmel G. Patterns of alcohol consumption and social consequences. Results from an 8-year follow-up study in Switzerland. Addiction (1999) 94:899–912.[CrossRef][Web of Science][Medline]
Rehm J., Room R., Graham K., et al. The relationship of average volume of alcohol consumption and patterns of drinking to burden of disease: An overview. Addiction (2003) 98:1209–1228.[CrossRef][Web of Science][Medline]
Robson K. Peer Alienation: Predictors in Childhood and Outcomes in Adulthood (2003) accessed 21-6-2007. Available at: http://www.iser.essex.ac.uk/pubs/workpaps/pdf/2003-21.pdf.
Rossow I., Romelsjo A. The extent of the ’prevention paradox’ in alcohol problems as a function of population drinking patterns. Addiction (2006) 101:84–90.[CrossRef][Web of Science][Medline]
Royston P. Multiple imputation of missing values. The STATA Journal (2004) 4:227–241.
Rubin D. B. Multiple Imputation for Nonresponse in Surveys (1987) New York: John Wiley and Sons. 1–288.
San José B., Van Oers H. A. M., Van De Mheen H. D., et al. Drinking patterns and health outcomes: occasional versus regular drinking. Addiction (2000) 95:865–872.[CrossRef][Web of Science][Medline]
Soderstrom C. A., Gordon S., Kufera J., et al. The accuracy of the CAGE, the brief Michigan alcoholism screening test, and the alcohol use disorders identification test in screening trauma center patients for alcoholism. The Journal of Trauma (1997) 43:962–969.
Stata Corp. Stata statistical Software: Release 8. [V8.2] (2003) College Station, TX: Stata Corp LP.
Van Der Vorst H., Engels R. C., Meeus W., et al. Parental attachment, parental control, and early development of alcohol use: a longitudinal study. Psychology of Addictive Behaviors (2006) 20:107–116.[CrossRef][Web of Science][Medline]
Vogel-Sprott M. Defining "light" and "heavy" social drinking:Research implications and hypotheses. Quarterly Journal of Studies on Alcohol (1974) 35:1388–1392.[Web of Science][Medline]
Weitzman E. R., Nelson T. F. College student binge drinking and the "prevention paradox": implications for prevention and harm reduction. Journal of Drug Education (2004) 34:247–265.[CrossRef][Web of Science][Medline]
Wells J. E., Horwood L. J., Fergusson D. M. Drinking patterns in mid-adolescence and psychosocial outcomes in late adolescence and early adulthood. Addiction (2004) 99:1529–1541.[CrossRef][Web of Science][Medline]
Williams R. J., McDermitt D. R., Bertrand L. D., et al. Parental awareness of adolescent substance use. Addictive Behaviors (2003) 28:803–809.[CrossRef][Web of Science][Medline]
Wilson C. R., Sheritt L., Gates E., et al. Are clinical impressions of adolescent substance use accurate? Pediatrics (2004) 114:e536–e540.
Windle M. An alcohol involvement typology for adolescents: Convergent validity and longitudinal stability. Journal of Studies on Alcohol (1996) 57:627–637.[Web of Science][Medline]
Yeh M.-Y., Chiang I.-C., Huang S.-Y. Gender differences in predictors of drinking behaviour in adolescents. Addictive Behaviors (2006) 31:1929–1938.[CrossRef][Web of Science][Medline]
Zhang L., Welte J. W., Wieczorek W. F. The influence of parental drinking and closeness on adolescent drinking. Journal of Studies on Alcohol (1999) 60:245–251.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||