Table 2

Cohort studies included in review
Author, year, country Age/grade at start of study, waves, n Predictora Outcomea Variables accounted for in study design or analysis Analysis Results
Adrados, 1995 Spain Age 15-18+ Trust in parents. Alcohol initiation. Situation variables.b Linear regression analysis B= -.076, p≤.05
2 waves in 1 year
n= 614
Andrews et al., 1997 USA Age 11-15 Quality of parent-adolescent relationship (created by summing the subscales measuring the parent’s appraisal of the adolescent and the adolescent’s appraisal of his/her parent). Categorized as “current user” and “current nonuser”. Gender, age, marital status Generalized estimating equations Father:
6 waves in 5 years User: currently using “at least once in a while” and a rate of use of >0 times per month over previous 6 months. Relationship: β= -.03, ns
n= 657 Age*relationship: β= -.64, ns
Gender*relationship: β=-.29, ns
Relationship: β= -.04, ns
Nonuser: never used (report of never tried, along with a rate of zero times per month in last 6 months) or previous but not current users (report of quitting, along with a rate of zero times per month in last 6 months). Age*relationship: ns
β= -.27, p<.01:
Boys: β= -.04, ns
Girls: β= -0.31, p<.001
Aseltine & Gore, 2000 USA 9th-11th gradec 4 waves in 9 years Parental support: degree to which the parents make the child feel loved and wanted, trust the child, and the extent to which the child enjoys being with family members. Frequency of alcohol use during past 12 months (0= never to 7= every day). Gender, family structure, family’s standard of living, parental education, conflict with parents, life events, peer support, peer conflict. Hierarchical linear modeling Frequency of use:
n= 1208 Parent conflict: average frequency of arguments with mother and father. Frequency of heavy alcohol use (5 or more drinks in a row) (0= never to 4= 6 or more times). Support: B= .007, ns
Age*support: B= .024, p<.05
Conflict: B= .033, ns
Age*conflict: B= -.010, ns
Frequency of heavy use:
Support: B= -.004, ns
Age*support: B= .012, p<.05
Conflict: B= .054, ns
Age*conflict: ns
Barnes et al., 2000 USA Age 13-16 Support: behavior toward the adolescent indicating to her or him that she or he is valued and loved. A composite alcohol misuse index: ounces of alcohol from all beverages per day (based on frequency and quantity of consuming beer, wine, and liquor in the past year), times drunk in the past year, and frequency of having five or more drinks at a time during past year. Gender, age, race, parental alcohol misuse. Latent growth structural equation modeling ns
7 waves in 6 years
n= 506
Branstetter et al., 2011 USA 10th grade Maternal support: attachment, caregiving, and affiliation Frequency of alcohol use during past 30 days (1= never used to 8= use every day). Gender, friend’s substance use, friendship support, friendship negative interaction, previous alcohol use Linear regression analysis Maternal support:
2 waves in 1 year Maternal negative interaction: conflict, antagonism, and criticism. β= -.011 ns
n=166 Maternal negative interaction:
β= -.013 ns
Chuang et al., 2005 USA Age 12-14 Parental closeness: attachment, involvement, child-centeredness Categorized as: “use of alcohol” and “no use of alcohol”. Age, gender, race/ethnicity, parents’ education, treatment condition. Structural Equation Modeling ns
3 waves in 1 year
n= 959
Cohen et al., 1994 USA Cohort 1: Positive relationship: affectional interactions. Categorized as: “monthly user” (used alcohol in previous month) and “nonuser” (did not use alcohol in previous month). Gender, study classification (study or control), ethnic group. Logistic regression analysis Cohort 1:
5th grade 5th to 6th grade/ 6th to 7th grade/ 7th to 8th grade, RR (95% CI)=
4 waves in 3 years 97 (.65-1.44), ns/
n= 618 69 (.46-1.01), ns/
Cohort 2: 96 (.65-1.42), ns
7th grade Cohort 2:
3 waves in 2 years 7th to 8th grade/ 8th to 9th grade
n= 732 (95% CI)=
(.52-.82) p<.001/
67 (.50-.90) p <.01
Cookston & Finlay, 2006 USA 7th -12th grade Parent–child involvement: activities shared with parents, discussion with parents, parent–child closeness. Mean of 3 items: how often alcohol used in past year, how often 5 drinks in a row, how often gotten drunk. Gender, age, parent education, father status, previous alcohol use. Structural Equation Modeling Father’s involvement:
waves in 1 year β= -.04, ns
n= 2387 Mother’s involvement:
β= -.05, ns
Crawford & Novak, 2002 USA 10th grade Attachment: quality of child-parent relation. Number of times consumed alcohol in their lifetime. Gender, race, socioeconomic background, peer affiliation, participation in unstructured peer interaction, participation in structured activities, time spent with parents, parental monitoring, parental control, previous alcohol use. Ordinary least square (OLS) and l ogistic regression analysis Number of drinks in lifetime:
2 waves in 2 years Number of times for heavy drinking (5 or more drinks in a row) in past 2 weeks. B= .00, ns
n= 2506 (OLS analysis)/ n= 426 (logistic analysis) Onset of alcohol use. Times heavy drink:
Onset of heavy drinking. B= -.01, ns
Onset of alcohol use:
B= -.0072, ns
Onset of heavy drinking:
B= -.0872, p<.05
Danielsson et al., 2011 Sweden 7th grade Attachment: strong/secure emotional bonds to parents. Categorized as: “yes” or “no” (reported respectively did not report heavy episodic drinking on at least one occasion). Drinking friends, money to spend, smoking, parental provision of alcohol, bullying, truancy, time with parents, parental monitoring, previous heavy episodic drinking. Logistic regression analysis Boys: ns
2 waves in 2 years Measurement: “how often do you drink six cans of medium-strength beer, or four cans of normal beer, or four large bottles of strong cider, or a bottle of wine, or half a bottle of spirits on the same occasion?” Girls:
n= 1222 OR (95% CI)= 1.33 (0.85-2.09)
Donohew et al., 1999 USA 6th grade Positive family relations: e.g., gets along with their mother/father, has fun with parents, are happy at home. How many times alcohol used during past year (1= none to 7= 40 or more times). Sensation seeking, attitudes toward alcohol and drugs, peer sensation seeking, perceived alcohol/marijuana use by friends, peer alcohol/ marijuana use, perceived peer influence to use alcohol/ marijuana, previous alcohol/marijuana use. Structural equation modeling 8th to 9th grade:
3 waves in 2 years β= -.07, ns
n= 428 9th to 10th grade:
β= -.03, ns
Droomers et al., 2003 New Zealand Age 11 Attachment: three ubscales of communication, trust and alienation. Average amount consumed on a typical occasion categorized at each measurement as 25% highest amounts vs. 75% lowest amounts of alcohol use. Gender, mother’s attitude towards alcohol consumption of child, friends’ attitude towards alcohol consumption, noticeable alcohol problems in family, knowledge of child about alcohol (obtained by parents), intelligence. Generalized estimating equations High attachment:
5 waves in 10 years OR= 1.00
n= 1037 Medium attachment:
OR= 1.42, p<.05
Low attachment:
OR= 1.50, p<.05
Eisenberg et al., 2008 USA Mean age 12.8 Family connectedness: “How much do you feel your mother/father cares about you?” and “Do you feel you can talk to your mother/father about your problems?” Categorized as: “at least monthly use” and “less frequent or nonuse.” Race, SES, previous substance use, family meals. Logistic regression analysis Female:
2 waves in 5 years Measurement: how often used alcohol during past year (0= never to 5= daily). OR (95% CI)= 1.31 (0.89-1.95)
n= 806 Male:
OR (95% CI)= 0.93 (0.65-1.34)
Ennett et al., 2001 USA Age 12-14 Supportiveness: parent–child relationship as helping the adolescent when needed, providing encouragement and praise, and spending time together (parent report). Categorized as: “escalators” (at least sipped at T1 and increased drinking level at T2), “initiators” (did not sip at T1 and sipped at T2) and lifetime nondrinkers (never sipped or drank any alcohol). Gender, age, race/ethnicity, mother’s education, family structure, parent–child communication about tobacco and alcohol (rules, consequences, media), parental smoking, parental alcohol use, parental disapproval of tobacco and alcohol use, monitoring. Logistic regression analysis Initiation:
2 waves in 1 year OR = 1.78, ns
n= 476 Escalation:
OR = 1.71, ns
Flory et al., 2004 USA 6th grade Family relations: how close the participant felt to their parents or guardians and the quality of these relationships. Past month alcohol use (0= not drunk alcohol to 6= 40+). Gender. Latent class growth analysis (to identify subgroups) and analysis of variance (to test differences between subgroups) F=2.67, ns
6 waves in 10-12 years Three subgroups determined: early onset, late onset, non-users.
n= 481
Guilamo-Ramos et al., 2004 USA 7th -11th grade Maternal warmth: “Most of the time my mom is warm and loving toward me.” Heavy episodic alcohol consumption: frequency of drinking five or more drinks in a row in the past 12 months (0= never to 6= every day or almost every day). Gender, grade. Extension of generalized estimating equations Three way interaction by gender*grade*warmth: p<.001:
2 waves in 1 year predicted means at high/medium/low level of maternal warmth for boys:
n= 1420 1.29 REF/ 1.85, p<.05/ 2.41, p<.05
Predicted means at high/medium/low level of maternal warmth for girls:
1.58 REF/ 1.48, ns/ 1.38, ns
Gutman et al., 2011 USA Age 13 Positive identification: feeling close to parents, respecting parents, wanting to be the kind of person the parent is, doing things together. How many alcoholic drinks have you had in past 30 days (0= none to 3= one or more per week). Gender, SES, ethnicity. Hierarchical linear modeling Positive identification:
5 waves in 7 years Negative interactions: parents criticizing ideas; putting their needs above adolescents’ needs; having hit, pushed, grabbed, or shoved adolescent. B= -.106, p<.001
n= 1160 Negative family interactions:
This study: wave 1, 3, 4 and 5 B= .084, p<.001
Time-lagged positive identification:
B= -.107, p<.001
Time-lagged negative family interactions:
B= ns
Horton & Gil, 2008 USA Mean age 11 Parent–child communication/ connectedness and attachment: sharing private thoughts and feelings with mother/father. Frequency and level of alcohol use. SES, family structure, previous alcohol use. Linear regression analysis Parent–child communication/connectedness and attachment:
3 waves in 2.5 years Parental derogation/rejection: being disliked by, put down by, or of little interest to one’s parents. β= -.057, ns
n= 451 (boys) Parental derogation/rejection:
β= -.038, ns
Hung et al., 2009 Taiwan 5th grade Parental support: encourages, praises, consoles, cares when sick, listens, cares about what happens at school, and helps in solving problems. Categorized as: “first-time user” (never-user at T1 and ever-user at T2) and “never-user” (never at T1 and T2). Gender, area, parent’s marital status/living arrangement, household income, father’s and mother’s educational level, parental alcohol use, family conflict. Logistic regression analysis β= -.05, p<.01
2 waves in 1 year Measurement: have you ever used alcohol? (1= never to 6= every day in the past month). OR (95% CI)= .95 (.92-.99)
n= 1183
Kosterman et al., 2000 USA 5th grade Bonding to mother: sharing thoughts and feelings and desire to be the kind of person one’s mother is. Alcohol initiation: the first point at which a participant reported having “ever drunk beer, wine, whiskey, gin, or other liquor.” From 5th wave and follow-up question was revised to include “other than a sip or two.” Gender, race/ethnicity, previous marijuana initiation, parents’ proactive family management, parents’ alcohol use norms, associates’ alcohol use, participants’ alcohol use norms. Survival analysis B= -.06, ns
8 waves in 7,5 years
n= 808
Kuntsche et al., 2009 The Netherlands Age 14-17 Quality of parent-child relationship for both parents: e.g., “I tell my mother/father my problems and worries” and “my mother/father respects my feelings.” Total number of alcoholic drinks in previous week during weekdays and on weekends at home and outside the home. Gender Structural equation modeling Overall:
3 waves in 2 years β= -.07, ns
n= 364 Low quality group:
β=- .10, ns
High quality group:
β= -.22, p<.01
Latendresse et al., 2008 Finland Age 11-12 Parental warmth: perceived home atmosphere (e.g., “warm, caring,” creative, supportive). Drinking frequency (1= never to 9= daily). Zygosity, sex, family structure, relational tension, shared activities, autonomy granting, parental discipline, parental monitoring, previous alcohol use. Multiple mediation modeling Parental warmth:
3 waves in 6 years Relational tension between adolescents and their parents (e.g., “unjust,” “argumentative”). β= .00, ns
n= 4731 Relational tension:
β= .02, ns
Mogro-Wilson, 2008 USA 7th to 12th grade Parental warmth: “Most of the time, my father/mother is warm and loving toward me”. Combination of frequency of alcohol drinking and frequency of drunkenness (1= never to 7= everyday or almost every day). Income, peer alcohol use, place of birth, language spoken at home. Structural equation modeling ns
2 waves in 1 year
n= 1887
Paschall et al., 2004 USA Age 11-21 Parent-adolescent closeness: “How close do you feel to your dad/mom?” and “How much do you think he/she cares about you?” Categorized as: “frequent” (more than once per month), “infrequent” (once per month or less) and “none.” Gender, age, race, mother’s educational level, personal income, work intensity, mother’s involvement, alcohol use before age 14, past-year heavy drinking (frequent/infrequent). Logistic regression analysis Parent adolescent closeness:
2 waves in 1 year Parent-adolescent conflict: serious argument in past 4 weeks. Measurement: frequency of drinking five or more drinks in a row in the past 12 months (0= never to 6= every day or almost every day). Frequent:
n= 4135 OR (95% CI)= .85 (.65-1.11), ns
OR (95% CI)= .82 (.68-1.00), ns
Parent-adolescent conflict: ns
Shelton & Van den Bree, 2010 USA 7th or 8th grade Parent–child relations: e.g., “Most of the time, your mother/father is warm and loving toward you”. A composite alcohol use index: five items about frequency and quantity. Age, gender, maternal smoking and drinking, BMI, previous alcohol use. Structural equation modeling Boys:
2 waves in 1 year Pubertal timing:
n= 2538 Early: B= .01, ns
On-Time: B= .07, p<.05
Late: B= .04, ns
Pubertal timing:
Early: B=.16, <.05
On-Time: B=.09, <.05
Late: B= .00, ns
Simons-Morton, 2004 USA 6th grade Parental conflict (e.g., I have a parent who is hard for me to get along with, with whom I am often angry). Categorized as initiators (participants who reported drinking in past 30 days at T2 and no drinking at T1) and no-drinkers (no drinking at T1 and T2). - Logistic regression analysis OR (95% CI)= 1.48 (.98-2.23)
2 waves in 1 year
n= 1009
Van der Vorst et al., 2006 The Netherlands Age 11-14 Attachment: relative degree of perceived parental security. Combination of alcohol frequency and intensity. Gender. Structural equation modeling Boys:
3 waves in 1 year Alcohol frequency: How often in past 4 weeks (1= every day to 6= have not been drinking). T1-T2: β= -.007, ns
n= 1012 Alcohol intensity: T2-T3: β= -.076, p<.05
How many glasses in past week, during weekdays, during weekends and inside and outside the home. Girls:
T1-T2: β= -.038, ns
T2-T3: β= -.008, ns
T1-T2: β= -.042, ns
T2-T3: β= -.024, ns
Wu et al., 2006 USA Age 10-13 Maternal warmth and supportiveness: mother-child relationship, mutual trust and understanding, closeness (parent report). A child was considered an alcohol user if child or parent reported consumption of a unit of alcohol (not just sips). - Analysis of variance Maternal warmth and supportiveness:
3 waves in 5 years Parental discipline: various forms of punishment, including physical and verbal abuse, and withholding of affection (parent report). Measurement: lifetime and past-year alcohol use. Use (mean): 2.4
n= 1119 No use (mean): 2.3
Parental discipline:
Use (mean): 0.5%
No use (mean): 0.6%

Abbreviation: SES=socio-economic status.

a adolescent report, unless otherwise indicated; b no further specification; c 5th grade= age 10-11; d analysis sample unclear.

Visser et al.

Visser et al. BMC Public Health 2012 12:886   doi:10.1186/1471-2458-12-886

Open Data