The association between parental attitudes and alcohol consumption and adolescent alcohol consumption in Southern Ireland: a cross-sectional study

A cross-sectional survey was undertaken. This involved distributing a questionnaire to a sample of second-level students in the Kanturk-Mallow local electorate area (Southern Ireland). The survey was carried out during the academic year 2014–15.

The sample consisted of students in their final 2 years of second level education from a local electorate area in Southern Ireland, which has an estimated general population of 28,404 and a total secondary school population of 2119 across 6 year groups. According to the 2011 national census this area has a breadth of socioeconomic classes and incomes [34]. All eight secondary schools in the locality were included yielding a sample population of 982 students in their penultimate and final year.

These 982 students were asked to bring an envelope home to their parents containing a consent form and two parent surveys. Parents completed their survey and returned it in the sealed envelope provided. The consent forms were separated from the parent questionnaires on collection. This consent form outlined the aim of the study, the nature of the survey and provided detail on who was organising the research. Furthermore, it highlighted the voluntary nature of the research and an individual’s right to discontinue at any point. The front page of the survey included information on the aim, the nature of participation, who the researchers were and the benefits of completing the questionnaire. Parents consented both for themselves and their child. The child also signed the consent form in the presence of their parent or guardian. Students were surveyed in class. Prior to completion students were again advised of the aim of the study and the voluntary nature of participation. The surveys were confidential; each parent and pupil had a corresponding identification number on their survey to link results. The final page of both parent and pupil questionnaires provided information on individuals and organisations to contact in case of distress or upset as a result of taking the survey. Data collection occurred over a 2 week period during November 2014.

Topics in both parent and pupil surveys included alcohol use, self-reported height and weight, smoking status, mental health and well-being. The study utilised previously validated questions. Consumption questions were attained from the Alcohol Use Disorders Identification Test for Consumption (AUDIT-C) [35]. AUDIT-C takes the first three questions of the AUDIT questionnaire which focus on frequency and volume of alcohol consumption. The guidelines on low-risk alcohol consumption are lower in adolescents than those for parents, reflecting their increased vulnerability to alcohol-related harm. Based on previous national and international research, hazardous drinkers were defined as an AUDIT-C score of 5 or more amongst parents [4] and 4 or more amongst adolescents [35]. Parental attitude to alcohol questions were included in the parent survey [36, 37]. Questions on adverse consequences described in CLAN were utilised in the adolescent survey [38]. Personal wellbeing was measured using the Personal Wellbeing Index (PWI) in both adolescents [39] and parents [40]. Parental attitude questions were taken from Smyth et al. and Research New Zealand papers [36, 37]. Second-hand effect, smoking and body mass index questions were taken from the Survey of Lifestyle, Attitudes and Nutrition in Ireland (SLAN) [41] and the College Lifestyle Attitudinal National (CLAN) Survey [38]. Body mass index (BMI) was estimated from self-reported height and weight [42].

IBM Statistical Package for Social Science (SPSS) Statistics 20 was used for statistical analysis. AUDIT-C scores were calculated according to instrument guidelines [43, 44]. Analysis included descriptive, frequency and binary logistic regression analysis. Parental attitude questions were measured on a 5-point Likert scale. To increase statistical power, strongly disagree, disagree and neutral were combined as were strongly agree and agree. Hazardous alcohol consumption of the adolescent, as described using the AUDIT-C, was the dependent variable. A stepwise multivariate logistic regression was undertaken to investigate the association between parent consumption, behaviour and demographics and adolescent alcohol consumption. The first model investigated the association between adolescent alcohol consumption and parental sociodemographic factors. The second model included parental attitude. The final model included consumption, attitude and socio-demographic information.

The median age of adolescents was 17 years. 55.8 % were female. Over one-third of adolescents reported being hazardous drinkers (34.2 %), more males (39.6 %; 95 % CI 31.9–47.3 %)) than females (29.9 %; 95 % CI 23.5–36.2 %)). 68.2 % of these hazardous drinking adolescents were under the age of 18. These results are shown in Table 1.

More mothers (55.8 %) took part in the survey than fathers (44.2 %). Forty-seven percent of all parents surveyed were hazardous drinkers. The median age was 50 years for fathers and 48 years for mothers. Fathers were more likely to be hazardous drinkers (50 %) compared to mothers (46.8 %). These results are shown in Table 2.

Over 90 % of parents disagreed with allowing their adolescent to get drunk. However, almost one-fifth of parents said they would not be worried if their adolescent consumed four pints of alcohol once a month. A large number of parents (42.8 %) agreed with permitting their adolescent to drink on special occasions and almost 18.2 % believed it would be okay for another parent to provide alcohol to their adolescent under supervision.

When fathers’ attitude toward alcohol was investigated further in univariate analysis it was found that if an adolescent is a hazardous drinker, their father is more likely to agree that it’s ‘Okay for their adolescent to get drunk sometimes’ (14.8 % vs. 2.8 %). Similarly, fathers of hazardous drinking adolescents were more likely to agree that ‘Getting drunk is part of having fun as an adolescent’ (13.5 % vs. 3.9 %) or that ‘It’s okay for pupils to drink on special occasions’ (65.2 % vs 29.8 %), when compared to adolescents that do not drink hazardously. Furthermore, if the adolescent is a hazardous drinker their father is more likely to ‘Allow another parent to supply their adolescent with alcohol’ (32.6 % vs. 9 %). These results are shown in Table 3.

The mother of an adolescent reporting hazardous drinking is more likely to agree that “it’s okay for their adolescent to get drunk sometimes” (5.9 % vs. 1.4 %) or that “getting drunk is part of having fun as an adolescent” (9.2 % vs. 2.8 %). Furthermore, 47.5 % of mothers of hazardous drinking teenagers agreed that adolescents should not drink at all compared to 63.3 % of mothers with non-hazardous drinking teenagers. Mothers who felt it was ok for their child to drink on special occasions were more likely to have hazardous consuming teenagers (62.5 %). These mothers were more likely to agree with introducing alcohol in the home (39 %) and not be concerned if there adolescent “drank four pints once a month” (27.4 %). These results are shown in Table 3.

A stepwise multivariate logistic regression was undertaken to investigate the association between parent consumption, behaviour and demographics on adolescent alcohol consumption. Two hundred twenty-one individuals are included in the final, fully adjusted model. The fully adjusted model reports that adolescents whose fathers reported hazardous alcohol consumption were almost three times as likely to report hazardous alcohol consumption themselves (O.R = 2.90, 95 % CI: 1.32–6.35). Parental attitude toward drinking on special occasions was also positively associated with adolescent hazardous alcohol consumption. The father permitting drinking on special occasions results in adolescents being over four times more likely to report hazardous alcohol consumption (O.R = 4.60, 95 % CI: 1.19–11.18). This is over three times more likely if mothers report the same attitude (O.R = 3.79 95 % CI: 1.31–10.94). These results are shown in Table 4.

The findings advocate the need for further public policy. In tackling alcohol consumption a variety of policy measures are required. Previous research has noted the importance of a minimum age for consumption. Exposure to alcohol in adolescence has detrimental effects on brain development and intellectual capabilities while it also increases the likelihood for later alcohol dependence [51, 52]. However, the current study notes that the majority of hazardous adolescent drinkers (68.2 %) were under the legal age of consumption. Therefore there is an urgent need to enhance our understanding of how adolescents acquire alcohol alongside improving enforcement of the law regarding under age alcohol sales. Moreover, previous research has noted the importance of restricting alcohol advertising and marketing which is aimed directly at adolescents [53‐55]. The promotion of alcohol in areas such as sport, social media and music is increasingly prevalent [55]. Well-established evidence shows alcohol marketing increases the likelihood that adolescents will start to drink or will drink more if they are already using alcohol [54]. The evidence surrounding minimum unit pricing highlights its protective effects [56]. In addition, introduction of a parent programme to equip parents with the development of a culture of nurturing behaviours, establishing boundaries, and parental monitoring has shown promising results [48].

This study had a number of strengths and weaknesses. The strengths of this study were that an electorate area in Southern Ireland was sampled and this is unlikely to differ significantly from the general population. Furthermore, every student in the final two years of second level education within the electorate area were invited to participate. The proportions of male and female participants was broadly similar to national figures where males represent 48 % of the population compared to 52 % of females [57]. In addition, age demographics were similar to national statistics [57]. Finally, the proportion of numbers who participated in the research across the penultimate and final years were broadly similar to national levels [57]. Moreover, our findings regarding parental and adolescent hazardous drinking levels were similar to recent national research [1, 58]. Both parent and adolescent surveys utilised previously validated questions from international studies. This study included two parent questionnaires so that both the influence of the mother and the father could be recorded.

However, a low response rate (37 %) was attained. Selection bias cannot be ruled out as no information on those who chose not to complete the survey is available. Furthermore, response rates for parents are unavailable due to the fact that population size information is unavailable. In addition, due to low response numbers particular characteristics could not be investigated due to infrequency of reporting i.e. single mothers, education status etc. It is possible that the survey may have produced socially desirable responses, certain results such as BMI were different from similar previous studies [41, 59‐61]. However the results of participants were confidential. Finally, as parent and adolescent alcohol consumption were measured simultaneously, reverse causation cannot be dismissed [62].