 
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
20-June-2003, Vol 116 No 1176
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Smoking behaviour and expectations among Auckland
adolescents
Judith McCool, Linda Cameron, Keith Petrie and Elizabeth
Robinson
Smoking uptake during adolescence continues to present a
perplexing public health issue in New Zealand. Despite the impact of public
health initiatives aimed at reducing smoking uptake and increasing quit rates
among adults, a resistant core of adolescents continue to smoke. Research
testifies that over the past decade there have been important changes in the
epidemiology of adolescent smoking; specifically, the ages at first uptake and
at establishment of persistent smoking behaviour have, over time,
decreased.1 Smoking experimentation remains a
consistent characteristic of adolescence, and as the child moves through the
period of adolescence the likelihood of smoking increases. Some of this
behaviour can be explained in terms of the social, cognitive and physical
developmental changes that typically occur between the ages of 12 and 16
years.2,3
To date there are few New Zealand studies that compare the
smoking behaviour, expectations, and familial smoking behaviour of Form Two
(mean age 12 years) and Form Six (mean age 16 years) students. Adolescents from
these two age groups were assessed in the present study for the following
reasons. As the age of smoking onset continues to decrease, the pre-adolescent
years are now widely accepted as a critical stage in the development of
formative smoking behaviour.1,2 Similarly, the
mid-teen years are also critical in terms of establishing persistent smoking
behaviours. Previous studies of adolescent smoking in New Zealand,
published4 and unpublished (Bandranayake and
McCool, 1997, unpublished data), have predominantly addressed the smoking
behaviour of Form Four students. Although these studies offer a valuable tool
for monitoring trends in adolescent smoking, we are currently lacking
information on smoking initiation during the early adolescent years. This
information is critical for the development of effective initiatives aimed at
reducing smoking uptake among young people.
MethodsSample
A stratified sampling strategy was used to obtain two representative
samples of Auckland Form Two and Form Six students. The combined sample was
stratified according to school decile rank, gender and ethnicity, which are
factors widely accepted to be important predictors of current and future
smoking.1,2 As a result, a two-stage sampling
procedure was undertaken. Two secondary and three primary or intermediate
schools from each of the ten decile ranks (1–10) were randomly selected
and invited to participate in the survey. In total, 10 secondary schools and 15
primary or intermediate schools participated in the study. Two primary schools
declined to participate on the grounds that the timing of the study was
inconvenient. Primary and intermediate schools were over-invited in order to
ensure that a comparable number of students as from secondary schools would be
included in the study. Not all schools that agreed to participate were included
in the survey due to the high response rate. Schools that had not arranged a
date for the survey to be undertaken at the stage of sample saturation were not
included in the final sample. In total, 68% of intermediate and primary schools
and 76% of secondary schools selected took part in the
survey.
Sample profile Of the total sample, 48% of students (n = 1464) were from Form Two classes and 51.8% (n = 1576) were from Form Six. The median age of Form Two students was 12 years; within the Form Six sample the median age was 16 years. In the majority of cases students opted to identify with one ethnic group (2793, 92%); a smaller proportion identified with two (215, 7%) or more (11, 0.4%) ethnic groups. Twenty two (0.7%) students did not identify with any nominated ethnic group. Procedure The survey was conducted early in the first school term (February 2000). Participation in the survey required passive consent, which enabled parents to provide input on the study if desired whilst preserving the representativeness of the sample and integrity of the study. All students were granted consent to participate in the survey. The principal researcher distributed the questionnaires to students during class time or school assembly and was present during the completion of the questionnaires. All schools were surveyed within a two-month period. Student attendance rates on the days of the survey were not formally collected. However, the principal researcher emphasized to each school liaison staff member the importance of selecting a survey date on which the majority of students would be present at school (ie, avoiding school extra-curricular events, work experience etc). Measures The smoking-behaviour and parental-smoking items were developed from standardized smoking-behaviour questions in previous studies assessing the current and susceptible smoking status of a population (Bandranayake and McCool, 1997, unpublished data).2,4 Smoking behaviour was assessed with the item ‘Have you ever smoked a cigarette?’ (1 = yes, 2 = no). Daily smoking was assessed through the item ‘How often do you smoke now?’ (five-point scale: 1 = at least once a day, 5 = never). A separate variable for daily smoking was subsequently created where 1 = daily smoker and 0 = non-daily smoker. Smoking expectations were assessed using the item ‘How likely is it that you will smoke a cigarette in the next year?’ (five-point scale: 1 = yes, definitely, 5 = definitely not). Items were re-coded to create a dichotomous variable where 1 = low smoking expectations and 0 = high smoking expectations. Results of the analyses of this outcome variable were unaffected by changing from a continuous to a dichotomous variable. The dichotomous variable was created by combining responses so that ‘probably not’ and ‘definitely no’ = 0, and ‘probably yes’ and ‘definitely yes’ = 1. Parental smoking behaviour was assessed with the item ‘Do either of your parents smoke?’ (1 = yes, 2 = no). Smoking inside the home was assessed with the item ‘Do people smoke inside your house?’ (1 = yes, 2 = no, 3 = sometimes). Smoking inside was subsequently re-coded to create a dichotomous variable for analysis (eg, smoking permitted inside home (yes or sometimes = 1, no = 0)). Logistic regression analyses were used to assess age, gender, smoking status, and ethnicity differences in tobacco use, daily smoking, parental smoking, and smoking inside the home. Across each measure interaction effects were assessed between age level (Form) and the other independent variables. Where an interaction effect was observed for age level the groups were assessed independently. Through analysing the samples independently, comparisons between the Form Two and Form Six samples, across the range of outcome measures, were assessed. ResultsAnalyses were conducted to assess
the effect of age level, gender, ethnicity, smoking status, school decile rank,
parental smoking and smoking inside on ever-smoking status, daily smoking and
smoking expectations for the future. Where an interaction effect was observed
between Form and gender, ethnic group, decile rank, parental smoking or smoking
inside the home, the two Forms were assessed independently. Table 1 presents the
data from the logistic regression analyses for each outcome measure. Where an
interaction effect was identified further analyses to explain the result of
these effects are also presented. Throughout the study a significance level of
0.01 was assumed to control for Type 1 errors.
Ever smoker Table 1
presents data from the logistic regression analyses for independent variables on
ever-smoker status. Consistent with expectation, Form Six students (64.4%) were
more likely to report having smoked in the past compared with Form Two students
(28.8%). As an interaction effect was observed between Form and gender (OR =
1.1, 95% CI 1.0–1.2, p <0.01), the two age groups were subsequently
assessed independently. A difference in reports of ever-smoking behaviour was
identified between boys and girls within the Form Two sample, with more boys
(13.6%) smoking than girls (10.3%), which approached statistical significance, p
<0.05. When the Form Six sample was assessed, the observed difference between
male (36%) and female (37%) ever-smoking behaviour was not statistically
significant. Across the total sample, a significant difference between ethnic
groups was also identified, with Asian students reporting lower ever-smoking
rates compared with all other ethnic groups. This pattern was also observed when
the groups were assessed separately. Fewer Form Two Asian students (2.4%)
reported having tried a cigarette in the past compared with Pakeha/European
(13.3%), Maori (20.6%), Pacific (15.2%) and other ethnic groups (7.1%). Within
the Form Six sample, fewer Asian students (21.4%) were ever smokers compared
with Pakeha/European (43.9%), Maori (44.1%), Pacific (33.4%), and students from
other ethnic groups (28.7%).
School decile rank was also found to be significantly
associated with ever-smoking behaviour, with students from higher-decile schools
more likely to report having smoked in the past (28%) compared with those from
lower-decile schools (23%). Parental smoking status was also associated with
ever-smoking behaviour. Students who reported that their parents were smokers
(31%) were more likely to have smoked in the past compared with those who did
not (21%). Similarly, students who reported that smoking was permitted inside
their home (38.3%) were more likely to be ever smokers than those who did not
(60%).
Table 1. Smoking behaviour and expectations among Form
Two and Form Six students
*among smokers;
†mean
(SD). NB: Data for ten students missing.
Daily
smoking Table 2 presents data
from the logistic regression analyses for independent variables on daily smoking
behaviour. Daily smoking rates were significantly higher among Form Six (10.2%)
than Form Two students (4.6%). Contrary to expectation, a significant effect for
gender was not detected within either the Form Two or Form Six samples. A
significant difference in daily smoking rates was observed, with Asian students
(3.8%) significantly less likely to report being a current smoker compared with
Pakeha/European (8.4%), Maori (5.6%), Pacific (8.6%), and students from other
ethnic groups (7.3%).
Table 2. Summary of logistic regression analyses for
daily smoking
A significant main effect was noted for school decile rank,
whereby students from a higher-decile-rank school were more likely to be daily
smokers. When the two groups were assessed independently, this effect was
consistent for the Form Two sample (OR = 2.2, 95% CI 1.2–3.9, p <0.01)
but was not evident within the Form Six sample. An interaction effect was also
identified between form and smoking inside the home (OR = 2.5, 95% CI
1.1–5.4, p <0.01); therefore, the groups were assessed independently.
Although a significant main effect was observed within the Form Two sample (p
<0.01), no significant main effect was identified for the Form Six sample.
Accordingly, students who reported that smoking was permitted in their home were
more likely be daily smokers than those who did not.
Smoking expectations
Table 3 presents data from the logistic regression analyses for
independent variables on students’ expectations of smoking in the future.
A significant main effect for age level was observed, with 36.5% of Form Six
students compared with 17% of Form Two students reporting they anticipated being
a smoker in the future. Significant differences in expectations of smoking in
the future were also identified between all ethnic groups. Specifically, fewer
Asian students (12%) expected to be smokers in the future compared with
Pakeha/European (33%), Maori (34%), Pacific (26%) and students from other ethnic
groups (17.4%).
Students who reported higher smoking expectations for the
future were more likely to belong to higher-decile schools (52.7%, p <0.001).
When the groups were assessed independently, this pattern of effect was
consistent for the Form Two students (65%) (OR = 2.1, 95% CI 1.5–2.8, p
<0.001), but was not evident within the Form Six sample (47%). Having a
parent who smokes was also found to be associated with positive smoking
expectations, p <0.01. This effect was found within the Form Six sample (OR =
0.67, 95% CI 0.52–0.86, p <0.01), but not within the Form Two sample.
In addition, students who reported that smoking was permitted in their home
(18%) were more likely to have positive smoking expectations for the future than
those who did not (80%, p <0.001).
Table 3. Summary of logistic regression analyses for
smoking expectations
DiscussionThis paper reported on the
relationships of age level, gender, smoking status, school decile rank,
ethnicity, parental smoking and smoking within the home with Form Two and Form
Six students’ current smoking, daily smoking, and smoking expectations for
future. Before discussing the implications of these results, the limitations of
this study design are acknowledged. Specifically, the sample was derived from a
regional population rather than a national sample, which would have enabled
useful comparative analyses with existing national databases, such as the
Wellington and the ASH surveys of Form Four students (Laugesen and Scragg, 2002,
ASH New Zealand, unpublished data).5,6 However,
this study, which assessed students from schools in the greater Auckland region,
provides a useful, detailed picture of trends within a specific socio-geographic
area. Another concern was related to school attendance records, which were not
formally collected on the school survey days. It is possible that this omission
may result in conservative reports of smoking behaviour, particularly among the
Form Six students, who may be absent for reasons such as sports events or work
experience. In addition, the data were not assessed as a cluster sample; the
school effect could not be assessed. Decile rank was included as an independent
variable within the models to control for school socioeconomic status.
This survey is one of the few conducted in New Zealand that
provides evidence of early smoking initiation and daily smoking behaviour among
Form Two students and Form Six students. Tobacco use among younger adolescents
was evident, with a considerable proportion (29%) of Form Two students having
already smoked, and 5% who reported smoking on a daily basis. Consistent with
previous studies of smoking among older adolescents in New
Zealand,4–8 this study showed
substantially higher rates of smoking among older adolescents (Form Six students
in relation to Form Two students). Daily smoking rates, an important measure of
established smoking behaviour among young people, increased by 25% between Form
Two and Form Six. Essentially, this study provides further empirical evidence of
the magnitude of developmental changes that occur between early adolescence and
the mid-teen years that have implications for the conceptualization and design
of smoking prevention policy and initiatives.
This study found that the effect of gender was dependent on
age for the outcome measure of ever smoker only, with a greater proportion of
Form Two males reporting that they had smoked in the past compared with females.
This result may be an effect of under-reporting of smoking status, as the
tobacco-use measure used the standard ‘ever’ and ‘never’
smoking items and should have included the option ‘even just a puff’
to account for those who have tried smoking but not smoked an entire cigarette.
This discrepancy may mean that ever-smoker rates are possibly underestimated,
especially within the younger age group. Although recent New Zealand studies
have found that girls are significantly more likely to smoke than boys, this was
not evident within the present study.5–7
It has been reported that daily smoking among 14- and 15-year-old females
significantly exceeded that for boys in the 1992 and 1997 surveys and in the
most recent survey conducted in 2001 (Laugesen and Scragg, 2002 unpublished
data).4 It is possible that the effect for
gender that was found only within the Form Two sample may be a factor associated
with the different ages of the samples and the demographic profile of this
Auckland regional sample (ie, different ethnic distribution of sample).
Moreover, this finding reiterates the potential for intervention within this
younger population group through reinforcing the benefits of remaining smoke
free.
Consistent with previous studies, Form Two and Form Six
Maori and Pakeha/European students reported higher ever- and daily-smoking rates
compared with students from all other ethnic groups. Conversely, Asian students
reported lower ever-smoking behaviour, daily smoking and smoking expectations.
Previous research has consistently found that Maori report higher smoking rates
compared with other ethnic groups.6,7 This
discrepancy in smoking rates among the Maori population has been partially
explained in terms of the impact of the colonization of New Zealand by the
Europeans, during which time tobacco products were introduced to the indigenous
Maori population.8 The impact of tobacco use on
the health status of Maori has been profound and, accordingly, should continue
to be a priority in terms of smoking cessation and prevention initiatives aimed
at both young people and adults. In addition, the number of Pacific students
ever and daily smoking increased by two thirds between Form Two and Form Six,
suggesting a need for interventions targeted appropriately to these
groups.
Expectations of future smoking varied significantly by age
level, ethnicity and smoking status. Being a sixth former and a current smoker
increased the expectation of being a smoker in the future, a finding which was
consistent with that identified by Laugesen and Scragg (2002, unpublished data).
In this study smokers were found to be significantly more likely to intend to
smoke in the future compared with non-smokers. Similarly, older adolescents were
more likely to be smokers and, therefore, more likely to expect that they will
smoke in the future or at least hold ambivalent attitudes towards their
likelihood of smoking in the future. Romer et al found that young people start
smoking with the intention to smoke for only a short period, expecting they can
quit when desired in the future.9 Qualitative
studies have provided useful analyses of the relevance of an adolescent’s
social world in the development of attitudes towards tobacco use. Specifically,
older adolescents are identified as being more likely to hold ambivalent
attitudes towards their personal smoking expectations and smoking in general,
generated in part through the increased prevalence of smoking among this age
group.10,11
Recent studies have emerged that challenge arguments
regarding the overriding effect of parental smoking on adolescent
smoking.12–14 It is suggested that
parental smoking is most influential as a predictor to adolescent smoking only
during the period of early adolescence. After this period, friends’
smoking attitudes and behaviours emerge as a stronger predictor of adolescent
smoking. In addition, smoking inside the home may also reflect the acceptability
of smoking within the family context. Overall, the higher level of smoking
acceptability identified in this study among Maori (30%) and Pacific students
(21%) supports the relevance of familial smoking norms as an important factor in
the pathway to smoking uptake among young
people.6 The sharp increase in smoking between
Form Two and Form Six Asian students may also suggest shifting patterns of
acceptable social and cultural behaviours among older Asian teenagers. This
differential effect may also reflect the transition between parental to peer
networks as predictive of smoking behaviour. Accordingly, public health
initiatives aimed at reducing adult smoking should continue to be supported as
an ongoing strategy to reduce adolescent smoking uptake.
School decile rank was assessed as a proxy measure of
socioeconomic status of the schools. The effect of decile rank on smoking
behaviour measure revealed that within the Form Two sample only students from
higher-decile-rank schools were more likely to be ever smokers, daily smokers
and have higher smoking expectations for their future. A recent study by Scragg
and colleagues reported a positive association between amount of pocket money
and cigarette smoking in both male and female
students.15 Although this relationship was
found to be independent of socioeconomic status, it is possible that students
from higher socioeconomic groups have access to cash to purchase cigarettes, or
alternatively socialize with adolescents or other adults (including parents) who
purchase cigarettes. Reeder et al also assessed school variables including
decile rank and sex composition, and concluded that these factors were
significantly associated with daily smoking. However, no effect was observed for
current smoking behaviour. Accordingly, it was suggested that differences in
‘school culture’ (including the school’s smoke-free policy)
are possibly associated with socioeconomic factors, which may affect the smoking
behaviour of students. Similarly, peer group and self-image factors, such as
sub-cultural factors and weight-control issues, may also need to be
considered.16
Despite the introduction of the New Zealand Smoke-free
Environments Act (1990), which has initiated smoke-free schools and an increase
in the price of cigarettes, the prevalence of smoking among adolescent
sub-populations remains high.17 Evidence
suggests that current tobacco-control initiatives, primarily those aimed at
reducing adult smoking and increasing the price of tobacco, have been effective
in slowing the prevalence of smoking among young
people.6,7 Future adolescent smoking research
needs to address the social and cognitive developmental differences between
adolescent girls and boys across these age
groups.3,8 The importance of continuing
research into the beliefs and perceptions young people hold towards tobacco
(usage, image, acceptability) can only complement and support the impact of
existing public policy initiatives aimed generally at the adult
population.
Author information:
Judith McCool, Postdoctoral Research Fellow, Department of Health Psychology,
University of Auckland; Linda Cameron, Senior Lecturer, Department of
Psychology, University of Auckland; Keith Petrie, Associate Professor,
Department of Health Psychology, University of Auckland; Elizabeth Robinson,
Biostatistician, School of Population Health, University of Auckland,
Auckland
Acknowledgements: We
thank the National Heart Foundation of New Zealand for their financial support
of this research. We are especially grateful for the assistance offered by
school staff members who helped with the administration of this study. This
study would not have been possible without the patience and generosity of the
3041 students who participated in this research.
Correspondence: Dr
Judith McCool, Department of Health Psychology, Faculty of Medical and Health
Sciences, University of Auckland, Private Bag 92019, Auckland. Fax: (09) 373
7013; email: j.mccool@auckland.ac.nz
References:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|