  |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
12-December-2003, Vol 116 No 1187
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Parental smoking and related behaviours influence adolescent
tobacco smoking: results from the 2001 New Zealand national survey of 4th form
students
Robert Scragg, Murray Laugesen and Elizabeth
Robinson
Most adult tobacco smokers began smoking during their
teenage years. Previous research has contrasted the relative importance of peer
influences from teenage friends with those from parents in determining whether
or not teenagers smoke. For example, findings from the National Longitudinal
Study of Adolescent Health in the United States showed that risk of adolescent
smoking is influenced more by the smoking behaviour of friends than of
parents,1 consistent with an earlier review
concluding that friends were a more important determinant of smoking behaviour
than parents.2 In contrast, a Norwegian study
reported that, while smoking by friends was an important predictor of smoking
behaviour during adolescence, smoking by mothers was the most important
long-term predictor of smoking as students progressed into
adulthood.3 A US cross-sectional study found
that adolescents of parents who quit smoking were less likely to be
smokers,4 while parental smoking was the
strongest predictor of smoking in a school sample of low smoking prevalence
(0.3% weekly smokers) from eastern China.5
Cross-sectional surveys from Australia have consistently reported that parental
(or family) smoking is a risk factor for adolescent
smoking.6–8
Inconsistent findings about the strength and effect of
parental smoking on the risk of adolescent smoking have emerged from previous
New Zealand studies. A Wellington study at two co-educational secondary schools
found that smoking by friends, siblings and parents were all significantly
associated with student smoking, although parental smoking had the weakest
association.9 Repeat surveys at a Wairoa high
school reported that maternal smoking was associated with student smoking in
both 1975 and 1989.10 A larger national survey
of 4th form students at 99 schools in 1992 also found that parental smoking was
associated with an increased risk of student
smoking.11 These findings contrast with results
from cohort studies in Dunedin and Christchurch, collected in the 1980s, which
show no effect of parental smoking behaviour on adolescent smoking after
controlling for smoking by friends.12,13 A
Rotorua study, which did not examine parental smoking, reported that student
smoking was associated with peer smoking in males (Maori and European) and in
Maori females, and with sibling smoking in European
females.14
The influence of parents on adolescent smoking is not
confined to the direct effect of their own smoking behaviour. Australian studies
have found that the amount of pocket money provided to children and parental
supervision of adolescent leisure time were both related to adolescent
smoking.7,8 Recent results from the 2000
national 4th form survey in New Zealand showed that the provision of more than
$10 per month pocket money explained 30% of smoking by girls and 25% by
boys.15 Parents are also a major source of
cigarettes for adolescents.16
Since 1997, annual surveys of 4th form students (aged
14–15 years) have been carried out in New
Zealand.17 In 1999, the survey was extended to
all schools with 4th form students, so that the survey sample (about 30 000
students each year) has ethnic subgroups with a wide variation in smoking
prevalence. This variation ranges sevenfold, from very low levels in Asian girls
to the highest levels in Maori girls.18 The
purpose of this study is to examine, in detail, the role of parental smoking on
adolescent smoking. In particular, we wish to determine whether:
MethodsDetails of previous national
surveys of tobacco smoking and purchasing by 4th form students, carried out in
November of 1992, and yearly during 1997–2000, have been
reported.11,15,17,18 All New Zealand schools
with 4th form students were invited to participate in a further survey carried
out in November 2001. The school response rate was 71.9% (332 out of 462
approached).
Students anonymously answered a one-page questionnaire
on age, sex, ethnicity (self-assigned) and smoking behaviour (frequency of
smoking, source of cigarettes). Smokers were asked if they acquired their
cigarettes from any of the following sources: bought themselves, received from a
family member, or received from a friend or someone else. Students were also
asked about parental smoking, whether people smoked inside the home, and how
much pocket money they received in a usual month (30 days). The Ministry of
Education classification of schools by socioeconomic decile (from the low of 1
to high of 10) was used to code students for socioeconomic status
(SES).19 Consent for the survey was obtained
from school principals in place of parents. The Ministry of Health Auckland
Ethics Committee gave permission to survey without formal referral to the
Committee.
A total of 31 002 questionnaires were returned by
schools with 43 696 students on their rolls (70.9% student response). Analyses
were restricted to 29 271 students who were 14 and 15 years old, with known sex,
ethnicity, student smoking and parental smoking status. Excluded were students
of: age 13 years (n = 273), age 16 years (455), other ages (26) or unknown
(128); unknown sex (91); unknown ethnicity (288); unknown student smoking status
(343); and unknown parental smoking status (127).
All statistical analyses were made using SUDAAN
(Release 7.5.6, 2000), which corrects standard errors and confidence intervals
for any design effect from clustering of students by school. Unconditional
logistic regression and logit models for ordinal and nominal outcomes were used
to estimate adjusted odds ratios, which were converted to relative
risks.20 In ethnic comparisons,
‘other’ students (n = 444) have been combined with 19 812 European
students. The population attributable risk was calculated by estimating the
attributable proportion for the exposed cases within each exposure category
using standard methods.21
ResultsThe survey sample contained 14 930
girls (Maori 2563, Pacific Islands 948, Asian 1171, European 10 248) and 14 341
boys (Maori 2442, Pacific Islands 898, Asian 993, European 10 008). The
prevalence of daily smoking varied with ethnicity. In girls there was a tenfold
variation, from Maori 34.4%, Pacific 19.4%, European 11.4% to Asian 3.3%; and
two- to threefold variation in boys, from Maori 19.1%, Pacific 14.1%, European
10.0% to Asian 7.3%.
There were 3977 (13.6%) students who had both parents as
smokers, 7807 (26.7%) who had one smoking parent, and 17 487 (59.7%) with both
parents as nonsmokers. Table 1 shows how parental smoking was related to other
variables. The distribution of sex did not vary with parental smoking category
(p >0.05). However, students with both parents smoking were more likely to be
Maori (37.2%), compared with students with one parent smoker (23.3%) and with
neither parent smoking (9.7%); while the proportions of Asian and European
students were each highest for students with neither parent smoking (8.7% and
76.3%, respectively), compared with the other two parental smoking
categories.
Students with both parents smoking were more likely to be at
a school in the lowest two deciles (15.6%) compared with students with one
parent smoker (12.0%) and with neither parent smoking (6.1%). Students with both
parents smoking were more likely to have smoking in the house (74.3%), be given
>$50 pocket money per month (38.5%), and to be daily smokers (31.6%), than
students with one parent smoker (51.1%, 34.7% and 18.3%, respectively) and with
neither parent smoking (11.3%, 28.0% and 7.2%, respectively). In contrast, the
percentages of students smoking less than daily did not vary greatly with
parental smoking (Table 1).
Table 1. Relationship between number of parents in the
house who smoke and other variables
Logistic regression was used to run multivariate models to
investigate the effects of the variables in Table 1, plus parental smoking and
student age, on the risk of being a daily smoker. Ethnicity was found to have
strong interactions (p <0.0001) with parental smoking, pocket money and sex,
so ethnicity-specific analyses, for males and females combined, were carried out
(Tables 2 to 5). Age 15 years was most strongly related to risk of daily smoking
in European students (Table 5) compared with Maori, Pacific or Asian students
(Tables 2 to 4). The risk of daily smoking associated with female sex, compared
with male, was most increased in Maori students (relative risk (RR) = 1.88,
Table 2), followed by Pacific (RR = 1.45, Table 3) and European (RR = 1.20,
Table 5), but decreased in Asian students (RR = 0.43, Table 4). School SES
decile was generally inversely associated with risk of daily smoking in all
ethnic groups, particularly among European students where the relative risk was
50% higher in deciles 1 and 2 compared with deciles 9 and 10 (Table 5). Parental
smoking, particularly by both parents, compared with neither parent smoking,
showed the greatest effect on risk of daily smoking among Asian students (RR =
6.64, Table 4), a moderate effect among European (RR = 3.11, Table 5) and
Pacific (RR = 3.05, Table 3) students, and lowest effect among Maori (RR = 1.74,
Table 2). The relative risk associated with living in a house where people
smoked was also highest in Asian students (RR = 2.99, Table 4) compared with
other ethnicities. The amount of pocket money was positively associated with
risk of daily smoking in all ethnic groups, with the relative risks for the
highest pocket money category (>$50 in the last 30 days), compared with the
lowest ($0–10), being higher for Asian (RR = 3.32, Table 4) and European
(RR = 2.45, Table 5) students than for Maori (RR = 1.47, Table 2) and Pacific
(RR = 1.46, Table 3).
Table 2. Maori students – adjusted relative risk
(95% confidence intervals) of daily smoking associated with demographic,
parental and related variables
*adjusted for all other variables in the table;
calculated from odds ratios estimated by logistic regression
Table 3. Pacific Island students – adjusted
relative risk (95% confidence intervals) of daily smoking associated with
demographic, parental and related variables
*adjusted for all other variables in the table;
calculated from odds ratios estimated by logistic regression
Table 4. Asian students – adjusted relative risk
(95% confidence intervals) of daily smoking associated with demographic,
parental and related variables
*adjusted for all other variables in the table;
calculated from odds ratios estimated by logistic regression
Table 5. European students – adjusted relative
risk (95% confidence intervals) of daily smoking associated with demographic,
parental and related variables
*adjusted for all other variables in the table;
calculated from odds ratios estimated by logistic regression
The data in Tables 2 to 5 show that the relative risks of
daily smoking by students associated with parental smoking, smoking in the house
and amount of pocket money remain significantly different from 1.00 when
adjusting for each other, and therefore they have separate effects on the risk
of daily smoking. Of particular note, the effect of parental smoking remains
independent of the adolescent smoking risk associated with smoking in the
home.
The factors associated with the source of cigarettes were
examined in student smokers (Table 6). In the questionnaire they were asked
‘Where do you get your cigarettes?’ and could choose one or more of
the following options: ‘I buy them myself’, ‘From a family
member’, and ‘From a friend or someone else’. Smokers were
categorised into three groups according to the following priority system: the
first group included any students who indicated that they bought cigarettes for
themselves (n = 2719), the second included any remaining students who recorded
that they obtained cigarettes from a family member (n = 1759), and the third
included those receiving cigarettes from a friend or someone else (n = 6884).
Smokers who did not answer any of these options (n = 211) were excluded from
these analyses.
Table 6. Adjusted relative risk of buying cigarettes,
or getting them from a family member, compared with getting them from a friend
or someone else, among smokers
*adjusted for all other variables in the table;
calculated from odds ratios estimated by logistic regression
Students who bought cigarettes were compared with those who
received them from a friend or someone else (Table 6). In this table, row
percentages for the sources of cigarettes – buying themselves, from a
family member, or from a friend or someone else – are shown for each
exposure level. Unadjusted relative risks of daily smoking can be calculated
from the ratio of percentages; for example, the unadjusted relative risk of
15-year-olds buying cigarettes, compared with the reference category 14 years,
is 25.7% / 21.8% = 1.18. However, all relative risks shown in Table 6 are
adjusted for all other variables in the table. Fifteen-year-old students were
more likely to purchase cigarettes than 14-year-olds; as were Maori, Pacific and
Asian smokers compared with Europeans. Students at low SES decile schools were
less likely to purchase than those at the schools in the highest two deciles (9
and 10). There was a dose-response relationship between the number of smoking
parents and the risk of students purchasing their own cigarettes, with students
of both parents smoking being 63% more likely to do so than students of
non-smoking parents. The amount of pocket money was also positively associated
with the risk of purchasing cigarettes, that risk being 74% higher for students
receiving >$50 per month, compared with students receiving
<$10.
The risk of student smokers receiving cigarettes from a
family member, when compared with those who received them from a friend or
someone else, was higher for Maori compared with all three other ethnic groups,
highest in low SES decile schools, and two and a half times higher if both
parents smoked than if neither parent smoked. However, the amount of pocket
money was unrelated to the risk of receiving cigarettes from family (Table
6).
Collectively, the analyses in Table 6 indicate that amount
of pocket money is a risk factor for the self-purchasing of cigarettes, and that
self-purchasing by student smokers, or receiving cigarettes from family members,
is more common in families where both parents smoke. The public health
significance of the combined effect of parental smoking, the related parental
behaviours of pocket money amount and the decision about whether people smoke in
the house, were examined in Table 7 by calculating ethnicity-specific
attributable risks for exposure to these three variables, either separately or
combined. The cut-off point for high pocket money was arbitrarily set at >$20
per month, equivalent to >$5 per week. The proportion of students exposed to
one or more of these three risk factors was highest for Maori students (91%),
followed by Pacific (80%), European (76%) and Asian (69%). However, the relative
risk associated with this combination variable was highest for Asian students
(RR = 14.74), so that the attributable risk was highest for Asian students
(91%), followed by European (67%), Maori (68%) and Pacific (55%). For all ethnic
groups combined, 67% of daily smoking could be explained by combined exposure to
one or more of the following factors: parent smoking, pocket money >$5 per
week, and smoking in the house.
Table 8 contains relative risks for daily smoking in Maori,
Pacific and Asian students, compared with European, which show the contribution
to the increased smoking risk in Maori and Pacific students due to ethnic
differences in exposure to various risk factors. In comparison with the relative
risk of daily smoking adjusting for age and sex only, further adjusting for
parental smoking decreased the increased risk of daily smoking in Maori students
by about 40% (change in RR from 2.51 to 1.89). The increased risk in Pacific
students decreased by one third (change in RR from 1.56 to 1.36). Additional
adjustment for parental smoking had little effect on the relative risk for daily
smoking in Asian students compared with European (from 0.47 to 0.51).
Additional adjustment for school SES decile also decreased
relative risks for daily smoking, compared with Europeans, in Maori from 2.51 to
2.26, and in Pacific students from 1.56 to 1.31. Thus, ethnic differences in
parental smoking and school SES decile both partly explain the increased risk of
daily smoking in Maori and Pacific students compared with European.
Table 7. Adjusted relative risk of daily smoking by 4th
form students associated with exposure to: parental smoking and/or receipt of
pocket money >$20 per month and/or smoking in the house, by
ethnicity
*adjusted for age and sex, calculated from odds ratios
estimated by logistic regression;
†percentage of attributable smokers in
each ethnic group
Table 8. Adjusted relative risks of daily smoking in
Maori, Pacific and Asian 4th form students, compared with European
*calculated from logistic regression odds
ratios
DiscussionThe results of this study indicate
that parental behaviours, including smoking, the amount of pocket money provided
to children, and whether people smoke in the home, explain a significant
proportion (67%) of daily smoking by adolescents, and are primary determinants
of the elevated smoking prevalences in Maori and Pacific students. The
dose-response associations with daily adolescent smoking observed for parental
smoking and amount of pocket money (Tables 2 to 5) support the possibility that
these associations are causal.
Our observation of an association between adolescent and
parental smoking is consistent with previous research, both international and in
New Zealand.3–7,9–11 A novel
finding from our study is the variation in parental effect between ethnic
groups, with the strength of the effect, which was highest in Asian students and
lowest in Maori, being inversely related to the prevalence of student smoking in
the subgroup (Tables 2 to 5). This finding is consistent with a recent
publication from China that reported parental smoking was the strongest
predictor of teenage smoking in a student sample where only 0.3% were regular
smokers (weekly or more often).5 Our
observation that parental smoking is most strongly associated with daily
smoking, rather than with less frequent smoking (Table 1), may explain why the
Dunedin cohort study, which defined children as smokers if they had smoked at
any time in the last two years, failed to report an independent effect from
parental smoking.12
This study has also shown that parental smoking behaviour is
associated with other factors that increase the risk of adolescent smoking.
First, parents who smoke are more likely to give high amounts of pocket money
(>$50 per month) to their children (Table 1), while the amount of pocket
money is a risk factor for adolescent smoking in all four ethnic groups (Tables
2 to 5). The latter finding confirms previous research on the positive
association between amount of pocket money and risk of adolescent
smoking.7,8,15,22,23 Second, parents who smoke
are more likely to allow smoking in the house, which is an independent risk
factor for daily smoking (Tables 1 to 5). Third, parents who smoke are more
likely to provide cigarettes to their children or have children who purchase
their own cigarettes (Table 6). A recent, US, qualitative study of 68 adolescent
smokers provides insight into how parental smoking increases the risk of
adolescent smoking.16 In this study parents
were found to be the primary source of cigarettes for children at the onset of
smoking, since children often started smoking using half-smoked cigarettes left
in ashtrays by relatives, or by stealing cigarettes from their parents; while
the practice of students spending school lunch money, supplied by parents, on
purchasing cigarettes confirmed the importance of regular access to money in
increasing the risk of smoking.
Parental smoking was identified in this study as a major
factor explaining the increased smoking risk among Maori and Pacific
adolescents. This finding is consistent with an earlier report that
identification with Maori culture, which typically is provided to children by
parents, was a risk factor for smoking among Maori
students.24 Ethnic differences in socioeconomic
status also contributed to the increased risk of daily smoking in both Maori and
Pacific students.
The threats to the validity of this study include its
cross-sectional design, which cannot distinguish cause and effect. However,
while it is possible that adolescent smoking behaviour could determine the
amount of pocket money received, rather than the other way around, we can be
certain that parental smoking precedes adolescent smoking in all or most cases.
Any error in the measurement of student smoking status by our questionnaire is
likely to have been non-differential, given the cross-sectional study design, in
which case we may have underestimated relative risks associated with daily
smoking. Further, measurement error is likely to have been contained in our
measure of parental smoking, which did not allow for single-parent and
extended-family households. Another study weakness is our inability to control
for the effects from students peers, which were not recorded in the
questionnaire and represent a further limitation of the study. Peer smoking
could be a confounder of parental smoking only if they were associated with each
other. Previous New Zealand studies have examined the separate effects of
parental and peer smoking on adolescent
smoking,9 but only the Dunedin and Christchurch
cohort studies have controlled for the effect of peer
smoking.12,13 Further research is required to
clarify this in the New Zealand context, given the important public health
implications of our findings.
The very high proportion of daily adolescent smoking
explained by parental smoking and related behaviours suggests that parents have
a central role in the prevention of adolescent smoking. Despite the conflicting
evidence about the relative importance of peer and parental influences on
smoking initiation in adolescence, preventive efforts against adolescent smoking
have focused on minimising the harm caused by fellow students, while the
potential role of parents has been neglected.25
However, our findings suggest efforts that target the role of parents should be
pursued. These could include health promotion strategies, such as television
campaigns that advise parents about the possible benefits of banning smoking in
the home.25 The data in Table 1 indicate that
in two thirds of homes that allow smoking indoors, one or both parents are
nonsmokers. Thus, any media campaign against smoking inside homes is likely to
be well received so that some reduction should be achievable. Other health
promotion strategies include advising parents not to provide cigarettes to their
children, and limiting pocket money, which, if it is going up in smoke, will
find favour with most parents, as few wish to pay for their children to
smoke.15 In addition, efforts to support
parents in attempts to stop smoking, such as the Quit campaigns, should be
strengthened.4
Author information:
Robert Scragg, Associate Professor in Epidemiology, School of Population Health,
University of Auckland; Murray Laugesen, Public Health Physician, Health New
Zealand; Elizabeth Robinson, Biostatistician, School of Population Health,
University of Auckland, Auckland.
Acknowledgements:
The survey was carried out by Action on Smoking and Health (ASH). The New
Zealand Ministry of Health provided funds. We thank the students and staff from
the participating schools.
Correspondence:
Associate Professor Robert Scragg, Epidemiology and Biostatistics Section,
School of Population Health, University of Auckland, Private Bag 92019,
Auckland. Fax: (09) 373 7624; email: r.scragg@auckland.ac.nz
References:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|