  |
|
|
||||
|
||||||
| Journal of the New Zealand Medical Association,
24-October-2003, Vol 116 No 1184
|
||||||
|
Differences in patterns of alcohol consumption between Maori
and non-Maori in Aotearoa (New Zealand)
Dale Bramley, Joanna Broad, Ricci Harris, Papaarangi Reid
and Rod Jackson, for the Alcohol Burden of Disease and Disability
Group
The relationship of alcohol consumption with morbidity and
mortality is well documented and is influenced by the pattern of
drinking;1–3 however, no such data exist
specifically for Maori.
The intent of this paper is to investigate relative
differences in patterns of consumption between Maori and non-Maori rather than
to document differences in the prevalence of consumption; therefore, we sought
all major available studies whether or not they
were based on representative population samples.
The results of this paper will inform future research that
examines the association between Maori alcohol consumption patterns and
health.
MethodsSurveys
Data from five large New Zealand surveys conducted since 1988 were
re-analysed by sex and age group to examine differences between Maori and
non-Maori. Inclusion criteria for studies included access to the data sets
comprising individual records, recording of self-identified ethnicity data,
adequate numbers of Maori and standardised reporting of alcohol consumption. As
the focus was on relative differences between Maori and non-Maori, it was
unnecessary for studies to be based on representative population samples as long
as all Maori and all non-Maori in the study populations selected had equal
opportunities for inclusion. Included were two national (New Zealand Health
Survey 1997 and the Sleep Survey 1999) and three population-specific studies
(Fletcher Challenge /University of Auckland Survey 1992, NZ Blood Donors Health
Study 1998–1999 and the Workforce Diabetes Survey 1988–1990).
Details of sampling, inclusion criteria, and data collection procedures are
available.3–8 It is unfortunate that the
National Alcohol Surveys conducted by Whariki and the Alcohol and Public Health
Research Unit at University of Auckland were not available (see
discussion).
Demographic
classification All data were
collected from the respondents at interview (or, in the case of the Sleep
Survey, by self-completed questionnaire). Ethnicity was classified and coded
according to self-identified ethnicity. Where more than one ethnic group was
recorded the New Zealand Census hierarchical categorisation of ethnic group was
used to classify all responses to one prioritised ethnic
group.9 In all analyses Maori were the
reference group to which non-Maori were compared.
Sex and age were recorded at interview/survey. Those
aged less than 18 years (because the law does not allow people aged under 18
years to buy liquor) or over 75 years (because numbers of people aged over 75
years were low) were excluded to achieve an age range of 18 to 74 years. In two
surveys this age range was further limited by the survey design: the Sleep
Survey (30–60 years only) and the Workforce Diabetes Survey (40–74
years only). Results are presented by sex and age group where age groups are
defined as 18–34 years, 35–49 years and 50–74 years.
Measures of alcohol
consumption Four different measures of alcohol consumption were
considered where available. Each survey asked how frequently participants drank
alcohol, and provided several response categories. In general, any participant
reporting drinking alcohol never or less often than monthly was regarded as a
non-drinker, those drinking more often than monthly were classified as drinkers.
Participants in the Sleep Survey were classified as non-drinkers if they drank
never or less often than once a week, because no longer time category was
available. Participants in the New Zealand Health Survey were classified as
non-drinkers if they reported not drinking alcohol in the last year.
Based on the reported frequency of drinking, we
calculated the number of days a year on which the respondent drank alcohol. Two
methods of calculation were employed: in the first we allocated zero days for
non-drinkers in order to represent frequency of consumption for all the
population; in the second we excluded non-drinkers from analyses in order to
represent frequency of consumption for drinkers only.
Where the survey obtained information about volume of
alcohol drunk on a typical occasion, we estimated the grams of ethanol consumed
per occasion. Where possible, we estimated the quantity of alcohol using
listings provided by the New Zealand Department of Food and Agriculture,
otherwise we used either 10 or 12 grams per standard drink according to the
methods used by the investigators. Volume was not calculated for non-drinkers
and they were omitted from these analyses.
Frequency of drinking and typical volume data were not
available for those in the Workforce Diabetes Survey. For the few participants
in the NZ Health Survey whose consumption information was missing, corresponding
data from the National Nutrition Survey has been
used.10
Averaged daily alcohol consumption (in grams) was
calculated from the product of the proportion of days on which drinking occurred
and the volume consumed on a typical occasion. Again, non-drinkers were
omitted.
Statistical
methods Within each survey, the
proportion of participants in each sex and age group classified as non-drinkers
was reported. Logistic regression was used to estimate the relative odds that
non-Maori were drinkers, compared with Maori. Ethnicity (in binary form) and age
(as a continuous variable to adjust for the different distributions of age
within the surveys) were the only predictors in these regressions. Summary
estimates by sex and age group were obtained in similar manner, with the
addition of a variable to indicate which survey the individual was from. An
interaction term between survey and ethnicity was used to test for
heterogeneity, then removed for reporting summary estimates for each
sex–age group.
Mean frequency of drinking is reported as mean number
of days a year on which alcohol is drunk. To estimate relative frequency of
drinking for non-Maori, we used Poisson regression models with an offset of 365
days and scaled for the deviance to correct for slight over-dispersions where
present. Summary estimates were obtained in a similar manner, again with the
addition of variables to represent sex and age group, age as a continuous
variable and which survey the individual was from. Again, an interaction term
between survey and ethnicity was used to test for heterogeneity.
For both volume of alcohol (grams) drunk on a typical
occasion, and averaged daily volume consumed, means for each group defined by
survey, sex and age are shown, together with the relative volume drunk by
non-Maori compared with Maori. Log transformations of alcohol measures were
modelled using generalised linear regression models to obtain summary estimates,
with tests of heterogeneity as above.
The model coefficients for non-Maori relative to Maori,
and their 95% confidence limits, were back-transformed to obtain relative
estimates of consumption for non-Maori compared with Maori.
In all models, persons with missing data for the
variable of interest were excluded from analyses. Results in which the 95%
confidence interval does not include 1.00 are regarded as statistically
significant. In producing the summary estimates, where significant heterogeneity
was found, box-plots for non-Maori and Maori were compared to determine its
extent and source.
ResultsThere was a total of 44 830 people
in the combined study population. Of these, 24 484 (54.6%) were males; 13 174
(29.4%) were aged 18–34 years; 18 478 (41.2%) aged 35–49 years; and
13 178 (29.4%) aged 55–74 years.
Of all participants, 6926 (15.4%) were Maori (3535 men) and
37 904 non-Maori (20 949 men). In the 50- to 74-year age group, a smaller
proportion was Maori compared with non-Maori (26.2% of Maori compared with 30.0%
of non-Maori). Table 1 shows the numbers included from each survey, by
ethnicity, sex and age group (click here to
view tables and figures). The largest survey was the NZ Blood Donors Health
Study, which contributed 17 437 study participants, 38.9% of the total. The
Sleep Survey contributed the largest number of Maori participants
(3194).
Figure 1 shows the percentage of participants who were
classified as non-drinkers (click here to
view tables and figures). In all age groups, except men aged 18–34
years, a significantly higher proportion of Maori were non-drinkers. Table 2
reports the likelihood of non-Maori participants being drinkers relative to
Maori (click here to view tables and
figures). As age increased, so did the strength of this relationship for
both men and women. Among one group at comparatively high risk of heart disease,
men aged 50–74 years, non-Maori were about 90% more likely to be drinkers
than Maori men (odds ratio = 1.89, 95% confidence interval (CI)
1.55–2.31).
Figure 2 (click here to
view) shows the frequency of alcohol consumption (days a year) for Maori and
non-Maori, inclusive of non-drinkers. Across all the groups examined, non-Maori
were more frequent consumers of alcohol relative to Maori, reporting drinking on
over 50% more days of the year. This effect was more marked with increasing age
(Table 3, click here to view), ranging
from a low among men aged 18–34 years (relative frequency (RF) = 1.47, 95%
CI 1.46–1.49) to a high among women aged 50–74 years of more than
double the frequency (RF = 2.27, 95% CI 2.24–2.30).
Table 4 and Figure 3 (click here to view) report frequency of
alcohol consumption (days a year) for Maori and non-Maori among drinkers. Again,
a similar pattern was seen though there was a lower degree of
association.
The volume of alcohol consumed on a typical occasion
reported by non-Maori was consistently less than Maori (Table 5, Figure 4, click here to view). For every study this
was apparent in each age group for both sexes; summary measures consistently
estimated the size of the difference at 35–42% less. Non-Maori men aged
35–49 years reported drinking about 40% less alcohol on a typical occasion
than Maori men in the same age group (relative volume (RV) = 0.59, 95% CI
0.56–0.62).
Regardless of survey, age group and sex, there were few
statistically significant differences in average daily consumption between Maori
and non-Maori (Figure 5, click here to
view). Table 6 (click here to view)
shows average daily consumption of alcohol among non-Maori drinkers relative to
Maori drinkers. When all surveys were combined, only two summary estimates
reached significance, both for the youngest age group: men (relative daily
volume (RDV) = 0.90, 95% CI 0.82–0.98) and women (RDV = 0.73, 95% CI
0.67–0.81).
In many of the models estimating overall effects, there were
significant interactions between the variables representing survey and
Maori/non-Maori ethnicity, suggesting heterogeneity in the association between
Maori and the various measures of alcohol in the different surveys. Generally,
this heterogeneity would lead to rejection of summary estimates based on such
models. However, examination of plots of alcohol consumption by sex, age group
and survey by Maori/non-Maori, showed that almost all such differences were
small. Given the narrow confidence intervals about the estimates for all but the
final alcohol measure (averaged daily consumption) and the general consistency
of the directions of effect, we concluded that large numbers and use of
continuous variables were providing power to detect very small differences that
could be ignored for the purpose of this paper, ie, to compare consumption
between Maori and non-Maori. Consequently, all summary measures have been
reported as the best estimate of the relationship between non-Maori and Maori
drinking patterns.
DiscussionThis research has been undertaken
using a kaupapa Maori framework,11 whereby the
study analysis was undertaken from a Maori perspective. This is distinct from
other methodologies that may ‘minoritise’ Maori with insufficient
data quantity or quality to undertake analyses necessary to inform Maori health
development. In this framework, Maori are the reference population and are
compared with non-Maori. Where appropriate, this type of analysis enables
disparities to be identified and their elimination prioritised. This analytic
approach is consistent with a Treaty of Waitangi framework.
This analysis is unique in that it combines data from a
number of studies that have been undertaken in Aotearoa in recent years. In
total, there were 6926 Maori participants in the studies used for this analysis
(and 37 904 non-Maori), and this is therefore the largest published analysis of
alcohol consumption in Aotearoa for Maori. Our objective was to describe
relative differences in patterns of consumption between Maori and non Maori so
we included all available major studies whether or not they were based on
representative population samples. Therefore, the relative differences in
drinking patterns identified, but not the absolute levels of consumption, are
likely to be generalisable to the population of Aotearoa.
The main findings from this research are that Maori are less
likely to drink alcohol, drink less often, but drink more on a typical drinking
occasion, when compared with non-Maori. These differences in drinking patterns
combine such that average alcohol consumption a day among Maori and non-Maori is
similar.
Between all studies, national and non-national, there is
marked consistency of results. This has occurred in spite of differences in
methodology between individual studies, supporting the validity of the pooled
estimates.
Maori, therefore, have markedly different patterns of
alcohol consumption to non-Maori. These different patterns of alcohol
consumption have implications for health.
Most estimates of this relationship have used measures of
average daily alcohol consumption and not pattern of alcohol
consumption.12 Using the average alcohol
consumption variable for Maori therefore may not produce accurate associations.
Other studies have shown that specific drinking patterns may have independent
effects on health that are not explained by total
consumption.13 This may also be true for Maori.
Further research regarding the relationship between Maori alcohol consumption
patterns and health risk is therefore needed.
Discussion of alcohol consumption by ethnicity has been a
sensitive topic for some years. This has in part been due to a ‘victim
blaming’ interpretation14,15 of data that
have been published where there is no acknowledgement that risk behaviours of
individuals are socially patterned. Such an approach is generally
counterproductive to dialogue and progress in reducing harm from alcohol
consumption.
Historically, alcohol became readily available to Maori in
the early nineteenth century.16 Maori initially
showed a strong aversion to drinking alcohol. Discriminatory legislation
regarding consumption of alcohol by Maori was not removed until 1948 with the
passage of the Licensing Amendment Act, which repealed the previous law
prohibiting consumption of alcohol by Maori in public
bars.16
While this study demonstrates the different drinking
patterns between Maori and non-Maori, research from the 1996/1997 New Zealand
Health Survey also indicates that Maori adults are more likely than non-Maori to
have potentially ‘hazardous drinking’
patterns.17 Such patterns carry with them
higher risk to physical or mental health. Furthermore, ‘hazardous
drinking’ is socially patterned in that it is associated with
socioeconomic gradients that disproportionately affect
Maori.18 Such drinking patterns and their
associated health problems are not unique to Maori and are similar among other
indigenous populations that have experienced
colonisation.19,20
In recent times there have been health promotion initiatives
to raise awareness among Maori about the harm associated with hazardous drinking
patterns in the expectation that those ‘at risk’ will change
attitudes and behaviours.21 These efforts have
occurred in an environment of increasing availability of alcohol and
re-introduction of multimedia advertising of alcohol. It is not yet known
whether these factors are associated with increased uptake of alcohol,
especially among younger and more vulnerable populations. At the same time,
there has been limited policy development aimed at addressing the socioeconomic
gradient that continues to marginalise Maori into the most deprived
echelons of our society within which
hazardous drinking is more prevalent.
To date, few published studies have compared Maori and
non-Maori alcohol consumption patterns. The most comprehensive is Te Ao
Waipiro.22 In that study, the median frequency
of drinking amongst Maori was lower, and the median annual volume of consumption
for Maori males was higher, than for all males. The median quantity consumed per
occasion was much higher among Maori than for the total population. Our results
are consistent with those study findings, although our analyses differ in that
we compared Maori with non-Maori, whilst in Te Ao Waipiro Maori were compared
with the total population, which included Maori.
Alcohol data from the National Alcohol Surveys, including
the 2000 National Maori Alcohol Survey conducted by the Whariki Research Group
and the Alcohol and Public Health Research Unit, were not included in this
analysis as the kaupapa of these surveys were to firstly present Maori-specific
analyses and then determine further analyses through a process of consultation
with Maori. This process was underway, though not complete, at the time of
writing this paper.
The 1997 National Nutrition Survey, which used a subset of
the participants of the NZ Health Survey plus additional Maori and Pacific
people, reported that Maori men had a higher mean daily alcohol intake (25 g)
than European and others (19 g). Among women, Maori had similar intakes to
European and others (8 g and 9 g
respectively).10 These findings are not
directly comparable as they represent alcohol intake over a single 24-hour
period and include non-drinkers.
A report from Te Puni Kokiri in 2000, using data from the
National Nutrition Survey, stated that Maori are considerably less likely to be
moderate drinkers (as opposed to
non-drinkers or hazardous drinkers) than non-Maori (46% and 66% respectively)
and relatively higher proportions of Maori are either non-drinkers or hazardous
drinkers.23 While our results are not directly
comparable, they are broadly consistent.
There are a number of potential sources of bias that may
occur with our study methodology. These include combining data from studies
conducted during different time periods (1988–2001) and the possibility
that drinking patterns may have changed over that period; the use of a mixed
group of studies, only some of which were population-based; and the use of
different instruments for measuring alcohol consumption. However, the similarity
of our results to those from the individual studies is reassuring and suggests
that these factors are likely to have caused only minimal bias.
In summary, non-Maori and Maori have markedly different
alcohol consumption patterns. This must be considered when determining the
relationship between Maori alcohol consumption and health risk or when applying
methods of risk estimation based on average consumption levels.
Author information:
Dale M Bramley (Nga Puhi), Senior Lecturer; Joanna B Broad, Research Fellow,
EPIQ Group, Section of Epidemiology and Biostatistics, School of Population
Health, University of Auckland; Ricci Harris (Ngati Kahungunu, Ngati Raukawa,
Kai Tahu), Public Health Medicine Registrar; Papaarangi Reid (Te Rarawa), Public
Health Physician, Te Ropu Rangahau Hauora a Eru Pomare, Wellington School of
Medicine and Health Sciences, University of Otago, Wellington; Rod Jackson, Head
of Section of Epidemiology and Biostatistics, School of Population Health,
University of Auckland, Auckland. On behalf of the Alcohol Burden of Disease and
Disability Group, University of Auckland: Shanthi Ameratunga, Dale Bramley,
Joanna Broad, Jennie Connor, Rod Jackson, Patricia Metcalf, Robert Scragg and
Sue Wells.
Acknowledgements:
This paper is part of a series of papers that has been commissioned by the
Alcohol Advisory Council of New Zealand.
Data sets for analysis were provided from Robert Scragg and
Patricia Metcalf (Workforce Diabetes Survey), Stephen MacMahon, Robyn Norton and
Shanthi Ameratunga (New Zealand Blood Donors Health Study), Rod Jackson, Stephen
MacMahon and Robyn Norton (Fletcher-Challenge University of Auckland Heart and
Health Study), Ricci Harris, Papaarangi Reid and Philippa Gander (Sleep Survey),
and the Ministry of Health (National Nutrition Survey). We particularly
appreciate the willingness of the investigators of these studies to allow use of
their data, and their advice and assistance.
We thank Jennie Connor, Robert Scragg, Shanthi Ameratunga,
Elizabeth Robinson, Robyn Norton, Margaret Geddes, Mike MacAvoy and Helen
Moewaka Barnes for their advice and comments on earlier drafts of this
paper.
Correspondence: Dr
Dale Bramley, Division of Community Health, University of Auckland, Private Bag
92019, Auckland. Fax: (09) 441 8957; email: dale.bramley@waitematadhb.govt.nz
References:
|
|
||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|