Increasing socioeconomic inequalities in male cirrhosisof the liver mortality: Australia 1981 – 2002

JAKE M. NAJMAN1,2, GAIL M. WILLIAMS2, & ROBIN ROOM3

1School of Social Science, University of Queensland, St Lucia, Brisbane, Queensland, Australia, 2School of Population

Health, University of Queensland, Brisbane, Queensland, Australia, and 3Centre for Social Research on Alcohol and Drugs,

Stockholm University, Stockholm, Sweden

AbstractLiver cirrhosis mortality is an indicator of harms associated with high levels of alcohol consumption. There is good evidence thatchanges in political and economic systems can lead to changing patterns of liver cirrhosis mortality. Socioeconomic inequalities inliver cirrhosis mortality have been periodically reported but there are few studies of changes in socioeconomic inequalities in livercirrhosis mortality over time. This paper examines changes in socioeconomic inequalities in liver cirrhosis mortality in Australiafor the period 1981 – 2002. Age standardised, liver cirrhosis mortality rates were calculated for occupational groupings forAustralia 1981 – 2002. Occupations were grouped into non-manual and manual categories, and there was imputationfor missing data. Despite decreasing overall liver cirrhosis mortality rates over time, liver cirrhosis mortality continues to accountfor about 3% of all deaths. Manual workers have consistently experienced liver cirrhosis mortality rates which are twice or morethe rate experienced by non-manual workers. These inequalities appear to have increased in recent years and currently appear tobe at historic highs (manual workers have mortality rates of about 2.5 times those of non-manual workers). Increasingsocioeconomic inequalities in liver cirrhosis mortality in Australia suggest that lower SES groups have, over time, increased theirlevel of harmful alcohol consumption relative to middle and higher SES groups. It is suggested that this might be attributed to arelative improvement in the affordability of alcohol over time. [Najman JM, Williams GM, Room R. Increasingsocioeconomic inequalities in male cirrhosis of the liver mortality: Australia 1981 – 2002. Drug Alcohol Rev2007;26:273 – 278]

Key words: changes over time, increasing inequalities, liver cirrhosis, mortality, socioeconomic.

Introduction

For countries which have time series data available,

it appears that socioeconomic inequalities in all-

cause mortality have been increasing, at least in

recent years [1 – 3]. It is not clear why these increased

inequalities are being observed but one possibility is

that there has been increasing socioeconomic inequal-

ities in lifestyle related diseases (tobacco, alcohol, diet

and physical activity) [4]. There is some evidence to

suggest that those in higher socioeconomic groups have

more enthusiastically adopted lifestyle options which

confer better health. It is not clear whether this

tendency for higher socioeconomic groups in recent

years to adopt a healthier lifestyle extends to patterns of

alcohol use, particularly in view of the younger ages at

which the highest levels of alcohol consumption are

observed.

While liver cirrhosis has a number of causes, high

levels of alcohol consumption are estimated to account

for between 40 and 90% of all liver cirrhosis deaths [5].

Arguably overall levels of alcohol consumption are a

major predictor of alcohol-related harms [6]. Liver

cirrhosis mortality, from this perspective, is likely to be

a marker for a number of harms experienced by those

who report high levels of alcohol consumption.

World Health Organisation data from around 1990

points to massive variations in population alcohol

consumption and consequently disparities in deaths

attributable to liver cirrhosis [7]. Upper extreme coun-

tries like Hungary and Romania had death rates at that

time which were ten times the rate observed in other

Received 11 April 2006; accepted for publication 21 November 2006.

Jake M. Najman, PhD, School of Social Science, University of Queensland, St Lucia, Brisbane, Queensland, Australia, Gail M. Williams, PhD,School of Population Health, University of Queensland, Brisbane, Queensland, Australia, Robin Room, PhD, Centre for Social Research onAlcohol and Drugs, Stockholm University, Stockholm, Sweden. Correspondence to Professor Jake M. Najman, Schools of Population Health andSocial Science, University of Queensland, St Lucia 4072, Queensland, Australia. Tel: þ61 7 3365 5180. E-mail: j.najman@uq.edu.au

Drug and Alcohol Review (May 2007), 26, 273 – 278

ISSN 0959-5236 print/ISSN 1465-3362 online/07/030273–06 ª Australasian Professional Society on Alcohol and Other Drugs

DOI: 10.1080/09595230701247699

countries, e.g. New Zealand and Ireland. Comparisons

of liver cirrhosis mortality for a number of eastern and

western European countries over time show evidence of

substantial fluctuations, in part reflecting changing social

and economic forces that influences patterns of alcohol

consumption [8]. The impact of changes in levels of

alcohol consumption is further evidenced by USA time

series data which shows increasing death rates attributed

to liver cirrhosis from 1935 to the early 1970s [9]. It is

now accepted that the introduction of some elements of a

market economy led to a substantial increase in alcohol

consumption in Russia [10]. A recent substantial

increase in liver cirrhosis mortality in Britain, possibly

contingent on a liberalisation of alcohol sales, further

suggests that alcohol consumption patterns are respon-

sive to a number of social policy changes [11]. The

decline in liver cirrhosis mortality in a number of

countries since the mid 1970s is attributed, not only to

a (modest) decline in per capita alcohol consumption,

but also to improved treatment services.

There are a number of papers which describe

socioeconomic inequalities in liver cirrhosis, but none

which describe how these socioeconomic inequalities

may have changed over time. These papers are

consistent in indicating that there is an inverse social

class/socioeconomic gradient in alcohol-related mortal-

ity rates [12,13]. Socioeconomic inequalities in alcohol-

related diseases are amongst the largest of any cause of

death [14]. British data referring to the years 1991 –

1993 suggest that unskilled workers have liver cirrhosis

mortality rates which are about four times those

experienced by professionals [12]. This social class

gradient of mortality reverses the patterns found

decades earlier in Britain [15], when prices of alcohol

were relatively much higher and access to alcohol

somewhat more restricted. The SES inequalities in

alcohol-related mortality appear to be consistent with

evidence of alcoholism and alcohol-related diagnoses in

a population [16]. It appears that as patterns of alcohol

consumption change over time socioeconomic inequal-

ities in alcohol-related mortality (primarily liver cirrho-

sis) also change. For example, in Sweden there is

evidence that, over the period 1967 – 1980, there was a

substantial increase in levels of moderate to high

alcohol consumption by manual workers (from 11 –

21%) and a possible decline in the same level of

consumption from (22 – 10%) by medium and high

level non-manual workers [17]. These changes were

consistent with changing socioeconomic inequalities in

rates of hospitalisations for alcohol-related diagnoses,

but not for rate ratios of mortality attributable to

alcohol-related diagnoses. Given the way patterns of

alcohol consumption have changed in different coun-

tries over time [18], both the magnitude and direction

of socioeconomic inequalities in liver cirrhosis mortality

may be country and time-period specific.

This paper examines male deaths attributed to

cirrhosis of the liver (ICD9-571) in the Australian

population over the period 1981 – 2002.

Method

Cause-specific deaths for all the years 1981 – 2002

(inclusive) were obtained from the Australian Bureau of

Statistics. To obtain population estimates we used

population counts for relevant census years (1981,

1986, 1991, 1996 and 2001) and, based upon a

constant population growth, estimated age-occupation

specific population numbers for each of the non-census

years. Rates were standardised to the 1996 age

distribution of the Australian population.

Occupations were aggregated into non-manual and

manual categories of workers. This leaves a category of

those not in the workforce, which introduces possible

numerator/denominator bias into comparisons of work-

ers over time. Numerator/denominator bias occurs

because of the different ways occupational data are

collected in the census and at death certification. At the

census current occupation is recorded, while on the

death certificate it is usual or most recent occupation

(differs by state jurisdiction). With changes in employ-

ment patterns over time (e.g. decrease in manual

workers as a proportion of all workers) and varying

levels of unemployment, there may be a decrease in

those who, at census, are categorised as manual workers

but not an equivalent decrease to those classified as

manual workers at death (see certification [3] for a

more detailed discussion). Thus comparisons over time

are subject to a bias which presents as a progressive

increase in the age standardised mortality rate for

manual workers compared to non-manual workers.

To correct for numerator/denominator bias there is a

need to adjust for those who are not in the workforce.

The Australian Bureau of Statistics publishes periodic

workforce surveys of the ‘usual’ work category of those

not currently in the workforce. Such data are disag-

gregated by the age and gender of the person not in the

workforce. The data are generally consistent in

indicating that about 60% of those not in the workforce

would be categorised as manual workers if they were

employed. We have redistributed, for each data

collection, those not in the workforce according to the

proportions suggested by the workforce surveys.

Following this process of imputation the age stan-

dardised mortality rates have been recalculated and

both unadjusted and imputed calculations are provided

in the tables which follow.

Results

Table 1 presents details of male cirrhosis of the

liver mortality deaths for Australia over the period

274 J. M. Najman et al.

1981 – 2002. The number of deaths declined from

1980s to the 1990s, though the differences are not

great. As a percent of all deaths, liver cirrhosis

continues to account for about 3% of male deaths for

males aged 15 – 64 years. Despite the similar numbers

and proportion of deaths attributed to liver cirrhosis,

the age standardised mortality rate is about half the rate

it was in the early 1980s.

Table 2 provides details of the changes in age

standardised mortality rates for manual and non-

manual workers over time. Looking at the non-imputed

data, the decline in liver cirrhosis mortality for manual

and non-manual workers is of a similar magnitude (see

trend of decline). The imputed data suggest a some-

what different pattern with the annual decline in

manual mortality rates being 2.67% compared to a

3.24% annual decline for non-manual workers (p for

difference in trend¼ 0.06). The evidence based upon

the imputed data suggests that death rates for non-

manual workers have declined more rapidly than for

manual workers.

Table 3 presents the manual/non-manual ratios

comparing the rates of mortality decline over time.

While the non-imputed and imputed comparisons

differ somewhat, particularly in the early 1980s, they

are similar in the early 2000s indicating that cirrhosis

death rates for manual workers are about two and a half

times those of non-manual workers. Indeed by the early

part of the 21st century the difference in cirrhosis of the

liver mortality rates between manual and non-manual

workers were the highest we have recorded.

Table 4 presents details of the number of additional

manual workers who die each year because of the

difference in cirrhosis of the liver mortality rates

experienced by manual and non-manual workers. It is

derived by applying the difference in manual and non-

manual mortality rates to the number of manual

workers in the Australian population each year. Over

the period 1981 – 2002 there were some 4317 manual

workers whose deaths could be attributed to their

manual workforce status. Based upon an estimated life

expectancy of 65 years each of these workers had their

lives shortened by an average of 31.7 years.

Discussion

Our findings confirm a decreasing rate of liver cirrhosis

deaths over the period 1981 – 2002 in Australia,

but with liver cirrhosis continuing to account for

about 3.3% of male deaths in the 15 – 64 age group.

Table 1. Cirrhosis of the liver mortality (ICD9-571), Australiamales aged 15 – 64, 1981 – 2002

Number ofdeaths

Percent ofall deaths

Age standardised mortalityrate (per 100,000)

1981 655 3.2 17.31982 654 3.1 16.81983 597 2.9 15.71984 618 3.1 15.11985 571 2.8 13.71986 581 2.9 13.51987 601 3.0 13.71988 576 2.9 12.91989 523 2.7 11.61990 526 2.7 11.41991 492 2.7 10.41992 533 2.9 11.21993 454 2.6 9.61994 463 2.6 9.61995 465 2.7 9.71996 499 2.9 10.21997 482 2.8 9.31998 468 2.7 8.91999 508 3.0 9.42000 458 2.7 8.32001 474 2.9 8.02002 543 3.3 9.2

Table 2. Cirrhosis of the liver (ICD9-571), age standardisedmortality rate

Non-imputed Imputed

Manual Non-manual Manual Non-manual

1981 17.01 8.53 19.23 8.671982 16.02 7.69 19.19 8.141983 15.17 7.34 17.57 7.381984 14.84 7.56 17.44 7.661985 14.50 7.49 15.77 7.081986 17.15 7.43 16.26 6.501987 18.41 8.16 16.42 7.081988 16.29 8.07 15.13 7.051989 15.69 7.98 13.43 6.491990 16.27 7.44 13.58 6.091991 15.16 7.83 11.86 6.351992 16.16 8.14 12.83 6.611993 14.92 6.22 11.31 5.091994 15.88 6.12 11.47 4.891995 14.72 7.88 10.10 6.051996 16.48 6.38 11.92 5.211997 13.22 6.07 11.12 5.081998 13.08 4.63 11.43 3.981999 11.60 6.24 10.83 5.442000 11.93 4.49 10.22 4.032001 10.84 4.32 10.36 4.052002 12.80 5.07 11.40 4.64

Non-imputed Trend 1981 – 2002Manual 71.78 (71.37, 72.20)Non-manual 71.82 (71.26, 72.38)p(difference between trends)¼NSImputed Trend 1981 – 2002Manual 72.67 (72.17, 73.17)Non-manual 73.24 (72.91, 73.57)p(difference between trends)¼ 0.06

Increasing socioeconomic inequalities in liver cirrhosis mortality 275

This decline is in contrast to the increases observed in

Britain over a similar time period [11].

Socio-economic inequalities have remained a persis-

tent feature of male liver cirrhosis mortality in Australia

– generally manual workers have age-standardised death

rates which are about twice or more those of non-

manual workers. There is little available Australian data

which would enable a longer time period comparison.

Najman used occupational status data [19] to assess

socioeconomic inequalities in liver cirrhosis mortality in

Australia for males over the period 1965 – 1967. He

found that the higher status occupations had a lower age

standardised (per 100,000 population) mortality rates

(status A¼ 9.7: status B¼ 8.7: status C¼ 8.7) than the

lowest status occupations (status D¼ 13.2). These SES

inequalities in liver cirrhosis mortality are lower that

those we have noted in more recent years. The evidence

suggests, with regard to the imputed data, that in recent

years socioeconomic inequalities in male liver cirrhosis

mortality have increased, with manual workers having

age standardised rates of cirrhosis of the liver that are

now about 2.5 times those of non-manual workers. This

difference is larger than that observed for socioeco-

nomic inequalities in all-cause mortality, suggesting that

differential patterns of alcohol consumption make a

disproportionate contribution to socioeconomic in-

equalities in mortality.

A number of explanations of (increasing) socio-

economic inequalities in liver cirrhosis mortality are

plausible. Sustained heavy levels of alcohol consump-

tion remain the main cause of liver cirrhosis [20].

In Australia the decline in liver cirrhosis mortality

since the mid 1970s is partly attributable to a decline in

per capita alcohol consumption (based upon alcohol

taxation data) since a peak of about 13 litres per person

aged 15 and older in the early 1970s down to about 10

litres in the early 1990s [21]. National survey data

shows very little change in the proportion of males

consuming alcohol at high-risk levels since 1989 – 1990

[22]. This decline has plateaued with average con-

sumption remaining at about 10 litres per person aged

15 and older since 1991. It is interesting that Australian

liver cirrhosis mortality rates have also plateaued since

the early 1990s.

This plateauing of liver cirrhosis deaths is despite a

massive increase, in Australia, in rates of hepatitis C,

leading, in some cases, to cirrhosis of the liver. While

cirrhosis is not a common outcome of hepatitis C, the

increase in hepatitis C, with at least 200,000 cases

nationally, may be contributing to the overall rate of

liver cirrhosis deaths, but only in recent years. Given

the chronic debilitating course of hepatitis C leading to

liver cirrhosis, this may be impacting on observed

mortality inequalities.

While some alcohol treatment services in Australia

are privately funded, the majority are part of the public

Table 3. Cirrhosis of the liver (ICD9-571), age standardisedmortality ratios over the period 1981 – 2002, Australian males

Non-imputed ratio M:NM Imputed ratio M:NM

1981 1.99 2.221982 2.08 2.361983 2.07 2.381984 1.96 2.281985 1.94 2.231986 2.31 2.501987 2.26 2.321988 2.02 2.151989 1.97 2.071990 2.19 2.231991 1.94 1.871992 1.99 1.941993 2.40 2.221994 2.59 2.351995 1.87 1.671996 2.58 2.291997 2.18 2.191998 2.83 2.871999 1.89 1.992000 2.66 2.542001 2.51 2.562002 2.52 2.46

Table 4. Cirrhosis of the liver (ICD9-571), death attributable tomanual occupational status, Australian males 1981 – 2002

Attributabledeaths

manual males

AttributableYLL manual

males

Mean YLLper manual

worker

1981 262 7988 30.51982 271 8246 30.41983 250 7477 29.91984 245 7635 31.21985 215 6892 32.11986 249 7715 31.01987 242 7436 30.71988 207 6402 30.91989 179 5636 31.51990 198 6295 31.81991 149 4822 32.41992 169 5447 32.21993 169 5048 29.91994 183 5663 30.91995 112 3489 31.21996 190 5744 30.21997 172 5632 32.71998 218 6716 30.81999 158 5228 33.12000 159 5230 32.92001 159 5230 32.92002 160 5229 32.7

Total (1981 –2002)

4317 135,199 31.7

276 J. M. Najman et al.

(government funded) system of health care. It is not

easy to see why unequal access to alcoholism treatment

services might account for our findings. The advent of

liver transplantation as a relatively routine procedure

raises the possibility that increasing socioeconomic

inequalities in liver cirrhosis mortality are a conse-

quence of the unequal delivery of these treatment and/

or transplantation services. Over the period 1989 – 1998

the number of liver transplants performed in Australia

increased from 100 – 167 [23]. These transplants are

almost entirely performed in public (government

funded) hospitals. While the cost of a transplant is

not likely itself to lead to inequalities in access to a

transplant, some co-occurring conditions such as a HIV

infection, continuing alcohol abuse or some co-

occurring disease states will mean that some cannot

qualify to obtain a transplant. It may be that persons in

lower SES groups are disproportionately likely to be

disqualified from obtaining a transplant. Based crudely

on the numbers of transplants recorded over time as a

proportion of deaths (transplants are for both males and

females; deaths in this study are male deaths only) it

seems unlikely that increasing socioeconomic inequal-

ities are a consequence of the improved availability of

liver transplantation services.

With the Australian decline in male per capita

alcohol consumption it may well be that the proportion

of high-risk drinkers in manual and non-manual

employment may have altered. Manual workers may

constitute a group with higher rates of ‘binge’ or heavy

and persistent drinking. There is some international

evidence suggesting that there are socioeconomic

differences in both the type and pattern of alcohol

consumed, although these vary from one society to

another. In Australia, males in the lowest socioeco-

nomic groups are over-represented at the higher risk

levels of consumption [22]. While beer is the traditional

working-class drink in Australia, it has gradually been

losing ground to wine. For Australia in 1977 beer

constituted 61.5% of all the alcohol consumed. Some

20 years later beer comprised just under 57% of all

alcohol consumed, with increased wine consumption

largely accounting for the difference. Popular accounts

of the above change focus on increased wine-drinking

by the affluent, but this is not the whole story. With no

tax beyond sales tax on wine in Australia, cask wine is

by far the cheapest form of alcohol available [24], and

thus is often the beverage chosen by poor and margin-

alized drinkers. While spirits and beer prices came close

to keeping pace with general inflation in Australia in the

20 years after 1974, wine prices rose at only two-thirds

of the rate for spirits and beer [25,26]. Thus the relative

availability to the poorest drinkers of the cheapest form

of alcohol increased.

Finally it is possible that changes in national

economic conditions have led to increased inequalities

in alcohol consumption. Over the period 1994 – 1995

and 2002 – 2003 Australia has experienced good

economic growth with welfare policies narrowly tar-

geted to the most economically disadvantaged. Harding

[27] has found that the average private household

income increased by 165% and 53% for the lowest and

next lowest deciles of household incomes (1994 – 1995

to 2002 – 2003). This contrasts with an average house-

hold income increase of about 35% to other decile

households over the same period. It is possible that

even a modest increase in the incomes of those in the

lowest income groups has led to an increase in ‘at risk’

levels of alcohol consumption by the most economically

disadvantaged. Makela [28] has previously found that

socioeconomic groups differed in their patterns of

alcohol-related mortality in response to an economic

boom. The lowest SES groups had a greater increase in

alcohol-related mortality during the boom. It would be

an interesting unanticipated outcome of improved

economic circumstances experienced by some econom-

ic groups, particularly by the most economically

disadvantaged, if their increased incomes (or increased

affordability of alcohol) led to increased consumption

of alcohol and perhaps other potentially unhealthy

products. If liver cirrhosis mortality is a marker for

other alcohol related harms then it is likely that there

have been increasing socioeconomic inequalities in

these other harms as well.

References

[1] Davey Smith G, Dorling D, Gordon D, Shaw M. The

widening health gap: what are the solutions? Crit Public

Health 2003;9:453 – 74.

[2] Mackenbach JP, Bos B, Andersen O, Cardano M, Costa G,

Harding S, Reid A, Hemstrom O, Valkonen T, Kunst AE.

Widening socioeconomic inequalities in mortality in six

Western European countries. Int J Epidemiol 2003;

32:830 – 7.

[3] Williams GM, Najman JM, Clavarino A. Correcting for

numerator/denominator bias when assessing changes in

occupational class inequalities in mortality in Australia

1981 – 2002. Bull WHO 2006;84:198 – 203.

[4] Najman JM, Toloo G, Siskind V. Socioeconomic status and

changes in health risk behaviours in Australia: 1989 – 1990

to 2001. Bull WHO 2006;84:976 – 84.

[5] Dufour M, Stinson FS, Caces MF. Trends in cirrhosis

morbidity and mortality: United States, 1979 – 1988. Sem

Liver Dis 1993;13:109 – 25.

[6] Norstrom T, Ole-Jørgen S. Alcohol and mortality: metho-

dological and analytical issues in aggregate analyses.

Addiction 2001;96(Suppl. 1):S5 – S17.

[7] La Vechhia C, Levi F, Lucchini F, Franceschi S, Negri E.

Worldwide patterns and trends in mortality from

liver cirrhosis 1955 to 1990. Ann Epidemiol 1994;4:

480 – 6.

[8] Corrao G. Liver cirrhosis mortality trends in Eastern

Europe, 1970 – 1989. Analyses of age, period and cohort

effects and of latency with alcohol consumption. Addict Biol

1998;3:413 – 22.

Increasing socioeconomic inequalities in liver cirrhosis mortality 277

[9] Singh GK, Hoyert DL. Social epidemiology of chronic liver

disease and cirrhosis mortality in the United States, 1935 –

1997: trends and differentials by ethnicity, socioeconomic

status and alcohol consumption. Human Biol 2000;72:

801 – 20.

[10] Nemtsov AV. Alcohol-related harm and alcohol consump-

tion in Moscow before, during and after a major anti-

alcohol campaign. Addiction 1998;93:1501 – 10.

[11] Leon DA, McCambridge J. Increases in liver cirrhosis mor-

tality rates in Britain are the largest in Western Europe: An

analysis of routine data 1950 – 2002. Lancet 2006;367:52 – 6.

[12] Harrison L, Gardiner E. Do the rich really die young?

Alcohol-related mortality and social class in Great Britain,

1988 – 1994. Addiction 1999;94:1871 – 80.

[13] Hemstrom O. Alcohol-related deaths contribute to socio-

economic differentials in mortality in Sweden. Eur J Public

Health 2002;12:254 – 62.

[14] Martikainen P, Makela P, Koskinen S, Valkonen T. Income

differences in mortality: a register-based follow-up study of

three million men and women. Int J Epidemiol 2001;

30:1397 – 1405.

[15] Terris M. Epidemiology of cirrhosis of the liver: national

mortality data. Am J Public Health 1967;57:2076 – 88.

[16] Hemmingsson T, Lundberg I, Romelsjo A, Alfredsson L.

Alcoholism in social classes and occupations in Sweden. Int

J Epidemiol 1997;26:584 – 91.

[17] Romelsjo A, Lundberg M. The changes in the social class

distribution of moderate and high alcohol consumption and

of alcohol-related disabilities over time in Stockholm

County and in Sweden. Addiction 2006;101:1307 – 23.

[18] Ramstedt M. Per capita alcohol consumption and liver

cirrhosis mortality in 14 European countries. Addiction

2001;96(Suppl. 1):S19 – S34.

[19] Najman JM. A social epidemiology of Australia, using

mortality data: 1965 – 1967. Thesis, Degree of Doctor of

Philosophy, University of New South Wales, 1978.

[20] U.S. Department of Health and Human Services. Healthy

People 2010, 2nd ed. Washington, DC: US Government

Printing Office; 2000.

[21] Australian Institute of Health and Welfare. Australia’s

Health 2004. Canberra: AIHW; 2004.

[22] Australian Bureau of Statistics. Occasional Paper: Health

risk factors – a guide to time series comparability from the

National Health Survey. Australia Publication number

4826.055.001. Canberra: Australian Bureau of Statistics;

2004.

[23] NSW Health. Selected specialty and statewide service

plans No.3: Liver transplantation. Sydney: NSW Health

Department; 2002.

[24] Stockwell T, Crosbie D. Supply and demand for alcohol

in Australia: relationships between industry structures,

regulation and the marketplace. Int J Drug Policy 2001;

12:139 – 52.

[25] Australian Bureau of Statistics. Consumer Price Index.

Australia, CPI: Alcohol and tobacco, weighted average of

eight capital cities, time series spreadsheet. Cat. no. 6401.0.

2005. Available from: http://www.abs.gov.au/AUSSTATS/

abs@.nsf/DetailsPage/6401.0Dec%202005?OpenDocument

[26] Chan HS, Bettington N. Demand for wine in Australia:

systems versus single equation approach. Armidale, NSW:

University of New England, Graduate School of Agricultur-

al and Research Economics, Working Paper Series 2001 – 5.

Available from http://www.une.edu.au/gsare/publications/

AREwp01-5.PDF 2001.

[27] Harding A. Recent trends in income inequality in Australia.

Conference on Sustaining Prosperity: New Reform

Opportunities for Australia; 2005 Canberra; NATSEM,

University of Canberra.

[28] Makela P. Alcohol-related mortality during an economic

boom and recession. Contemp Drug Prob 1999;26:

369 – 90.

278 J. M. Najman et al.