9.3 Contribution of smoking to health inequality

Show / hide chapter menu

As is the case elsewhere in the world, ill health and rates of premature death in Australia show a clear gradient across socioeconomic groups.72

People who are poorer or disadvantaged in other ways generally suffer more illness and reduced quality of life and die earlier than people who are better off. The social gradient holds regardless of how socioeconomic disadvantage is measured.72

People who are disadvantaged tend to have a poorer diet, are more likely to be overweight and to suffer from diabetes, especially so among Indigenous communities.72 However, with or without such additional risk factors, current smokers are much less likely than non-smokers to be in good health, and the incidence of numerous diseases is significantly higher among smokers and recent ex-smokers than among long-time ex-smokers and never-smokers.73, 74 With social differentials in smoking during pregnancy, smoking prevalence, cigarette consumption, duration of smoking and exposure to environmental tobacco smoke must contribute substantially to socioeconomic differentials in health status and mortality.

This section outlines data on relative rates of poor health, disease, mortality and life expectancy across SES groups, and also presents estimates of the contribution of smoking to these health disparities.

9.3.1 Socioeconomic position, reported health status and smoking

People who live in disadvantaged areas are much less likely to assess their own health as excellent or good.72 Results from the ABS National Health Surveys suggest that rates of arthritis, bronchitis and emphysema and diabetes are higher in less advantaged groups.30, 31

In 2004–05, only 45% of smokers participating in the ABS's National Health Survey reported their overall health as very good or excellent, compared to 57% of ex-smokers and 60% of non-smokers. At all ages, the proportion of current smokers who rated their health as fair or poor was substantially higher than for those that never smoked. In 2004–05 smokers (those still alive who had not already died from smoking-related diseases) were twice as likely to report suffering from bronchitis, and reported asthma was also more common among smokers.58

9.3.2 Socioeconomic position and illnesses known to be caused by smoking

Hospitalisation for cardiovascular disease shows a clear socioeconomic gradient, with rates for males in the most disadvantaged groups 24% higher than for males in the most advantaged groups and rates for females in the most disadvantaged group 32% higher than rates for females in the most advantaged groups.

Chronic kidney disease is also more common in lower SES groups, especially among Indigenous Australians.75

9.3.3 Socioeconomic disparities in death rates from diseases known to be caused by smoking

People from lower socioeconomic groups have higher rates of mortality overall and for most causes of death.76

For males living in the most disadvantaged areas compared to males living in the least disadvantaged areas, death rates were 1.8 greater overall. They were: 1.5 times greater for cancer and two times higher for lung cancer; 2.1 times higher for circulatory disease, 1.9 times higher for stroke, 2.8 times higher for a disease of the respiratory system. Among women, death rates were 1.5 times as high overall, 1.7 times higher for lung cancer, 2.3 times higher for circulatory disease; 1.8 times higher for stroke and 2.4 times higher for respiratory disease. Smoking is a significant risk factor for all of these diseases.77

9.3.4 Socioeconomic disparities in health-adjusted life expectancy

As part of research on preventable causes of disease conducted for the Australian Government, researchers at the University of Queensland examined differentials in the burden of disease across socioeconomic groups.74

At birth, those in the lowest socioeconomic quintile can expect to die at least three years earlier than those in the highest economic quintile (79.6 compared to 82.7 years). Adjusting for ill health, those in the lowest quintile can expect to live four years less than those in the highest quintile. By the age of 60, those in the lowest quintile can expect 15% fewer years of health-adjusted life than those in the highest quintile.

Researchers estimated that, for the year 2003, more than 2.6 million disability-adjusted life years (DALYS) were lost in Australia. DALYs were calculated for each of the five socioeconomic quintiles.

After adjusting for age, loss rates were 31.7% higher in the lowest SES quintile than in the highest.74 Rates of burden were higher for most causes, but particular for mental disorders and cardiovascular disease. Per head of population, rates of burden were 26.5% higher in remote areas than in major cities.74, 78

9.3.5 Quantifying the contribution of smoking to socioeconomic differentials in health status

So how much of the differential in disease rates between higher and lower socioeconomic groups can be attributed to smoking?

Estimates of the contribution of smoking to social inequality vary quite widely.

Prabhat Jha and colleagues report in a four-country study that most social inequalities in adult male mortality during the 1990s were due to smoking.79 In the UK, Jarvis and Wardle used an 'indirect method' to estimate that tobacco caused about two-thirds of the difference in risk of death across social class in men age 35–69 years.80 Bobak and colleagues reported similar results for Canada, Poland and the USA, and contended that eliminating smoking would halve the social gradient in mortality among men.81 Commentator Michael Marmot has been quite critical of these sorts of estimates82 and authors of these studies have generally acknowledged the limits of indirect estimation.

Blakely and Wilson and their colleagues have used direct methods to estimate the contribution of smoking to socioeconomic and ethnic inequalities in mortality in New Zealand. For the 1996–99 period, smoking contributed 21% to the gap between 45–74-year-old men with post-school qualifications and those with none. The corresponding figure for women was 11%.83. But other work suggested that only 5–10% of the larger inequality in mortality between Maori and non-Maori individuals was due to smoking, despite large differences in smoking prevalence.84 This estimate contrasted with a much greater estimated contribution by the Ministry of Health using Jha and colleagues' indirect method.85

A study by Siahpush, English and Powles86 estimated that in Australia smoking could account for just over one third of the excess deaths that would otherwise be attributed to lower levels of education. Data on deaths among men aged between 40 and 69 taking part in a prospective cohort study in Melbourne between 1990 to 1994 showed that the association between education and mortality was greatly weakened after adjustment for smoking and the aetiologic fraction for low level of education was reduced from 16.5% to 10.6%.

It is likely that indirect estimates of the contribution of tobacco smoking overestimate the importance of smoking by failing to take account of higher-than-average prevalence of other behavioural and other risk factors in low SES populations. Direct methods however may underestimate the importance of smoking because they don't take into account the long-term impact of smoking during pregnancy and the impact of smoking and exposure to tobacco smoke on diseases other than the ones for which epidemiological data are readily available. They also may not take account of the effects of spending on tobacco products on financial security and inter-generational poverty, which may help to perpetuate continuing high smoking rates in the children of smokers. These issues are explored further in Sections 9.4 to 9.8 below.

9.3.6 Are tobacco-related differentials in health status widening?

In the United States the socioeconomic gap in life expectancy appears to be worsening. In people who had more than 12 years of education, life expectancy in the 1990s was about a year and a half greater than it was in the 1980s. In less educated people, life expectancy increased by only half a year. Much of the growing mortality gap can be attributed to the higher levels of decline in smoking-related diseases such as lung cancer and chronic obstructive pulmonary disease in more-advantaged groups.87 Study authors attribute this to the larger declines in smoking prevalence in more advantaged compared to less advantaged groups that have been evident for some time in the United States.

The situation for Australia is much less clear.

A study published by the Australian Institute of Health and Welfare in 2006 indicated that death rates for cardiovascular disease reduced in all socioeconomic groups between 1999 and 2003. There was a decrease in the size of the gap between the rates of death between upper and lower SES groups for coronary heart disease and cardiovascular disease as a whole but an increase in the relative effect of disadvantage (the proportion by which the lowest SES group was higher than the highest SES group) for coronary heart disease, stroke and cardiovascular disease as a whole.88