Aborigines and tuberculosis: why they are at risk ¹

Tuberculosis was declared a global emergency by the World Health Organization in 1993. It is essentially a social disease which occurs mostly in people who live in poverty and overcrowding. There are predisposing host risk factors that increase the chance that the individual, once infected, will progress to tuberculosis disease. The infected individual carries the bacilli for variable lengths of time. Should the person's immune system become less vigilant, susceptibility to disease increases. Known risk factors for acquiring infection and progression to disease include contact with an infectious person, low body weight, alcohol abuse, diabetes mellitus, renal failure and human immunodeficiency virus (HIV) infection.		general topic - about tuberculosis
There is no firm evidence of tuberculosis among Aborigines before the arrival of Europeans. High rates of tuberculosis in Aborigines have been documented in the post-arrival period with most reports coming from early this century. Early reports suggested that Aborigines were a group of people with no immunity to tuberculosis, who were not at great risk while they maintained relatively isolated life styles but who, on contact with outsiders, were vulnerable to tuberculosis. This vulnerability was then compounded by overcrowding and poor living conditions. 5		specific topic - tuberculosis and Aborigines
Although comparatively little contemporary data about disease prevalence are available, it is possible to examine known risk factors for tuberculosis. This can be considered an important task while there is the possibility that one identifiable group in Australia is particularly vulnerable to tuberculosis. This study aimed to throw light on this public health concern by researching the available data on the relationship between tuberculosis and Aborigines. We aimed: to ascertain the available data; to determine morbidity and mortality, to ascertain the extent of known risk factors and to consider the public health implications of our findings.		research aims
Method We began by identifying available data on tuberculosis in Australian Aborigines. We identified morbidity and mortality data available from the Northern Territory (NT), considered available data about known risk factors for tuberculosis and the occurrence of these risk factors in Aborigines. We examined the usual risk factors for tuberculosis, with the exceptions of silicosis, gastrectomy, jejunoileal bypass and miscellaneous immunosuppressive illnesses for which we could not find information specific to Aborigines.		- the sentences in yellow in the following paragraphs outline the steps taken to gain the data which would provide answers to the research questions
We calculated the age-standardised rate ratio for hospitalisation for tuberculosis for Aborigines compared with non-Aborigines in the NT for 1979 to 1988. Hospital separation diagnosis is the diagnosis when the patient leaves by discharge, death, transfer to another hospital or discharge against medical advice. Numerator data were obtained from the NT Department of Health and Community Services and denominator data from the Australian Bureau of Statistics. Indirect age-standardisation was done by calculating the number of cases that would have been expected in Aborigines each year if Aborigines had the same rate as the same-sex non-Aborigines averaged over the years 1979 to 1988, and then taking a ratio of the observed to the expected rates. 5
The indirectly age-standardised rate ratio of death from tuberculosis for NT Aborigines (1979 to 1991) was compared with the total Australian rate for 1991. We examined 'tuberculosis' (ICD-9 codes 010-018) and 'late effects of tuberculosis' (ICD-9 code 137) separately.' Indirect standardisation was chosen because of the small number of deaths in the Aboriginal population, and the Australian population was chosen as the reference population because it was available (unlike the morbidity data) and because the non-Aboriginal NT population had too few deaths to use as a baseline. We grouped the years 1979 to 1991 into three-year groups with the exception of the most recent years (1988 to 1991), for which we grouped four years.
Ninety-five per cent confidence intervals (CI) were calculated using the method described by Bailar and Ederer. 8
Lastly, we looked at data for HIV-AIDS in Aborigines and examined the potential for problems with tuberculosis control based on what was known about HIV and sexually transmissible diseases in Aborigines. We used the hypothetical and rather simple model proposed by Nunn et al. to estimate the effect HIV could have on tuberculosis rates in Aborigines. 9
Results Current data Current data were sparse. Thomson reported evidence of decreasing rates of tuberculosis in Aborigines, noting that in 1982 the rate was 26.3 per 100 000 for all forms of tuberculosis. This was still about three times the Australian rate. He quoted an annual rate of 135 per 100 000 in the NT in 1969 to 1973, declining to 30.2 in 1982. In Queensland, the rate fell from an average rate of 151.9 per 100 000 in 1950 to 1972 to 26.2 per 100 000 in 1982.'°		Incidence of tuberculosis for Aborigines and non-Aborigines across Australia
Beilby et al. noted that in 1985 there were 36 cases of tuberculosis in Aborigines in South Australia, a rate of 20.8 per 100 000, which was three times the rate in the rest of South Australia without age- and sex-adjustment. Subsequent age-standardisation for the years 1978 to 1988 gave an annual rate of 91.6 (Cl 56.2 to 127.0) for males and 38.6 (Cl 16.7 to 60.5) for females. Beilby et al. also undertook Mantoux test surveys in four communities and found age-standardised prevalence rates of infection of 6.2, 10.6, 21.0 and 33.5 per cent in the four communities. The authors noted that there might be problems with both denominator and numerator identification and that the rural-urban gradient was conflicting for infection and disease reports. There did not appear to be any obvious cause for this conflict, although one suggestion of the authors was nutrition-related immunosuppression affecting Mantoux-positive rates.		Incidence of tuberculosis in SA
Krause and Britton reported that the rate of tuberculosis in Aborigines in the NT was 114 per 100 000 in 1989, with seven deaths (17 per cent of Aboriginal notified cases). They reported that the rate had fallen to 60 per 100 000 in 1991, and that there was over 90 per cent compliance with treatment.		Incidence of tuberculosis in NT
Queensland health authorities reported that the rate of tuberculosis (per 100 000) in indigenous Australians was 25.8 in 1991, 11.1 in 1990, 20.2 in 1989, 12.1 in 1988, 20.4 in 1987, 24.5 in 1986, and 22.2 in 1985. This compares with between 3.1 and 4.5 per 100 000 for all persons in Queensland over the same period. 15 Krause and Britton reported that the rate of tuberculosis in Aborigines in the NT was 114 per 100 000 in 1989, with seven deaths (17 per cent of Aboriginal notified cases). They reported that the rate had fallen to 60 per 100 000 in 1991, and that there was over 90 per cent compliance with treatment.		Incidence of tuberculosis in QLD
In the three years 1989 to 1991, there were 155 cases of tuberculosis notified in the NT, of which 109 (70 per cent) were in Aborigines. The NT contributes about 3 to 8 per cent of tuberculosis in Australia, and represents about 1 per cent of the total population. Aborigines represent about 24 per cent of the NT population.
The rate ratios of hospital separation for Aborigines in the NT compared with non Aborigines are shown for tuberculosis in Figures 1 and 2. Both male and female Aborigines had an estimated 3 to 29 times the rate of hospital separation of non-Aboriginal Territorians. Because separations reflect events rather than the actual number of individuals, this could be a slight overestimation of cases, or, because notifications may not be as accurate as hospital separation data, it may be an underestimation. For this reason we (informally) compared the total number of notifications from the annual reports of the NT Department of Health with the total number of hospital separations. No formal evaluation could be done because the notification data were not available by unit records and data for different periods were published. However, for the years 1978 to 1988, excluding 1981, there were 325 notifications and 329 hospital separations, indicating an overall correlation between the two data sets. However, this correlation cannot be checked by year, or by individual record.		pointing out a problem in data collection and the step taken to'get around' the problem
Mortality Table 1 shows the rate ratio of death from tuberculosis and the late effects of tuberculosis for Aboriginal males and females in the NT compared with the Australian rates. The estimates of the rate ratio of death are high, although the small numbers are reflected in the wide confidence intervals.		mortality data from NT
Risk factors Table 2 summarises risk factors for tuberculosis and what is known about the occurrence of the risk factors in Aborigines. This table emphasises the prevalence of known risk factors for tuberculosis in Aborigines. Of particular concern is the higher risk of sexually transmissible diseases with the associated implications for the risk of HIV in Aborigines. The advent of HIV in Aborigines is likely to lead to morbidity from tuberculosis.		extent of known risk factors for tuberculosis
Little data about the occurrence of tuberculosis and risk factors are available. However, Patel et al. during a study of central Australian Aborigines from 1984 to 1986 inclusive, found nine new cases of tuberculosis, of which three were in diabetic people. Assuming that the other six were not diabetic and using the estimated Aboriginal population of the whole central Australian region, we calculated the rate of tuberculosis for both the diabetic and the non-diabetic group, estimated the crude rate ratio and calculated 95 per cent confidence intervals. The crude rate ratio of tuberculosis in diabetic Aborigines compared with non-diabetic Aborigines is 15 (CI 3 to 44). The wide confidence intervals reflect the small numbers of people with both diseases, even though the rate ratio is high. These are the only data addressing the issue of diabetes and tuberculosis in Aborigines that we could find, but they highlight the importance of diabetes in a population predisposed to tuberculosis.		expands on the issue of risk factors
What could the future hold? Nunn et al. used a simple model to predict what the effect of HIV could be on a tuberculosis-infected population.Using the same argument and assuming that in the NT there are approximately 40 000 Aborigines, of whom 60 per cent are aged 15 or older, and assuming that in the adult population 25 per cent are Mantoux positive, there are 6000 Mantoux-positive adult Aborigines in the NT. This rate of Mantoux positivity seems reasonable considering the results from Beilby et al. and the rate of tuberculosis in the NT. Although HIV is thought to be uncommon in Aborigines at the moment, the rapid changes in HIV prevalence in Africa and Thailand, and the pre-existing risk factors for HIV in Aborigines, indicate that the prevalence could change rapidly. Assuming that 10 per cent of the adult population contracted HIV, about 600 people would have concurrent tuberculosis and HIV infection. Assuming a 10 per cent annual rate of progression of tve tuberculosis. This would be, over the total population of the NT, 60 in 40 000, or 150 per 100 000 in addition to those people contracting tuberculosis who were not HlV-infected. This remains a frightening prospect, and one that would cause an enormous drain on health services, as well as the great personal and social costs of such an increase.		this paragraphs looks at the public health implications of tuberculosis in an Aboriginal population - the effect HIV could have on tuberculosis rates in Aborigines
Discussion Aborigines already have high rates of infection and disease from tuberculosis. They suffer from social disadvantage which in turn exposes them to the risks of poverty, underweight, homelessness, overcrowding and poor nutrition. As well, they have a higher than average prevalence of risk factors for tuberculosis, such as diabetes mellitus, renal failure, alcohol abuse and smoking. Although as yet largely untouched by HIV-AIDS, Aborigines are in considerable danger of being affected by the AIDS epidemic given their high rates of sexually transmissible diseases. If HIV spreads to Aboriginal people, it follows that the tuberculosis-HlV interface will become a major public health problem. We have no knowledge about the heterogeneity of the population for either tuberculosis or risk factors for tuberculosis; it is likely that there are significant urban-rural differences at the very least. In particular, we found little evidence of tuberculosis in Aborigines in Western Australia, New South Wales or Victoria.		- morbidity rates for Aborigines - predisposition factors - risk factors and Aborigines - public health implications of the findings - comments on the limitations of the study and the data
Data about Aboriginal health remain of poor quality. Many routinely collected data sets fail to identify Aboriginality adequately. Census enumeration of denominator data is also imperfect. There are few data that provide more than a glimpse of the potential problem for Aborigines and tuberculosis. There is no potential to be confident about the size of the problem, nor to calculate risk versus benefits of preventive therapy. It is not possible to compare Aborigines with others with respect to the time taken for diagnosis of active disease, the effectiveness of contact tracing and the extent to which preventive therapy is used in skin-test-positive people. Completion rates of treatment have improved substantially in the NT, from a low of 34 per cent in the mid-eighties to 79 per cent in 1989, 88 per cent in 1990, 8:) per cent in 1991 and 100 per cent in 1992 (unpublished data, V. Krause, 1994). More information is necessary to develop specific public health programs. At the very least, it is essential that more information about tuberculin-skin-test positivity be obtained to consider preventive therapy for people infected with tuberculosis.		- suggested directions for future research
When the model proposed by Nunn et al. is applied to the Aboriginal population of the NT, it is clear that HIV, among other things, has the potential to have an enormous impact on tuberculosis rates in the future. The most important issues in tuberculosis control in countries such as Australia are the rapid and accurate diagnosis of disease, ensuring compliance with effective medication and ensuring that appropriate contact tracing measures are undertaken. These issues are more critical in high-risk groups such as Aborigines, and are compounded by insufficient information on compliance, outcomes of treatment, results of contact tracing and multi-drug resistance. Tuberculosis is a curable and potentially a preventable disease. A national tuberculosis program must recognise this at-risk group of Aboriginal people and support measures necessary to carry out effective control. With the prospect of HlV, and what it could mean for this already disadvantaged group of people, it is time to consider tuberculosis screening and, when appropriate, preventive therapy for Aboriginal people.		implications of research recommendations of the study

¹ AUSTRALIAN JOURNAL OF FUBLIC HEALTH 1995 Vol 19 No. 5 p.487-91
⁵Charles Kerr Department of Public Health and Community Medicine, University of Sydney
⁸Aileen J. Plant National Centre for Epidemiology and Population Health, Australian National University, Canberra
⁹Victoria L. Krause and John R. Condon Northern Territory Department of Health, Darwin