Data on hospital separations are often used as indicators of morbidity. However, as Deeble et al. (1998: 46) point out, these are imperfect measures as high rates may reflect not only serious morbidity but inadequate primary care or specialist services (especially in areas where Indigenous people are the predominant population). Low rates, on the other hand, may simply be the result of difficulties of access. In either event, the decision to hospitalise is often subjective and based on different perceptions of the need for hospital care on the part of doctors and health workers.

Cunningham and Beneforti (2000) have produced a major study of Indigenous and non-Indigenous hospital statistics that has greatly assisted in the analysis of morbidity. Despite implementation of the National Aboriginal and Torres Strait Islander Health Information Plan, as well as a series of framework agreements involving the federal, state and territory governments, inadequate identification of the Indigenous population within hospital records remains a major constraint for analysis. As a consequence, comparisons of the Indigenous population with other Australians using hospital statistics will underestimate the true differences between the two populations.

Notwithstanding under-reporting, and after adjusting for age differences, almost twice as many hospital separations were reported in 1997–98 for those identifying as Indigenous than would have been expected if they had experienced the same rates as the total population. Higher rates of Indigenous hospitalisation were also reported in the Northern Territory, Western Australia and South Australia, although the extent to which this reflects jurisdictional differences in the completeness of Indigenous identification in hospital records is unknown. At the same time, hospitalisation rates also appear to be highest in remote areas (more than twice the non-Indigenous rate), while rates for Indigenous people in rural and metropolitan centres were still between 1.5 and 1.7 times higher than for the rest of the population.

One striking observation in the Cunningham and Beneforti analysis, which is of direct relevance for the present study, is that fully 98 per cent of separations identified as Indigenous in 1997–98 occurred in public hospitals, compared with only 68 per cent of non-Indigenous separations. This partly reflects the under-identification of Indigenous patients in private hospitals, although it is consistent with the pattern of highest public health spending among the lowest income groups.

By far the largest reason for hospitalisation among those identified as Indigenous in 1997–98 was regular and repeat visits for dialysis. A sense of the much greater burden of hospital care due to dialysis for Indigenous people compared with the rest of the population is provided by the standardised morbidity ratio (SMR) for this cause of 6.7 for males and 11.2 for females.^[3] Use of SMRs as a guide to other major causes of Indigenous hospitalisation reveals relatively high ratios (>2.0) for endocrine/nutritional and metabolic disorders, infectious diseases, respiratory diseases, and diseases of the skin and subcutaneous tissue. However, as a proportion of all Indigenous separations, complications of pregnancy and childbirth, respiratory diseases and injury dominated for women, while the most common causes for men were respiratory diseases and injury.