As discussed above, the estimates of per capita health expenditure by income group have a great deal of sampling variability and therefore high standard errors. We are therefore very constrained in our ability to determine whether there are genuine differences in health expenditure by income and whether there are differences between the Indigenous and non-Indigenous populations.

In an attempt to overcome this uncertainty, further analysis of usage of health services by equivalent income groups was conducted by focusing on those who reported no usage in the two weeks prior to interview. This is motivated in part by the fact that estimated proportions of people by Indigenous origin and income group will have less sampling error and therefore be more stable than the estimated total health expenditure by Indigenous origin and income group (Table 5.1). Another advantage of this shift in focus is that it provides a direct measure of Indigenous utilisation of health services, albeit one that does not capture the intensity of usage of respective services.

For the Indigenous population, there is no statistically significant relationship between the usage of health services and equivalent income, irrespective of the equivalence scale used.

A similar pattern is revealed for the non-Indigenous population. The only difference is that the proportion utilising health services in the lowest income group is significantly smaller than for the highest income group for the raw and per capita income measures. For example, using the raw family income groups, the proportion utilising health services rises from 51.5 per cent for the lowest income quintile to 54.3 per cent for the highest income quintile.

Note: The utilisation rates are based on the services described in Appendix A. The standard errors of the estimates of expenditure are presented in parentheses. See Appendix D for details of their calculations.

Note that there are substantial and significant differences between the Indigenous and non-Indigenous populations in the usage of health services. Low-income Indigenous people are much more likely than non-Indigenous people in the same income category to report not having used any health services (except for those classified using raw family income). The differential in usage of health services of Indigenous and other Australians is about 14 percentage points for the Henderson, new OECD and per capita scales, and is 7.3 percentage points for the raw family income scale. Given that less than 10 per cent of Indigenous people are aged 55 and over, compared with around 25 per cent of other Australians, demographic factors are likely to play a major role in explaining this differential. However, demographics cannot be the whole story because the proportion of older people (aged 55 years or more) in the bottom quintile of the per capita income measure is only marginally different between Indigenous and other Australians.

For the higher income groups, the estimated proportion of Indigenous people not using health services is also larger than for non-Indigenous people, but the differential is only statistically significant for the raw family income and Henderson scales. The relatively small difference between usage of health services between high-income Indigenous and other Australians (at least by this measure) is largely driven by the standard errors of the respective estimates, with non-Indigenous estimates being more reliable.

To summarise, in contrast to the earlier analysis, the level of Indigenous usage (or lack of usage) of health services across income groups appears to be similar to that of other Australians. Given the apparently weak relationship between income and the proportion without health expenditure, differences in the relationship between income and expenditure for Indigenous and non-Indigenous populations must be driven by the amount expended by those who spent some money on health.^[17] In the next section we explore one possible explanation for this observation, in the context of establishing the relationship between self-assessed health status and equivalent income.