Aboriginal Injury-related Hospitalisation 1991/92 - Data Issues
![[Previous]](/images/arr_prev.gif)
![[Next]](/images/arr_next.gif)
![[Top]](/images/arr_top.gif)
Data Issues
Little is known about the overall reliability of the identification of
Aboriginal and Torres Strait Islander peoples in hospital data collections. It
is likely to vary from place to place and may vary from age to age: "Several
validation studies have shown that Aboriginal and Torres Strait Islander
peoples are often significantly under enumerated in health related data
collections and that the extent of under enumeration varies between State and
Territories, regions, even between different hospitals..." (Woodward &
Bhatia, 1996).
The extent of under identification in hospital separation data in Victoria has
recently been estimated: "For the first year of mandatory reporting of
Aboriginality of hospital patients there was a significant increase in the
number of Koori admissions. There were 2683 Koori admissions to public
hospitals reported in 1992-93. This increased to 4212 Koori admissions in
public hospitals for 1993-94"(Information on Koori Illness in the Community
Koori Health Unit of the Department of Human Services Victoria. World Wide Web
site address: http://hna.ffh.vic.gov.au/phb/hdev/koori/kh4.html). This
represents a 57 percent increase which suggests probable substantial under
identification in the 1991/92 Victorian data used in the present report. Even
if other states do not experience such a large discrepancy, it is likely that
the level of underestimation is high and therefore comparative rate ratios
between Aboriginal and Torres Strait Islander peoples and non-Aboriginal people
will underestimate the real difference in risk levels.
Identification may also vary from cause to cause. There is a large body of
evidence to suggest that violence to females in the non-Aboriginal community is
poorly identified (e.g. National Committee on Violence Against Women, 1992).
This may be less so for Aboriginal and Torres Strait Islander females where
both the patient and hospital staff may be more willing to record an incident
as violent.
In addition, little is known about how appropriate the International
Classification of Diseases (ICD) (World Health Organization, 1977) external
causes codes are for coding injury occurrences among Aboriginal and Torres
Strait Islander peoples. For example, it is not known how traditional
punishment practices are coded. They could be classified as legal intervention
in terms of traditional law; they may be coded as violence, especially if the
hospital is not aware of the reason behind the injury; or they may be recorded
as being accidental if this is the explanation given by the injured person.
In this paper, Aboriginal and Torres Strait Islander peoples were defined as
all persons who had been definitely identified as such in the hospitalisation
data set. All other persons including those whose Aboriginality was coded as
unknown were included in the comparison non-Aboriginal population group. This
is a conservative approach that may have included some Aboriginal and Torres
Strait Islander cases in the comparison group which therefore may have
underestimated some differences between the Aboriginal and Torres Strait
Islander peoples and non-Aboriginals. It is not possible to separately identify
Torres Strait Islander peoples nor to consider Aboriginal peoples with
differing cultural affiliations. Consequently there is no consideration of
variations in injury patterns within the Aboriginal and Torres Strait Islander
populations.
The best estimates of Aboriginal and Torres Strait Islander populations
have been published by the Australian Bureau of Statistics (ABS). Furthermore,
the ABS has studied the validity of the estimates of the Aboriginal and Torres
Strait Islander populations (Benham & Howe, 1994). It appears that there
has been a change in the underlying propensity of an individual to identify as
an indigenous Australian. Nonetheless, it is uncertain whether the
identification of the Aboriginal and Torres Strait Islander populations is
complete.
In this report, 1991 Census counts of the Aboriginal and Torres Strait Islander
populations were used as denominators for calculating rates of hospitalisation.
The more accurate five-year age group estimates of Benham and Howe are higher
than the 1991 Census counts across all age groups; the percentage by which the
1991 Census counts underestimate the more accurate estimates of Benham and Howe
are shown in Table 2.
Because the rates of Aboriginal and Torres Strait Islander injury presented in
this report are calculated using the 1991 Census population, they will be
higher than the rates that would have been obtained if the Benham and Howe
population estimates had been used. The degree of overestimation is
proportional to the level of underestimation of the populations in any specific
age and sex group. Differences in rates obtained using these two estimates of
the Aboriginal and Torres Strait Islander populations are, in general, small in
comparison to the differences observed between either of these rates and rates
for the non-Aboriginal population. The non-Aboriginal population was calculated
by subtraction of the Aboriginal and Torres Strait Islander population from the
total 1991 population of Australia (excluding the Northern Territory).
Table 2: Percentages by which the 1991 Census counts of the Aboriginal and
Torres Strait Islander population are less than the more accurate experimental
estimates of Benham and Howe
| Age group |
Males |
Females |
Persons |
| 0-4 |
4.5 |
4.2 |
4.3 |
| 5-9 |
4.5 |
4.2 |
4.4 |
| 10-14 |
3.3 |
3 |
3.2 |
| 15-19 |
8.1 |
6.3 |
7.2 |
| 20-24 |
13.6 |
5.2 |
9.5 |
| 25-29 |
14.3 |
5.8 |
10.1 |
| 30-34 |
10.1 |
3.6 |
6.8 |
| 35-39 |
11.2 |
3.9 |
7.5 |
| 40-44 |
4.9 |
2.5 |
3.7 |
| 45-49 |
4.2 |
5.2 |
4.7 |
| 50-54 |
3.9 |
1.8 |
2.8 |
| 55-59 |
4.3 |
6.1 |
5.2 |
| 60-64 |
5.3 |
2.7 |
3.9 |
| 65+ |
2.4 |
4 |
3.3 |
| Total |
7.4 |
4.4 |
5.9 |
Injury hospitalisation data from the Northern Territory were not
available at the level of detail required. The Northern Territory has the
highest proportion of Aboriginal people in its population of any State or
Territory (Table 3). The effect of not having hospitalisation data from the
Northern Territory is that the number of injury-related hospitalisations in
Australia is underestimated and it is likely that the underestimation of
Aboriginal and Torres Strait Islander peoples' cases is greater than that for
the non-Aboriginal population.
Table 3: Aboriginal and Torres Strait Islander Census count, 30 June 1991
| State or territory |
Number of Aboriginal and Torres Strait Islander persons |
Percentage of total state or territory population |
| New South Wales |
75,020 |
1.3 |
| Victoria |
17,890 |
0.4 |
| Queensland |
74,214 |
2.5 |
| South Australia |
17,239 |
1.2 |
| Western Australia |
44,082 |
2.7 |
| Tasmania |
9,461 |
2.0 |
| Northern Territory |
43,273 |
26.1 |
| Australian Capital Territory |
1,616 |
0.6 |
| Australia (includes other territories) |
282,979 |
1.6 |
Source: Australian Bureau of Statistics, 1994
Hospitalisation data do not measure the number of injury incidents but provide a count of the
number of hospital episodes that are injury-related. Different criteria for hospital admission in
different hospitals will result in similar cases being admitted in one setting and not in another.
Admission practices are highly policy sensitive (for example, they have been affected by the
introduction of DRG's) and may change over time because of the policy climate rather than an
underlying change in injury incidence, making time series of injury data difficult to interpret.
These differences may be systematic across states, making between state comparisons difficult and
impacting on the Australian aggregate.
The geographical distribution of the Aboriginal and Torres Strait Islander peoples and
non-Aboriginal population is very different. It is likely that practices relating to the
hospitalisation of Aboriginal and Torres Strait Islander peoples and non-Aboriginals differ because
of this. For example, remote dwelling Aboriginal and Torres Strait Islander peoples are less likely
to have local access to an inpatient hospital and therefore may be less likely to be admitted for
injuries of moderate severity than a person living in a less remote area. This is not certain,
however, because if the Aboriginal and Torres Strait Islander peoples are transported to a hospital
a long distance from where they live, they may be more likely to be admitted to that hospital than
a person with an identical condition who lives locally.
Age-adjusted rates have been calculated in addition to crude rates to control for the effect of
differences in the proportions of people of different ages (and different injury risks) in the
populations that are compared. Direct age standardisation was used, taking the 1991 population of
Australia, excluding the Northern Territory, as the standard.
The injury categories presented in this report are based on standard aggregations of the ICD9
external cause (E-code) classification. The data set is comprised of all cases where at least one
E-code was reported. The E-code equivalents for the causes examined in detail are presented in the
Definition Box at the top of each section. Further information is available from the National
Injury Surveillance Unit.
|