Incorporating the AIHW National Injury Surveillance Unit
The distribution of selected causes of injury death by Statistical Subdivision
This publication was compiled by NISU from Unit Record data supplied by the ABS, State Registrars of Death and Coroners. Please acknowledge data obtained from this publication as follows:
An Atlas of Injury Death in Australia 1990-1992, NISU Adelaide 1995.
Section 1 - Introduction
The purpose of this atlas.
National injury statistics provide an insight into the overall incidence of injury. Individual States produce separate analyses which show significant variation between States. Systematic variations are not however restricted to the State level. NISU's Bulletin on the spatial distribution of injury deaths in Australia: Urban Rural and remote Areas showed significant difference in injury patterns between areas. This injury Atlas considers an even finer geographical dissection of Australia. It aims to provide more information about variations from the national average with a view to ensuring that those who plan policies and prevention strategies can take into account these differences.
The National Injury Surveillance Unit has prepared this atlas to test the feasibility of the routine production of such a publication covering the peri censal triennium. It is therefore not a comprehensive coverage of all injury types or rosk and age groups. The topics and the geographic unit (the statistical subdivision) chosen for mapping, are those where there were sufficient numbers for meaningful interpretation and in which there was national interest.
The atlas should be considered as a pilot project, and comments on its usefulness and suggestions for improving it are welcome.
Population data were obtained from the Australian Bureau of Statistics. The data in this report were released on disc from the ABS in September 1994 from the 1991 Census of Population and Housing. Populations are based on the Statistical subdivision (SSD) of usual residence. Definitions of SSD were published by the Australian Bureau of Statistics in 1991 Census Geographic Areas Cat No. 2905. SSDs are contiguous geographic areas made up of a number of statistical local areas. Their boundaries are determined by the Australian Bureau of Statistics and State and Territory governments.
Death data are obtained in unit record form from the Australian Bureau of Statistics. In these tables, each death has been reported according to the Statistical Subdivision of usual residence and the calendar year in which it was registered. Of the 23,127 deaths registered in 1990, 1991 or 1992, 5.5% occurred before 1990.
Standardised mortality ratios
Age adjustment of data is used to calculate overall population rates which take into account the age distribution of the underlying population. This permits more reliable comparisons to be made between States with different age profiles and over time as age profiles shift. Age adjustment is to the 1991 Australian population for which usual residence at Statistical local area has been recorded. Standardised mortality ratios have been calculated using the indirect method. In data referring to young males, adjustment has been to the male population Australia aged 15 to 29 years.
Use of data based on small numbers of deaths
The tables in this report have been abridged to suppress rate information based on 3 or fewer cases. Care must be taken when interpreting rates based on small numbers of deaths. Information on methods for comparing population based rate data when case numbers are small can be obtained from NISU at NISU.
Information in the mortality collection originates with coroners, medical practitioners and persons familiar with the deceased, is recorded by State and Territory Registrars of Births, Deaths and Marriages, and is further processed by the ABS. The key data item for present purposes, 'Cause of death', contains the four digit International Classification of Diseases (ICD9) external causes code (E-Code) attributed by ABS. The cause groups included in this report are created by the aggregation of E-Codes.
Definitions of causesMotor vehicle occupants
This is defined as cases where the cause of death is in the E-Code range 810-825/.0, 810-825/.1
This is defined as cases where the cause of death is in the E-Code range E800-E869 , E880-E929
This is defined as cases where the cause of death is in the E-Code range E800-999 and therefore includes all accidents poisoning and violence.
This is defined as cases where the cause of death is in the E-Code range E950-E959
Interpretation of Maps.
Injury death has been mapped according to the SMR of the SSD. It is common when mapping to calculate ranges so that a constant number of areas fits into each range. In this atlas however the same aggregations of SMRs have been used across all maps. The key to the colour-coding of SMR ranges is repeated on the first map of each set. The number of SSDs in each SMR range differs between the sets of maps, and is shown in the keys. This approach means that the same colour indicates the same SMR range throughout the atlas. The ranges were chosen to reflect the magnitude of difference from the Australian SMR of 100. The centre range, for example, includes SSDs with an SMR which ranges from one and a half times lower and higher than the Australian SMR to one and a half times higher (66.8 - 150)
Statistical Subdivisions are chosen by the ABS for administrative reasons. They are aggregations of local government areas or Statistical Local Areas. The death rate and other characteristics such as age distribution and socioeconomic status within subsections of the SSDs may vary considerably. Differences between smaller areas are therefore masked and some sensitivity is lost. The largest differences between SSDs are likely to occur where SSDs cover a relatively homogenous population.
Maps should therefore be interpreted with care. They provide a guide to injury patterns in small regions which need to be interpreted in the light of other knowledge about the region. Their principal use is to reveal consistent patterns which suggest that the injury experience of some types of regions is different from that in other regions. Such patterns may suggest avenues for prevention.
Areas with small populations will have only few injury deaths, even over a three year period. In these areas even a small variation in the number of deaths would cause large changes in the death rate. For this reason the tables in this report provide information on the number of deaths on which statistics have been calculated and summary information on the populations of each SSD are also provided. Confidence intervals around SMRs and crude rates have not been shown. The information in the tables is sufficient to permit the calculation of confidence intervals and the estimation of probability of difference between SSDs where this is deemed necessary. Confidence intervals are largest for the areas with the smallest populations, therefore it is possible to have more statistical certainty about densely populated areas. It is important however not to refrain from making judgements about sparsely populated areas on the grounds of statistical uncertainty when a close look at the maps indicates that there are consistent patterns of higher injury death rates from some causes in the less densely populated areas. NISU Bulletin 8 The Spatial distribution of injury deaths in Australia: Urban, Rural and Remote Areas shows that rural and remote areas experience higher injury death rates. While differences between individual sparsely populated areas need to be treated with caution, it is important to consider them in the light of broader patterns revealed by the maps.
Using the Atlas
Each set of maps shows data for Australia with enlargements for each capital city. The corresponding tables present the equivalent data sorted in descending SMR order. Two final sections contain tables which show the whole range of indicators sorted in the traditional ABS order of presenting SSDs within State or Territory and the populations of SSDs.
The Atlas allows the reader to identify interesting patterns on the map, obtain additional information from the accompanying data table, including the name of the SSD and then look up the pattern of injury for that SSD and its population in the concluding sections.
Here you can do some simple searches on Statistical SubDivisions and Statistical Local Areas.
If you are not sure of the SSD or SLA name, try the name of the local council or suburb for the area you want to find. In many cases the Local Government Area name is the same as the Statistical Local Area name.
SSD QueryThe first option allows you to enter a Statistical Sub Division name and find what Statistical Local Areas are grouped under that Statistical Sub Division.
SLA QueryThe second option allows you to enter a Statistical Local Area name (or part of the name) and find out in which Statistical Sub Division that name is grouped, what other Statistical Local Areas fall into the same group, and display the selected data for those SSDs.