Cancer data in Australia

There are several technical considerations of which to be aware when interpreting cancer incidence presented by country and region of birth. Further information to aid interpretation is provided in appendices B to D.

Data quality can limit the precision of findings

The country of birth (COB) cancer incidence data within this report has some data quality issues and so the precision of rates can often be in question. However, even acknowledging uncertainties associated with data quality, the data clearly suggests that incidence rates are influenced strongly by the country of birth and that the cancer risk for different populations, overall and by type, are often very different. We encourage those using the data to take the time to understand the data quality. Appendix B discusses the nature of the data quality issues.

What statistics are available?

This report provides cancer incidence for selected COBs and Regions of Birth (ROBs). The Standard Australian Classification of Countries 2016 is used as the basis from which COBs and ROBs are derived for this report. This report refers to Minor Regions of Birth as regions of birth.

The selected COBs and ROBs have an average minimum population of 20,000 or more people living in Australia between 2006 and 2020. Note that some ROBs are comprised of only one COB (for example, Australia, New Zealand, Ireland) and are presented in both contexts.

For each COB and ROB, the leading 20 cancers diagnosed for the 2006–2010, 2011–2015, 2016–2020 and 2006–2020 periods (by sex) have been released. These data accompany and inform this report and are available through a Tableau data visualisation and Excel spreadsheets are within the Data section.

Comparing incidence rates between countries or regions of birth

The report generally does not discuss specific countries or regions of birth as having the highest or lowest incidence rate. Records with missing place of birth information create uncertainty about what the actual rates would be with complete country of birth information. This, and the similarity of some rates, often prevents any such certainty that specific COBs or ROBs have the highest or lowest incidence. In addition, in some instances, the highest or lowest ROB/COB may change depending on whether the WHO or Australian 2001 standard population is used.

Rather than focussing on specific COBs or ROBs that are the highest or lowest, discussion focuses on COBs and ROBs that, in relative terms, are within tiers of higher or lower incidence. The use of tiers encourages the identification of COBs or ROBs where incidence appears reasonably different and may indicate a different risk for the population.

Making international comparisons

This report uses the International Agency for Research on Cancer (IARC) tool Cancer Today 2022 to make international comparisons. World age-standardised rates in Cancer Today are similarly age-standardised to the WHO world population and expressed per 100,000, and so are comparable to the WHO age-standardised rates used in this report.

A notable point of difference is the definition of ‘All cancers combined’. In the Australian Cancer Database (ACD), ‘All cancers combined’ includes notifiable rare types of non-melanoma skin cancer (NMSC) but excludes the highly common NMSC basal and squamous carcinomas of the skin. Cancer Today ‘All cancers’ estimates include all types of NMSC, whereas Cancer Today ‘All cancers excluding NMSC’ excludes all types of NMSC including (both common and rarer types).

Given the comparatively high incidence of basal and squamous cell carcinomas in the NMSC category, ‘All cancers excluding NMSC’ is considered a closer representation to the ACD definition of ‘All cancers combined’ and is used for comparisons throughout this report.

Complexities using regions of country or region of birth statistics

Time series

The ROB data accompanying this report includes 2006–2020 incidence rates and three 5-year periods within. The various time periods are provided primarily to establish whether various cancer incidence rates and comparisons appear to be generally representative. 

These data are less suited for time series analysis due to the small number of time periods, data quality issues (that is, the completeness of country of birth records and imputation may influence rates) and the population may be prone to changes in migration trends over time which may make finer level comparisons within the same general population less reliable.

Should comparisons across time-periods be undertaken, it is important to recognise that the population composition for a ROB can change relatively quickly across periods due to migration trends and changes in cancer composition and rates may be partly due to the changes in the composition of the population.

Population composition

Most ROBs are comprised of several different COBs. The demographic composition of Australian residents born in each ROB is related to the migration of different populations to Australia over time and mortality, and may not be representative of the demographic population in that region during the time periods being considered. 

ROB data has been released primarily because it will have larger populations than COB and this should generally lead to more stable and representative reporting than generally occurs with the smaller COBs. However, the ROB’s are in some instances predominantly represented by one or two main COBs. Where this occurs, the ROB’s cancer composition by type and incidence rates will be biased towards the rate and cancer composition of the most populous COBs. Where the ROBs population composition within the ACD is different to the region, the cancer incidence trends occurring within the region may be of less value understanding the ROB cancer incidence trends within Australia. This is particularly true for the regions and cancers where incidence trends within the region differ from country to country.

Where this report compares ROBs, both Central Asia and Melanesia are excluded from the report. These ROBs are both relatively small in terms of the population living in Australia and are comprised predominantly of one COB. Around 87% of the population born in Central Asia was born in Afghanistan while for those born in Melanesia, around 86% were born in Papua New Guinea. Cancer incidence is discussed separately for both of these ROBs within this report, with the discussion more focussed on the COB cancer incidence within the ROB.

Smaller populations

Analysis of the incidence rates for COBs/ROBs becomes more challenging for smaller populations as a small change in cancer case numbers can have a relatively large impact on incidence rates. While the incidence rate is technically accurate, it may not be representative. The provision of data for three 5-year periods helps to gauge if cancer incidence rates are generally representative of the population with consistently high or low rates providing a level of reassurance.

Sex composition may impact on COB and ROB comparisons and rates

Comparisons between regions and countries of birth as well as rankings may be influenced by the sex composition of the migrated population. In general, males have higher rates of cancer than females and the composition of a population by sex may influence rates for persons.

The sex composition for ROBs and COBs can be very different. For example, in 2006–2020 the Thailand-born population in Australia had a higher proportion of females (68%) while the Pakistan-born population had a higher proportion of males (60%). For the Thailand-born population, it is possible that some cancers more prominent in females will rank higher for the population (that is, persons data) because of the proportion of females in the population and the opposite may occur for the Pakistan-born population.

Comparing all cancers combined between a population with higher proportions of males compared to a population with higher proportions of females may see the comparison impacted by genuine differences in cancer incidence but also the difference in the sex composition of the different populations.