Using linked data to inform targeted skin cancer screening in Australia

Data source

The data source used for this project was the Person Level Integrated Data Asset (PLIDA). The PLIDA is a partnership among Australian Government agencies to develop a secure and enduring approach for combining information on health, education, government payments, income and taxation, employment, and population demographics to create a comprehensive picture of Australia over time. The PLIDA is managed under the custodianship of the Australian Bureau of Statistics (ABS), and more details can be found on their website (ABS, n.d.).

Creating the cohort

To create the cohort, data from the 2016 Census of Population and Housing were linked to death registrations from state and territory registrars of births, deaths, and marriages using Version 8 of the Person Linkage Spine.

At the time of the analysis in PLIDA, 26 September 2025, death registration data for the project was available up until 31 December 2022. The use of the 2016 Census of Population and Housing allowed for the follow-up of about 76 complete months of death registrations (from 09 August 2016 to 31 December 2022).

The linkage rates between the spine and death registrations were good, with about 99% of all cancer deaths that occurred between 09 August and 31 December 2022 linking to the spine.

For the 2016 Census of Population and Housing, those who identified as overseas visitors on Census night were excluded. Once those people were excluded, the linkage rates between the spine and the 2016 Census of Population and Housing were good, with about 90% of all records linking to the spine.

Once each data source was linked to the spine separately, they were joined using the person linkage spine. About 80% of all cancer deaths that occurred between 09 August 2016 to 31 December 2022 mapped to a record on the 2016 Census of Population and Housing, when joined via the spine. Deaths (counts and rates) in this report are therefore an underestimate. Any exact duplicate records were then de-duplicated prior to analysis.

Defining events and priority populations

In this project, deaths from melanoma, non-melanoma skin cancer (NMSC) and all cancers combined were the ‘event’ of interest. Deaths were defined based on the underlying cause of death and coded using the International Statistical Classification of Diseases and Related Health Problems 10^th Revision (ICD-10) as follows:

• Melanoma – ICD-10 code C43.
• NMSC – ICD-10 code C44. NMSC includes basal cell carcinoma, squamous cell carcinoma and rare skin cancers such as sebaceous carcinoma.
• All cancers combined – ICD-10 codes C00–C97, D45–D46, D47.1 and D47.3–D47.5.

These codes were used to capture deaths where the underlying cause was skin cancer or all cancers combined, but they might not capture all deaths due to skin cancers where the underlying cause of death was coded to a cancer of unknown primary site.

Variables used to derive priority population groups were extracted from the 2016 Census of Population and Housing (a list is provided in Table 1).

Source: ABS (2016b).

Where possible, priority population groups were presented by standard classifications. However, in consideration of our obligation to ensure privacy and confidentiality, and to minimise the risk of unreliable rates due to small counts, it was sometimes necessary to further combine categories. To minimise the risk of biasing any rates, we combined categories where trends in mortality were similar.

Aboriginal and Torres Strait Islander (First Nations) people

First Nations people were defined as any respondent who identified as either Aboriginal or Torres Strait Islander on Census night. It should be noted that Indigenous status is collected through self-identification and changes to identification impacts the count of First Nations people. The 2021 Census of Population and Housing saw a 25% increase in self-identification of First Nations people, compared to the 2016 Census (ABS 2022). The figures presented in this report are therefore likely to be underestimates for First Nations people compared to the figures produced from the 2021 Census.

Age group

Age group was based on the age reported on Census night and was recorded in single years.

People from culturally and linguistically diverse (CALD) backgrounds

People from CALD backgrounds are presented in this report through their ancestry and region of birth.

Where possible, ancestry was categorised according to the broad and narrow groups listed in the Australian Standard Classification of Cultural and Ethnic Groups (ASCCEG), 2016 (ABS 2016a).

People could provide up to two responses for ancestry and, where this occurred, they were assigned against both groups. Consequently, the sum of narrow ancestry groups will not necessarily equal the total for the broad group, and the sum of broad groups will not necessarily equal the Australian total.

Where possible, region of birth was categorised according to the major and minor groups listed in the Standard Australian Classification of Countries (SACC), 2016 (ABS 2016c).

People with disability

The 2016 Census of Population and Housing included an item on whether an individual has a profound or severe disability. The definition used for this item includes people who need assistance in their day to day lives with any or all the following core activities – self-care, mobility or communication because of a disability, long-term health condition (lasting six months or more), or old age.

If a person’s only answer to the reason question was ‘old or young age’ and that person was under 40, then they were coded to, ‘Does not have need for assistance with core activities’.

Occupation

Where possible, occupation was categorised according to the major and sub-major groups listed in the Australian and New Zealand Standard Classification of Occupations (ANZSCO), 2013, version 1.2 (ABS 2013).

Occupation referred to the main job held by employed people in the week prior to Census night and was applicable to all employed persons aged 15 years or older.

People who were unemployed, not in the labour force, had a labour force status of ‘not stated’ or were under 15 years of age were considered not applicable for this item.

People living in rural and remote areas

Rural and remote regions are presented in this report through Remoteness Areas (2016) boundaries. Data by Primary Health Network (2017) boundaries are also included in the supplementary data tables. All results are based on the Statistical Area Level 2 (SA2) (2016) boundary of usual residence at Census night.

Remote and very remote geographic areas have been combined for the purpose of this report. Primary Health Network (PHN) areas cover 31 geographic areas across Australia, and 2017 boundaries are defined by the Australian Government Department of Health, Disability and Ageing.

Measures calculated for rural and remote areas were compiled by applying a geographic concordance to the unit record data. Where records crossed more than one boundary, they were attributed to each area based on the percentage of the SA2 population that fell within that respective area.

Sex

The sex of an individual was recorded as being either male or female in the 2016 Census of Population and Housing.

People living in lower socioeconomic areas

Socioeconomic outcomes are presented in this report through the Socio-Economic Indexes for Areas (SEIFA) – Index of Relative Socio-Economic Disadvantage (IRSD) (2016). Data by Highest Level of Educational Attainment (HEAP) are also included in the supplementary data tables.

Highest Level of Educational Attainment was coded according to the Australian Standard Classification of Education (ASCED), 2001 (ABS 2001). People who were under 15 years of age were considered not applicable for this item.

People were assigned to a SEIFA IRSD quintile based on the SA2 2016 boundary of usual residence at Census night. Population-based Australian cut-offs for SEIFA quintiles have been used in this report and are calculated by dividing SEIFA areas into 5 groups in such a way that the population in each group is approximately equal. As SEIFA measures the characteristics of an area rather than individuals, the population in the most disadvantaged population-based quintile (‘1 – lowest’) is the 20% of the population living in the most disadvantaged areas, rather than the most disadvantaged 20% of the population.

Age-standardised rates

Age-standardised rates are hypothetical rates that would have been observed if the populations studied had the same age distribution as the standard population. This facilitates comparisons between populations with different age structures. This adjustment is important as mortality from cancer varies with age.

The direct method of age-standardisation was applied to the data. Age-standardised rates were derived by calculating crude rates by 5-year age groups of 0–4 to 85+, unless otherwise stated. These rates were then given a weight that reflected the age composition of the standard population, that is, the Australian Standard Population as of 30 June 2001. It is important to consider the age groupings used to derive the age-standardised rates, as differences in the age grouping used will mean that results between priority populations are not directly comparable.