The Australian Urban Health Indicators (AusUrb-HI) project, a collaboration between NCRIS facilities, AURIN, PHRN, and researchers, aims to improve understanding of urban and regional health. This dataset, developed with Cancer Council Queensland and the Australian Cancer Atlas, provides spatial indicators of breast cancer outcomes in Queensland, exploring associations between various determinants and cancer patterns. The dataset’s goal is to identify spatial variations in breast cancer outcomes within Queensland, and motivate further research into their drivers.
This dataset includes seven spatial indicators of breast cancer outcomes for Queensland women, modeled using Bayesian statistics, covering the period of 2000-2019. These indicators, presented as Standardized Incidence Ratios (SIRs) with 95% credible intervals and posterior probability differences (ppd), are at the Statistical Area Level 2 (SA2). The dataset provides modelled statistical indicators and geographic data, covering breast cancer outcomes, spatial epidemiology, health disparities, and cancer surveillance at the SA2 level (2016 ASGS) in Queensland, Australia. The intended application is for identifying geographical disparities in breast cancer, informing public health planning, and enabling further research.
The methodology uses linked health datasets from the Queensland Cancer Register, BreastScreen Queensland, and the Queensland Hospital Admitted Patients Data Collection. Bayesian spatial models with Besag-Yorke-Mollié (BYM) priors were applied, using a Poisson distribution with the log of expected counts as an offset, and incorporating spatial terms for neighboring regions. Models were validated using trace plots, density plots, Moran’s I statistics, and Geweke statistics.
The data is structured as a table, with each row representing an SA2 region, and columns representing indicator values and associated metrics. The complete attribute dictionary for the dataset is shown below:
- sa2: Five digital sa2 code using 2016 ASGS
- sa2name: Name of each sa2 area
-
urban: Categorical variable with three levels: 1. brisbane/gold coast; 2. Cairns; 3. Rest of Queensland, with the first two representing the urban areas in Queensland. Categorised based on sa3 name.
-
sir_incidence: Standardised incidence ratios (SIR) showing the variation in incidence of breast cancer. SIR is using the total Queensland cohort as the reference population.
- low_cri_incidence: The lower bound of the 95% credible interval for the sir_incidence.
- high_cri_incidence: The higher bound of the 95% credible interval for the sir_incidence.
-
ppd_incidence: The posterior probability difference of sir_incidence outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_localised: Standardised incidence ratios (SIR) showing the variation in proportion of invasive breast cancers that were diagnosed when localised. SIR is using the total Queensland cohort as the reference population.
- low_cri_localised: The lower bound of the 95% credible interval for the sir_localised.
- high_cri_localised: The higher bound of the 95% credible interval for the sir_localised.
-
ppd_localised: The posterior probability difference of sir_localised outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_survive5yrs: Standardised incidence ratios (SIR) showing the variation in surviving at least five years among women with a diagnosis of invasive breast cancer. SIR is using the total Queensland cohort as the reference population.
- low_cri_survive5yrs: The lower bound of the 95% credible interval for the sir_survive5yrs.
- high_cri_survive5yrs: The higher bound of the 95% credible interval for the sir_survive5yrs.
-
ppd_survive5yrs: The posterior probability difference of sir_survive5yrs outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_hospitalization: Standardised incidence ratios (SIR) showing the variation in hospitalization rates among women with a diagnosis of invasive breast cancer. SIR is using the total Queensland cohort as the reference population.
- low_cri_hospitalization: The lower bound of the 95% credible interval for the sir_hospitalization.
- high_cri_hospitalization: The higher bound of the 95% credible interval for the sir_hospitalization.
-
ppd_hospitalization: The posterior probability difference of sir_hospitalization outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_surgery: Standardised incidence ratios (SIR) showing the variation in rates of breast cancer surgery among women with a diagnosis of invasive breast cancer. SIR is using the total Queensland cohort as the reference population.
- low_cri_surgery: The lower bound of the 95% credible interval for the sir_surgery.
- high_cri_surgery: The higher bound of the 95% credible interval for the sir_surgery.
-
ppd_surgery: The posterior probability difference of sir_surgery outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_screen: Standardised incidence ratios (SIR) showing the variation in prevalence of regular breast cancer screening among women with a diagnosis of invasive breast cancer, SIR is using the total Queensland cohort as the reference population.
- low_cri_screen: The lower bound of the 95% credible interval for the sir_screen.
- high_cri_screen: The higher bound of the 95% credible interval for the sir_screen.
-
ppd_screen: The posterior probability difference of sir_screen outcome, with ppd>0.6 showing a significant difference from the Queensland average.
-
sir_optexp: Standardised incidence ratios (SIR) showing the variation in optimal breast cancer experience among women with a diagnosis of invasive breast cancer. SIR is using the total Queensland cohort as the reference population.
- low_cri_optexp: The lower bound of the 95% credible interval for the sir_optexp.
- high_cri_optexp: The higher bound of the 95% credible interval for the sir_optexp.
- ppd_optexp: The posterior probability difference of sir_optexp outcome, with ppd>0.6 showing a significant difference from the Queensland average.