Spatial variability in breast cancer incidence, care and outcomes
This project was funded by the Health Research Council of New Zealand in partnership with Breast Cancer Foundation New Zealand.
Rationale: Disparities exist in breast cancer patients in New Zealand, which are expected to remain, or to increase for decades, and are likely to be due to variations in access to and quality of cancer care. Personalising these disparities by locating them geographically may help determine where to invest scarce resources for the greatest improvement in cancer care. This research aims to facilitate this by creating interactive cancer maps at different geographical levels.
Aims:
- to create maps of breast cancer incidence and mortality for the whole country; and
- to create maps of breast cancer diagnosis and treatment for the nine District Health Boards (DHBs) covered by the national breast cancer register.
Data sources used:
- Integrated data infrastructure (IDI) from Statistics New Zealand for incidence and mortality atlases
- National breast cancer register which consolidate four regional registers (Auckland, Waikato, Wellington and Christchurch) and covers about 63% of all breast cancer registrations in New Zealand for diagnosis and treatment atlases. Only the two contiguous regional registers (Auckland and Waikato) were mapped.
Geographical level:
- District Health Baord (DHB, 20 units)
- Super Data Zone (SDZ, 111 units)
- Intermediate Data Zone (IDZ, 595 units)
We developed intermediate data zones (IDZs) to ensure that the base geography of analyses has sufficiently large and comparable populations across New Zealand to support robust spatial statistical analyses of breast cancer data. IDZs used New Zealand data zones, which have a population range of 500 to 1,000, as a base building blocks. IDZs were built according to the following zone design criteria: they (a) have geographical contiguity; (b) respect higher geographies (e.g. District Health Board); (c) consider environmental factors (e.g. topography, railways, motorways, large water bodies); (d) respect urban and rural areas where possible; (e) have a population range of 5,000 – 10,000 for all ages based on the census 2013 population; and (g) have as much internal homogeneity as possible in terms of socioeconomic deprivation (IMD).
Similarly, we developed super data zones (SDZs) using IDZs as building blocks to deal with small-number issues, and yet provide sub-regional details. SDZs have a population range of 26,000 – 50,000 based on the census 2013 population.
Measures displayed in the Atlases:
- Rates (crude rates)
- Age standardised rates (calculated directly using the WHO standard age composition as the reference)
- Hot/cold spots (using Getis-Ord Gi* Hot Spot Analysis)
- Clusters and Outliers (using Anselin Local Morans I Cluster and Outlier Analysis)