The average ‘cancer burden’ is measured in the Atlas by two separate concepts. First, the risk of being diagnosed with cancer and second, the excess risk of dying from that cancer within five years of diagnosis.
Excess risk of dying from a cancer
When diagnosed with a cancer, a person may ask “Am I likely to die from this?”
There are many ways to answer this question. The statistical models used for the Australian Cancer Atlas are based on the numbers of deaths within 5 years of diagnosis of the cancer. Hence, the 5-year excess death rate is reported.
Click for further information
Excess death rate is calculated from survival, or how long a person lives after being diagnosed with cancer. One of the most common measures used for survival is a concept known as ‘net survival’, which considers only the deaths that are specifically associated with the cancer diagnosis. When using data from cancer registries, as the Australian Cancer Atlas does, ‘net survival’ is typically estimated by comparing the death rate among people diagnosed with cancer to the death rate among people of the same age without cancer, expressed as a percentage. A 5-year relative survival of 60%, for example, means that cancer patients are 60% as likely to survive five years compared to the general population. For calculation purposes, the excess death rate is equal to (100-relative survival). So, a relative survival estimate of 60% translates to a 40% excess death rate.
What does this mean in practice? A 40% 5-year excess death rate means that people diagnosed with cancer are 40% more likely to die within five years of diagnosis than the general population, after accounting for age at diagnosis.
To compare the excess death rate between areas across Australia, the excess death rate for a specific area is compared to the Australian average over the same period. This is known as the Excess Hazard Rate, or EHR for short. The EHR indicates whether the number of observed number of deaths due to that cancer is higher (shown in orange/red) or lower (shown in blue) than the Australian average.
Click for further information
In the Australian Cancer Atlas, an EHR value greater than 1 (shown in orange/red) mean that, overall, people living in the area have a higher risk of dying from their cancer within five years of diagnosis than the Australian average, while values below 1 (shown in blue) mean that, overall people living in the area have a lower risk of dying from their cancer within five years of diagnosis than the Australian average.
The EHR in the Australian Cancer Atlas reflects the overall risk of dying from a cancer within a specific geographic area for people “at-risk” during 2006-2014. All the estimates presented in the Australian Cancer Atlas are about the average cancer burden for all people living within a specific geographical area. Since people are different, with different lifestyles and habits, this overall risk will not reflect the risk for every individual living in that area. The aggregated Australian cancer numbers (new cases per year, rate per 100k) are reported for the time period 2006-2014.
‘Uncertainty’ of excess death estimates
A single number is not enough to understand the excess deaths in an area. Each estimate has a degree of uncertainty around it. If an estimate has a high level of uncertainty, then there is less confidence of what the true estimate is.
Areas with smaller numbers of cancers diagnosed are generally more likely to have more uncertainty around their estimates of excess deaths, and therefore lower confidence.
In the main map for the Australian Cancer Atlas, the level of uncertainty is shown by the level of transparency in the map colours. The more uncertain the estimates are, the higher the transparency. This means that even if the EHR for an area is larger than one, if there is a high level of uncertainty, then the high transparency will mean it looks similar to the Australian average (that is, yellow). The transparency effect in the map can be turned on and off.
Information about the uncertainty around the EHR estimates can also be found in the V-plots and wave plots.