Life Expectancy Comparison

Team name & bio

• Clare Coleman - Problem Owner
• Belinda Bennetts
• Rachel Bunder
• Marzieh Morovatpasand

The Problem

The Australian Bureau of Statistics (ABS) perform calculations to find the Life Expectancy of Aboriginal and Torres Strait Islanders. This is an important figure used in the Closing the Gap initiative which aims to reduce the disadvantages for Aboriginal and Torres Strait Islanders. Currently the life expectancy for all Australian men is 79.7 years and for Australian women it is 83.1 years. For Aboriginal and Torres Strait Islander men the life expectancy is 69.1 years and 73.7 years for women. This is is a 10.6 year difference for men and a 9.4 for women.

To perform these calculations the number of Aboriginal and Torres Strait Islanders who have died and their age are needed. This information is collected through death certificate information. Studies have been conducted linking death certificates with census data to try and estimate the accuracy of the number of these deaths.

In both situations the number of Aboriginals and Torres Strait Islanders are underestimated. A death certificates may not identify some one as being Aboriginal or Torres Strait Islander due to lack of information available at time of death. Census data is incomplete due to difficulties in reaching Aboriginal and Torres Strait Islanders and reluctance to be identified as Aboriginal or Torres Strait Islander. The ABS report that the census does not count 17.3% of Aboriginal and Torres Straight Island people.

To adjust for the underestimation, the ABS makes assumptions to about the levels of underestimation. The problem is in these assumptions. From the information found through the death certificate and census linkage information, the ABS determined that following percentages of Aboriginal and Torres Strait Islanders deaths are reported as such:

• 82% up to 14 years old
• 89% from the ages 15 to 59
• 78% for ages 60+

But we know that the picture is more complicated than this. Young Aboriginal and Torres Strait Islander men are more likely to be under represented in the Census and their linkage to death records was substantially different from other groups. Additionally, small changes to these assumptions can make large changes to the life expectancy. These changes in assumptions can mean we are currently over estimating the Life Expectancy for Aboriginal and Torres Strait Islanders. A dangerous assumption which can impact critical services which aim to improve Life expectancy.

There is no tool that simply compares life expectancy given different assumptions in a way that can convince others to change their underlying assumptions. Currently, the life expectancy calculation is done in a laboriously way which doesn't clearly show how a change in one assumption can impact the life expectancy. A tool which does this could impact many lives.

The Solution

The team has decided to create an interactive visualisation that can inform the general public on the magnitude of change in life expectancy by simply changing input parameters.

Application/Relevance

Each year the Prime Minister of Australia makes a report on the 'Closing the Gap' strategy which includes the life expectancy statistics for Aboriginals and Torres Strait Islanders. This application will help those who create the statistics that are reported more aware of the ramifications of the assumptions that are being made. This in turn will allow for more accurate life expectancy figures, which will allow for better targeted services to decrease the gap for Aboriginals and Torres Strait Islanders life outcomes.

Datasets

The two major reports used to create this application are:

ABS. (2013a). Death Registrations to Census Linkage Project - Methodology and Quality Assessment Australia 2011-2012. Canberra: Commonwealth of Australia. ABS. (2013b). Life Tables for Aboriginal and Torres Strait Islander Australians. Canberra: ABS.

Links

• Github repository
• Demo site hosted on GitHub Pages
• Axure Prototype(Note, you can select different graphs for men and women)
• Youtube video

Tech stack

Javascript and d3js for the webpage. Axure for the mockup. After Effects for the video.

Tradeoffs/analysis

What went well, what would you have done better?

What went well:

• Clare now have something to take to a conference next month. The product will be used in her talk.
• Have a fully functional phase 1 (see future functionality)
• We worked well together

what didn't go so well:

• Not being able to fully link the java script to the Belinda's mock up
• We needed more time for the video

Future functionality

We propose a three phase process to complete this project.

The first phase is what we have completed in this weekend: a website which allows the user to change underlying assumptions and compare the life expectancy calculated from these assumptions with the official life expectancy as calculated by the Australian Bureau of Statistics (ABS). The assumptions made by the ABS are also shown.

The next phase is to allow for multiple comparisons. Instead of having just one set of assumptions, we want to be able to compare multiple sets of assumptions to the official life expectation. This means we need the ability to add more comparison columns and still be able to adjust the assumptions in each.

The final phase is to extend this comparison to problems beyond life expectancy. The life expectancy problem is characterised by multiple inputs with a single output. Many problems have this structure. We would love the ability to change the number of inputs we are considering as well as the underlying model.