Oral Presentation Australian Epidemiology Association ASM 2018

Temporally changing loss of life expectancy for people with cancer in WA (#69)

Cameron Wright 1 , Rachael Moorin 1
  1. Health Systems and Health Economics, School of Public Health, Faculty of Health Sciences , Curtin University, Perth, Australia, Western Australia , Australia

Background: Cancer survival has improved in Western Australia (WA) over recent decades. This improvement has been in the context of increasing general population life expectancy. Estimated loss of life expectancy is a useful cancer survival measure.

Aim: To estimate the absolute and proportionate loss of life expectancy at the median age of diagnosis for cancer overall, female breast, colorectal, prostate and lung cancers, and melanoma in WA, in 1990, 2000 and 2010.

Methods: Person-level linked cancer registry and mortality data for first invasive cancer diagnoses (age 15 to 89 years) between 1983 and 2011, with follow-up to 31/12/2011, were used to estimate loss of life expectancy. Estimates were obtained using flexible parametric survival models, adjusted for sex (where relevant), age at diagnosis and year of diagnosis. Life expectancy of the general population was used as the reference, by constructing sex, age and year-specific life tables for WA.

Results: Marked reductions in loss of life expectancy were observed for cancer overall, female breast, colorectal and prostate cancers, and for melanoma. There was no such reduction for lung cancer. The median age at diagnosis was reasonably stable, except for prostate cancer, which had a reduction in median age at diagnosis from 74 years in 1990, to 66 years in 2010. The proportionate loss of life expectancy decreased from 33% to 7%.

Conclusion: These estimates facilitate assessment of population-level changes in loss of life expectancy following a cancer diagnosis. Stratifying by cancer stage would make these estimates more useful at an individual level.