Demographic, health and socioeconomic characteristics related to lung cancer diagnosis: a population analysis in New South Wales, Australia

Abstract

Abstract Background Lung cancer is a major cause of health loss internationally, and in Australia. Most of that loss is inequitably concentrated among vulnerable or disadvantaged people and amenable to prevention and earlier detection. In response, best practice lung cancer care considers peoples’ background, circumstances and care needs. Comprehensive, person level descriptions of demographic, health and discrete socio-economic disadvantage related factors are therefore required to inform best practice. We examine population wide correlations of demographic, health and socioeconomic characteristics with lung cancer diagnosis for use in cancer control programs, including screening. Methods A study of 5,504,777 (89.9%) adults living in New South Wales and participating in Australia’s Census in August 2016 with subsequent follow-up to the end of 2018. The Australian Bureau of Statistics’ (ABS) person-level integrated data asset linked census records with the NSW population cancer registry which includes primary site. Our study compared census participants who did not experience cancer in the follow-up period with those diagnosed with lung cancer, (n = 6160 and ICD10 C33-34). Outcomes are expressed as the adjusted relative odds (aOR) of incident lung cancer among adults in the community and measured using multi-variable logistic regression models. Validated ABS methods informed categorisation of social and economic variables. Results Multivariable comparison of those with lung cancer and those without a first cancer diagnosis (3276 lung cancers among 2,484,145 males; 2884 lung cancers among 2,944,148 females) showed associations with increasing age, varying ancestry, living alone (aOR = 1.30 95% CI 1.19–1.42 males; 1.24 95% CI 1.14–1.35 females), number of health conditions medicated, less than Year 12 education (aOR = 1.40 95% CI 1.30–1.51 males; 1.37 95% CI 1.27–1.48 females) and housing authority rental (aOR = 1.69 95% CI 1.48–1.94 males; 1.85 95% CI 1.63–2.11 females). Additional associations occurred among males with low income, disabilities before age 70, those unemployed and labouring occupations. As numbers of characteristics increased, so did the likelihood of lung cancer. Conclusion We provided a population wide description of characteristics relevant to lung cancer diagnosis. Deeper knowledge of these characteristics inform continuing development of lung cancer programs in prevention (e.g. tobacco control) and detection (e.g. lung cancer screening), then help prioritise targeted delivery of those programs.