David W Donnelly1, Anna Gavin. 1. Northern Ireland Cancer Registry, Centre for Public Health, Queen's University Belfast, Mulhouse Building, Grosvenor Road, Belfast BT12 6BJ, Northern Ireland, UK. d.donnelly@qub.ac.uk
Abstract
BACKGROUND: There is a well established relationship between cancer incidence and socio-economic deprivation. The strength of this relationship may be subject to the choice of deprivation index used. METHODS: A range of possibilities for measuring area-based deprivation in Northern Ireland are investigated. The relationship between each measure and cancer incidence is described using standardised incidence ratios and age-standardised rates fitted with a log-linear model. RESULTS: Standardised incidence ratios for lung cancer comparing the most to the least deprived deciles were greater using an income measure (371.1, 95%CI: 355.4-386.9) than an employment measure (321.1; 95%CI: 307.9-334.2). Income, employment and education measures gave similar results for breast, prostate and colorectal cancers. Standardised incidence ratios generated for all cancers (excluding non-melanoma skin) using income deciles based upon census output areas (142.4; 95%CI: 139.6-145.1) were larger than those generated using super output areas (133.0; 95%CI: 130.3-135.7) or electoral wards (130.4; 95%CI: 127.8-133.1). CONCLUSION: While the choice of time period for measuring area-based deprivation makes little difference to relative inequalities, smaller geographic units and income based deprivation measures can produce a stronger relationship between deprivation and cancer incidence than other measures.
BACKGROUND: There is a well established relationship between cancer incidence and socio-economic deprivation. The strength of this relationship may be subject to the choice of deprivation index used. METHODS: A range of possibilities for measuring area-based deprivation in Northern Ireland are investigated. The relationship between each measure and cancer incidence is described using standardised incidence ratios and age-standardised rates fitted with a log-linear model. RESULTS: Standardised incidence ratios for lung cancer comparing the most to the least deprived deciles were greater using an income measure (371.1, 95%CI: 355.4-386.9) than an employment measure (321.1; 95%CI: 307.9-334.2). Income, employment and education measures gave similar results for breast, prostate and colorectal cancers. Standardised incidence ratios generated for all cancers (excluding non-melanoma skin) using income deciles based upon census output areas (142.4; 95%CI: 139.6-145.1) were larger than those generated using super output areas (133.0; 95%CI: 130.3-135.7) or electoral wards (130.4; 95%CI: 127.8-133.1). CONCLUSION: While the choice of time period for measuring area-based deprivation makes little difference to relative inequalities, smaller geographic units and income based deprivation measures can produce a stronger relationship between deprivation and cancer incidence than other measures.
Authors: David W Donnelly; Avril Hegarty; Linda Sharp; Anne-Elie Carsin; Sandra Deady; Neil McCluskey; Harry Comber; Anna Gavin Journal: J Cancer Epidemiol Date: 2013-12-22