Geographical Variation In Stage At Diagnosis Of Colorectal, Lung And Ovarian Cancers In England, 2008-2013

Chiara DI GIROLAMO, London School of Hygiene & Tropical Medicine, United Kingdom
MULLER P. 1 , MORRIS M. 1 , RACHET B. 1

1 Cancer Survival Group, Department of Non-Communicable Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UK

Purpose: Stage at diagnosis is a key predictor of cancer survival and a key component for cancer control policy. Clinical Commissioning Groups (CCGs), which plan health services in England, are required to report the proportion of tumours diagnosed at early stages as an indicator of their cancer management quality. However, information on stage in population-based registries often remains incomplete. We seek to investigate the recent variation in stage at diagnosis of colorectal, lung and ovarian cancers between CCGs in England.
Methods: Clinical and socio-demographic information on cancer cases was retrieved from various population-based sources. CCG characteristics were obtained from official publications and sources. Multiple imputation was performed to derive the likely stage when missing. Random intercept models were fitted to quantify variations in stage between the 211 CCGs and identify patient and CCG factors which could explain between-CCG variation.
Results: The study included 99,942 patients diagnosed with colorectal cancer (2010-2012), 203,215 with lung cancer (2008-2013), 14,641 with ovarian cancer (2011-2013). Among those with known stage, proportions of stage I-II were 41.6% for colorectal, 21.1% for lung, 33.9% for ovarian cancer. Variation in the stage distribution exists between CCGs and the results describe the potential roles in this variation of individual factors (deprivation, tumour characteristics and comorbidities), and system features at CCG level (availability of specialists and tests, and percentage of budget spent on cancer).
Conclusions: A thorough understanding of the extent and the nature of geographic variations in stage at diagnosis is relevant from a cancer policy perspective. These findings will help identify individual and system factors responsible for differences in CCG outcomes and hence inform public health interventions aimed at shifting diagnosis to earlier stages, ultimately improving cancer survival.