1. Risk of skin cancer following phototherapy for neonatal jaundice: retrospective cohort study David H Brewster, 1,2 Janet S Tucker, 3,4 Michael Fleming, 1 Carole Morris, 1 Diane L Stockton, 1 David J Lloyd, 3,4 Sohinee Bhattacharya, 3,4 James WT Chalmers 1,2 1 Information Services Division, NHS National Services Scotland 2 University of Edinburgh 3 University of Aberdeen 4 Aberdeen Maternity Hospital Scottish Cancer Taskforce Workshop: Better Awareness of Cancer Symptoms 22 January 2010 Does late presentation explain the apparent survival deficit for cancer in Scotland?

2. Outline Background – selected results from the EUROCARE-4 study Potential explanations for survival variations Evidence for more advanced disease at diagnosis in the UK Potential explanations for this England’s National Awareness and Early Diagnosis Initiative (NAEDI) Conclusions

4. Some factors to consider in population-based survival comparisons Data quality factorsTumour-related factors Population coverageExtent of disease Completeness of ascertainmentSite (and sub-site) of tumour Accuracy of registrationTumour morphology Completeness of follow-upTumour biology ‘Death certificate only’ registrations Host factorsHealth care-related factors AgeScreening SexDiagnostic facilities Socio-economic status Treatment facilities Race/EthnicityQuality of treatment Co-morbidityFollow-up care Mortality from other causes Behaviour

6. Absolute excess death rates (breast cancer)

7. text

8. Important observation The survival deficit (the excess mortality) in England is mainly in the older patients... and mainly in the short term after diagnosis Source (last three slides): Professor Henrik Møller, Kings College London and Thames Cancer Registry. See also: Møller H, Sandin F, Bray F, Klint A, Linklater KM, Purushotham A, Robinson D, Holmberg L. Breast cancer survival in England, Norway and Sweden: A population-based comparison. Int J Cancer 2010 (in press).

9. Registry (No of cases) Model 1 (sex + age + site) Model 2 (model 1 + stage) Model 3 (model 2 + surgery – resected cases only) Model 4 (model 3 + staging procedures† – resected cases only) Mersey (207) 1.151.101.010.99 Thames (176) 1.41*1.37*1.251.19 *P<0.05 †Staging procedures = no of LNs examined and liver imaging EUROCARE high resolution study of colorectal cancer: Relative risk of death within 3 years of diagnosis Source: Gatta et al. Gut 2000;47:533-8.

11. England’s response to these observations The National Awareness and Early Diagnosis Initiative (NAEDI) Announced in the English Cancer Reform Strategy (2007) Co-led by CR-UK and DoH Aim is to coordinate a programme of activity to support local interventions to raise public awareness of symptoms and signs of cancer, and to encourage people to present sooner Also encompasses a programme of research Much of the evidence underpinning NAEDI was published in a supplement to the British Journal of Cancer (3 December 2009).

12. Certainly there is some evidence that.. Public awareness of warning signs is low (esp. among males, younger people, lower SES, and ethnic minorities). Some patients present long after the onset of symptoms. GPs are sometimes slow to refer. Some reasons for pre-hospital delays have been identified. Sometimes, there are some perceived barriers to consulting. There are delays in hospital. Individual and community interventions may promote awareness and early presentation. BUT Delay is not synonymous with advanced stage – don’t forget tumour biology

13. Colorectal cancer diagnosed 2002: the delay-survival paradox

14. Some outstanding questions… Is awareness of cancer symptoms lower in the UK than in some other European countries? What will be the impact on patients with cancer of more patients without cancer coming forward? Do UK GPs perform any better or worse than their European counter-parts? Are delays longer in the UK, and if so, do they account for any of the survival deficit? What is the role of lifestyle factors in relation to stage and outcome? Can GPs really improve their referral performance? – the prevalence of cancer is relatively low among GP attendees, which inevitably means that the positive predictive value of symptoms is lower than in hospital.

15. Prevalence (%)PPV (%)Sensitivity (%)Specificity (%) 0.11.89095 1.015.49095 5.048.69095 50.094.79095 Effect of prevalence on positive predictive value (PPV) with constant sensitivity and specificity

16. SettingPPV (%) Single episode in the community 0.1 Reported to GP2-3 Referred to hospital 5-7 Positive predictive value of rectal bleeding for colorectal cancer

17. Conclusions Survival from major epithelial cancers seems to be lower in Scotland (and the UK) compared to all of the Nordic countries except Denmark The excess risk of death seems to occur early on and is more apparent in oldest age groups Other evidence suggests that, on average, UK patients may be presenting with more advanced disease at diagnosis But we don’t know for sure whether this is due to later presentation, later referral, delays in diagnosis or staging, or more aggressive disease We know that lifestyle factors can influence survival, but we don’t really know to what extent, if any, this contributes to European survival variations The reasons for reported survival differences seem most likely to be multifactorial