Impact of the UK colorectal cancer screening pilot studies on incidence, stage distribution and mortality trends.

OBJECTIVE: To assess the impact of the UK colorectal cancer guaiac faecal occult blood test screening pilot studies on incidence trends, stage distribution and mortality trends.

DESIGN: Ecological study.

SETTING: Scotland and the West Midlands.

DATA: We extracted anonymised colorectal cancer (ICD-10 C18-C20) registration (1982-2006) and death records (1982-2007), along with corresponding mid-year population estimates.

INTERVENTION: Residents of the screening pilot areas, in the age group 50-69 years, were offered biennial guaiac faecal occult blood test screening from 2000 onwards. Screening was not offered routinely in non-pilot areas until the start of the roll-out of the national screening programmes in England and in Scotland in 2006 and 2007, respectively.

MAIN OUTCOME MEASURES: We analysed trends in age-specific incidence and mortality rates, and Dukes' stage distribution. Within each country/region, we compared the screening pilot areas to non-screening pilot ('control') areas using Chi square tests and Poisson regression modelling.

RESULTS: Following the start of the screening pilots, as expected in the prevalent round of a new screening programme, in the pilot areas there was a short-lived increase in incidence of colorectal cancer among 50-69 year olds except for females in the West Midlands. A trend towards earlier stage and less advanced disease was also observed, with males showing significant increases in Dukes' A and corresponding decreases in Dukes' C in the screening pilot areas (all P < 0.03). With the exception of females in the West Midlands, mortality rates for colorectal cancer decreased significantly and at a faster rate in the populations invited for screening.

CONCLUSION: The existence of a natural control population not yet invited for screening provided a unique opportunity to assess whether the benefits of colorectal cancer screening, beyond the setting of a randomised controlled trial, could be detected using routinely collected statistics. Our analysis suggests that screening will fulfil its aim of reducing mortality from colorectal cancer.