Ascertainment and evaluation of interval cancers in population-based mammography screening programmes: a collaborative study in four European centres

Abstract

The purpose of the present study was to estimate the interval cancer (IC) rates in four population-based mammography screening programmes in four countries with different health-care environments, different access to cancer registry data, and different age groups of women invited. The screening programmes in Coimbra (Portugal), Dublin (Ireland), Stockholm (Sweden), and Turin (Italy) participated in the study. All cancer cases were searched for in cancer registries. IC rates and other outcome measures from the screening programmes were estimated and compared between the centres. Poisson regression model was used to estimate the proportional incidence based on IC rate in relation to expected total breast cancer incidence rate in the absence of screening. There was a more than tenfold difference in the number of invited women at the first round between the involved centres. The IC rates varied between 4.3 and 23.8 per 10,000 women screened. The levels of IC rates in relation to the estimated background incidence varied from 0.35 up to 0.46 depending on age groups involved in the programme, but did not differ significantly between three of the four involved centres. IC rates were quite similar between three of the four centres despite the differences in target population, invited ages, length of building-up of the programmes and different health-care organizations. Different access to complete cancer registry data is likely to explain the lower IC rates in the fourth centre.