After an organised cervical screening programme was introduced in Turin in 1992, the age-adjusted cervical cancer incidence ratio in 1992 -98 was 0.81 (95% confidence interval (CI) 0.59 -1.09) for invited vs not invited women and 0.25 (95% CI 0.13 -0.50) for attenders vs non attenders. An organised screening programme can further reduce cervical cancer incidence in an area where substantial spontaneous activity was previously present. There is clear evidence of efficacy for cervical cancer screening (IARC, 2005). High effectiveness and cost-effectiveness were obtained in Nordic countries, where cervical screening was organised from the outset (Laara et al, 1987). In the United Kingdom, a relevant reduction in cervical cancer mortality followed reorganisation (Sasieni and Adams, 1999;Peto et al, 2004), There is however little experience of the impact of moving from opportunistic activity to an organised programme.For many years, cervical cancer screening in Italy was almost only opportunistic (Segnan et al, 2000). An organised programme started in Turin in 1992, the main changes being (a) an active callrecall system, (b) protocols for diagnosis and treatment, (c) a failsafe system for both these phases and (d) monitoring and intensive quality assurance for every step of the screening process.In order to evaluate the impact of these changes on invasive cervical cancer incidence, we linked the screening registry to the local population cancer registry.
MATERIALS AND METHODSFrom June 1992, female Turin inhabitants aged 25 -64 years were invited, irrespective of their previous spontaneous screening history, for a Pap-test, with 3-year intervals for screen-negatives. Activation was progressive: in 1996, about 78% of the target population had been invited and in 1998, about 95%. The order of invitation was substantially random, with some stratification by area (women were enrolled by general practitioners, who were selected with balance by location of practice).About one-third of invited women attended the organised programme. Attenders mostly had not had spontaneous recent cytology. As a combined result of organised and opportunistic screening, we estimated a 74% 3-year overall coverage (Ronco et al, 1997) vs 40% before the start of the programme (Segnan et al, 1990).We computed cervical cancer incidence for women aged 24 -69 years in the following groups: not invited (not yet having received the first invitation in the organised programme), invited (having already received the first invitation), attenders (invited women with at least one cytology in the organised programme), nonattenders (invited women with no cytology in the organised programme).Person-years (py) at risk of cervical cancer for the last three categories were calculated using the computerised screening registry. It was not possible to consider the screening history outside the organised programme. Each woman contributed to the 'invited' cohort from first invitation to the end of follow-up, that is, the earliest among (a) diagnosis of cervical cancer, ...