Time trends of process and impact indicators in Italian mammography screening programmes--1996-2004

Since 1990, the Italian Group for Mammography Screening (GISMa) has been promoting the development of new organised programmes and performing a yearly systematic survey of data activity. The screening extension has increased over time, reaching an overall 76.4% of coverage in 2005. The geographical extension is still heterogeneous, with a higher distribution in Northern and Central Italy compared with Southern and Insular Italy, where the screening activity was implemented only recently. Notwithstanding the continuity in implementation, the actual coverage reached only 50.3% of the target population, due to a reduced flow of invitations over time as a consequence of a chronic lack of invested resources and of well-plannedpolicies. The overall Italian rate for crude attendance was above the acceptable 50% standard even though a North-South trend is still confirmed; in Southern/Insular Italy participation was still inadequate (<40%) and did not reach the standard considered acceptable. Participation was higher in centralised programmes compared with those without regional coordination (+5-8%). The time trends for the other key performance indicators showed good average performance: the benign/malignant surgical biopsy ratio (B/M ratio) progressively decreased, reaching an 0.25 ratio (both for first and subsequent screening) in 2004; overall detection rate, detection rate for in situ and small cancers (< or =10 mm) showed a good trend, reaching 6.7% per hundred, 0.7% per hundred, and 1.6% per hundred, respectively, for the first screening, and 5.1% per hundred, 0.9% per hundred, and 1.7% per hundred for the subsequent screening in 2004. The only exception was the referral rate (RR) at first screening, which exceeded standards (> 7% in 2002-2004). Data comparing activity volume and programme duration were also analysed. In programmes with greater activity (average test number: 22,506) the referral rate for the first screening was higher, but still within acceptable standards. 6.3% per hundred; RR: 1.01 (0.98-1.04). This performance is compensated by better specificity and sensitivity: Positive Predictive Value (PPV). 12.8; RR:1.16 (1.05-1.27); overall detection rate: 8.1% per hundred, RR 1.19 (1.07-1.31). An improvement in quality with the increase of programme experience is evident. programmes with more than 6 years of activity, compared with newer programmes, show a recall rate in first screening of 6.5%; RR. 0.87 (0.84-0.89), a PPV of 1.7, RR:1.61 (1.48-1.75) andan overall detection rate of 7.6% per hundred, RR:1.41 (1.29-1.55). These results are consistent with those observed in other European programmes and encourage to explore new analysis strategies. The website of the National Centre for Screening Monitoring (ONS) is http://www.osservatorionazionalescreening.it.