In the Netherlands several screening scheme, follow-up, administration and financing protocols/guidelines were changed in 1996 after evidence suggested their performance was suboptimal. The target age group was enlarged from 35-53 to 30-60 years, and the screening interval prolonged from 3 to 5 years. All eligible women now receive a screening invitation. Non-attendees are systematically sent reminders. Reimbursement of opportunistic smears was stopped. The classification criteria of ASCUS smears was made more restrictive, while the follow-up recommendations were simplified. Follow-up on negative smears lacking endocervical cells was stopped. We assessed changes since 1996 in various indicators: coverage rates, follow-up compliance, the number of smears, and the short-term interval cancer rates.

Methods

Information on all cervix uteri tests registered until 31st March 2004 was retrieved from the Dutch nationwide registry of histo- and cytopathology (PALGA). The performance of the new protocols and guidelines was based on programme smears from 2003, and compared to those from 1994. The interval cancer rates were based on invasive cases and person-years at risk accrued in the period 1998-2002, and compared to those from the period 1994-1997.

Results

Five-year coverage rates in the added age groups improved, becoming closer to those observed among the 35-53 year-old women in both periods (80%). The frequency of advice for follow-up decreased by about 80% (from 19% to 3% per screening round). Fourteen percent more women with an abnormal smear were followed-up, and the time spent in follow-up decreased. Opportunistic screening decreased sharply. These changes led to a 20% decrease in the annual number of smears made. Preliminary analyses showed that the interval cancer rates within the first screening round after the 1996 changes did not increase above the pre-1996 levels.

Conclusions

The 1996 policy changes resulted in an equally good screening effectiveness, but this was achieved with a less intensive screening programme. Our results suggest that various policy instruments can motivate individuals and their physicians to abstain from screening outside of the recommended schedule. Further, test specificity can be improved by evidence-based optimization of the cut-off point for test positivity. It appears that in the Netherlands this was achieved without an associated loss in programme sensitivity.