Increasing Cervical Cancer Screening Coverage: A Randomised, Community-based Clinical Trial

Abstract

Background: Opportunistic cervical cancer screening can lead to suboptimal screening coverage. Coverage could be increased after a personalised invitation to the target population. We present a community randomized intervention study with three strategies aiming to increase screening coverage. Methods: The CRICERVA study is a community-based clinical trial to improve coverage of population-based screening in the Cerdanyola SAP area in Barcelona. A total of 32,858 women residing in the study area, aged 30 to 70 years were evaluated. A total of 15,965 women were identified as having no registration of a cervical cytology in the last 3.5 years within the Public Health data base system. Eligible women were assigned to one of four community randomized intervention groups (IGs): (1) (IG1 N = 4197) personalised invitation letter, (2) (IG2 N = 3601) personalised invitation letter + informative leaflet, (3) (IG3 N = 6088) personalised invitation letter + informative leaflet + personalised phone call and (4) (Control N = 2079) based on spontaneous demand of cervical cancer screening as officially recommended. To evaluate screening coverage, we used heterogeneity tests to compare impact of the interventions and mixed logistic regression models to assess the age effect. We refer a "rescue" visit as the screening visit resulting from the study invitation. Results: Among the 13,886 women in the IGs, 2,862 were evaluated as having an adequate screening history after the initial contact; 4,263 were lost to follow-up and 5,341 were identified as having insufficient screening and thus being eligible for a rescue visit. All intervention strategies significantly increased participation to screening compared to the control group. Coverage after the intervention reached 84.1% while the control group reached 64.8%. The final impact of our study was an increase of 20% in the three IGs and of 9% in the control group (p<0.001). Within the intervention arms, age was an important determinant of rescue visits showing a statistical interaction with the coverage attained in the IGs. Within the intervention groups, final screening coverage was significantly higher in IG3 (84.4%) (p< 0.001). However, the differences were more substantial in the age groups 50-59 and those 60+. The highest impact of the IG3 intervention was observed among women 60+ y.o with 32.0% of them being rescued for screening. The lowest impact of the interventions was in younger women. Conclusions: The study confirms that using individual contact methods and assigning a fixed screening date notably increases participation in screening. The response to the invitation is strongly dependent on age.