Improving cervical cancer preventive strategies in Australia

Abstract

Background: Screening for cervical cancer reduces its mortality and morbidity. Despite high participation in the cervical cancer screening program in Australia, most cervical cancers occur in non-attendees. Randomised controlled trials have shown that offering HPV self-sampling to women who had not been screened for a long time results in higher participation than reminding them to attend for a Pap test. However, substantial heterogeneity between studies indicated that locally conducted trials are necessary. The overall aim of this thesis was to determine whether HPV self-sampling is an acceptable, feasible and effective strategy to improve participation in the cervical cancer screening program in Australia. Methods: A randomised controlled trial (iPap) was conducted in Victoria, Australia. Never- and under-screened women, 30-69 years, not pregnant and without a history of hysterectomy were eligible. Women were randomly allocated in a 7:1 ratio to either receive an HPV self-sampling kit or a reminder/invitation to attend for a Pap test (current policy), stratified by screening status (never- versus underscreened). The primary outcome measure was participation, as indicated by returning a self-sampling swab or undergoing a Pap test (6 months); the secondary outcome, restricted to the self-sampling arm, was the proportion of women with a positive HPV test who had further follow-up (6 months from receipt of the results letter). The PCR based Roche cobas® 4800 test was used to measure presence of high-risk HPV DNA. All materials were focus group tested. The use of the dry flocked swab (dry sample) with the cobas® 4800 test was validated prior to the trial. Questionnaires were mailed to women in the trial to document their experience with and preference for self-sampling in future and reasons for non-participation. Systematic reviews, meta-analyses and meta-regression were conducted before and after the inclusion of the iPap trial to explore sources of heterogeneity. Results: The systematic review and meta-analysis of trials showed that HPV self-sampling improved participation in cervical screening over reminders to have a Pap test when the kits were directly mailed to women; however, there was large between-study heterogeneity (I2 = 97.5%) that was not explained by any study level factors. Dry samples showed almost perfect agreement with wet samples for the detection of high-risk HPV using the cobas® 4800 test [kappa: 0.85 (95% CI: 0.78 – 0.92)]. Participation in the iPap trial was higher for the self-sampling arm: 20.3% versus 6% for never-screened women (absolute difference 14.3%, 95% CI: 12.6%-16.1%; p<0.001) and 11.5% versus 6.4% for under-screened women (absolute difference 5.1%, 95% CI: 3.4% - 6.8%; p<0.001). Overall, 75.7% (95% CI: 67.8% - 82.6%) had the recommended follow-up test. Participants stated embarrassment, lack of time, pain and discomfort as common reasons for not having a Pap test. They found home-based self-sampling less embarrassing and less uncomfortable, and more convenient than their last Pap test experience; although, many were unsure about the test accuracy. Most participants preferred to take the self-sample at home in the future because it was simple and did not require a doctor’s appointment. The main reason for not doing the self-sample test was that women had had a hysterectomy. Inclusion of the iPap trial results in the meta-analysis and meta-regression did not alter the heterogeneity or explain it. Conclusion: HPV self-sampling is an acceptable, feasible and effective strategy to increase participation in cervical cancer screening program in Australia.