Vaccination against human papillomavirus (HPV) is highly effective and is expected to reduce substantially the burden of cervical cancer (CC), but cervical cancer screening (CCS) remains necessary for females who were not covered by the vaccination plans, those who remain non-adherent to vaccination, and for early detection of the cervical lesions not prevented by the vaccine. However, the sustainability of CCS programs will be challenged by the expectedly dramatic reduction in frequency of cervical precancerous lesions. A paradigm shift for CCS programs, towards a substantial improvement of efficiency, and ability to reach those more in need of screening, is essential to maintain their cost-effectiveness. We will implement a state-of-the-art approach to population-based CCS, with potential to broaden screening coverage, lower costs and increase efficiency, compared to the current standard of care, by including self-sampling and state-of-the-art molecular methods for triage.
The investigation is based on a pragmatic parallel cluster randomized controlled trial (RCT), including nearly 6,000 participants (2,000 in each arm), implemented at Local Health Unit Gaia e Espinho (ULSGE) in collaboration with the Research Center of Portuguese Institute of Oncology of Porto (CI-IPOP). Family doctors will be randomly allocated to one of the three study arms, and all females from their lists of patients who are eligible for screening, as defined by the current guidelines for the organized screening program, will be invited.
CCS screening will be based on high-risk HPV (HrHPV) testing, according to the Portuguese CCS program. Referral to colposcopy will follow the standard of care (based on HrHPV and liquid based cytology \[LBC\]), regardless of the strategy used for CCS; females testing non-HPV16/18 HrHPV-positive in self-sampling will undergo a new sample collection by a health professional, so that a LBC evaluation can be performed to support the decision of referral to colposcopy, as defined by the standard of care. Deoxyribonucleic acid (DNA) methylation will be conducted in parallel.
Intention to treat analysis will be the primary strategy of analysis for comparison between the groups, based on crude comparisons, or controlling for confounding through multivariable binary logistic regression, when an imbalance between the distribution of confounders between study arms is observed. Methylation accuracy estimates will be computed with 95% confidence intervals. The cutoffs for each biomarker will be those defined in our previous studies.
The sample size was defined to address the primary objectives that involve the comparison of study arms, with a 1:1:1 allocation ratio, considering a statistical power of 90% and a design effect of 1.1 (assuming an average cluster size of 50 and an intracluster correlation coefficient of 0.002). For a non-inferiority hypothesis, considering an alpha (one-sided) of 5% and assuming an adherence of 50% and a non-inferiority margin of 5%, a sample of 5,700 females is needed.
Family doctors from a total of 10 to 20 primary health care units are expected to be enrolled to achieve the sample size needed, depending on the adherence of the medical doctors and the number of eligible females in their lists of patients.
The study will be initially implemented in 2 to 3 primary health care units, to achieve an expected sample of at least 600 participants, for piloting and adjustment of the overall sample size based on the empirical data on the proportions of adherent females in each arm.
The assessment of the adherence according to the different strategies to incorporate self-sampling in the CCS program will be complemented by a qualitative study to assess barriers and facilitators to the implementation of CCS and adherence to CCS in different formats. This will involve the family doctors who have collaborated in the study and females selected among those invited for screening in the different study arms, based on focus groups and individual semi-structured interviews.
Also, a quantitative assessment of the invited females experiences with CCS within the study and determinants of adherence will be conducted on a subsample including approximately 10% of the invited females, who will be asked to answer a structured questionnaire.
