Abstract
Background: Human papillomavirus (HPV) testing is recommended for cervical screening, and HPV self-testing (self-collected sample) increases uptake by under-screened women and people with a cervix. Whether a universal offer of HPV self-testing to all eligible people at primary care practices increases screening coverage is unknown. We aimed to assess whether a universal offer of HPV self-testing for cervical screening achieved non-inferior screening coverage compared with usual care (offer of cervical cytology by vaginal speculum examination).
Methods: This cluster-randomised, non-inferiority trial was done in 14 primary care practices in Te Tai Tokerau Northland, Aotearoa New Zealand. Random sampling selected practices stratified by practice size and rurality. Practices were randomly assigned (1:1) to intervention (self-testing) or comparison (clinician-taken cytology sampling; seven per group). The study statistician who did the random sampling was masked to the identity and location of the practices. Participants were not randomly assigned. Women and people with a cervix aged 24·5–69·9 years who were due or overdue for cervical screening and attending participating practices were eligible for inclusion. In the intervention group, the universal offer of HPV self-testing included community engagement, identifying a key clinician champion, practice-level training, ongoing education for practices, and an offer of self-testing. If declined, clinician-taken sampling was offered. Participants in primary care practices in the comparison group were offered usual care (clinician-led cervical cytology by vaginal speculum examination). The primary outcome was the number of eligible people having a cervical screen divided by the number of age-eligible people registered in study practices in the time period, defined as screening coverage. If non-inferiority was evident (lower bound of the 95% CI for the risk difference was above the non-inferiority limit of –10), superiority of the intervention over the comparison was assessed with a zero difference. The primary outcome was assessed using an unadjusted generalised linear mixed model according to the intention-to-treat principle, including all age-eligible women and people with a cervix regardless of screening status or hysterectomy. An internal data safety committee with independent chair monitored safety. This study was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12621001675819), and is complete.
Findings: Between Feb 8, 2022, and Sept 8, 2023, 22 661 people were assessed for eligibility, 22 511 of whom were included in the main analysis: 10 727 (47·7%) people in the intervention group and 11 784 (52·3%) people in the comparison group. The mean age was 47·7 years (SD 13·5) and 9765 (43·4%) participants were Māori. The mean proportion screened was significantly higher in the intervention group (35·6% [SE 2·4]) than in the comparison group (24·9% [2·0]; absolute risk difference 10·8% [95% CI 4·7–16·9]; p=0·0006) over a period of 14 or 19 months, depending on when the practices were recruited, showing both non-inferiority and superiority of the intervention over the comparison. No adverse events were identified.
Interpretation: Based on the findings of superior screening coverage, countries with organised cervical screening programmes should consider implementing a universal offer of HPV self-testing to increase screening coverage and reduce preventable morbidity and mortality from cervical cancer.