Opportunistic offering of self-sampling to non-attenders within the English cervical screening programme: a pragmatic, multicentre, implementation feasibility trial with randomly allocated cluster intervention start dates (YouScreen).

BACKGROUND

Self-sampling has game-changing potential to tackle the declining participation and inequities seen in many organised cervical screening programmes. Wide variation in uptake between settings and mode of kit offer highlight the importance of local piloting. Furthermore, harnessing the benefits of self-sampling in real-world settings has been surprisingly challenging. The YouScreen study estimated the impact of offering self-sampling to non-attenders within the English Programme and evaluated large-scale opportunistic offering of self-sampling in primary care.

METHODS

A pragmatic modified stepped-wedge implementation feasibility trial with randomly-allocated cluster intervention start dates at primary care practices in England (133 participating, 62 non-participating). Eligible women were aged 25-64 years and ≥6 months overdue for screening ("non-attenders"). Between January 13, 2021 and 30 November, 2021 self-sampling kits were distributed to non-attenders via an opportunistic offer in primary care when they consulted for any reason and direct mailout to those unscreened 15-months after routine invitation. Primary outcomes were the proportion of non-attenders screened each month; change in coverage; and uptake (90 days). YouScreen is registered with ISRCTN:12759467.

FINDINGS

8338 women provided self-samples following recruitment between January 13, 2021 and 30 November, 2021. Self-samples were returned from 65.5% (6061/9248) who accepted an opportunistically offered kit and 12.9% (2777/17,604) directly-mailed kits. Responders were representative of the ethnically diverse and deprived underlying non-attendee population (64% ethnic minority groups, 60% from the two most deprived national quintiles). The self-sampling intervention resulted in a 22% (95% CI 18-26) increase in non-attenders screened per month (per-protocol analysis) and 12% (95% CI 9-15) (intention-to-treat analysis). Change in coverage at participating (mean intervention duration 7.5 months) vs non-participating practices was 1.6% (95% CI 0.4-2.8). Adverse effects were not formally collected.

INTERPRETATION

Opportunistically offering self-sampling to under-screened women in primary care could increase coverage in England and potentially reach underserved populations.

FUNDING

North Central London and North East London Cancer Alliance.