Abstract Background Cohort studies represent an important research-tool, yet participant retention is difficult. SIREN followed a cohort of healthcare workers across 135 NHS sites in the UK during a pandemic. We designed an evolving retention programme, underpinned by our Participant Involvement Panel (PIP). We aimed to evaluate cohort retention over time and identify learnings. Methods Mixed method evaluation of our evolving retention programme in the 12 and 24-month cohort. We described cohort retention by demographics and site, using odds ratios from logistic regression. Withdrawal reasons were collected by survey. We collected participant feedback via cross-sectional survey using a behavioural science approach. We conducted focus groups with research teams and conducted thematic analysis. Results 37,275 (84.7%) participants at 135 sites completed 12-months of follow-up. 12,635 (85.5%) participants at 87 sites completed 24-months. Retention increased with age (24-months: >65 vs < 25 years OR = 2.92; 95%CI: 1.78-4.88; p < 0.001) and was highest in the Asian and Black ethnic groups (24-months OR = 1.78; 95%CI: 1.42-2.25; p < 0.001, and OR = 2.12; 95%CI: 1.41-3.35; p < 0.001, respectively, Ref=white). SIREN scored highly in participant feedback. Lessons for future studies include the need to monitor changing participant motivation; using inclusive and comprehensive communication; acknowledging the contributions of participants and investing in the skillset of research teams. Conclusions Despite challenges to running a large multicentre cohort study in an evolving pandemic, retention in SIREN has remained very high. Some factors impacting retention were outside of study control, including the altruistic motivation of participants. Participant feedback was positive, suggesting SIREN methodology will provide useful learning for future retention programmes. Site feedback provided learning for future multicentre research studies, offering insights into key facilitators and common challenges. Key messages • Cohort retention is difficult. Mixed methods evaluation shows SIREN achieved high retention while following a cohort of healthcare workers across 135 NHS sites in four UK nations during a pandemic. • Lessons learned include the need to monitor changing participant motivation; using inclusive communication; acknowledging participants contributions and investing in the research team skillset.