Aims. COVID-19 necessitated abrupt changes in trauma service delivery. We compare the demographics and outcomes of patients treated during lockdown to a matched period from 2019. Findings have important implications for service development. Methods. A split-site service was introduced, with a COVID-19 free site treating the majority of trauma patients. Polytrauma, spinal, and paediatric trauma patients, plus COVID-19 confirmed or suspicious cases, were managed at another site. Prospective data on all trauma patients undergoing surgery at either site between 16 March 2020 and 31 May 2020 was collated and compared with retrospective review of the same period in 2019. Patient demographics, injury, surgical details, length of stay (LOS), COVID-19 status, and outcome were compared. Results. There were 1,004 urgent orthopaedic trauma patients (604 in 2019; 400 in 2020). Significant reductions in time to theatre and LOS stay were observed. COVID-19 positive status was confirmed in 4.5% (n = 18). The COVID-19 mortality rate was 1.8% (n = 7). Day-case surgery comprised 47.8% (n = 191), none testing positive for COVID-19 or developing clinically significant COVID-19 symptoms requiring readmission, at a minimum of 17 days follow-up. Conclusion. The novel split-site service, segregating suspected or confirmed COVID-19 cases, minimized onward transmission and demonstrated improved outcomes regarding time to surgery and LOS, despite altered working patterns and additional constraints. Day-surgery pathways appear safe regarding COVID-19 transmission. Lessons learned require dissemination and should be sustained in preparation for a potential second wave or, the return of a “normal” non-COVID workload. Cite this article: Bone Joint Open 2020;1-9:568–575