For cost-saving purposes, children and young people with juvenile idiopathic arthritis (JIA) are being switched (for non-medical reasons) from biological originators to biosimilars. Here, we aimed to investigate those who switched from an anti-tumour necrosis factor (TNF) originator to a biosimilar, regarding drug survival and disease activity, compared with a matched cohort who continued the originator. This analysis included all patients in the UK JIA Biologics Register switching directly from an anti-TNF originator to a biosimilar of the same product. All patients were matched (age, sex, disease duration, calendar year of when patients started originator therapy, line of therapy, and International League of Associations for Rheumatology [ILAR] category) to patients continuing the originator. For those matched successfully, a Cox proportional hazard model assessed whether drug persistence differed between those who switched compared with those who continued the originator. Overall change in the 71-joint juvenile arthritis disease activity score and the proportion of patients with a clinically important worsening score (by ≥1·7 units) after 6 months was compared between cohorts. This analysis was designed to address a priority research area set by our patient partners. There were 224 children and young people with non-systemic JIA (139 [62%] were female, and 85 [38%] were male) identified as switching from a biological originator to a biosimilar of the same product from Jan 1, 2017, to July 7, 2023. 143 (64%) patients were originally on adalimumab, 56 (25%) on etanercept, and 25 (11%) on infliximab. Of these, 164 patients were matched successfully to those continuing the originator. There was no evidence that patients switching were more likely to stop treatment compared with those continuing the originator, with a hazard ratio of 1·46 (95% CI 0·93-2·30). Of the 51 patients in the biosimilar group who stopped treatment, 18 (35%) switched back to the originator (14 in the first year), 28 (55%) started a different biological drug, and five (10%) discontinued all treatment by the last follow-up. Of the 87 matched patients with available disease activity, there was no evidence that JADAS-71 worsened more after 6 months, with an odds ratio of 0·71 (95% CI 0·34-1·51; p=0·38). In this matched comparative effectiveness analysis, children and young people with JIA switched from originators to biosimilars. Disease activity was similar between patients switching compared with those continuing the originator. Three quarters of patients were still receiving their biosimilar after 1 year, with switching back to originator uncommon, at only 9% after 1 year, suggesting good tolerability of non-medical switching in this patient population. This information is reassuring to clinicians and patients regarding the effect of non-medical biological switching. British Society for Rheumatology, Versus Arthritis, and National Institutes for Health Research Manchester Biomedical Research Centre.