BACKGROUND: Procalcitonin is a biomarker specific for bacterial infection, with a more rapid response than other commonly used biomarkers, such as C-reactive protein, but it is not routinely used in the National Health Service. OBJECTIVE: To determine if using a procalcitonin-guided algorithm may safely reduce duration of antibiotic therapy compared to standard of care in hospitalised children with suspected or confirmed infection. DESIGN: A pragmatic, multicentre, open-label, parallel two-arm, individually randomised controlled trial with internal pilot phase, qualitative study and health economic evaluations. SETTING: Paediatric wards or paediatric intensive care units within children's hospitals (n = 6) and district general hospitals (n = 9) in the United Kingdom. PARTICIPANTS: Children aged between 72 hours and 18 years admitted to hospital and being treated with intravenous antibiotics for suspected or confirmed bacterial infection. INTERVENTIONS: Procalcitonin-guided algorithm versus usual standard care alone. MAIN OUTCOME MEASURES: Coprimary outcomes were duration of intravenous antibiotic use and a composite safety measure. RESULTS: Between 11 June 2018 and 12 October 2022, 1949 children were recruited: 977 to the procalcitonin group [427 female (43.7%), 550 male (56.3%)], and 972 to the usual care group [478 female (49.2%), 494 male (50.8%)]. Duration of intravenous antibiotics was not significantly different between the procalcitonin group (median 96.0 hours) and the usual care group (median 99.7 hours) [hazard ratio = 0.96 (0.87, 1.05)], and the procalcitonin-guided algorithm was non-inferior to usual care [risk difference = -0.81% (95% confidence interval upper bound 1.11%)]. At clinical review, a procalcitonin result was available for 81.8% of the time, which was considered as part of clinical decision-making 66.6% of the time, and the algorithm was adhered to 57.2% of the time. Incremental cost-effectiveness ratio per duration of intravenous antibiotics hour avoided from bootstrapped samples was £467.62 per intravenous antibiotic hour avoided. Cost analysis of complete cases was also higher in the procalcitonin arm for all age groups, and for children aged 5 years and over. The intervention is not cost-effective as it is more expensive with no significant improvement in intravenous antibiotic duration. LIMITATIONS: Robust antimicrobial stewardship programmes were already implemented in the lead recruiting sites, and adherence to the algorithm was poor. Clinicians may be reluctant to adhere to biomarker-guided algorithms, due to unfamiliarity with interpreting the test result. CONCLUSIONS: In children hospitalised with confirmed or suspected bacterial infection, the addition of a procalcitonin-guided algorithm to usual care is non-inferior in terms of safety, but does not reduce duration of intravenous antibiotics, and is not cost-effective. In the presence of robust antimicrobial stewardship programmes to reduce antibiotic use, a procalcitonin-guided algorithm may offer little added value. FUTURE WORK: Future trials must include an implementation framework to improve trial intervention fidelity, and repeated cycles of education and training to facilitate implementation of biomarker-guided algorithms into routine clinical care. TRIAL REGISTRATION: This trial is registered as ISRCTN11369832. FUNDING: This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 15/188/42) and is published in full in Health Technology Assessment; Vol. 29, No. 16. See the NIHR Funding and Awards website for further award information.