Background: Paediatric health systems across high-income countries are facing avoidable adverse outcomes and increasing demands and costs. The aim of this study was to compare the effect of an enhanced usual care model with that of an integrated health-care model that offers local health clinics for general paediatric problems and early intervention and care for children and young people with tracer conditions. Methods: In this pragmatic two-arm cluster randomised controlled trial, we compared the Children and Young People's Health Partnership (CYPHP) model of care versus enhanced usual care (EUC) among children registered at general practices in south London, UK. The CYPHP trial intervention was delivered between April 1, 2018, and June 30, 2021, and children younger than 16 years during the intervention period and registered at study practices on June 30, 2021, were included in the analysis. A restricted randomisation (1:1) following a computer-generated sequence was done by a masked independent statistician at the level of general practice cluster, stratified by borough (Lambeth or Southwark). Cluster allocation and data collection were masked, with unmasking of trial statisticians before analysis. The CYPHP model comprised all elements of EUC (electronic decision support, a primary care hotline, health checks, self-management support and health promotion, and resilience building and mental health first aid) plus local child health clinics delivered by paediatricians and general practitioners, and a nurse-led early intervention service for children with tracer conditions (asthma, eczema, and constipation). Primary outcomes were non-elective admissions (NELA; admissions coded as an emergency) among the whole trial population up to June 30, 2021, and paediatric quality of life (Pediatric Quality of Life Inventory [PedsQL]) among participants with tracer conditions at 6 months after recruitment. Secondary outcomes were primary and secondary care use, child mental health, parental wellbeing, standardised symptom scores for asthma, eczema, and constipation, health-care quality, and child absences from school and parent absences from work. The trial was registered on ClinicalTrials.gov, NCT03461848, and is complete. Findings: The trial was conducted between April 1, 2018, and Dec 31, 2021. In total, 23 general practice clusters, consisting of 70 practices with 97 970 registered children, were randomised to CYPHP (n=11) or EUC (n=12). We found no effect, at the population level, of CYPHP versus EUC on non-elective admissions during the intervention period (adjusted mean incidence rate ratio [IRR] 1·00 [95% CI 0·91 to 1·10], p=0·99). Among children with tracer conditions, we found no difference in paediatric quality of life (PedsQL score) at 6 months (adjusted mean difference –0·033 [95% CI –0·122 to 0·055], p=0·46). As a secondary outcome, among children with tracer conditions and requiring care, NELA rates at 12 months did not differ between the CYPHP and EUC groups (66·1 per 1000 person-years vs 75·3 per 1000 person-years; adjusted mean IRR 0·87 [0·61–1·22], p=0·42). In children requiring care, a statistically significant improvement was observed in eczema symptoms at 6 months from baseline in the CYPHP group versus the EUC group (adjusted mean difference –1·370 [–2·630 to –0·122], p=0·032). Quality of asthma care significantly improved among children in the CYPHP group compared with children in the EUC group. No significant improvement was seen for all other secondary outcomes. Interpretation: Although the CYPHP trial found a null effect for the primary outcomes, we found clinically important improvements in some secondary outcomes including care quality. Previous research has shown that large-scale system change requires time to observe a potential positive effect. Funding: Guy's and St Thomas Charity, the Lambeth and Southwark Clinical Commissioning Groups, and Evelina London Children's Hospital.