OBJECTIVES To compare the efficacy, safety and cost utility of CSII with MDI during the first year following diagnosis of TID in paediatric patients. DESIGN Pragmatic, open, multicentre, parallel group, randomised, controlled trial. SETTING15 paediatric National Health Service (NHS) diabetes services in England and Wales. PARTICIPANTSPatients between 7 months and 15 years of age, newly diagnosed with TID were eligible to participate. Patients with a sibling with TID, those who took medications or had additional diagnoses that may influence glycaemic control were ineligible. INTERVENTIONSParticipants were randomised, stratified by age and centre, to start treatment with CSII or MDI within 14 days of diagnosis. Starting doses of aspart (CSII and MDI) and glargine or detemir (MDI) were calculated according to weight and age, and titrated according to blood glucose measurements according to local clinical practice.MAIN OUTCOME MEASURESPrimary outcome :glycaemic control (HbA1c) at 12 months. Secondary outcomes: percentage of patients in each treatment arm with HbA1c within the national target range; incidence of severe hypoglycaemia and diabetic ketoacidosis (DKA); change in height and body mass index (BMI) standard deviation score (SDS); insulin requirements (units/kg/day); partial remission rate (insulin dose adjusted HbA1c (IDAAC) < 9), PedsQL score; cost-utility based on the incremental cost per Quality-Adjusted Life-Year (QALY) gained and an NHS costing perspective. RESULTS294 participants were randomised and 293 included in intention to treat analyses (CSI=144 and MDI=149). At 12 months mean HbA1c was comparable with clinically unimportant differences : 60.9mmol/mol in CSII participants and: 58.5mmol/mol in MDI participants, mean difference (CSII-MDI) 2.4mmol/mol (95% -0.4 to 5.3), p=0.09. Achievement of HbA1c <58mmol/mol was low: 66 (46%) of 143 CSII participants and 78 (55%) of 142 MDI participants, RR 0.84 95%CI (0.67 to 1.06). Incidence of severe hypoglycaemia and DKA were low in both groups. Sixty-eight adverse events (AEs) (14 serious) were reported during CSII treatment and 25 AEs (8 serious) during MDI treatment. Parents, but not children, reported superior PedsQL scores for those treated with CSII. CSII was more expensive than MDI by £1,863 (95% CI £1,620 to £2,137) per patient with no additional QALY gains, -0.006 (95% CI -0.031 to 0.018). CONCLUSIONDuring the first year of TID no clinical benefit of CSII over MDI was identified in the UK setting and treatment with either regimen was suboptimal in achieving HbA1c thresholds. CSII was not cost-effective.