One thousand, eight hundred and fifty‐seven patients underwent magnetic resonance imaging following the establishment of a structured sedation programme. Forty‐eight of these patients came from the intensive care unit with a secure airway and were therefore excluded from any further analysis. Oral sedation was to be given to children aged 5 years and below. For children ≥ 6 years old, oral sedation could be given only if their level of co‐operation was judged to be inadequate by the referring physician. Oral sedation consisted of chloral hydrate 90 mgkg^−1^ (maximum 2.0 g) orally with or without rectal paraldehyde 0.3 ml.kg^−1^. All magnetic resonance imaging requests for children who failed oral sedation as well as those referred for general anaesthesia from the outset were reviewed by a consultant anaesthetist who then allocated patients to undergo the procedure with either general anaesthesia or intravenous sedation. Scans requiring intravenous sedation or general anaesthesia were performed in the presence of a consultant anaesthetist. Intravenous sedation consisted of either a propofol 0.5 mgkg^−1^ bolus followed by an infusion (maximum 3 mgkg^−1^.h^−1^) or midazolam 0.2–0.5 mgkg^−1^ boluses. General anaesthesia was given using spontaneous ventilation with a mixture of 66% nitrous oxide in oxygen and isoflurane following either inhalation (sevoflurane) or intravenous (propofol) induction. One thousand and thirty‐nine (57.4%) of the scans were done without sedation whereas 93 scans were performed during the consultant anaesthetist supervised sessions. Oral sedation failed in 50 out of 727 patients (6.9%). Eighty‐seven per cent of children aged 5 years and below needed sedation compared with 4.5% of those aged over 10 years. Two patients who had only received chloral hydrate developed significant respiratory depression. This structured sedation programme has provided a safe, effective and efficient use of limited resources.