Résumé : OBJECTIVE: To describe the use of paediatric “observation status” in the accident and emergency department (A&E). METHOD: A prospective survey was performed in 12 Belgian hospitals during 2 weeks straddling October and November 2010. All patients (<16 years) attending A&E were included. “Observation status” was defined when after the first medical evaluation, instead of hospitalization or home discharge, the situation required further observation of the patient. The clinicians in charge were asked at the start of the “observation” period to prognosticate whether the child would be discharged or admitted. RESULTS: Among 3220 children included in the study, the observation rate was 38.6%. The characteristics of these children were as follows. Median age: 5.0 years old (IQR: 1.7-11.3), boys: 53.5%. The median length of stay in A&E was 110 minutes (IQR: 65-175) and 14.3% were admitted as in-patient. The most common observations concerned orthopaedic, medical digestive and respiratory affections. The three main reasons for observation were additional procedures (69.0%), diagnostic determination (10.7%), and treatment testing (8.3%). Most of the observations (86.9%) were performed in a waiting room (not in a bed), 9.7% in an observation unit dedicated to children and 3.4% in an observation unit to adults. Only 4.1% of the observations satisfied the French paediatric short-stay definition (patient in a bed monitored hourly). Compared to the clinicians’ prognosticated rate, 1.3% less children were admitted as in-patients. CONCLUSION: Despite the absence of financial and regulatory frame in Belgium, observation is a frequent modality of care. Observation helped to refine diagnosis and treatments and to reduce in-patient admission rate. However, with regards to the French guidelines less than 5% could be qualified as paediatric A&E short-stay. Flexible observation units should be recommended. KEYWORDS: Emergency, paediatrics, short-stay, observation status, use, Belgium