TY - JOUR T1 - JournalScan JF - Emergency Medicine Journal JO - Emerg Med J SP - 554 LP - 555 DO - 10.1136/emj.20.6.554 VL - 20 IS - 6 AU - J P Wyatt AU - J Lee Y1 - 2003/11/01 UR - http://emj.bmj.com/content/20/6/554.abstract N2 - Previous studies to validate knee rules in children have shown sensitivities of 100% but wide confidence intervals have limited their applicability. This prospective, multicentre study is the largest yet to attempt to validate the best known knee rule in children. Children, aged between 2 and 16 years, presenting to five Canadian paediatric emergency departments with knee injuries sustained in the preceding seven days were assessed by physicians who reported presence or absence of a clinical indication for radiographs according to the Ottawa Knee Rule. Exclusion criteria included superficial lacerations and abrasions. Radiographs were requested at the physician’s discretion and interpreted by paediatric radiologists blinded to the knee rule result. A positive outcome was defined as any fracture. Of 750 children enrolled, 70 had a fracture. The Ottawa Knee Rule proved 100% sensitive (95% CI 94.9% to 100%), with a specificity of 42.8% (CI 39.1 to 46.5%). The authors identify several limitations with the study. Almost 389 (34%) eligible children were not enrolled, though the records of the 16 children with fractures found in this group indicate that Ottawa indications for a radiograph would have been met. Not all children had the criterion standard (knee radiograph) applied. It is likely, however, that any fractures missed in these 80 patients, asymptomatic on telephone follow up at 14 days, were clinically insignificant. Overall, 89% of children enrolled had radiographs performed, which seems intuitively to be rather high. The finding that 31% less radiographs would have been performed had the rules been strictly applied is from a post hoc analysis, which might tend to overestimate the reduction. Funding restrictions resulted in early termination of the study, a consequence of which was the small number of children enrolled in the 2–5 years age group. The wide CI (47.8–100%) around the sensitivity of the rule in … ER -