OBJECTIVES: To compare knee radiology requesting rates among junior doctors before and after the formal introduction of the Ottawa Knee Rules (OKR) in a UK emergency department (ED), and to test the validity of the OKR for decisions on the use of radiography for acute, isolated knee injuries. METHODS: All junior doctors in a district general hospital ED seeing adult patients with isolated knee injuries completed a questionnaire before and after the introduction of the OKR. All patients were followed up to obtain a final diagnosis. The outcome measures were: adherence to the OKR, the presence of a fracture and whether a radiograph had been requested. The results were analysed to determine the sensitivity, specificity, positive and negative predictive values of the OKR. Comparisons between the request rate for knee radiography before and after the introduction of the OKR were made. RESULTS: A total of 130 patients were enrolled and followed up over a 3-month period; 58 before and 72 after OKR introduction. The OKR had a sensitivity of 100% (71.8-100%), a specificity of 55.1% (46.1-64.1%), a positive predictive value of 18.5% (9.03-27.9%) and a negative predictive value of 100% (87.8-100%) for the detection of bony injury. The introduction of the OKR did not result in a significant reduction in the number of radiographs performed (58.6-55.6%; P= 0.726). DISCUSSION: This study shows the OKR to be a highly sensitive clinical guide with a high negative predictive value in the setting of a UK ED. It suggests that the reduction in radiograph requests seen elsewhere may not be as apparent in this setting.
OBJECTIVES: To compare knee radiology requesting rates among junior doctors before and after the formal introduction of the Ottawa Knee Rules (OKR) in a UK emergency department (ED), and to test the validity of the OKR for decisions on the use of radiography for acute, isolated knee injuries. METHODS: All junior doctors in a district general hospital ED seeing adult patients with isolated knee injuries completed a questionnaire before and after the introduction of the OKR. All patients were followed up to obtain a final diagnosis. The outcome measures were: adherence to the OKR, the presence of a fracture and whether a radiograph had been requested. The results were analysed to determine the sensitivity, specificity, positive and negative predictive values of the OKR. Comparisons between the request rate for knee radiography before and after the introduction of the OKR were made. RESULTS: A total of 130 patients were enrolled and followed up over a 3-month period; 58 before and 72 after OKR introduction. The OKR had a sensitivity of 100% (71.8-100%), a specificity of 55.1% (46.1-64.1%), a positive predictive value of 18.5% (9.03-27.9%) and a negative predictive value of 100% (87.8-100%) for the detection of bony injury. The introduction of the OKR did not result in a significant reduction in the number of radiographs performed (58.6-55.6%; P= 0.726). DISCUSSION: This study shows the OKR to be a highly sensitive clinical guide with a high negative predictive value in the setting of a UK ED. It suggests that the reduction in radiograph requests seen elsewhere may not be as apparent in this setting.