The objective of this study was to develop a decision rule for ordering x-rays in knee injuries. Phase I was a retrospective chart review of 201 consecutive patients receiving knee radiographs in the emergency department in a 10-month period. Logistic regression was performed on 11 clinical indicators to develop a clinical decision rule. Phase II was a prospective validation study on 133 consecutive patients with knee injuries. All patients received radiographs to validate the decision rule. Sensitivity, specificity, and misclassification rate were calculated. Logistic regression analysis found that a fall or blunt trauma mechanism yielded a logistic regression sensitivity of 92%, specificity of 57%, with a false-negative rate of 0.9%. The addition of inability to ambulate and age (younger than 12 or older than 50 years of age) yielded a sensitivity of 92% with a specificity of 63%. The prospective study found the combination of fall or blunt trauma with either inability to ambulate or age (younger than 12 to older than 50 years of age) was 100% sensitive, with a specificity of 79%. The misclassification rate was 20%. Using this decision rule, the number of x-rays taken could have been reduced by 78%. A larger multicenter validation study of this knee radiograph decision rule is needed before widespread clinical usage.