| Alonso LL and Thomas TB, 1995, USA | 110 consecutive patients in the emergency department with jaw pain | Diagnostic study | Patient ability to grasp a tongue blade/depressor between his teeth and hold the blade against a twisting motion, with the ability to crack the tongue blade on both sides defined a negative test | Sensitivity of tongue blade test: 45 positive of 47 fractures, sensitivity 95.7%, CI (85.5 to 99.5%) | No sample size estimates—sample used too small |
| Exclusion criteria: patients unable to cooperate because of age, language barrier, inability to open mouth because of pain, intoxication or head injury | Inability to crack tongue blade on both sides of the mandible defined positive test | Specificity of tongue blade test: 23 positive TBT of 63 non-fractures, specificity 63.5%, CI (50.4 to 75.3%) | “Gold standard” radiologist interpretation sensitivity measured in an other study in this hospital and was found to be only 95.5% |
| Single radiologist to report radiographs—no double reporting |
| Robert A et al, 1998, USA | 119 patients with jaw pain after trauma presenting at the emergency department | Diagnostic study | Patient ability to grasp a tongue blade/depressor between his teeth by and then to hold the blade against mild resistance by the examiner | Sensitivity of tongue blade test: 42 positive of 44 fractures, sensitivity 95% | No OPG despite panoramic series known to be more sensitive than mandibular series in detecting mandibular fracture |
| Single radiologist to report radiographs—no double reporting |
| No interobserver variability measurements |
| No CIs calculated |
| Incorrect use of χ2 test—number too small should have used Fisher’s exact test |
| Exclusion criteria: airway compromise, inability to perform or cooperate with clinical examination, edentulousness | Inability to hold tongue blade against resistance on either hemimandible defined as a positive test | Specificity of tongue blade rest: 25 positive TBT of 75 non-fractures, specificity 67% | No sample size estimates—sample used too small |