The aim of this study was to develop a short diagnostic test for pain-related craniomandibular disorders (CMD) based on the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD).Participants included 1,177 CMD patients and 896 general population subjects who were examined according to the RDC/TMD. This new diagnostic short test consisted of a combination of the least amount of RDC/TMD items that distinguished between patients and subjects with a sensitivity of > or =70% and a specificity of > or =90%. The diagnostic test items were selected from all available RDC/TMD items using best subset logistic regression.The question about the presence of facial pain achieved a sensitivity of 96% and a specificity of 95%. The lower limits of the confidence interval for test accuracy measures exceeded the postulated thresholds specified for test development. Assuming a CMD pain prevalence of 10% in the general population this short test resulted in a positive predictive value of 80% and a negative predictive value of >99%.A single question about facial pain is a strong predictor for a pain-related CMD disorder and could provide an effective CMD short test.