Purpose: Detection of early inflammatory back disease is often difficult. Certain clinical characteristics have been reported to increase the likelihood of its detection in referral patients, but the usefulness of these clinical characteristics has not been evaluated in an open population. In our study, we undertook to evaluate the value of the clinical history as a screening test for inflammatory back disease in a general population. Patients and methods: Twenty-three male patients with back pain of moderate duration and with clinical characteristics suggestive of inflammatory back disease were recruited by advertising and were studied by various means, including computed tomography (CT), scintigraphy, and radiography. Results: One patient had radiographic sacroiliitis. Two had positive results for the B27 antigen, and another had positive results for the cross-reacting HLA antigen B7. Eight patients had abnormal scintiscans of the sacroiliac joints. Twenty-one of 23 patients and 20 of 23 control subjects had abnormalities that were detected by CT. Repeat plain radiographs of the pelvis done 36 months after enrollment into the study did not uncover further evidence of sacroiliitis. Conclusion: These results indicate plain radiographic evidence of sacroiliitis will often not develop in patients with historical features suggestive of inflammatory back disease even with long-term evaluation, thus vitiating the specificity of these historical findings in men with back pain of relatively brief duration. © 1988 The American Journal of Medicine.