Aims. To evaluate physician compliance with the guidelines of the National Agency for Accreditation and Health Evaluation (ANAES) and the Consensus Conference on the use of medical imagery in low back pain and sciatica. Methods. We performed a retrospective study of 132 computed tomography scans (CTs) of the lumbar spine performed in one public and one private healthcare facility in the Haute-Vienne district, France. For each patient, the clinical findings, results of other investigations, prescriptions, and procedures reimbursed by the universal health insurance system were recorded. Results. Guidelines on imagery were followed in 2% of patients with chronic nonspecific low back pain. In 72% of patients, CT results had no influence on the subsequent clinical management. The guidelines were followed more often in patients with sciatica: 85% underwent CT more than 4 weeks after the initial painful episode. However, before CT was ordered, only 54% received appropriate initial treatment with analgesics, nonsteroidal anti-inflammatory drugs (NSAIDs), and/or muscle relaxants. Among these patients, 25% also received second-line medical therapy consisting of facet joint injection, conventional traction and, after the initial acute phase, physical therapy. In 39% of the sciatica patients, the imaging results had no effect on subsequent management. Among these patients, 12% underwent surgery for disk herniation. Conclusions. Ten years after the consensus conference and despite the publication of the ANAES guidelines, there is still a wide gap between observed practice and recommendations for optimal management. The consequences of this extend beyond unnecessary expenses for the universal health insurance system to include important deleterious effects on the patients. In particular, prompt appropriate management may help to avoid progression to chronic low back pain and unnecessary imaging studies and surgical procedures, which often have devastating social and occupational consequences.