A clinical study was undertaken to formulate better criteria for accurate diagnosis of the lumbar facet syndrome and for predicting treatment response to facet joint injection. Twenty-two consecutive patients with a clinical diagnosis of lumbar facet syndrome, made by conventional diagnostic criteria, who were then treated with facet joint injection, were reviewed for their treatment responses. New diagnostic criteria were formulated based on a scoring system derived from the values observed in this review study. The scoring system has a total of 100 points, allocated as follows: back pain associated with groin or thigh pain, 30 points; well-localized paraspinal tenderness, 20 points; reproduction of pain with extension-rotation, 30 points; corresponding radiographic changes, 20 points; and pain below the knee, -10 points. A score of 60 points or more indicates a very high probability of satisfactory response to facet joint injection (100% prolonged response in this study). When only the conventional criteria were used, the overall results of prolonged relief of pain after facet joint injection was 50%. A "scorecard" system is proposed that may give a higher degree of diagnostic accuracy and predictability of successful response to facet joint injection.