Little attention has been given to selecting treatments in clinical trials of acupuncture. Yet in order to perform objective tests of this procedure, it is crucial that the selected treatments are considered representative of the style of practice being tested. We examined 16 traditional Chinese medicine (TCM) acupuncture texts or treatment articles to determine the consistency of diagnosis and recommended treatment for chronic low-back pain. Although 24 diagnostic patterns were described by 1 or more texts, only 4 patterns were described by at least half of the texts. Most texts (12/16) described only 3 or 4 patterns. These could be categorized into 3 broad types: cold, damp, wind, heat channel obstruction patterns; kidney vacuity patterns (sometimes differentiated into yang and yin patterns); and blood (or blood and qi) stasis patterns. Several acupuncture points were recommended by most texts regardless of the diagnosis, whereas other acupoints were recommended for specific diagnostic patterns. There was, however, substantial variation between texts in recommended acupoints, with less than 20% of all acupoints recommended by half or more of the texts. This varibility will make it difficult to select TCM treatments for clinical trials of chronic low-back pain that have wide applicability. We believe that examining treatment patterns in actual clinical practice is crucial in this situation. We suggest that this method of selecting treatments should be part of the process used when selecting treatments for all clinical trials of acupuncture, regardless of the style of practice.