Identifying relevant subgroups of nonspecific low back pain patients may improve research efficiency and clinical outcomes. This article presents available data supporting the rationale behind a new pathoanatomic oriented classification system for low back pain. A systematic framework was used to assess issues of validity, reliability, feasibility, and generalizability. Results from studies examining elements of the 12 categories of the classification system are presented to allow the assessment to be made. Particular emphasis has been given to the evidence of diagnostic accuracy of the criteria based on data from history taking and clinical tests. The classification system was found to be potentially useful for research purposes. Further testing of the classification system is required.