Background: Analysis of large health care claims databases indicate that many tests and treatments are done for indications not supported by evidence, or in quantities in excess of those supported by evidence. Findings of this sort, coupled with variance in care for groups of cases of similar diagnoses, suggest that health care efficiency and effectiveness, key elements of quality, could be improved by adhering to practices supported by evidence of effectiveness.

Objectives: 1. To review and evaluate systematic reviews of tests and treatments for low back pain. 2. To present examples of care in excess of evidentiary support for low back complaints. 2. To suggest means to reduce inappropriate and variable use of health care resources.

Methods: We identified a number of systematic reviews and evidence-based guidelines for effective care of low back pain. We assessed them and collected recommendations and time frames for tests and treatments. We analyzed very large medical claims databases from the states of California, Texas and Ohio and compared the tests and treatments used per case to evidence of effectiveness.

Results: There is good evidence for the effectiveness or lack thereof of imaging, pharmaceutical treatment, physical modalities, and procedures for low back pain, radiculopathy, and other common diagnoses. Our studies revealed testing and treatment far in excess of evidentiary support. Higher levels of treatment are associated with worse outcomes, including disability absence. We also found some evidence of improvement in these patterns following organizational or public policy interventions, but in some cases it was not sustained.

Conclusion: Evidence-based best practices for back care are not uniformly applied. There is wide variation in practice patterns. Effective means to improve the consistency and appropriateness of practice are needed to improve outcomes.