Applying decision curve analysis to spine surgery

Spine surgery is expensive, invasive and associated with important risks. The potential benefits of intervention must be balanced against these harms and risks. Decision curve analysis is a statistical method for evaluating models, rules and tests to indicate patients for intervention. The net benefit of a proposed tool can be compared with the clinical default strategies of “treat all” vs. “treat none” across a range of clinically reasonable threshold probabilities, thereby demonstrating whether the use of a prediction model or diagnostic test is clinically useful. Here we discuss the current applications of decision curve analysis within the spine population.