Cost-Effectiveness of Spine Surgery: The Jury Is Out

Back pain affects more than 80% of persons and costs more than $100 billion annually in the United States (1). Although most patients with low back pain can be managed nonoperatively, spine surgery is an important option for patients with disabling herniated nucleus pulposus, lumbar spinal stenosis, or spondylolisthesis with stenosis. More than 650 000 surgical procedures are performed annually for these disorders in the United States, at a cost exceeding $20 billion (2, 3). Whether this investment provides good value is largely unknown.

In this issue, Tosteson and colleagues (4) investigate the cost-effectiveness of surgery for spinal stenosis. This work is critically important in the context of limited resources and seemingly unlimited costly therapies for chronic disease. The validity of cost-effectiveness analyses hinges on several questions, including 2 particularly pertinent to the work of Tosteson and colleagues. First, does the intervention improve health outcomes? Second, does the cost-effectiveness analysis incorporate all relevant costs, harms, and benefits (5–7)?

We address these questions by examining the design and results of the SPORT (Spine Outcomes Research Trial) studies that provide data for the analyses of Tosteson and colleagues (8–11). The scope, innovation, and significance of the SPORT studies are breathtaking. The SPORT studies compared outcomes of surgery with those of nonoperative therapy for the 3 most frequent indications for elective adult spine surgery: sciatica due to disc protrusion, neurogenic claudication due to lumbar spinal stenosis, and symptomatic degenerative lumbar spondylolisthesis. More than 3900 eligible persons at 13 sites were invited to participate (Table) in a randomized, controlled trial of surgery versus nonoperative treatment. Those unwilling to be randomly assigned were asked to join a parallel observational cohort study. Thirty-seven percent of eligible patients declined to participate, 36% agreed to enter the observational study, and 27% agreed to be randomly …