COST-EFFECTIVENESS OF THE MOST COMMONLY USED NON-SURGICAL TREATMENTS FOR NECK PAIN

Purpose: The most commonly prescribed neck pain treatments in Canada and the United States are cyclooxegenase-2 selective inhibitors, exercise, manipulation, mobilization, and non-steroidal anti-inflammatory drugs. Our objective was to evaluate the cost-effectiveness of these five non-surgical treatments for acute non-specific neck pain in adult patients.  

Method: A decision-analytic Markov model was constructed to simulate the costs and consequences of the clinical course of acute non-specific (mechanical) neck pain in a hypothetical cohort of patients. The perspective adopted was that of the health care payer and the analysis was conducted over a lifetime time horizon. Resource use and costs were derived from professional recommended fee schedules, the Ontario Health Insurance Plan Schedule of Benefits, and the Ontario Case Costing Initiative. Costs were expressed in 2008 Canadian prices. The impact of the beneficial and harmful treatment effects on health were expressed in quality-adjusted life years (QALYs). Costs and QALYs were discounted at 5% per annum. Model inputs included estimates of: 1) the clinical course of acute neck pain, 2) treatment effectiveness, 3) treatment-associated risk of cerebrovascular, cardiovascular, and gastrointestinal adverse events, 4) background risk of adverse events in the general population, 5) standard gamble utilities directly elicited from a sample of 220 neck pain patients, and 6) direct and out-of-pocket costs. Cost-effectiveness was determined by using a willingness-to-pay threshold of $50,000 per QALY. Uncertainty surrounding model parameters results was explored with probabilistic sensitivity analysis with 10,000 simulations.

Result: Manipulation was cost-effective compared to the least costly alternative non-steroidal anti-inflammatory drugs (incremental cost-effectiveness ratio [ICER] = $25,123 per QALY). The ICER for mobilization was $381,926 per QALY. Cyclooxegenase-2 selective inhibitors and exercise were subject to simple dominance. Results of probabilistic sensitivity analyses suggested that non-steroidal anti-inflammatory drugs were most likely cost-effective at a willingness-to-pay of less than $24,000 per QALY whereas manipulation was most likely cost-effective between $24,000 and $50,000 per QALY.

Conclusion: Manipulation is cost-effective compared to non-steroidal anti-inflammatory drugs at a conventional threshold of $50,000 per QALY. Under varying thresholds, non-steroidal anti-inflammatory drugs are cost-effective at less than $24,000 per QALY. Forthcoming analysis will identify the main drivers of the uncertainty surrounding the decision about which non-surgical acute neck pain treatment to adopt.