Purpose. The Bone and Joint Decade 2000-2010 Task Force on Neck Pain (Task Force) is conducting a best-evidence synthesis of the world literature on ‘neck pain', a common condition associated with significant disability. A parallel decision-analytic synthesis of the evidence on neck pain treatment was commissioned by the Task Force as part of its research mandate. Our purpose was to identify the best treatment amongst common non-invasive neck pain treatments. Methods. Life expectancy (in years) and quality-adjusted life expectancy (in quality-adjusted life years [QALYs]) associated with standard and Cox-2 inhibiting non-steroidal anti-inflammatory drugs (NSAIDs), exercise, mobilization, and manipulation were compared in a decision-analytic Markov model. Estimates of the course of neck pain, the background risk of adverse events in the general population, treatment effectiveness and risk were, wherever possible, drawn from the Task Force best-evidence synthesis. Patient-preferences for treatment outcomes were derived from a utility study of neck pain patients in Toronto (n=116) and Los Angeles (n=104). The model used two sets of assumptions about treatment effectiveness: 1) equivalent effectiveness, and 2) differential effectiveness, with estimates drawn from the best-available data (a single high-quality randomized trial). Results. When effectiveness was assumed to be equivalent, the model predicted the best treatment to be exercise (36.2 years, 27.3 QALYs). There were no important differences across treatments in losses in life expectancy and quality-adjusted life expectancy. The difference between the best treatment and lowest-ranked treatment (Cox-2 NSAIDs) was –4.5 days of life expectancy and –3.1 quality-adjusted life-days. Using effectiveness data from the trial, the model predicted mobilization to be the best treatment (27.3 QALYs). There were minor differences in quality-adjusted life expectancy across treatments. The difference between the highest and lowest ranked treatments was –7.3 quality-adjusted life-days. Average gains in quality-adjusted life expectancy associated with the effectiveness of these treatments were similarly small. For mobilization, exercise, and manipulation, average gains were 4.2, 3.9, and 1.7 quality-adjusted life days, respectively, compared to NSAIDs (the referent group). Conclusions. When the objective is to maximize life expectancy and quality-adjusted life expectancy, differences in expected value are too small to identify the best treatment amongst common, non-invasive neck pain treatments. The treatment decision about which treatment to provide should be based on patients' preferences for treatment and their attitudes toward risk.