OBJECTIVE Investigators reporting decompressive surgery to treat patients with cervical spondylotic myelopathy (CSM) have described inconsistent benefits. In the present study the authors used three types of outcomes instruments to assess the results of CSM surgery. METHODS The authors collected prospective baseline and 6-month follow-up data in a cohort of 62 patients with CSM. Data collection included those pertaining to demographics; symptoms; physical findings; myelopathy severity; health status measured with the Short Form-36; and health values according to the standard gamble, time trade-off, visual analog scale, and willingness to pay. Rank-order methods were used to compare surgical and nonsurgical patients, and multivariate regression techniques adjusting for baseline characteristics were performed to examine the effects of surgery. During the study period, 28 patients underwent surgery, 34 did not, and there were no baseline differences between the two groups in demographics, symptoms, myelopathy scores, health status, or health values (p > or = 0.120 in all domains); there was a greater prevalence of hand intrinsic muscle atrophy (p = 0.035) and Hoffmann sign (p = 0.006) in the surgery-treated group. Neither raw comparisons nor regression analyses showed a consistent surgery-related benefit. There were sporadic associations between worse outcomes and older patients, higher income, Babinski sign, longer duration of CSM symptoms, hand clumsiness, lower-extremity numbness, and multilevel surgery (p < or = 0.049 for all). CONCLUSIONS Analysis of results obtained in the 62 patients with CSM failed to show a surgery-related benefit despite the use of three classes of outcomes instruments. Patient demographics, symptoms, physical signs, and the surgical approach may explain some of the variation in outcomes in patients with CSM.