STUDY DESIGN: Prospective cohort study. OBJECTIVE: We estimated the cost-effectiveness of spinal cord stimulation (SCS) among workers' compensation recipients with failed back surgery syndrome (FBSS). SUMMARY OF BACKGROUND DATA: Randomized controlled trial (RCT) evidence suggests that SCS is more effective at 6 months than medical management for patients with FBSS. However, procedure costs are high and workers' compensation claimants often have worse outcomes than other patients. METHODS: We enrolled 158 FBSS patients receiving workers' compensation into three treatment groups: trial SCS with or without permanent device implant (n = 51), pain clinic (PC) evaluation with or without treatment (n = 39), and usual care (UC; n = 68). The primary outcome was a composite measure of pain, disability and opioid medication use. As reported previously, 5% of SCS patients, 3% of PC patients and 10% of UC patients achieved the primary outcome at 24 months. Using cost data from administrative databases, we calculated the cost-effectiveness of SCS, adjusting for baseline covariates. RESULTS: Mean medical cost per SCS patient over 24 months was $52,091. This was $17,291 (95% confidence intervals [CI], $4100-30,490) higher than in the PC group and $28,128 ($17,620-38,630) higher than in the UC group. Adjusting for baseline covariates, the mean total medical and productivity loss costs per patient of the SCS group were $20,074 ($3840-35,990) higher than those of the PC group and $29,358 ($16,070-43,790) higher than those of the UC group. SCS was very unlikely (<5% probability) to be the most cost-effective intervention. CONCLUSION: In this sample of workers' compensation recipients, the high procedure cost of SCS was not counterbalanced by lower costs of subsequent care, and SCS was not cost-effective. The benefits and potential cost savings reported in RCTs may not be replicated in workers' compensation patients treated in community settings.