OBJECTIVES: Estimate the return of investment (ROI) on bariatric surgery vs. conventional, non-surgical approach as treatment for morbid obesity from the Mexican public health system perspective in the long term. METHODS: The individual experience of a morbidly-obese patient was assessed using a discrete event simulation model built in Arena™. Patients were created with unique, randomly assigned clinical and epidemiologic characteristics, cloned and sent to either bariatric surgery (BS) or conventional treatment – pharmacologic treatment of associated comorbidities and lifestyle modifications (control arm). Evaluated comorbidities were type-2 diabetes, hypertension and hypercholesterolemia. Preoperative prevalences and up-to year 2 recovery rates were taken from published meta-analyses. 2- and 10-year prevalences were derived from incidence and recovery rates shown in SOS study; in-between prevalences were interpolated assuming exponential growth, thus allowing clinical state worsening in both arms. Additional assumptions include infrastructure restrictions, no perioperative complications and short term mortality. 5% of patients in control group were allowed to have surgery after year 5. Considered costs included the bariatric procedure and comorbidity-specific pharmacologic treatment, taken from public health institution’s DRGs. Simulation was run with 150 patients for 10 years and 10 iterations using a 4.5% annual discount rate. Results are shown in years and 2010 inflation-adjusted MXP. 95% confidence intervals were estimated. RESULTS: 10-year accumulated cost for a BS patient was $125,902 ($125,041-$ 126,763), and $259,413 ($258,098 - $260,728) for a control patient. ROI on BS was achieved on year 6.94 (6.88-7.00). Cost differences are due to the reduced resource utilization after surgery resulting from resolution of comorbidities. CONCLUSIONS: Investment in BS offsets its cost and is recouped within a reasonable time, thus allowing institutions to reduce the burden imposed by morbid obesity. Long-run data for other associated comorbidities is needed, as their inclusion in the analysis could decrease ROI.