OBJECTIVES: To compare the effects of the two currently used treatment methods for intractable epilepsy, i.e. surgical and pharmacological treatment, METHODS: From the Korean national health insurance data from 2002 to 2013, patients with epilepsy (G40) or status epilepticus (G41) identified as their chief diagnoses were selected. Among these patients, those that were treated with two or more types of anticonvulsants were defined as patients with intractable epilepsy. Among these patients, those that had epilepsy operations, such as temporal lobectomy, corpus callosotomy, non-temporal lobectomy, and multiple subpial transection, were selected. An exact matching was performed to select the comparative nonsurgical group, based on the characteristics of the surgically treated patients. Consequently, 1,443 patients in each of the surgical and nonsurgical groups were matched. Either descriptive or generalized estimating equation(GEE) models were used to model cost outcomes. RESULTS: Both surgical and nonsurgical groups spent similar medical costs until 12 months prior to surgery. However, the medical costs in the surgical group rapidly increased at the time of surgery (surgical group, KRW 16,009,867; nonsurgical group, KRW 1,317,113 for six months). Their medical costs subsequently decreased to a level lower than that of the nonsurgical group beginning one year after surgery (two years after surgery: surgical group, KRW 1,132,205; nonsurgical group, KRW 1,395,980 for six months). Similar trends were observed in the GEE analysis results, which revealed a significant decrease in the surgical group’s costs from 18 months after surgery (ref = nonsurgical group 12 months before surgeries, p < 0.05). CONCLUSIONS: The medical costs paid by the patients rapidly increased before and after surgery due to the surgeries themselves. The surgical group paid less medical costs after surgery with the passage of time after surgery. Greater efforts will be needed to demolish the economic barriers present in epilepsy treatments.