OBJECTIVES: Korea introduced a new positive list system in 2007 together with a price negotiation procedure. Importantly, these two systems are run by two different, independent organizations, namely the Health Insurance Review & Assessment  Service (HIRA) and the National Health Insurance Corporation (NHIC). HIRA reviews the cost-effectiveness data in submissions and makes listing decisions, then NHIC takes over and sets the reimbursement price via negotiations with manufacturers. The aim of this study is to compare the difference in price after cost-effectiveness appraisal by HIRA and price negotiation by NHIC, and to analyze the factors that NHIC has considered to determine the reimbursement price. METHODS: All 35 submissions made to the NHIC between August 2007 and June 2008 were reviewed. 19 submissions concluded with agreement, 15 failed and one case was suspended. In this review only 15 cases of successful negotiations were included. The level of the reimbursement price compared to the submitted price for both essential drugs and non-essential ones and factors affecting the final price were analyzed.  RESULTS: The discrepancy between reimbursement price and cost-effective price was about 12.33±11.44% on average. For 3 essential drugs, the price level was almost equal to the submitted price whereas the average level was 84.94 ±11.21% of the cost-effective price for non-essential drugs. The major factors affecting negotiations to determine the final price were narrowed down to total cost of substitutes, the foreign price, and the pharmaceutical budget impact.  CONCLUSIONS: Our findings have demonstrated that drug pricing within the new environment has been done independently of cost-effectiveness appraisal. The payer has exhibited limited bargaining power for essential drugs. Overall, 87.67% of the cost-effective price was accepted during price negotiations, and the total cost of substitutes, foreign prices and pharmaceutical budget impact were considered equally when fixing the reimbursement price. A limitation of this study is that the result may not be generalized because of insufficient cases.