The statistical significance of a phase-III clinical trial will guide clinical treatment in practice. However, the clinical trial-based strategy for treatment could be contaminated by the statistical errors. OBJECTIVES: The study purpose was to formulate the influence of the statistical errors on the cost-effectiveness of the subsequent treatment strategy based on the result of a phase-III clinical trial. METHODS: Consider a clinical trial for the two regimens: a new treatment A (TA) and an old treatment B (TB). Assuming the truth is that TA is more effective than TB, consider the subsequent three rational scenarios: 1) no significant difference between TA and TB with the type II error(?), then regard TA and TB as equivalent, 2) prefer TB to TA with the type III error(?), and 3) prefer TA to TB with the other probability (1????). Then the cost-effectiveness of the scenarios was theoretically evaluated in use of decision-analytic modeling. RESULTS: Using the variables?and?, the mathematical formula which estimates the cost-effectiveness ratio of the treatment model was proven to show an intermediate value between the cost-effectiveness ratios for TA and TB given in the phase-III clinical trial. If the type II error decreases, the subsequent cost-effectiveness ratio approaches to the better cost-effectiveness ratio given in the phase-III clinical trial. CONCLUSIONS: It was theoretically proven that rational treatment strategy according to the result of a phase-III clinical trial can improve the cost-effectiveness within the certain range even if the type II and the type III errors exist.