Abstract: Objective To evaluate the health economic benefits of various screening strategies for upper gastrointestinal cancer in Fujian Province and to determine the most effective approach. Methods A Markov model was developed from the perspective of the healthcare system, utilizing both survey and literature data. Four distinct screening strategies were proposed⁃screening every 2 years, 3 years, 5 years, and 10 years with no screening serving as the control. The life years (LY), quality⁃adjusted life years (QALY) and net present value (NPV) were served as the cost⁃effectiveness, cost⁃utility, cost⁃benefit evaluation indicators, respectively. The incremental cost⁃effectiveness ratio (ICER), incremental cost⁃utility ratio (ICUR), and NPV were calculated for each screening strategy. The economic viability of each strategy was assessed using the 2023 per capita GDP of Fujian Province (¥129,865) as the willingness⁃to⁃pay (WTP) threshold. One⁃way sensitivity analysis and probability sensitivity analysis were performed to assess the robustness of the results. Results Compared with no⁃screening strategy, the ICERs for screening strategies conducted every 10, 5, 3, and 2 years were ¥4,215.47/LY, ¥5,208.17/LY, ¥5,566.44/LY, and ¥5,940.07/LY, respectively. Similarly, the ICURs were ¥4,316.71/QALY, ¥5,397.35/QALY, ¥5,788.25/QALY, and ¥6,183.54/QALY, respectively, all of which were significantly below the WTP threshold.The NPVs of  four screening strategies were positive, with biennial screening emerging as the most cost⁃effective option. Sensitivity analysis showed that the biennial screening remained highly cost⁃effective across all reasonable parameter variations, achieving a cost⁃effectiveness probability of 90% and stabilizing when the WTP threshold exceeded ¥20,000 per QALY. Conclusions Upper gastrointestinal cancer screening in Fujian Province offers favorable economic benefits, with biennial screening identified as the optimal strategy.

Key words: Upper gastrointestinal cancer, Screening strategy, Health economic evaluation, Markov model