隨著全球生方活方式的演變，現代人常出現不良的生活習慣，如不均衡的飲食習慣跟缺乏運動等等，由於這些問題層出不窮，在各個年齡層都有可能發生，因而代謝症候群成為一個日益普遍的健康問題，這不僅增加了心血管疾病和糖尿病的風險，也與胃癌等腸胃道癌症的發展密切相關。然而，目前的腸胃道疾病診斷方法，如內視鏡檢查，雖然精確但存在著侵入性和高成本的問題，使其難以作為大規模的篩檢工具，基於這樣的背景，本研究的目標是開發一種更有效的預測模型，用於識別具有高風險IM的個體，這樣的方法不僅有助於早期發現高風險患者，還能改善臨床篩檢的效率，從而降低腸胃道疾病的發病率和死亡率
本研究的資料蒐集於林口長庚醫院，涵蓋2010-2014年期間的患者健檢相關資訊，並且透過統計分析、Hayes Process Macro中介分析以及機器學習技術來探討各種影響腸胃道疾病的因子，藉由分析大量臨床數據來建構預測模型，這些模型的預測準確性和穩定性將進一步應用於臨床篩檢中，從而實現對高危險群體的識別能力，並為後續的診斷和治療提供科學依據，此研究不僅具有理論上的意義，還具備實際應用價值，希望能夠幫助醫療機構優化資源配置，提升腸胃道疾病的篩檢效率，最終為公共健康做出貢獻。

With the evolution of global lifestyles, modern people often exhibit poor living habits, such as unbalanced diets and lack of exercise. These issues are prevalent across all age groups, leading to metabolic syndrome becoming an increasingly common health problem. This not only raises the risk of cardiovascular diseases and diabetes but also is closely related to the development of gastrointestinal cancers such as stomach cancer. However, current gastrointestinal disease diagnostic methods, such as endoscopy, are accurate but invasive and costly, making them challenging as large-scale screening tools. Against this backdrop, the goal of this study is to develop a more effective predictive model for identifying individuals at high risk of metabolic syndrome. Such a method will aid in early detection of high-risk patients and improve the efficiency of clinical screening, thereby reducing the incidence and mortality rates of gastrointestinal diseases.
Data for this study were collected from Linkou Chang Gung Memorial Hospital, covering patient health check-up information from 2010 to 2014. Through statistical analysis、Hayes Process Macro mediation analysis and machine learning techniques, the study explores various factors influencing gastrointestinal diseases. By analyzing a large volume of clinical data, predictive models will be constructed, and their accuracy and stability will be further applied in clinical screening. This will enable the identification of high-risk groups and provide scientific evidence for subsequent diagnosis and treatment. This research not only has theoretical significance but also practical application value, aiming to help medical institutions optimize resource allocation, enhance the efficiency of gastrointestinal disease screening, and ultimately contribute to public health.

論文審定書 i
摘要 ii
Abstract iii
圖次 vi
表次 vii
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 1
第二章 文獻探討 3
2.1 發炎組 4
2.2 代謝組 5
2.3 其他組 5
2.4 IM 6
2.5相關因子 6
2.6 中介分析 10
第三章 研究方法與步驟 12
3.1 研究對象與資料蒐集 12
3.2 資料挑選 13
3.3 研究步驟 13
3.4研究方法 15
3.4.1 統計分析 15
3.4.2 Hayes Process Macro 16
3.4.3 Serial Mediation 18
3.4.4 Random Forest & XGBoost 19
第四章 研究成果 21
4.1 各組別的迴歸分析 21
4.1.1 未分組、性別分組與中介因子 21
4.1.2 經性別、年齡的分組 28
4.2 發炎、代謝與其他組的迴歸分析 39
4.3 各組的中介分析 54
4.4 發炎組、代謝組與其他組的中介分析 61
4.5 各組的機器學習 65
第五章 討論 72
5.1 中介分析 72
5.1.1 經年齡、性別分組 72
5.1.2 發炎組、代謝組與其他組 73
5.2 機器學習 73
5.3 結論 76
參考文獻 80
附錄 83

1. Hao, Y., Wang, Y., Qi, M., He, X., Zhu, Y., & Hong, J. (2020). Risk factors for recurrent colorectal polyps. Gut and Liver, 14(4), 399.
2. Liu, B., Wen, P., Gu, X., Weng, R., & Liu, S. (2020). Elevated serum triglyceride predicts recurrence of colorectal polyps in patients with advanced adenomas. Lipids in Health and Disease, 19(1), 1-8.
3. Khaw, K.-T., Wareham, N., Bingham, S., Luben, R., Welch, A., & Day, N. (2004). Preliminary communication: glycated hemoglobin, diabetes, and incident colorectal cancer in men and women: a prospective analysis from the European prospective investigation into Cancer–Norfolk Study. Cancer Epidemiology Biomarkers & Prevention, 13(6), 915-919.
4. Tseng, P.-H., Lee, Y.-C., Chiu, H.-M., Chen, C.-C., Liao, W.-C., Tu, C.-H., Yang, W.-S., & Wu, M.-S. (2012). Association of diabetes and HbA1c levels with gastrointestinal manifestations. Diabetes care, 35(5), 1053-1060
5. Dashti, S. G., Viallon, V., Simpson, J. A., Karahalios, A., Moreno‐Betancur, M., English, D. R., Gunter, M. J., & Murphy, N. (2020). Explaining the link between adiposity and colorectal cancer risk in men and postmenopausal women in the UK Biobank: a sequential causal mediation analysis. International Journal of Cancer, 147(7), 1881-1894.
6. van der Poorten, D., Milner, K. L., Hui, J., Hodge, A., Trenell, M. I., Kench, J. G., London, R., Peduto, T., Chisholm, D. J., & George, J. (2008). Visceral fat: a key mediator of steatohepatitis in metabolic liver disease. Hepatology, 48(2), 449-457.
7. Forootan, M., Tabatabaeefar, M., Yahyaei, M., & Maghsoodi, N. (2012). Metabolic syndrome and colorectal cancer: a cross-sectional survey. Asian Pacific journal of cancer prevention, 13(10), 4999-5002
8. Hamaguchi, M., Hashimoto, Y., Obora, A., Kojima, T., & Fukui, M. (2019). Non-alcoholic fatty liver disease with obesity as an independent predictor for incident gastric and colorectal cancer: a population-based longitudinal study. BMJ open gastroenterology, 6(1), e000295.
9. Kune, G. A., Kune, S., & Watson, L. F. (1988). Colorectal cancer risk, chronic illnesses, operations, and medications: case control results from the Melbourne Colorectal Cancer Study. Cancer research, 48(15), 4399-4404.
10. Ren, H., Luu, H., Cai, H., Xiang, Y., Steinwandel, M., Gao, Y., Hargreaves, M., Zheng, W., Blot, W., & Long, J. (2016). Oral health and risk of colorectal cancer: results from three cohort studies and a meta-analysis. Annals of Oncology, 27(7), 1329-1336.
11. Momen‐Heravi, F., Babic, A., Tworoger, S. S., Zhang, L., Wu, K., Smith‐Warner, S. A., Ogino, S., Chan, A. T., Meyerhardt, J., & Giovannucci, E. (2017). Periodontal disease, tooth loss and colorectal cancer risk: Results from the Nurses' Health Study. International Journal of Cancer, 140(3), 646-652.
12. Ward, H. A., Murphy, N., Weiderpass, E., Leitzmann, M. F., Aglago, E., Gunter, M. J., Freisling, H., Jenab, M., Boutron‐Ruault, M. C., & Severi, G. (2019). Gallstones and incident colorectal cancer in a large pan‐European cohort study. International Journal of Cancer, 145(6), 1510-1516.
13. Cheong, C. Y., Yap, P., Yap, K. B., & Ng, T. P. (2023). Associations of Inflammatory, Metabolic, Malnutrition, and Frailty Indexes with Multimorbidity Incidence and Progression, and Mortality Impact: Singapore Longitudinal Aging Study. Gerontology, 69(4), 416-427.
14. Muñoz, J. J. F., & Gónzalez, J. M. G. (2017). El análisis de mediación a través de la macro/interfaz Process para SPSS. REIRE Revista d'Innovació i Recerca en Educació, 10(2), 79–88-79–88.
15. Abu-Bader, S., & Jones, T. V. (2021). Statistical mediation analysis using the sobel test and hayes SPSS process macro. International Journal of Quantitative and Qualitative Research Methods.
16. Fiedler, K., Schott, M., & Meiser, T. (2011). What mediation analysis can (not) do. Journal of Experimental Social Psychology, 47(6), 1231-1236.
17. Anothaisintawee, T., Lertrattananon, D., Thamakaison, S., Reutrakul, S., Ongphiphadhanakul, B., & Thakkinstian, A. (2017). Direct and indirect effects of serum uric acid on blood sugar levels in patients with prediabetes: a mediation analysis. Journal of diabetes research, 2017.
18. Joo, Y.-E., Park, H.-K., Myung, D.-S., Baik, G.-H., Shin, J.-E., Seo, G.-S., Kim, G. H., Kim, H. U., Kim, H. Y., & Cho, S.-I. (2013). Prevalence and risk factors of atrophic gastritis and intestinal metaplasia: a nationwide multicenter prospective study in Korea. Gut and Liver, 7(3), 303.
19. Goldani, H., Adami, F. S., Antunes, M. T., Rosa, L. H., Fassina, P., Grave, M. T. Q., & Dal Bosco, S. M. (2015). Applicatility of the visceral adiposity index (VAI) in the prediction of the components of the metabolic syndrome in elderly. Nutricion hospitalaria, 32(4), 1609-1615.
20. Zullo, A., Hassan, C., Romiti, A., Giusto, M., Guerriero, C., Lorenzetti, R., Campo, S. M., & Tomao, S. (2012). Follow-up of intestinal metaplasia in the stomach: When, how and why. World journal of gastrointestinal oncology, 4(3), 30.
21. Amato, M. C., & Giordano, C. (2014). Visceral adiposity index: an indicator of adipose tissue dysfunction. International journal of endocrinology, 2014(1), 730827.
22. André, Q. (2022). PyProcess-PyPI. https://pypi.org/project/PyProcessMacro/#history