Application of High-Dimensional Propensity Score Model and Marginal Structural Model for Unmeasured Confounding and Informative Censoring – Take Metabolic Risk Factors, Drugs, and Infection Risks as an Example = 深入探討以高維傾向性分數與邊緣性結構模型處理未測量干擾因子及非獨立設限 - 以代謝危險因子 藥物 與感染症風險為例