Continuous variable Categorical variables ChiSquare test T test

論文中常用的統計分析方法 Continuous variable Categorical variables • 兩組獨立樣本檢定 • 關聯性檢定 – Chi-Square test – T test • 兩組配對樣本檢定 – Paired t test • 三組獨立樣本檢定 – ANOVA – Mcnemar’s Test 9

Group (independent variable) 2組樣本t test， P < 0. 05表示有顯著 差異 (2組年齡平均值 不一樣) 卡方檢定， P < 0. 05表示有顯著 差異 (2組有DM的 比例不一樣) 主要比較有depressive symptoms與沒有 depressive symptoms的 CAD patients在不同變項 間有無差異 What is p-value ? 13

The Statistical Model-building in Epidemiologic Studies

Generalized Linear Models Model Equation Interpretation Linear (simple) Increase in outcome y mean value per unit increase in x 1, adjusted for all other variables in model Logistic Increase in log (odds) of outcome per unit increase in x 1, adjusted for all other variables in model Cox Increase in log (hazard) of outcome per unit increase in x 1, adjusted for all other variables in model Poisson Increase in log (hazard) of outcome per unit increase in x 1, adjusted for all other variables in model

RR & OR的判讀 Binary (Dichotomous) Outcome (e. g. mortality) 依據不同之研究設計 Prospective Relative risk Odds ratio Retrospective Hazard ratio (involving time) Odds ratio 22

Differences in mean levels of plasma lipids between categories of coffee consumption were tested by analysis of variance (ANOVA) after adjusting for the covariates included in model 2 (above). §調整了年齡、性別等效應後，比較喝不同咖啡杯數的人，其 Total cholesterol, LDL, HDL, Triglycerides其平均值是否有差異? §結果顯示 4組喝不同咖啡杯數的人，其Total cholesterol, LDL, HDL, Triglycerides的平均值並無明顯差異 (p value 皆大於 0. 05)。 31

多數據層面的統合分析 (Meta-analysis) + = 32 Data from http: //library. downstate. edu/EBM 2/2100. htm

Fig 1 | Random effects models of effect of calcium supplements on risk of total fracture. 結論：目前的最佳證據並不支持 補鈣可以降低骨折風險。 38 Calcium intake and risk of fracture: systematic review. BMJ 2015; 351: h 4580.