Transcript statistics
Basic Statistics Fabrício Rodrigues Santiago MD Vascular Surgeon/Board Certified by the Brazilian Society for Vascular Surgery Vascular Sonographer/Board Certified by the Brazilian College of Radiology- Brazilian Society for Vascular Surgery Master in Science– University of São Paulo Phd fellow- University of Goias Member of the Society for Vascular Surgery Member of the American College of Phlebology Basic Statistics NO DISCLOSURES Basic Statistics Basic Statistics Basic Statistics STATISTICS Don’t wanna know Don’t understand Scared Basic Statistics Basic Statistics Basic Concepts Hypothesis testing Ho = H1 Ho ≠ H1 Sample size and power calculations α=5% (Type I error) β=70-90% (Type II error) Basic Statistics Basic Concepts Types of Data Categorical Ordinal Nominal Basic Statistics Basic Concepts Objective Non parametric data Compare 2 independent groups Parametric data (normal distribution) t test Compare 2 dependent groups Paired t test Wilcoxon test Compare 3 or more independent groups Numerical Variable correlation variance analysis (ANOVA) Kruskal-Wallys Pearson correlation coefficient Sperman correlation coefficient Categorical variable association Qui-square Fisher’s test (for independent and small samples) Mann-Whitney test Basic Statistics Basic Concepts P value Confidence Interval Odds Ratio Relative Risk Clinical versus statistical significance P value Null Hypothesis Ho = H1 α= 5% Treatment A x Treatment B Treatment A x Treatment B 100 individuals for each arm 20 individuals for each arm P = 0.056 95% Confidence Interval Showed up to 20 percentage points “Absence of evidence is not evidence of absence” P = 0.034 Clinically significant Basic Statistics THE GOOD AND THE BAD Become more research-active More equipped ( teach or explain things to other people) Time consuming Skills (Numbers, Statistical packages) Critically evaluate published research and drug company literature Better understand and explain risk Justify Clinical Decisions Communicate risk to patients Basic Statistics Ecouraging ACP Basic Statistics P<0,001 Basic Statistics THANK YOU!