Fundamentals Tamuno Alfred, PhD Biostatistician DataCamp - PowerPoint PPT Presentation

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Fundamentals Tamuno Alfred, PhD Biostatistician

DataCamp Designing and Analyzing Clinical Trials in R What are clinical trials? Clinical trials are scientific experiments used to evaluate the safety and efficacy of one or more treatments in humans.

DataCamp Designing and Analyzing Clinical Trials in R Classifications

DataCamp Designing and Analyzing Clinical Trials in R Randomization

DataCamp Designing and Analyzing Clinical Trials in R Hierarchy of medical evidence

DataCamp Designing and Analyzing Clinical Trials in R Confounders

DataCamp Designing and Analyzing Clinical Trials in R Aim to achieve similar patient characteristics

DataCamp Designing and Analyzing Clinical Trials in R Blinding

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Let's practice!

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Types of data and endpoints Tamuno Alfred, PhD Biostatistician

DataCamp Designing and Analyzing Clinical Trials in R Conducting clinical trials Guidelines Trial protocols

DataCamp Designing and Analyzing Clinical Trials in R Statistical Analysis Plans

DataCamp Designing and Analyzing Clinical Trials in R Endpoints Prioritized into primary and secondary endpoints Safety and/or efficacy measures

DataCamp Designing and Analyzing Clinical Trials in R Continuous endpoints

DataCamp Designing and Analyzing Clinical Trials in R Continuous endpoints library(ggplot2) ggplot(data=exercise, aes(x=sbp_change)) + geom_histogram(fill="white", color="black") + xlab("SBP Change, mmHg")

DataCamp Designing and Analyzing Clinical Trials in R Continuous endpoints library(dplyr) library(magrittr) exercise %>% summarize(mean_sbp = mean(sbp_baseline), sd_spb = sd(sbp_baseline))

DataCamp Designing and Analyzing Clinical Trials in R Categorical endpoints library(dplyr) library(magrittr) finaldata %>% filter(!is.na(response)) %>% count(response, treatment) %>% mutate(pct = 100 * n / sum(n)) table(finaldata$response, finaldata$treatment)

DataCamp Designing and Analyzing Clinical Trials in R Composite endpoints Combine multiple outcomes Summarize as categorical variable

DataCamp Designing and Analyzing Clinical Trials in R Count endpoints Non-normal distribution Discrete values

DataCamp Designing and Analyzing Clinical Trials in R Survival endpoints

DataCamp Designing and Analyzing Clinical Trials in R Other collected data Dates of birth Dates of study visits Times of blood collection Ethnicity Gender Adverse events Year of diagnosis Concomitant medication …

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Let's practice!

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Basic statistical analysis Tamuno Alfred, PhD Biostatistician

DataCamp Designing and Analyzing Clinical Trials in R Statistical inference

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing Estimate treatment effect, e.g. difference in means Confidence interval, typically 95%

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing Estimate treatment effect, e.g. difference in means Confidence interval, typically 95% Test statistic p-value

DataCamp Designing and Analyzing Clinical Trials in R Hypothesis testing Estimate treatment effect, e.g. difference in means Confidence interval, typically 95% Test statistic p-value Compare to significance level α, typically 0.05

DataCamp Designing and Analyzing Clinical Trials in R Continuous: comparison of means Normal distribution Two-sample t-test t.test(pct.change~group, var.equal=TRUE, data=bmd)

DataCamp Designing and Analyzing Clinical Trials in R Continuous: comparison of means

DataCamp Designing and Analyzing Clinical Trials in R Continuous: comparison of distributions Non-normal distribution Wilcoxon Rank Sum test (aka Mann Whitney test) wilcox.test(outcome.variable~ group.variable, data=dataset)

DataCamp Designing and Analyzing Clinical Trials in R Continuous: comparison of distributions Non-normal distribution Null hypothesis Wilcoxon Rank Sum test (aka Mann Whitney test) wilcox.test(outcome.variable~ Alternative hypothesis group.variable, data=dataset)

DataCamp Designing and Analyzing Clinical Trials in R Binary: comparison of proportions Chi-squared test of independence Use Fisher’s Exact Test on small sample sizes table1<-table(care.trial$group, care.trial$recover) prop.test(table1, correct=FALSE)

DataCamp Designing and Analyzing Clinical Trials in R Binary: comparison of proportions

DataCamp Designing and Analyzing Clinical Trials in R Assumptions Independent groups Similar patient characteristics between groups

DataCamp Designing and Analyzing Clinical Trials in R Extensions Repeated measures data Three or more treatment groups

DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Let's practice!