DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Randomization methods Tamuno Alfred, PhD Biostatistician
DataCamp Designing and Analyzing Clinical Trials in R
DataCamp Designing and Analyzing Clinical Trials in R
DataCamp Designing and Analyzing Clinical Trials in R Simple Randomization set.seed(888) treatment <- c("A","B") simple.list <- sample(treatment, 20, replace=TRUE) cat(simple.list,sep="\n")
DataCamp Designing and Analyzing Clinical Trials in R Simple Randomization set.seed(888) treatment <- c("A","B") simple.list <- sample(treatment, 20, replace=TRUE) cat(simple.list,sep="\n") table(simple.list)
DataCamp Designing and Analyzing Clinical Trials in R Random Permuted Blocks
DataCamp Designing and Analyzing Clinical Trials in R Random Permuted Blocks library(blockrand) set.seed(888) block.list <- blockrand(n=20, num.levels = 2,block.sizes = c(2,2)) block.list
DataCamp Designing and Analyzing Clinical Trials in R Random Permuted Blocks For random block sizes block.list2 <- blockrand(n=20, num.levels = 2,block.sizes = c(1,2))
DataCamp Designing and Analyzing Clinical Trials in R Stratified Randomization Balance in patient characteristics may not be achieved in small studies Common strata include age group, geographical region and disease severity Generate a randomization list for each stratum
DataCamp Designing and Analyzing Clinical Trials in R Stratified Randomization
DataCamp Designing and Analyzing Clinical Trials in R Stratified Randomization over50.severe.list <- blockrand(n=100, num.levels = 2, block.sizes = c(1,2,3,4), stratum='Over 50, Severe', id.prefix='O50_S', block.prefix='O50_S')
DataCamp Designing and Analyzing Clinical Trials in R Extensions Three or more treatment groups Can accommodate other allocation ratios, e.g. 2:1 for active drug to placebo
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 Crossover, factorial and cluster randomized trials Tamuno Alfred, PhD Biostatistician
DataCamp Designing and Analyzing Clinical Trials in R Crossover trials
DataCamp Designing and Analyzing Clinical Trials in R Crossover trials Advantages Within-patient comparisons Eliminate between-patient variability Improved treatment effect precision Fewer patients need to be recruited
DataCamp Designing and Analyzing Clinical Trials in R Crossover trials Disadvantages Chronic, stable conditions Can only evaluate short-term effects There may be order effects Carryover effects Patient dropouts impact analysis
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs head(recovery.trial)
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs head(recovery.trial) recovery.trial %>% count(A, B)
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs recovery.trial %>% count(A, B, recover)
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs recovery.trial %>% count(A, B, recover) recovery.trial %>% group_by(recover) %>% filter(A=="Yes") %>% summarise (n = n()) %>% mutate(prop = n / sum(n))
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs Odds of recovery in A: 257/41 Odds of recovery in not A: 229/68 Odds of recovery in B: 241/58 Odds of recovery in not B: 245/51 Odds ratio of A vs. not A: (257/41)/(229/68) = 1.86 Odds ratio of B vs. not B: (241/58)/(245/51) = 0.86
DataCamp Designing and Analyzing Clinical Trials in R Factorial designs library(epitools) oddsratio.wald(recovery.trial$A, recovery.trial$recover)
DataCamp Designing and Analyzing Clinical Trials in R Cluster randomized trials
DataCamp Designing and Analyzing Clinical Trials in R Cluster randomized trials Example clusters: Schools Communities Factories Hospitals General/primary care practices Geographical regions
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 Equivalence and Non- inferiority Trials Tamuno Alfred, PhD Biostatistician
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials Demonstrate similar efficacy between two treatments Treatments may differ in convenience of administration, e.g. pill instead of injection Check efficacy of generic drugs
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials library(dplyr) library(magrittr) infection.trial %>% group_by(Treatment, Infection) %>% summarise (n = n()) %>% mutate(pct = (n / sum(n))*100)
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials prop.test(table(infection.trial$Treatment,infection.trial$Infection), alternative = "less", conf.level = 0.95, correct=FALSE) prop.test(table(infection.trial$Treatment,infection.trial$Infection), alternative = "greater", conf.level = 0.95, correct=FALSE)
DataCamp Designing and Analyzing Clinical Trials in R Equivalence trials prop.test(table(infection.trial$Treatment,infection.trial$Infection), alternative = "two.sided", conf.level = 0.90, correct=FALSE) Two-sided 90% confidence interval excludes the delta of 12%, therefore can claim equivalence at the 5% level.
DataCamp Designing and Analyzing Clinical Trials in R Non-inferiority trials
DataCamp Designing and Analyzing Clinical Trials in R Non-inferiority trials prop.test(table(infection.trial$Treatment,infection.trial$Infection), alternative = "less", conf.level = 0.975, correct=FALSE) One-sided 97.5% confidence interval includes the margin of 12%, therefore cannot claim non-inferiority at the 2.5% level.
DataCamp Designing and Analyzing Clinical Trials in R Caution There are various recommendations so clearly state: Whether using one or two-sided confidence intervals Significance level Lack of superiority does not imply equivalence
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 Bioequivalence trials Tamuno Alfred, PhD Biostatistician
DataCamp Designing and Analyzing Clinical Trials in R Bioequivalence trials Conducted to assess whether blood concentration profiles for two formulations of a drug are similar
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetics Pharmacokinetics (PK): The study of what the body does to a drug. ADME: Absorption Distribution Metabolism Excretion
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetic profiles
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetic profiles
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetic profiles Crossover designs are often used
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetic profiles AUC using the linear trapezoidal method
DataCamp Designing and Analyzing Clinical Trials in R Pharmacokinetic profiles AUC using the linear trapezoidal method library(PKNCA) pk.calc.auc(PKData$plasma.conc.n, PKData$rel.time, interval=c(0.25, 12), method="linear")
DataCamp Designing and Analyzing Clinical Trials in R Assessing bioequivalence 90% confidence interval of the ratios should be contained within 0.8 to 1.25
DataCamp Designing and Analyzing Clinical Trials in R Assessing bioequivalence
DataCamp Designing and Analyzing Clinical Trials in R DESIGNING AND ANALYZING CLINICAL TRIALS IN R Let's practice!
Recommend
More recommend
