Please enable JavaScript.
Coggle requires JavaScript to display documents.
Stratified Randomization for Clinical Trials (2. Potential advantages of…
Stratified Randomization for Clinical Trials
1. Why doing? Limitations of simple stratification
First, it can fail to assign the compared treatments to equal numbers of patients.
Blocking -->
Second, simple randomization can fail if it creates treated groups of patients that are unbalanced for critical features that are known or suspected to affect prognosis.
Stratification -->
Just as
blocking
addresses the first potential failure of randomization,
stratified randomization
is intended to prevent the second (imbalance for prognostic features). In practice, the two procedures are often used together when randomization is blocked within strata.
2. Potential advantages of stratification
PROTECTION AGAINST TYPE II ERROR (INCREASED POWER)
INCREASE EFFICIENCY
PROTECTION AGAINST TYPE I ERROR
FACILITATION OF INTERIM ANALYSIS
ASSURANCE THAT COMPARED GROUPS ARE SIMILAR WITH RESPECT TO KNOWN PROGNOSTIC FACTORS
PROTECTION AGAINST EFFECTS OF RECRUITMENT CENTRO DROPOUT
3. Choice of stratification factors
Fewer strata are better.
Of greater importance is that it can improve statistical efficiency by making it more likely that equal numbers of patients are assigned to each treatment
The third advantage to fewer strata is to assure equal distribution of the stratification factor between treatment groups (i.e., to achieve the purpose of stratification).
First, it simplifies randomization and trial administration.
Fourth, limiting the number of strata avoids the problem of multiple comparisons that may occur if investigators report outcome rates separately by numerous strata.
To assure parsimony, investigators are encouraged to select only those clinical variables that have a known and important effect on outcome risk or treatment responsiveness.
Suggest that the number of strata be maintained at less than
n/(B x 4)
, where n is the sample size at the first planned interim analysis and 4 is a safety factor that accounts for unequal distribution of patients among strata.
4. Special analytic procedures for stratified data
The analysis of results from trials with stratified randomization should take account of the stratification
Simply to include the stratification factor as a covariate in whatever multivariate model is chosen.
Failure to account for stratification in the analysis will result in an overestimation of the P value for a difference between endpoint rates in treatment groups
5. Potential disadvantage of stratified randomization
For trials of therapies that require emergency administration, however, stratification may pose an unwarranted administrative burden.
6. Alternatives to stratified randomization
minimization
poststratification