Table 1.
Clinical scenarios considered in the simulation study
Table 2.
The effect size based on a total sample size of N patients, a 5% significance level, balanced allocation across K strata, m independent endpoint components and 80% power of the stratified WL test
Fig 1.
Impact of allocation bias in stratified clinical trials with multi-component endpoints evaluated using the stratified WL test.
The T1Es under misspecification calculated for samples of randomization lists generated by different RPs in clinical trials with N = 32 patients, K balanced strata, m standard normally distributed uncorrelated endpoints and common allocation bias effect
for all
and
chosen as proportion
of the effect sizes
,
,
and
.
Fig 2.
Impact of allocation bias on the test decisions of the stratified WL test depending on the sample size and the number of strata.
T1Es under misspecification calculated for samples of randomization lists generated by different RPs in clinical trials with
patients,
balanced strata, m = 2 standard normally distributed uncorrelated endpoints and common allocation bias effect
for all
and
with
and
.
Fig 3.
Impact of allocation bias on the test decisions of the stratified WL test for increasing number of endpoint components aggregated into the multi-component endpoint.
Mean misspecified T1Es calculated for samples of randomization lists generated by different RPs in clinical trials with N = 32 patients, K = 2 balanced strata,
standard normally distributed uncorrelated endpoints and common allocation bias effect
for all
and
with
.
Fig 4.
Impact of allocation bias on the test decisions of the stratified WL test for correlated endpoint components aggregated into the multi-component endpoint.
Misspecified T1Es calculated for samples of randomization lists generated by different RPs in clinical trials with N = 32 patients, K = 2 balanced strata, m = 4 correlated normally distributed uncorrelated endpoints and common allocation bias effect
for
and
with
. All endpoint components are equally correlated by the factor
.
Table 3.
Key lessons learned for mitigating allocation bias in small sample stratified clinical trials with multi-component endpoints that are evaluated using the WL test.