Statistical Software
Evaluating predictive markers in a randomized trial with binary outcomes
Stuart G. Baker, 2014
Introduction
A predictive marker is a baseline variable in a randomized trial that is used to determine subgroups in which the effect of treatment is greater than average. This software uses a modified adaptive signature design to evaluate a randomized trial with a binary outcome and multiple baseline variables (possibly high dimensional). The software splits the data into training and test samples. For the training sample the software fits various benefit functions. A fundamental option is whether to use crossvalidation in the training sample to select the best set of benefit functions (Method 1) or directly fit models in training sample (Method 2). For the test sample the software computes benefit scores based on the benefit function and treatment effect in subgroups with benefit scores greater than cutpoints. The software plots estimated treatment effect versus cutpoint, which is similar to a tailoriented subpopulation treatment effect pattern plot.
Requirement
Mathematica Version 8 or later.
SetUp
copy  all files into some folder called "FOLDER" 
start  a new Mathematica session 
type  SetDirectory["FOLDER"] 
type  << trialfit.m (M File, KB) 
To run simulation:
type  trialFit[datasim, NewFitQ>True] 
To run hypothetical data based on microarray with risk difference benefit function:
type  trialFit[dataPC, Method>2, ModelSet>{{"RD","F"}}] 
To try on your own data,
type  trialFit[dataset, options] 
Options
 Default 
Explanation 

Method  1  1 is a crossvalidation of training sample 2 is fit model directly to training sample  NewFitQ  TRUE  New fitting or use stored result of previous fit  Split  0.5  Fraction split into test sample  ModelSet  "All"  Models used e.g. {{"RD","F"},{"Cadit","M"}}  MinTrainTestGroupSize  20  Minimum sample size of trainingtest sample  Num Cut  8  Number of cutpoints in test sample  ShowProgQ  FALSE  Show progress of fitting algorithm  ShowCVQ  FALSE  Show details of crossvalidation  ShowTabQ  FALSE  Show data and results tables  MaxBoot  20  Number of bootstrap iterations 
dataset={x0,x1,y0,y1,xname,datasetname}
x0  n x g matrix of baseline variables for randomization group 0  x1  n x g matrix of baseline variables for randomization group 1  y0  a length n list of binary outcomes (0 or 1) for randomization group 0  y1  a length n list of binary outcomes (0 or 1) for randomization group 1  xname  a length g list of names of baseline variables  datasetname  name of dataset 
Downloads
Download All (zip, 1.59MB)
trialfit.m (M File, 13KB) 
calls all files 
trialfitinputcheck.m (M File, 9KB) 
Checks input 
trialfitcv.m (M File, 3KB) 
Method 1: crossvalidation of training sample 
trialfitcvchoose.m (M File, 6KB) 
Method 1: choose benefit function from crossvalidation 
trialfitmodel.m (M File, 3KB) 
Method 2: fit models 
trialfitsplit.m (M File, 5KB) 
Split data into training and test samples 
trialfitlogit.m (M File, 6KB) 
Fit logistic regression 
trialfitplot.m (M File, 7KB) 
Plot results 
trialfitboot.m (M File, 3KB) 
Bootstrap test sample 
trialfitbootci.m (M File, 4KB) 
Boostrap confidence interval 
trialfitrd.m (M File, 4KB) 
Risk difference 
trialfitrdtrain.m (M File, 2KB) 
Risk difference: training sample 
trialfitrdtest.m (M File, 2KB) 
Risk diference: test sample 
trialfitrdb.m (M File, 5KB) 
Risk difference with boosting 
trialfitrdc.m (M File, 5KB) 
Risk difference common variables 
trialfitrdctrain.m (M File, 1KB) 
Risk difference common variables: training sample 
trialfitresp.m (M File, 4KB) 
Responder only 
trialfitresptrain.m (M File, 2KB) 
Responder only: training sample 
trialfitresptest.m (M File, 2KB) 
Responder only: test sample 
trialfitcadit.m (M File, 5KB) 
Cadit 
trialfitcaditrain.m (M File, 3KB) 
Cadit: training sample 
trialfitcadittest.m (M File, 2KB) 
Cadit: test sample 
trialfitcaditb.m (M File, 5KB) 
Cadit with boosting 
trialfitvote.m (M File, 4KB) 
Vote 
trialfitvotetrain.m (M File, 4KB) 
Vote: training sample 
trialfitvotetest.m (M File, 3KB) 
Vote: test sample 
trialfitmax.m (M File, 4KB) 
Max 
trialfitmaxtrain.m (M File, 10KB) 
Max: training sample 
trialfitmaxtest.m (M File, 3KB) 
Max: test sample 
trialfitboost.m (M File, 3KB) 
Boosting computations 
trialfitsim.m (M File, 5KB) 
Generate simulated data 
trialfitdatamicro.m (M File, 3KB) 
Create hypothetical example from prostate cancer (PC) microarray data 
trialfitrawmicro.m (M File, 4.76MB) 
Raw prostate cancer (PC) microarray data for hypothetical example 
trialfitfig.m (M File, 7KB) 
Schematic figure 
Disclaimer
This code is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the NCI or the individual developers be liable for any claim, damages or other liability of any kind. Use of this code by recipient is at recipient's own risk. NCI makes no representations that the use of the code will not infringe any patent or proprietary rights of third parties.
Last updated: July 30, 2014
Back to top


