MMMS (version 0.1)

get.subgroup: Searching for a treatment-specific subgroup

Description

Searching for a treatment-specific subgroup based on MMMS composite scores.

Usage

get.subgroup(score.info, data, lb = 20, ub = 80)

Arguments

score.info

The object returned by get.score().

data

A list containing input data used for calculating composite scores. The list should have at the least the following elements: $time, $event, $treat, and $covar, which are the same as those used for calculating the composite scores.

lb

A scalar indicating the lower bound of the search range for desired subgroup sizes in percentage (e.g. 20 means 20%).

ub

A scalar indicating the lower bound of the search range for desired subgroup sizes in percentage (e.g. 80 means 80%).

Value

A list with the following elements:

pct

All possible subgroup sizes within the desired range.

cutoff

All possible cutoffs for the composite scores (based on interaction terms) that give subgroup sizes within the desired range.

i.best

The index with respect to pct and cutoff that corresponds to the optimal subgroup.

fit.best

The fitted model based on the subgroup defined by the optimal cutoff.

pct.best

The subgroup size of the subgroup defined by the optimal cutoff.

cutoff.best

The optimal cutoff.

chisq.best

The chi-square statistic for the treatment-by-subgroup interaction for the subgroup defined by the optimal cutoff.

Details

This function is a function called by MMMS() to search for an optimal treatment-specific subgroup. A range of desired subgroup sizes can be specified. When no subgroup can be found within the range, values of NA will be returned.

References

Lin Li, Tobias Guennel, Scott Marshall, Leo Wang-Kit Cheung (2014) A multi-marker molecular signature approach for treatment-specific subgroup identification with survival outcomes. The Pharmacogenomics Journal. http://dx.doi.org/10.1038/tpj.2014.9

See Also

MMMS, get.score

Examples

  # load the dataset
  data(simdat)
  attach(simdat)
  
  # get composite scores using an interaction model
  score = get.score(time,event,treat,bio,covar,nfolds=5,alpha=0.5,
                    pos.direction=FALSE)  

  # get subgroup
  subgrp = get.subgroup(score,simdat,lb=20,ub=80)