summarize.subgroups: Summarizing covariates within estimated subgroups

In personalized: Estimation and Validation Methods for Subgroup Identification and Personalized Medicine

Summarizing covariates within estimated subgroups

Description

Summarizes covariate values within the estimated subgroups

Usage

summarize.subgroups(x, ...)

## Default S3 method:
summarize.subgroups(x, subgroup, ...)

## S3 method for class 'subgroup_fitted'
summarize.subgroups(x, ...)

Arguments

x	a fitted object from fit.subgroup() or a matrix of covariate values
...	optional arguments to summarize.subgroups methods
subgroup	vector of indicators of same length as the number of rows in x if x is a matrix. A value of 1 in the ith position of subgroup indicates patient i is in the subgroup of patients recommended the treatment and a value of 0 in the ith position of subgroup indicates patient i is in the subgroup of patients recommended the control. If x is a fitted object returned by fit.subgroup(), subgroup is not needed.

Details

The p-values shown are raw p-values and are not adjusted for multiple comparisons.

See Also

fit.subgroup for function which fits subgroup identification models and print.subgroup_summary for arguments for printing options for summarize.subgroups().

Examples

library(personalized)

set.seed(123)
n.obs  <- 1000
n.vars <- 50
x <- matrix(rnorm(n.obs * n.vars, sd = 3), n.obs, n.vars)


# simulate non-randomized treatment
xbetat   <- 0.5 + 0.5 * x[,21] - 0.5 * x[,41]
trt.prob <- exp(xbetat) / (1 + exp(xbetat))
trt01    <- rbinom(n.obs, 1, prob = trt.prob)

trt      <- 2 * trt01 - 1

# simulate response
delta <- 2 * (0.5 + x[,2] - x[,3] - x[,11] + x[,1] * x[,12])
xbeta <- x[,1] + x[,11] - 2 * x[,12]^2 + x[,13]
xbeta <- xbeta + delta * trt

# continuous outcomes
y <- drop(xbeta) + rnorm(n.obs, sd = 2)

# create function for fitting propensity score model
prop.func <- function(x, trt)
{
    # fit propensity score model
    propens.model <- cv.glmnet(y = trt,
                               x = x, family = "binomial")
    pi.x <- predict(propens.model, s = "lambda.min",
                    newx = x, type = "response")[,1]
    pi.x
}

subgrp.model <- fit.subgroup(x = x, y = y,
                             trt = trt01,
                             propensity.func = prop.func,
                             loss   = "sq_loss_lasso",
                             nfolds = 5)    # option for cv.glmnet

comp <- summarize.subgroups(subgrp.model)
print(comp, p.value = 0.01)

# or we can simply supply the matrix x and the subgroups
comp2 <- summarize.subgroups(x, subgroup = 1 * (subgrp.model$benefit.scores > 0))

print(comp2, p.value = 0.01)

personalized documentation built on June 28, 2022, 1:06 a.m.