acoarm: A function to estimate parameters in Cox proportional hazard...

In TwoPhaseInd: Estimate Gene-Treatment Interaction Exploiting Randomization

A function to estimate parameters in Cox proportional hazard models by augmented case-only designs for randomized clinical trials with failure time endpoints.

Description

This function estimates parameters of proportional hazards models with gene-treatment interactions. It employs classical case-cohort estimation methods, incorporating the case-only estimators. The method was published in Dai et al. (2016) Biometrics.

Usage

acoarm(data, svtime, event, treatment, BaselineMarker,subcohort, esttype = 1, 
augment = 1, weight=NULL, extra = NULL)

Arguments

data	A data frame used to access the following data.
svtime	A character string of column name, corresponds to one column of the data frame, which is used to store the failure time variable (numeric).
event	A character string of column name, corresponds to one column of the data frame, which is used to store the indicator of failure event (1: failure, 0: not failure).
treatment	A character string of column name, corresponds to one column of the data frame, which is used to store the binary vector of treatment variable (1: treatment, 0: placebo).
BaselineMarker	A character string of column name, corresponds to one column of the data frame, which is used to store a vector of baseline biomarker that is under investigation for interaction with treatment. The BaselineMarker variable is missing for those who are not sampled in the case-cohort.
subcohort	A character string of column name, corresponds to one column of the data frame, which is used to store the indicator of sub-cohort (1: sample belong to the sub-cohort, 0: not belong to the sub-cohort)
esttype	The option of estimation methods (1: Self-Prentice estimator, 0: Lin-Ying estimator).
augment	The indicator of whether subcohort was drawn from the placebo arm (augment=0), from the active treatment arm (augment=1), or from both arms (augment=2).
weight	If the genotype data are obtained through case-control sampling, weight is a vector of sampling weights (inverse of sampling probability) corresponding to rows of data. If the genotype data are obtained through case-cohort sampling, weight is NULL. If a vector of weights have been supplied by user, then esttype is automatically set to 0: Lin-Ying estimator.
extra	A string vector of column name(s), corresponds to more or more column(s) of the data frame, which is/are used to store the extra baseline covariate(s) to be adjusted for in addition to treatment and biomarker.

Details

The function returns point estimates and standard error estimates of parameters in the proportional hazards model. The method was published in Dai et al. (2015) Biometrics.

Value

beta	Estimated parameter
stder	Estimated standard error of parameter estimates
pVal	p value

Author(s)

James Y. Dai

References

J. Y. Dai, X. C. Zhang,C. Y. Wang, and C. Kooperberg. Augmented case-only designs for randomized clinical trials with failure time endpoints. Biometrics, DOI: 10.1111/biom.12392, 2016.

Examples

## Load the example data
data(acodata)
## ACO in placebo arm
rfit0 <- acoarm(data=acodata,
                 svtime="vacc1_evinf",
                 event="f_evinf",
                 treatment="f_treat",
                 BaselineMarker="fcgr2a.3",
                 subcohort="subcoh",
                 esttype=1,
                 augment=0,
                 weight=NULL,
                 extra=c("f_agele30","f_hsv_2","f_ad5gt18","f_crcm",
                 "any_drug","num_male_part_cat","uias","uras")) 
rfit0

## ACO in active arm
rfit1 <- acoarm(data=acodata,
                 svtime="vacc1_evinf",
                 event="f_evinf",
                 treatment="f_treat",
                 BaselineMarker="fcgr2a.3",
                 subcohort="subcoh",
                 esttype=1,
                 augment=1,
                 weight=NULL,
                 extra=c("f_agele30","f_hsv_2","f_ad5gt18","f_crcm",
                 "any_drug","num_male_part_cat","uias","uras")) 
rfit1

## ACO in both arms
rfit2 <- acoarm(data=acodata,
                 svtime="vacc1_evinf",
                 event="f_evinf",
                 treatment="f_treat",
                 BaselineMarker="fcgr2a.3",
                 subcohort="subcoh",
                 esttype=1,
                 augment=2,
                 weight=NULL,
                 extra=c("f_agele30","f_hsv_2","f_ad5gt18","f_crcm",
                 "any_drug","num_male_part_cat","uias","uras")) 
rfit2

TwoPhaseInd documentation built on March 18, 2022, 7:52 p.m.