A suite of descriptive and inferential methods designed to evaluate one or more biomarkers for their ability to guide patient treatment recommendations. Package includes functions to assess the calibration of risk models; and plot, evaluate, and compare markers. Please see the reference Janes H, Brown MD, Huang Y, et al. (2014)
This R package includes a suite of descriptive and inferential methods designed to evaluate one or more biomarkers for their ability to guide patient treatment recommendations. relevant functions are:
trtsel_measuresfor evaluating the performance of a user-specified marker-based treatment rule
trtselfor creating trtsel objects
plot.trtselfor plotting risk curves and more
evaluate.trtselfor evaluating marker performance
calibrate.trtselfor assessing model calibration
compare.trtselto compare two trtsel objects.
To dowload the package from CRAN, type:
To download and install the most recent version of the package directly from github, type:
if (!require("devtools")) install.packages("devtools")devtools::install_github("TreatmentSelection", "mdbrown")
added plot.type 'selection impact'. see ?plot.trtsel for more information.
fixed a small bug that caused warnings when plotting a trtsel object created from providing fixed risk estimates.
eval.trtsel: modeled estimates of event rates calculated for subcohort designs had a bug. Calculations now are correct.
eval.trtsel: also fixed bug where event rates werent calculated when bootstraps = 0.
eval.trtsel: fixed event rate labeling bug that swapped trt all for trt none.
plot.trtsel: added new plot.type that plots event rate under different marker based treatment assignment rules vs. F_delta(v).
plot.trtsel: the mean trt effect line for subcohort designs was not weighted correctly. this has been fixed for continuous and discrete markers.
plot.trtsel: fixed a bug building plots for subcohort designs with discrete markers
fixed various small things so that the package now has no warnings from R CMD Check
added manual files for tsdata_scc and tsdata_cc
calibrate.trtsel: fixed bug when calibrating models fit using stratified case-control data.
Note: This is a large update with many changes. Although, the estimation methods remain the same, many of the functions inputs have changed. This is to accomodate multivariate models and survival outcomes.
compare.trtselnow implements a bootstrap bias-correction method.
Two key additions:
One minor additions: