In bmckay/ceaR: Cost-Effectiveness Analysis Tools for R

The package ceaR is designed to facilitate cost-effectiveness analyses of individual-level data. In most cases, the interested user will have a dataset of individuals that fall into one of two (or more) treatment or intervention conditions. At a minimum, each individual should have a cost, effect, and an indicator of the intervention or condition of which he or she is a member. There may also be one or more covariate values for each individual including baseline costs or effects, demographic, environmental, or disease-specific factors.

Setup

The cea_setup function is used to define the variables representing costs, effects, and intervention condition as well as any covariate variables that should be included in analyses and whether these covariates are applied to both the cost and effect analyses or just one or the other. Further, an optional setting eff_more_better can be set to False in cases where more of the effect is not a better outcome for an individual. An example would be if the effect is a binary variable where a 1 represents death or disease.

As an example, the dataset clintrial_cea will be used... The formula method can be used where the baseline costs (blcost) and QALYs (blqaly) will be covariates with cost and qaly, respectively.

library(ceaR)
testmodel <- cea_setup(cost | qaly ~ blcost | blqaly, intv = "treat", 
                       clintrial_cea)

The object testmodel is of class ceamodel and includes a data frame including only the variables from clintrial_cea including in the function call. There are also references to the cost, effect, intervention, and covariate variables; cost and effect formulas, and the number of individuals in the dataset.

Incremental CEA

To conduct an incremental cost-effectiveness analysis, the next step is to pass the testmodel object to the ceamodel_incremental function. This function will estimate delta_e and delta_c between conditions. The analyses are conducted based on a regression model for both costs and effects. When covariates are not included in the model, the results are equivalent to taking the differences in the mean costs (and effects) by condition across individuals in the sample. The results of this analysis are attached to the object of class ceamodel that is passed to the function.

testmodel <- ceamodel_incremental(testmodel)

bmckay/ceaR documentation built on May 23, 2019, 9:01 p.m.