prioritize_covariates: prioritize_covariates

In lendle/hdps: High-dimensional propensity score algorithm

Description

Returns the indexes of columns covars ordered by the possible amount of counfounding each column could adjust for.

Usage

prioritize_covariates(outcome, treatment, covars, return_bias = FALSE,
  keep_NaNs = FALSE)

Arguments

outcome	binary vector of outcomes
treatment	binary vector of treatments
covars	matrix or data.frame of binary covariates.
return_bias	Should the calculated multiplitive biases be returned with the ordered indexes? Defaults to FALSE
keep_NaNs	If the calculated multiplicitive bias for a covariate is NaN, should its index be included in the returned vector? Defaults to FALSE

Details

For each covariate in covars, the potential multiplicitive bias is calculated as described in Schneeweiss et al. (2009). The column indexes are then put in descending order of the absolute value of the log of the multiplicitive bias.

If return_bias==TRUE, the returned vector of indexes includes the multiplicitive biases sorted in the same order and stored in an attribute called "bias_m".

If outcome has no variation for a particular value of a covariate, then the multiplicitive bias is calculated as NaN. If keep_NaNs==FALSE, then the column indexes of such covariates are not included in the returned vector.

Value

Vector of column indexes of covars.

Author(s)

Sam Lendle

References

Schneeweiss, S., Rassen, J. A., Glynn, R. J., Avorn, J., Mogun, H., & Brookhart, M. A. (2009). High-dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology (Cambridge, Mass.), 20(4), 512.

lendle/hdps documentation built on May 9, 2019, 8:34 a.m.