Marginal structural models

Causal models fitted in the pseudo-population created by stabilized inverse-probability-of-treatment weights (Robins, Hernan and Brumback, 2000), so their coefficients are causal rather than confounder-adjusted associations.

Usage

morie_msm_linear(y, treatment_history, weights = NULL) 

morie_msm_logistic(y, treatment_history, weights = NULL) 

morie_msm_poisson(y, treatment_history, offset = NULL, weights = NULL) 

morie_msm_negative_binomial(y, treatment_history, alpha = 1, offset = NULL,
  weights = NULL) 

morie_msm_cox_marginal(time, event, treatment_history, weights = NULL) 

morie_msm_accelerated_failure(time, event, treatment_history, weights = NULL) 

morie_msm_gmm_estimator(y, treatment_history, instruments = NULL, weights = NULL) 

morie_msm_time_varying_exposure(y, exposure_history, weights = NULL)

Arguments

y

outcome vector

treatment_history

matrix of treatments, one row per subject

exposure_history

matrix of exposures, one row per subject

time

observed follow-up times

event

event indicator, 1 for an event and 0 for censored

weights

stabilized inverse-probability-of-treatment weights

offset

optional log-exposure offset

alpha

negative-binomial overdispersion parameter

instruments

optional instrument matrix for the GMM estimator

Value

A list holding the fitted quantities for the method called: point estimates, fitted values and any diagnostics it produces. See the individual method descriptions above for the components each returns.

Details

These functions mirror, in R, the implementations in the Python package morie; the two are checked against each other by the parity tests in tests/testthat/.

Examples

set.seed(1)
A <- matrix(rbinom(60, 1, 0.5), 30, 2)
y <- 1 + 0.8 * rowMeans(A) + rnorm(30, 0, 0.3)
morie_msm_linear(y, A)
#> $estimate
#> [1] 0.3538458
#> 
#> $beta
#> [1] 1.0859866 0.3538458
#> 
#> $beta_a
#> [1] 0.3538458
#> 
#> $a_bar
#>  [1] 0 1 1 1 1 2 2 1 2 0 1 1 2 1 2 1 1 1 1 2 1 1 1 0 0 0 0 1 2 0
#> 
#> $fitted
#>  [1] 1.085987 1.439832 1.439832 1.439832 1.439832 1.793678 1.793678 1.439832
#>  [9] 1.793678 1.085987 1.439832 1.439832 1.793678 1.439832 1.793678 1.439832
#> [17] 1.439832 1.439832 1.439832 1.793678 1.439832 1.439832 1.439832 1.085987
#> [25] 1.085987 1.085987 1.085987 1.439832 1.793678 1.085987
#>