Estimate NDE/NIE for BKMR(plus TE)

Estimate NDE/NIE for BKMR(plus TE)

mediation.bkmr(
  a,
  astar,
  e.m = NULL,
  e.y,
  fit.m = NULL,
  fit.y = NULL,
  fit.y.TE = NULL,
  X.predict.M = NULL,
  X.predict.Y = NULL,
  effects = "all",
  m.quant = c(0.1, 0.5, 0.75),
  m.value = NULL,
  alpha = 0.05,
  sel,
  seed,
  K
)

Arguments

a

exposure variables at current level

astar

exposure variables at counterfactual level

e.m

effect modifier for the mediator variable

e.y

effect modifier for the outcome variable

fit.m

model fit regressing mediator on exposures and confounders on mediator

fit.y

model fit regressing outcome on exposures, effect modifiers, mediator and confounders on outcome

fit.y.TE

total effect model fit regressing outcome on exposures, effect modifiers and confounders on outcome

X.predict.M

counfounders for mediator

X.predict.Y

counfounders for outcome

effects

type(s) of effects that users want to output

m.quant

values of the quantile that the mediator is set to

m.value

values that the mediator is set to

alpha

1-confidence interval

sel

a vector selecting which iterations of the fit should be retained or inference

seed

the random seed to use to evaluate the code

K

number of samples to generate for each MCMC iteration in YaMastar calculation

Value

A list contaning the sample prediction for TE, NDE, NIE and their summary statistics

Details

For guided examples, go to https://zc2326.github.io/causalbkmr/articles/BKMRCMA_QuickStart.html

Examples

if (FALSE) {
library(causalbkmr)
dat <-  cma_sampledata(N=300, L=3, P=3, scenario=1, seed=7)

A <- cbind(dat$z1, dat$z2, dat$z3)
X <- cbind(dat$x3)
y  <- dat$y
m  <- dat$M

E.M <- NULL
E.Y <- dat$x2

Z.M <- cbind(A,E.M)
Z.Y <- cbind(A, E.Y)
Zm.Y <- cbind(Z.Y, m)

set.seed(1)
fit.y <- kmbayes(y=y, Z=Zm.Y, X=X, iter=5000, verbose=TRUE, varsel=FALSE)
#save(fit.y,file="bkmr_y.RData")

set.seed(2)
fit.y.TE <- kmbayes(y=y, Z=Z.Y, X=X, iter=5000, verbose=TRUE, varsel=FALSE)
#save(fit.y.TE,file="bkmr_y_TE.RData")

set.seed(3)
fit.m <- kmbayes(y=m, Z=Z.M, X=X, iter=5000, verbose=TRUE, varsel=FALSE)
#save(fit.m,file="bkmr_m.RData")

X.predict <- matrix(colMeans(X),nrow=1)
astar <- c(apply(A, 2, quantile, probs=0.25))
a <- c(apply(A, 2, quantile, probs=0.75))

e.y10 = quantile(E.Y, probs=0.1)
e.y90 = quantile(E.Y, probs=0.9)

sel<-seq(2500,5000,by=5)

#' mediationeffects.ey10  <- mediation.bkmr(a=a, astar=astar, e.y = e.y10, fit.m=fit.m, fit.y=fit.y, fit.y.TE=fit.y.TE, X.predict.M=X.predict, X.predict.Y=X.predict, alpha=0.05, sel=sel, seed=22, K=10)
mediationeffects.ey10$est
plotdf <- as.data.frame(mediationeffects.ey10$est)
plotdf["Effect"] <- rownames(plotdf)
ggplot(plotdf, aes(Effect, mean, ymin = lower, ymax = upper ))  +
  geom_pointrange(position = position_dodge(width = 0.75))  +  coord_flip()

}