UpdateQMediation: UpdateQMediation

In podTockom/medltmle: Longitudinal Mediation Analysis with Time-varying Mediators

UpdateQMediation

Description

Update the initial fit of Q using clever covariates.

Usage

UpdateQMediation(
  Qstar.kplus1,
  logitQ,
  combined.summary.measures,
  cum.g,
  cum.q.ratio,
  working.msm,
  uncensored,
  intervention.match,
  is.deterministic,
  msm.weights,
  gcomp,
  observation.weights
)

Arguments

Qstar.kplus1	Estimate of the expectation with respect to the distribution of the node one ahead of the current node given the past (dimension: n x num.regimes).
logitQ	logit predicted values of Q for the current node, as estimated by Estimate (dimension: n x num.regimes).
combined.summary.measures	TO DO (dimension: n x num.measures x num.regimes, where num.measures=num.summary.measures + num.baseline.covariates).
cum.g	Cumulative g estimate up to the most recent AC node. (dimension: n x num.regimes x num.measures)
cum.q.ratio	Cumulative Q estimate ratio of following abar.prime regime and following abar regime. (dimension: n x num.regimes x num.measures)
working.msm	Working marginal structural model.
uncensored	Uncensored samples.
intervention.match	Samples with exposure that matches intervention (dimension: n x num.regimes).
is.deterministic	Logical variable indicating samples deterministic due to death or deterministic Q.
msm.weights	Marginal structural model weights (dimension: n x num.regimes).
gcomp	Logical indicating whether to use Gcomp instead (no updating if TRUE).
observation.weights	Sample weights.

Value

Returns updated estimate of Q, unless gcomp=TRUE or no samples left for estimation purposes. The output also includes the offset (initial estimate of Q for the current node), fluctuation model, and the clever covariate.

podTockom/medltmle documentation built on Aug. 9, 2022, 9:32 a.m.