mvshape fits a multivariate meta-analysis for groups of the exposure (e.g. deciles).
Usage
mvshape(
y = y,
x = x,
covar = NULL,
study = NULL,
ngrp = 10,
refgrp = 1,
family = "gaussian",
float = FALSE,
method = "reml"
)Arguments
- y
outcome.
- x
exposure.
- covar
data.frame with covariates.
- study
study variable.
- ngrp
number of quantiles of the exposure.
- refgrp
reference group.
- family
the glm family (options: gaussian and binomial).
- float
floating point variances.
- method
meta-analysis method.
Value
List of multivariate meta-analysis results for each group.
- results
data.frame of results: q is the quantile group, xbeta is the mean of x in each quantile, xse is the standard error of the mean of x in each quantile, beta is the regression coefficient of association between y and each quantile of x, se is the standard error of the regression coefficient of association between y and each quantile of x
- varcor
variance-covariance matrix
- xmin
miniumum value of the exposure
- xmax
maximum value of the exposure
- family
family used in the analysis
Author
James Staley jrstaley95@gmail.com
Examples
# Data
y <- rnorm(5000)
x <- rnorm(5000, 10, 1)
c1 <- rbinom(5000, 1, 0.5)
c2 <- rnorm(5000)
study <- c(rep("study1", 1000), rep("study2", 1000), rep("study3", 1000), rep("study4", 1000), rep("study5", 1000))
covar <- data.frame(c1 = c1, c2 = c2, study = study)
# Analyses
res <- mvshape(y = y, x = x, covar = covar[, c("c1", "c2")], study = study, family = "gaussian")