metaLong

Longitudinal Meta-Analysis with Robust Variance Estimation and Sensitivity Analysis

Installation

# From CRAN (once released)
install.packages("metaLong")

# Development version from GitHub
remotes::install_github("causalfragility-lab/metaLong")

Overview

metaLong provides tools for synthesising evidence from studies that report outcomes at multiple follow-up time points. Core functions:

Function	Purpose
ml_meta()	Pool effects at each time point with RVE + Tipton correction
ml_sens()	Time-varying ITCV sensitivity analysis
ml_benchmark()	Benchmark ITCV against observed covariates
ml_spline()	Nonlinear spline time trend
ml_fragility()	Leave-k-out fragility analysis
ml_plot()	Combined publication figure
sim_longitudinal_meta()	Simulate longitudinal meta-analytic data

Basic usage

library(metaLong)

dat  <- sim_longitudinal_meta(k = 20, times = c(0, 6, 12, 24), seed = 42)
meta <- ml_meta(dat, yi = "yi", vi = "vi", study = "study", time = "time")
sens <- ml_sens(dat, meta, yi = "yi", vi = "vi", study = "study", time = "time")
ml_plot(meta, sens_obj = sens)

References

• Frank (2000) doi:10.1177/0049124100029002003
• Hedges, Tipton & Johnson (2010) doi:10.1002/jrsm.5
• Tipton (2015) doi:10.1037/met0000011

Author

Subir Hait, Michigan State University (haitsubi@msu.edu) ORCID: 0009-0004-9871-9677