In martinlaw/dcv: Meta-Analysis on Postoperative Pneumonia

title: "Subgroup analysis" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{my-vignette} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8}

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(metafor)
initial.data <- read.csv(file="../data/secondary-outcomes.csv")
head(initial.data)
dat <- escalc(measure="MD", data=initial.data, m1i=m1i, sd1i=sd1i, n1i=n1i, m2i=m2i, sd2i=sd2i, n2i=n2i, slab=study)

Subgroup analysis

Data unclear:

Multiple/range of values given for tidal volume
No PEEP values greater than zero (dichotomisation is at zero)
Li et al. results are split in two in the primary outcome, but not clear if tidal volume etc. are the same in both sets of results.

Tidal volume (>6 vs <=6mL/kg)

PEEP (<0 vs 0)

Anesthesia type (TIVA vs inhaled)

anes.tiva <- rma(yi, vi, data=dat, subset=(anesthesia=="tiva" & outcome1=="pO2"))
anes.inhaled <- rma(yi, vi, data=dat, subset=(anesthesia=="inhaled" & outcome1=="pO2"))
anes.comp <- data.frame(estimate = c(coef(anes.tiva), coef(anes.inhaled)), stderror = c(anes.tiva$se, anes.inhaled$se),
                       meta = c("tiva","inhaled"), tau2 = round(c(anes.tiva$tau2, anes.inhaled$tau2),3))
rma(estimate, sei=stderror, mods = ~ meta, method="FE", data=anes.comp, digits=3)