Rangkuman RScript Meta-analysis

Berikut ini adalah code R dari kelas Meta-analysis di Risetku Academy. Code berikut bisa dijalankan melalui RStudio (Download & Install RStudio here). Silakan copy-paste di RScript RStudio Anda!

• Munculkan panel RScript dengan cara:
  • Windows: CTRL + SHIFT + N
  • Mac: Command ⌘ + Shift + N
• Eksekusi code per-baris dengan cara
  • Windows: CTRL + Enter
  • Mac: Command ⌘ + Return

Binary Data

#Install dan Load package terlebih dahulu
if (!require("remotes")) {
  install.packages("remotes")
}
remotes::install_github("MathiasHarrer/dmetar")
library(meta)

#Load data dikotom untuk meta analisis
#Disediakan oleh package dmetar
data(DepressionMortality)
View(DepressionMortality)


#Lakukan pooling effect untuk data binary / dikotom
m.bin <- metabin(event.e = event.e, 
                 n.e = n.e,
                 event.c = event.c,
                 n.c = n.c,
                 studlab = author,
                 data = DepressionMortality,
                 sm = "RR",
                 method = "MH",
                 fixed = TRUE,
                 random = TRUE,
                 method.tau = "PM",
                 title = "Depression and Mortality")

#Tampilkan hasil pooling size effect
summary(m.bin)

#Tampilkan forest plot 
forest(m.bin)

#Heterogeneity
#Subgroup Analysis tidak dilakukan (K per grup <2)
#Meta-regresi tidak dilakukan (K per grup <2)

#Robustness
#Melakukan leave-one-out meta
m.bin.leave <- metainf(m.bin)
forest(m.bin.leave)

#Melakukan cumulative meta
m.bin.cum <- metacum(m.bin,
                     sortvar = -n.c)
forest(m.bin.cum)

#Menampilkan Funnel plot
funnel(m.bin)

#Menampilkan Bias Publikasi
metabias(m.bin,
         method.bias = "Begg",
         plotit = FALSE)

#Menampilkan Contour-enhanced Funnel plot
col.contour = c("gray75", "gray85", "gray95")
legend(x = 0.1, y = 0.1, 
       legend = c("p < 0.1", "p < 0.05", "p < 0.01"),
       fill = col.contour)

funnel(m.bin,
          studlab = FALSE,
          xlim = c(0.05, 40),
          contour = c(0.9, 0.95, 0.99),
          col.contour = col.contour)

#Melakukan Trim-and-Fill analysis
m.bin.tf <- trimfill(m.bin)
summary(m.bin.tf)

#Menampilkan contour-enhanced Funnel plot + Trim-and-Fill analysis
funnel(m.bin.tf,
       studlab = FALSE,
       xlim = c(0.05, 40),
       contour = c(0.9, 0.95, 0.99),
       col.contour = col.contour)

Continous Data

#Install dan Load package terlebih dahulu
if (!require("remotes")) {
  install.packages("remotes")
}
remotes::install_github("MathiasHarrer/dmetar")
library(meta)

#Load data untuk kontinu
#Disediakan oleh package dmetar
data(SuicidePrevention)

#Lakukan pooling effect untuk data kontinu
m.cont <- metacont(n.e = n.e,
                   mean.e = mean.e,
                   sd.e = sd.e,
                   n.c = n.c,
                   mean.c = mean.c,
                   sd.c = sd.c,
                   studlab = author,
                   data = SuicidePrevention,
                   sm = "MD",
                   fixed = FALSE,
                   random = TRUE,
                   method.tau = "SJ",
                   title = "Suicide Prevention")

#Tampilkan hasil pooling size effect
summary(m.cont)

#Menampilkan forest plot 
forest(m.cont,
       layout = "RevMan5")

##### Heterogenitas #####
#Subgroup analysis berdasarkan risiko bias
update.meta(m.gen, 
            subgroup = age_group)

#Meta regresi
m.cont.reg <- metareg(m.cont, ~pubyear)
m.cont.reg

bubble(m.cont.reg)

##### Robustness #####
#Melakukan leave-one-out meta
m.cont.leave <- metainf(m.cont)
forest(m.cont.leave)

#Melakukan cumulative meta
m.cont.cum <- metacum(m.cont,
                     sortvar = pubyear)
forest(m.cont.cum)

#Menampilkan Funnel plot
funnel(m.cont)

#Menampilkan Bias Publikasi
metabias(m.cont,
         method.bias = "Egger",
         plotit = FALSE)
#Dibatalkan: K<10 

Generic Inverse Variance Data

#Install dan Load package terlebih dahulu
if (!require("remotes")) {
  install.packages("remotes")
}
remotes::install_github("MathiasHarrer/dmetar")
library(meta)

#Load data Generic IV untuk meta analisis
#Disediakan oleh package dmetar
data(ThirdWave)

#Lakukan pooling effect untuk data generic IV
#TE dalam SMD dan seTE adalah Standard error
m.gen <- metagen(TE = TE,
                 seTE = seTE,
                 studlab = Author,
                 data = ThirdWave,
                 sm = "SMD",
                 fixed = FALSE,
                 random = TRUE,
                 method.tau = "REML",
                 hakn = TRUE,
                 title = "Third Wave Psychotherapies")

#Tampilkan hasil pooling size effect
summary(m.gen)

#Tampilkan forest plot 
forest(m.gen)

##### Heterogenitas #####
#Subgroup analysis berdasarkan risiko bias
update.meta(m.gen, 
            subgroup = RiskOfBias)

#Meta regresi
#Meta-regresi data kategorik
m.gen.reg <- metareg(m.gen, ~RiskOfBias)
summary(m.gen.reg)
bubble(m.gen.reg, studlab = TRUE)

#Meta-regresi data kontinu
#Karena pada data tidak ada variabel kontinu, kita tambahkan tahun publikasi
year <- c(2014, 1998, 2010, 1999, 2005, 2014, 
          2019, 2010, 1982, 2020, 1978, 2001,
          2018, 2002, 2009, 2011, 2011, 2013)

m.gen.reg <- metareg(m.gen, ~year)
summary(m.gen.reg)
bubble(m.gen.reg, studlab = TRUE)

##### Robustness #####
#Melakukan leave-one-out meta
m.gen.leave <- metainf(m.gen)
forest(m.gen.leave)

#Melakukan cumulative meta
m.gen.cum <- metacum(m.gen,
                     sortvar = year)
forest(m.gen.cum)

##### Publication bias #####
#Menampilkan Funnel plot
funnel(m.gen)

#Menampilkan Bias Publikasi
metabias(m.gen,
         method.bias = "Egger",
         plotit = TRUE)

#Menampilkan Contour-enhanced Funnel plot
col.contour = c("gray75", "gray85", "gray95")
legend(x = 0.1, y = 0.1, 
       legend = c("p < 0.1", "p < 0.05", "p < 0.01"),
       fill = col.contour)

funnel(m.gen,
       studlab = FALSE,
       contour = c(0.9, 0.95, 0.99),
       col.contour = col.contour)

#Melakukan Trim-and-Fill analysis
m.gen.tf <- trimfill(m.gen)
summary(m.gen.tf)

#Menampilkan contour-enhanced Funnel plot + Trim-and-Fill analysis
funnel(m.gen.tf,
       studlab = FALSE,
       contour = c(0.9, 0.95, 0.99),
       col.contour = col.contour)