# Meta-analysis with SMDs
  # FROM: https://wviechtb.github.io/metadat/reference/dat.normand1999.html 

library(metafor)
# copy data into 'data' and examine data
data <- dat.normand1999
data

# using the escalc() function, calculate mean differences and corresponding sampling variances
data <- escalc(measure="MD", m1i=m1i, sd1i=sd1i, n1i=n1i, m2i=m2i, sd2i=sd2i, n2i=n2i, data=data, , slab=source)
data 	# last two columns are the calculated yi and vi values  

# meta-analysis of mean differences using a random-effects model
result <- rma(yi, vi, data=data)
result

# using the escalc() function, calculate standardised mean differences (SMD) and corresponding sampling variances
data <- escalc(measure="SMD", m1i=m1i, sd1i=sd1i, n1i=n1i, m2i=m2i, sd2i=sd2i, n2i=n2i, data=data, , slab=source)
data 	# last two columns are the calculated yi and vi values  

# meta-analysis of mean differences using a random-effects model
result <- rma(yi, vi, data=data)
result

# plots
forest(result)
funnel(result)

###########################

# Meta-analysis with risk ratios
  # FROM: https://wviechtb.github.io/metadat/reference/dat.bcg.html 

# copy data into 'data' and examine data
data <- dat.bcg
data

# using the escalc() function, calculate log risk ratios and corresponding sampling variances
  # same with odds ratios (OR) by changing measure to "OR")
data <- escalc(measure="RR", ai=tpos, bi=tneg, ci=cpos, di=cneg, data=data, slab=author)
data 	# last two columns are the calculated yi and vi values

# meta-analysis of risk reatios using a random-effects model
result <- rma(yi, vi, data=data)
result

# average risk ratio with 95% CI
predict(result, transf=exp)

# plots
forest(result)
funnel(result)

###########################

# Meta-analysis with correlation coefficients
  # FROM: https://wviechtb.github.io/metadat/reference/dat.mcdaniel1994.html 

# copy data into 'data' and examine data
data <- dat.mcdaniel1994 
data

# calculate r-to-z transformed correlations and corresponding sampling variances
data <- escalc(measure="ZCOR", ri=ri, ni=ni, data=data, slab=study)

# meta-analysis of the transformed correlations using a random-effects model
result <- rma(yi, vi, data=data)
result

# plots
forest(result)
funnel(result)

########################### 

# A meta-analysis using hazard ratios and 95% confidence intervals
  # Data from Steurer et al. (2006)
  # https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD004270.pub2/full 

# Create vectors for the hazard ratios and 95% CI lower and upper limits from each study
study <- c("FCG on CLL 1996", "Leporrier 2001", "Rai 2000", "Robak 2000")
HR <- c(0.55, 0.92, 0.79, 1.18)
lower.HR <- c(0.28, 0.79, 0.59, 0.64)
upper.HR <- c(1.09, 1.08, 1.05, 2.17)
data <- cbind(study, HR, lower.HR, upper.HR)
data <- as.data.frame(data)
data$HR=as.numeric(data$HR); data$lower.HR=as.numeric(data$lower.HR); data$upper.HR=as.numeric(data$upper.HR)

# Calculate yi and vi from the HR and 95% CI values entered into the data frame created above
data$yi = log(data$HR)
data$vi = ((log(upper.HR) - log(lower.HR))/3.92)^2
data

library(metafor)
# Run meta-analysis:
result <- rma.uni(yi = data$yi, vi = data$vi, slab = study)	
result

# Generate plots based on the R object 'result'
plot(result, addpred = TRUE, showweights = TRUE, header = TRUE, transf = exp, qqplot = TRUE) 
  # or: forest and funnel plots can be generated separately:
forest(result, addpred = TRUE, order = "obs", showweights = TRUE, header = TRUE, transf = exp)	
					# ordered by observed effect sizes of included studies
funnel(result)
funnel(result, level = c(90, 95, 99), shade = c("white", "gray55", "gray75"), refline = 0, legend = TRUE) 		
					# contour-enhanced funnel plot

# trimfill method for assessing publication bias
trimfill(result) 			

# Statistical assessment of publication bias (funnel plot asymmetry):
ranktest(result)
regtest(result)

tes(result)				# test of excess significance "tes"

install.packages("numDeriv")
library(numDeriv)
sel <- selmodel(result, type="power")		# fitting selection models (selmodel) to identify the model of 
								# potential publication bias in a meta-analysis
sel								# displays the selection model test result