Meta of Pediatric Aneurysmal Subarachnoid Hemorrhage
Niels Pacheco
2024-05-20
——–1. Meta analysis of proportions, by subgroup, of those who obtained good outcomes after treament (subgroup) ———
metaprop1_PAH <- read.table(text = "
Study_ID Sample subgroup suc_treat complications
'Mbaye 2021' 24 'Microsurgery' 10 14
'Skoch 2017-S' 9 'Surgery' 9 0
'Skoch 2017-E' 7 'Endovascular' 7 0
'Slator 2019-S' 4 'Surgery' 4 0
'Slator 2019-E' 16 'Endovascular' 14 2
'Aryan 2006' 50 'Endovascular+surgery' 44 10
'Stiefel 2008' 12 'Microsurgery+Endovascular' 11 1
'Kakarla 2010' 48 'Surgery' 19 29
'Mehrotra 2012' 63 'Surgery' 44 19
'Deora 2016' 44 'Microsurgery' 31 13
'Vaid 2008' 27 'Surgery' 21 6
'Herman 1991' 16 'Surgery' 11 5
'Storrs 1982' 29 'Surgery' 16 13
'Alawi 2014-S' 200 'Surgery' 187 13
'Alawi 2014-E' 920 'Endovascular' 905 15
", header = TRUE, stringsAsFactors = FALSE)
print(metaprop1_PAH)## Study_ID Sample subgroup suc_treat complications
## 1 Mbaye 2021 24 Microsurgery 10 14
## 2 Skoch 2017-S 9 Surgery 9 0
## 3 Skoch 2017-E 7 Endovascular 7 0
## 4 Slator 2019-S 4 Surgery 4 0
## 5 Slator 2019-E 16 Endovascular 14 2
## 6 Aryan 2006 50 Endovascular+surgery 44 10
## 7 Stiefel 2008 12 Microsurgery+Endovascular 11 1
## 8 Kakarla 2010 48 Surgery 19 29
## 9 Mehrotra 2012 63 Surgery 44 19
## 10 Deora 2016 44 Microsurgery 31 13
## 11 Vaid 2008 27 Surgery 21 6
## 12 Herman 1991 16 Surgery 11 5
## 13 Storrs 1982 29 Surgery 16 13
## 14 Alawi 2014-S 200 Surgery 187 13
## 15 Alawi 2014-E 920 Endovascular 905 15# Install the meta package if not already installed
if (!require(meta)) {
install.packages("meta")
library(meta)
}## Loading required package: meta## Loading required package: metadat## Loading 'meta' package (version 7.0-0).
## Type 'help(meta)' for a brief overview.
## Readers of 'Meta-Analysis with R (Use R!)' should install
## older version of 'meta' package: https://tinyurl.com/dt4y5drs# Assuming 'df' is your DataFrame loaded as previously shown
# Meta-analysis of proportions with study labels
# Calculate the pooled proportion using metaprop function, labeling each study by the First Author
meta_analysis1 <- metaprop(event = suc_treat, n = Sample, data = metaprop1_PAH,
sm = "PLO", method.tau = "DL",
prediction = FALSE, comb.fixed = FALSE,
comb.random = TRUE, studlab = metaprop1_PAH$Study_ID,
byvar = subgroup
)
# Summary of the meta-analysis
summary(meta_analysis1)## proportion 95%-CI %W(random) subgroup
## Mbaye 2021 0.4167 [0.2211; 0.6336] 7.5 Microsurgery
## Skoch 2017-S 1.0000 [0.6637; 1.0000] 4.2 Surgery
## Skoch 2017-E 1.0000 [0.5904; 1.0000] 4.2 Endovascular
## Slator 2019-S 1.0000 [0.3976; 1.0000] 4.1 Surgery
## Slator 2019-E 0.8750 [0.6165; 0.9845] 6.4 Endovascular
## Aryan 2006 0.8800 [0.7569; 0.9547] 7.4 Endovascular+surgery
## Stiefel 2008 0.9167 [0.6152; 0.9979] 5.5 Microsurgery+Endovascular
## Kakarla 2010 0.3958 [0.2577; 0.5473] 7.7 Surgery
## Mehrotra 2012 0.6984 [0.5698; 0.8077] 7.8 Surgery
## Deora 2016 0.7045 [0.5480; 0.8324] 7.6 Microsurgery
## Vaid 2008 0.7778 [0.5774; 0.9138] 7.3 Surgery
## Herman 1991 0.6875 [0.4134; 0.8898] 7.1 Surgery
## Storrs 1982 0.5517 [0.3569; 0.7355] 7.6 Surgery
## Alawi 2014-S 0.9350 [0.8914; 0.9649] 7.7 Surgery
## Alawi 2014-E 0.9837 [0.9733; 0.9908] 7.8 Endovascular
##
## Number of studies: k = 15
## Number of observations: o = 1469
## Number of events: e = 1333
##
## proportion 95%-CI
## Random effects model 0.8170 [0.6557; 0.9128]
##
## Quantifying heterogeneity:
## tau^2 = 2.3618 [0.8052; 5.2929]; tau = 1.5368 [0.8973; 2.3006]
## I^2 = 93.2% [90.4%; 95.2%]; H = 3.84 [3.22; 4.57]
##
## Test of heterogeneity:
## Q d.f. p-value
## 206.04 14 < 0.0001
##
## Results for subgroups (random effects model):
## k proportion 95%-CI tau^2
## subgroup = Microsurgery 2 0.5725 [0.2915; 0.8134] 0.5863
## subgroup = Surgery 8 0.7502 [0.5464; 0.8822] 1.3432
## subgroup = Endovascular 3 0.9561 [0.8067; 0.9913] 1.4859
## subgroup = Endovascular+surgery 1 0.8800 [0.7576; 0.9451] --
## subgroup = Microsurgery+Endovascular 1 0.9167 [0.5868; 0.9884] --
## tau Q I^2
## subgroup = Microsurgery 0.7657 5.18 80.7%
## subgroup = Surgery 1.1590 64.65 89.2%
## subgroup = Endovascular 1.2190 7.88 74.6%
## subgroup = Endovascular+surgery -- 0.00 --
## subgroup = Microsurgery+Endovascular -- 0.00 --
##
## Test for subgroup differences (random effects model):
## Q d.f. p-value
## Between groups 10.28 4 0.0359
##
## Details on meta-analytical method:
## - Inverse variance method
## - DerSimonian-Laird estimator for tau^2
## - Jackson method for confidence interval of tau^2 and tau
## - Logit transformation
## - Clopper-Pearson confidence interval for individual studies
## - Continuity correction of 0.5 in studies with zero cell frequencies# To visualize the results, you can plot a forest plot
meta::forest(meta_analysis1, layout = "JAMA")
1. Meta analysis of proportions, by subgroup, of those who obtained good outcomes after treament (subgroup)
——–2. Meta analysis of proportions, by subgroup, of those who obtained complications after treament (subgroup) ———
Same data as the previous one
# Calculate the pooled proportion using metaprop function, labeling each study by the First Author
meta_analysis2 <- metaprop(event = complications, n = Sample, data = metaprop1_PAH,
sm = "PLO", method.tau = "DL",
prediction = FALSE, comb.fixed = FALSE,
comb.random = TRUE, studlab = metaprop1_PAH$Study_ID,
byvar = metaprop1_PAH$subgroup
)
# Summary of the meta-analysis
summary(meta_analysis2)## proportion 95%-CI %W(random) subgroup
## Mbaye 2021 0.5833 [0.3664; 0.7789] 7.5 Microsurgery
## Skoch 2017-S 0.0000 [0.0000; 0.3363] 4.2 Surgery
## Skoch 2017-E 0.0000 [0.0000; 0.4096] 4.1 Endovascular
## Slator 2019-S 0.0000 [0.0000; 0.6024] 4.1 Surgery
## Slator 2019-E 0.1250 [0.0155; 0.3835] 6.4 Endovascular
## Aryan 2006 0.2000 [0.1003; 0.3372] 7.6 Endovascular+surgery
## Stiefel 2008 0.0833 [0.0021; 0.3848] 5.4 Microsurgery+Endovascular
## Kakarla 2010 0.6042 [0.4527; 0.7423] 7.7 Surgery
## Mehrotra 2012 0.3016 [0.1923; 0.4302] 7.8 Surgery
## Deora 2016 0.2955 [0.1676; 0.4520] 7.7 Microsurgery
## Vaid 2008 0.2222 [0.0862; 0.4226] 7.3 Surgery
## Herman 1991 0.3125 [0.1102; 0.5866] 7.1 Surgery
## Storrs 1982 0.4483 [0.2645; 0.6431] 7.6 Surgery
## Alawi 2014-S 0.0650 [0.0351; 0.1086] 7.7 Surgery
## Alawi 2014-E 0.0163 [0.0092; 0.0267] 7.8 Endovascular
##
## Number of studies: k = 15
## Number of observations: o = 1469
## Number of events: e = 140
##
## proportion 95%-CI
## Random effects model 0.1903 [0.0924; 0.3517]
##
## Quantifying heterogeneity:
## tau^2 = 2.2707 [0.7831; 5.1483]; tau = 1.5069 [0.8850; 2.2690]
## I^2 = 93.2% [90.3%; 95.2%]; H = 3.82 [3.21; 4.55]
##
## Test of heterogeneity:
## Q d.f. p-value
## 204.59 14 < 0.0001
##
## Results for subgroups (random effects model):
## k proportion 95%-CI tau^2
## subgroup = Microsurgery 2 0.4275 [0.1866; 0.7085] 0.5863
## subgroup = Surgery 8 0.2498 [0.1178; 0.4536] 1.3432
## subgroup = Endovascular 3 0.0439 [0.0087; 0.1933] 1.4859
## subgroup = Endovascular+surgery 1 0.2000 [0.1111; 0.3333] --
## subgroup = Microsurgery+Endovascular 1 0.0833 [0.0116; 0.4132] --
## tau Q I^2
## subgroup = Microsurgery 0.7657 5.18 80.7%
## subgroup = Surgery 1.1590 64.65 89.2%
## subgroup = Endovascular 1.2190 7.88 74.6%
## subgroup = Endovascular+surgery -- 0.00 --
## subgroup = Microsurgery+Endovascular -- 0.00 --
##
## Test for subgroup differences (random effects model):
## Q d.f. p-value
## Between groups 8.60 4 0.0719
##
## Details on meta-analytical method:
## - Inverse variance method
## - DerSimonian-Laird estimator for tau^2
## - Jackson method for confidence interval of tau^2 and tau
## - Logit transformation
## - Clopper-Pearson confidence interval for individual studies
## - Continuity correction of 0.5 in studies with zero cell frequencies# To visualize the results, you can plot a forest plot
meta::forest(meta_analysis2, layout = "JAMA")
——–3. Meta analysis of odds ratios for risk of successful treatment surgery vs endovascular ———
data_comparison <- read.table(text = "
Study_ID sample_surg suc_surg comp_surg sample_endo suc_end comp_end
'Skoch, 2017' 9 9 0 7 7 0
'Slator, 2019' 4 4 0 16 14 2
'Alawi, 2014' 200 187 13 920 905 15
", header = TRUE, stringsAsFactors = FALSE)
print(data_comparison)## Study_ID sample_surg suc_surg comp_surg sample_endo suc_end comp_end
## 1 Skoch, 2017 9 9 0 7 7 0
## 2 Slator, 2019 4 4 0 16 14 2
## 3 Alawi, 2014 200 187 13 920 905 15# Load the necessary library
library(meta)
# Assuming mrs_odds is already loaded with your data
# You might need to check the names of your columns with names(mrs_odds)
# Meta-analysis of odds ratios
meta_analysis3 <- metabin(
event.e = data_comparison$suc_end,
n.e = data_comparison$sample_endo,
event.c = data_comparison$suc_surg,
n.c = data_comparison$sample_surg,
data = data_comparison,
studlab = paste(data_comparison$Study_ID),
sm = "OR", # Specify summary measure as Odds Ratio
method.tau = "DL", # DerSimonian-Laird estimator for tau^2
comb.fixed = FALSE, # Random effects model
comb.random = TRUE, # Include random effects
prediction = FALSE # No prediction interval by default
)
# Summary of the meta-analysis
summary(meta_analysis3)## OR 95%-CI %W(random)
## Skoch, 2017 NA 0.0
## Slator, 2019 0.6444 [0.0259; 16.0534] 14.4
## Alawi, 2014 4.1943 [1.9631; 8.9614] 85.6
##
## Number of studies: k = 2
## Number of observations: o = 1156 (o.e = 943, o.c = 213)
## Number of events: e = 1126
##
## OR 95%-CI z p-value
## Random effects model 3.2014 [0.8715; 11.7609] 1.75 0.0796
##
## Quantifying heterogeneity:
## tau^2 = 0.3650; tau = 0.6041; I^2 = 20.4%; H = 1.12
##
## Test of heterogeneity:
## Q d.f. p-value
## 1.26 1 0.2622
##
## Details on meta-analytical method:
## - Inverse variance method
## - DerSimonian-Laird estimator for tau^2
## - Mantel-Haenszel estimator used in calculation of Q and tau^2 (like RevMan 5)
## - Continuity correction of 0.5 in studies with zero cell frequencies# To visualize the results, you can plot a forest plot
meta::forest(meta_analysis3, layout = "JAMA")
——–4. Meta analysis of odds ratios for risk of complications in surgery vs endovascular ———
# Load the necessary library
library(meta)
# Assuming mrs_odds is already loaded with your data
# You might need to check the names of your columns with names(mrs_odds)
# Meta-analysis of odds ratios
meta_analysis4 <- metabin(
event.e = data_comparison$comp_end,
n.e = data_comparison$sample_endo,
event.c = data_comparison$comp_surg,
n.c = data_comparison$sample_surg,
data = data_comparison,
studlab = paste(data_comparison$Study_ID),
sm = "OR", # Specify summary measure as Odds Ratio
method.tau = "DL", # DerSimonian-Laird estimator for tau^2
comb.fixed = FALSE, # Random effects model
comb.random = TRUE, # Include random effects
prediction = FALSE # No prediction interval by default
)
# Summary of the meta-analysis
summary(meta_analysis4)## OR 95%-CI %W(random)
## Skoch, 2017 NA 0.0
## Slator, 2019 1.5517 [0.0623; 38.6542] 14.4
## Alawi, 2014 0.2384 [0.1116; 0.5094] 85.6
##
## Number of studies: k = 2
## Number of observations: o = 1156 (o.e = 943, o.c = 213)
## Number of events: e = 30
##
## OR 95%-CI z p-value
## Random effects model 0.3124 [0.0850; 1.1475] -1.75 0.0796
##
## Quantifying heterogeneity:
## tau^2 = 0.3650; tau = 0.6041; I^2 = 20.4%; H = 1.12
##
## Test of heterogeneity:
## Q d.f. p-value
## 1.26 1 0.2622
##
## Details on meta-analytical method:
## - Inverse variance method
## - DerSimonian-Laird estimator for tau^2
## - Mantel-Haenszel estimator used in calculation of Q and tau^2 (like RevMan 5)
## - Continuity correction of 0.5 in studies with zero cell frequencies# To visualize the results, you can plot a forest plot
meta::forest(meta_analysis4, layout = "JAMA")