軟件meta分析命令-2018朱教授原創(chuàng)

1、R軟件Meta分析命令集合此命令由醫(yī)藥加朱教授編輯原創(chuàng)，付出大量勞動(dòng)，請(qǐng)學(xué)員不要外傳！1、計(jì)數(shù)資料Meta分析 (RR，OR， RD，Peto-OR)注釋命令加載meta包library(meta)調(diào)取表格data-read.table(file.choose(),header=T,sep=,)瀏覽和編輯導(dǎo)入數(shù)據(jù)edit(data)執(zhí)行計(jì)數(shù)資料meta分析metabin(case1,tot1,case0,tot0, data=data, sm=RR, method=MH, comb.fixed=T,comb.random=F, studlab=study)將meta結(jié)果命名為meta88文件m

2、eta88-metabin(case1,tot1,case0,tot0,data=data, method=MH,sm=RR,comb.fixed=T,comb.random=F,studlab=study)作森林圖forest(meta88)做拉貝圖labbe(meta88)做加爾布雷斯圖radial(meta88)做博雅圖baujat(meta88)敏感性分析metainf(meta88)敏感性分析制圖forest(metainf(meta88)漏斗圖funnel(meta88)剪補(bǔ)法分析并作圖trimfill(meta88)funnel(trimfill(meta88)偏倚回歸分析met

3、abias(meta88,method=linreg)回歸分析metareg(meta88,TMN+AGE)注意：發(fā)表偏倚計(jì)算需要10個(gè)以上研究數(shù)目，否則報(bào)錯(cuò)。其他：exlude=1， byvar=，subset=, method=,title=, complab=, outclab=, method=MH,Inverse,GLMM,Peto。發(fā)生率低于1%為稀有事件，選取Peto-OR。2、計(jì)量資料Meta分析注釋命令調(diào)取表格data-read.table(file.choose(),header=T,sep=,)編輯和瀏覽數(shù)據(jù)edit(data)做計(jì)量meta分析metacont(n1,m

4、1,sd1,n2,m2,sd2,data=data, sm=MD,comb.random=T,comb.fixed=F, studlab=study,byvar=race)命名為meta99meta99-metacont(n1,m1,sd1,n2,m2,sd2,data=data,sm=MD, comb.random=T,comb.fixed=F,studlab=study,byvar=race)作森林圖forest(meta99)做拉貝圖labbe(meta99)做加爾布雷斯圖radial(meta99)做博雅圖baujat(meta99)敏感性分析metainf(meta99)敏感性分析制

5、圖forest(metainf(meta99)漏斗圖funnel(meta99)剪補(bǔ)法分析并作圖trimfill(meta99)funnel(trimfill(meta99)偏倚回歸分析metabias(meta99,method=linreg)回歸分析metareg(meta99,TMN+AGE)3、比率資料注釋命令調(diào)取表格data-read.table(file.choose(),header=T,sep=,)編輯和瀏覽數(shù)據(jù)edit(data)做meta分析meta66=metaprop(events,total,studlab=study,sm=PFT,data=data) meta66

6、作森林圖forest(meta66)做拉貝圖labbe(meta66)做加爾布雷斯圖radial(meta66)做博雅圖baujat(meta66)敏感性分析metainf(meta66)敏感性分析制圖forest(metainf(meta66)漏斗圖funnel(meta66)剪補(bǔ)法分析并作圖trimfill(meta66) funnel(trimfill(meta66)偏倚回歸分析metabias(meta66,method=linreg)回歸分析metareg(meta66,TMN+AGE)Available methods=Inverse variance；generalised li

7、near mixed model (GLMM)Logit transformation (sm=PLOGIT, default) Log transformation (sm=PLN) Freeman-Tukey Double arcsine transformation (sm=PFT) Arcsine transformation (sm=PAS) Raw, i.e. untransformed, proportions (sm=PRAW) 生存資料、泊松資料、相關(guān)系數(shù)和均數(shù)等（4 , 5, 6, 7, 8）數(shù)據(jù)類型命令data-read.table(file.choose(),heade

8、r=T,sep=,)edit(data)4、生存資料metagen(logHR, Se, sm=REML,data=data,comb.fixed=T,comb.random=F, studlab=study) (Log Hazard Ratios, Standard errors)Generic Inverse Variance Method:REML，ML，HS，SJ，HE，EB5、事件和人時(shí)-兩組比較(密度、頻次)metainc(event.e,time.e,event.c,time.c,sm=IRR,data=data, comb.fixed=F,comb.random=T,studl

9、ab=study,method=Cochran)泊松資料Note: Incidence Rate Ratio (sm=IRR), Incidence Rate Difference (sm=IRD) Method=MH, Inverse, Cochran, GLMM6、事件和人時(shí)-單組metarate(events,times,sm=IRLN,data=data, comb.fixed=T,comb.random=F, studlab=study,method=Inverse)泊松資料Note: Log transformation (sm=IRLN, default) Square root

10、 transformation (sm=IRS) Freeman-Tukey Double arcsine transformation (sm=IRFT) No transformation(sm=IR) Method= Inverse, GLMM7、相關(guān)系數(shù) + 樣本數(shù)metacor(cor, n, sm=zcor, data=data, comb.fixed=T,comb.random=F, studlab=study)sm=ZCOR or COR Method:Inverse Variance Weighting 8、均數(shù)和標(biāo)準(zhǔn)差metamean(n, mean, sd, sm=MRA

11、W,data=data,comb.fixed=T,comb.random=F, studlab=study)Raw, i.e. untransformed, means (sm=MRAW, default) Log transformed means (sm=MLN) Method:Inverse Variance methodinverse variance weightingNote: Calculation of fixed and random effects estimates for meta-analyses with continuous outcome data; inver

12、se variance weighting is used for pooling. Calculation of fixed effect and random effects estimates (risk ratio, odds ratio, risk difference, or arcsine difference) for meta-analyses with binary outcome data. Mantel-Haenszel, inverse variance, Peto method, and generalised linear mixed model (GLMM) a

13、re available for pooling. For GLMMs, the HYPERLINK :25095/library/metafor/html/rma.glmm.html rma.glmm function from R package metafor (Viechtbauer 2010) is called internally. Restricted maximum-likelihood estimator (method.tau=REML) Maximum-likelihood estimator (method.tau=ML) Hunter-Schmidt estimator (method.tau=HS) Sidik-Jonkman estimator (method.tau=SJ) Hedges estimator (method.tau=HE) Empirical Bayes estimator (method.tau=EB).Meta-analysis of incidence rates (metainc)Generic inverse variance meta-analysis (metagen) Meta-analysis of single correlations (metacor) Meta-analysis of single mea