博客

NMA(网状meta分析)stata简易教程(2)网状图的千变万化

有很多童鞋问到stataNMA的作图过程网状图的变化问题,在此做一下解答。

相关问题如果有疑惑,请先查看前期网状meta分析相关推文,猴哥在此不详述。

 Open the data editor window:

#输入

Edit

 

The variable study contains the ID number of each study. The variables t1 and t2, r1 and r2, n1 and n2 contain the name of the treatments (abbreviated), the number of events and the total number of patients in arms 1 and 2. For each multi-arm study with T number of arms there are observations (with the same study ID) equal to the number of pairwise comparisons that this study provides. The variables logOR and selogOR contain the log-odds ratio and its standard error for the comparison t1 vs. t2. Finally, the variable allocation_concealment contains the judgments for the risk of bias regarding the appropriate conduct of allocation concealment (U=unclear, L=low).

变量study:研究的ID号。

变量t1:干预措施1;变量t2:干预措施2。

变量r1:事件发生数1;变量r2:事件发生数2。 R表示事件发生数

变量n1:病人总数1;变量n2:病人总数2。

对于每一个研究有T个观察结果,T=T×(T-1)/2,即双臂研究有2组对比,三臂研究有3组对比,四臂研究有6组对比。

变量logOR:t1 vs. t2的对数差异比。

变量selogOR:logOR的标准误。

变量allocation_concealment:根据分配隐藏得到的偏倚风险的判断(U=unclear, L=low)。

 

##预处理

. use "F:\【Network Meta-Analysis - 网状meta分析】使用STATA软件绘图\example_datasets\acute mania efficacy pairwise.dta"

 

. networkplot t1 t2

##预处理

 

Produce a network plot in which both nodes and edges are weighted according to the number of studies evaluating each treatment and direct comparison respectively:

 

创建network plot,节点:该药物相关研究的数目,边界线:两者间直接对比的研究的数目。点=涉及该药物的研究数目  线=两药物比较的研究的数目

networkplot t1 t2

 

Weight the nodes according to the number of patients that have received each treatment:

节点:接受该干预措施的病人数量。

gen ntot=n1+n2

networkplot t1 t2, nodew(ntot)

 

networkplot t1 t2

 

 

Calculate the control group risk for studies including the placebo:

计算研究内对照组的风险(包含安慰剂在内)。

gen cgr=r1/n1 if t1=="PLA"    #如果第1组t1 r1 n1为 pla 的话,产生cgr 为率p为 p1

replace cgr=r2/n2 if t2=="PLA"   #如果第2组t2 r2 n2为 pla 的话,产生cgr 为率p为 p2

 

### cgr为control 安慰剂的平均风险

 

 

and weight the edges according to to the mean control group risk for all comparisons versus placebo:

边界线:所有试验组vs.安慰剂的平均风险。

networkplot t1 t2, nodew(ntot) edgew(cgr mean)

 

Increase the thickness of all edges and change the aspect ratio of the plot to make it more legible:

增加边界线的厚度,改变节点的纵横比使其更清晰。

 

//Networkplot CMD 

 networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(0.8)

 networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(2)

 networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(1)

 networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(0.8)

 

Replace the abbreviated names of the treatments with the full names:

干预措施的简称换成全称。

networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(0.8) lab(Aripiprazole Asenapine Carbamazpine Divalproex Haloperidol Lamotrigine Lithium Olanzapine Paliperidone Placebo Quetipaine Ripseridone Topiramate Ziprasidone)

Note that the names of the treatments need to be given in alphabetical order.

注意按字母顺序排列。

14种药物进行进一步的命名:

 

Color the edges according to the adequacy of allocation concealment in the majority of trials:

边界线上色:根据分配隐藏(大多数试验)。

networkplot t1 t2, nodew(ntot) asp(0.8) lab(Aripiprazole Asenapine Carbamazpine Divalproex Haloperidol Lamotrigine Lithium Olanzapine Paliperidone Placebo Quetipaine Ripseridone Topiramate Ziprasidone) edgecol(by allocation_concealment)

 

分配隐藏 亮黄色为UPPER ,    高偏倚风险试验  黄牌警告  绿色为 LOW 低偏倚风险试验

 

Color the edges according to the adequacy of allocation concealment of the trial with the higher risk of bias in each comparison:

边界线上色:根据分配隐藏(高偏倚风险试验)。

networkplot t1 t2, nodew(ntot) asp(0.8) lab(Aripiprazole Asenapine Carbamazpine Divalproex Haloperidol Lamotrigine Lithium Olanzapine Paliperidone Placebo Quetipaine Ripseridone Topiramate Ziprasidone) edgecol(by allocation_concealment max)

 

Create a numeric variable for the adequacy of allocation concealment and specify different colors for each bias level:

为分配隐藏创建数字变量,为不同的偏倚风险设定不同的颜色。

gen alloc_new=1 if allocation==”L”

replace alloc_new=2 if allocation==”U”

networkplot t1 t2, nodew(ntot) edgew(cgr mean) edgesc(2.5) asp(0.8) lab(Aripiprazole Asenapine Carbamazpine Divalproex Haloperidol Lamotrigine Lithium Olanzapine Paliperidone Placebo Quetipaine Ripseridone Topiramate Ziprasidone) edgecol(by alloc_new min) bylev(2) bycol(magenta maroon)

 

### 分析后得:L 低 为1 ,高u 为2,从min到max ,min为洋红色,颜色略浅,低风险,好   max为深红色,颜色略深,高风险

 

 

 

本文由 GCBI学院 作者:其明技术专家 发表,转载请注明来源!

热评文章

发表评论