Skip to main content

Table 1 Sample from \(\hat{R_{\mathbbm{T}}\ }\) and \(\hat{H_{\mathbbm{T}}\ }\) (panel a), ranking matrix (panel b) and hierarchy matrix (panel c) of the hypothetical network of three treatments t1, t2 and t3 of Fig. 1

From: Answering complex hierarchy questions in network meta-analysis

Panel a

Panel b

Ranking matrix: Summary of the \(\hat{H}_{T}\) sample to show uncertainty in the ranking of each treatment

Panel c

Hierarchy matrix: Estimated probability mass function of treatment hierarchy

Sample from \(\hat{{R}_{T}}\)

Sample from \(\hat{{H}_{T}}\)

  

\({{r}_{{t}_{1}}}^{\ast}={3}\)

\({{r}_{{t}_{2}}}^{\ast}={2}\)

\({{r}_{{t}_{3}}}^{\ast}={1}\)

t3

t2

t1

 

1st

2nd

3rd

hl

\({p}_{{h}_{l}}\)

\({{r}_{{t}_{1}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={1}\)

\({{r}_{{t}_{3}}}^{\ast}={3}\)

t2

t1

t3

t1

25%

50%

25%

{t3, t1, t2} (h1)

25%

\({{r}_{{t}_{1}}}^{\ast}={1}\)

\({{r}_{{t}_{3}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={3}\)

t1

t3

t2

t2

30%

40%

30%

{t2, t1, t3} (h2)

25%

\({{r}_{{t}_{1}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={3}\)

\({{r}_{{t}_{3}}}^{\ast}={1}\)

t3

t1

t2

t3

45%

10%

45%

{t3, t2, t1} (h3)

20%

\({{r}_{{t}_{1}}}^{\ast}={1}\)

\({{r}_{{t}_{2}}}^{\ast}={2}\)

\({{r}_{{t}_{3}}}^{\ast}={3}\)

t1

t2

t3

    

{t1, t2, t3} (h4)

20%

\({{r}_{{t}_{1}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={1}\)

\({{r}_{{t}_{3}}}^{\ast}={3}\)

t2

t1

t3

    

{t2, t3, t1} (h5)

5%

\({{r}_{{t}_{1}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={3}\)

\({{r}_{{t}_{3}}}^{\ast}={1}\)

t3

t1

t2

    

{t1, t3, t2} (h6)

5%

\({{r}_{{t}_{1}}}^{\ast}={1}\)

\({{r}_{{t}_{2}}}^{\ast}={2}\)

\({{r}_{{t}_{3}}}^{\ast}={3}\)

t1

t2

t3

      

\({{r}_{{t}_{1}}}^{\ast}={2}\)

\({{r}_{{t}_{2}}}^{\ast}={1}\)

\({{r}_{{t}_{3}}}^{\ast}={3}\)

t2

t1

t3

      

\({{r}_{{t}_{1}}}^{\ast}={3}\)

\({{r}_{{t}_{2}}}^{\ast}={2}\)

\({{r}_{{t}_{3}}}^{\ast}={1}\)

t3

t2

t1

      
   

…

        
   

…

        
   

…

       Â