The data used in this study comes from GRISELDA
(Group of Researchers Investigating Specific Efficacy of individual Drugs for Acute depression) [1].
Double-blind RCTs of the acute phase treatment with
**21 antidepressants**
for adult patients (≥18 years old) with a primary diagnosis of
**major depression**
were
**systematically**
searched and included. About one in four trials included in the dataset comes from unpublished data, and two come from both published and unpublished sources.

In the
**network meta-analysis,**
we use two primary outcomes: efficacy and acceptability.
**Efficacy**
was measured by the response rate of the intervention, defined as the proportion of patients
who showed 50% or greater reduction on a validated depression severity scales from baseline.
**Acceptability**
was measured by the all-cause discontinuation rate, defined as the proportion of patients
who withdrew early due to any reasons.

[1] Cipriani A, Furukawa TA, Salanti G, et al, 2018. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet 391, 1357-1366.

Please wait. The calculations of network meta-analysis may take up to several minutes.

The network includes the following treatments:

Please wait. The calculations of network meta-analysis may take up to several minutes.

The network includes the following treatments:

(1) The size of the node is proportionate to the number of randomized patients, and the thickness of the line is proportionate to the number of studies.

(2) The
*netmeta*
package in R adopts generalized
Cochran’s Q statistic
to assess the assumption of homogeneity and consistency. It is calculated as the weighted sum of squared differences between
individual study effects and the pooled effect across studies, with the weights being those used in the pooling method.
This Q statistic can be decomposed
in a sum of within-design Q statistics (used to assess heterogeneity among direct comparisons)
and one between-designs Q statistic (used to assess inconsistency between direct and indirect evidence).
We provide both Q statistics and their p values for the test.

(3) The I-squared statistic describes the percentage of variation across studies that is due to heterogeneity rather than chance.

(4) In random-effects meta-analysis, the extent of variation among the effects observed in different studies (between-study variance) is referred to as tau-squared . tau-squared is the variance of the effect size parameters across the population of studies and it reflects the variance of the true effect sizes. The square root of this number is referred to as tau.

Please wait. The calculations of network meta-analysis may take up to several minutes.

Please click the plot to see a zoomed one with the optimal ranges of x and y axis. (The zoomed plot will be displayed below.)

(1) Drugs in the upper right area are better balanced in both efficacy and acceptability.

(2) The size of the node is proportionate to the number of randomized patients.

(3) If placebo-controlled trials are included, the plot will display 2 sets of result. The solid circles indicate results from full dataset, while the empty circles indicate results from only published data.

Please wait. The calculations of network meta-analysis may take up to several minutes.

Please wait. The calculations of network meta-analysis may take up to several minutes.

Download

**Note: **
(1) The lower left half shows efficacy, and the upper right half shows acceptability.

(2) OR>1 indicates that the column-defined drug is favored.

**Caution: **
The league table cannot be constructed if the drugs in the efficacy network and the acceptability network do not match. So please make sure of it in the 'Basic Information', before you try to download the league table.

Select the 2 two drugs that you would like to compare the relative effect:

**Note: **
Please wait a moment for the result, since it needs time to process. The OR is with reference to Drug B.

Please wait. The calculations of network meta-analysis may take up to several minutes.

Please wait. The calculations of network meta-analysis may take up to several minutes.

**Note: **
The comparison-adjusted funnel plot is used to check if there were significant differences between precise and imprecise trials.
If placebo-controlled trials are included, the plot shows all the direct comparisons with placebo.
If only head-to-head trials are included, the plot shows only the studies directly compared with fluoxetine, which is the most frequently used active controls.

We assessed the confidence in the evidence for a series of consecutive network meta-analyses every 5 years since 1990, for head-to-head trials. We believe that it should be as important
to see how credible the evidence are, as with the relative effect estimates. We evaluated it at 3 levels: RCT-level, comparison-level and drug-level. At
**RCT-level, **
we assessed the risk of bias for each RCT based on Cochrane Collaboration risk of bias tool. At
**comparison-level, **
we evaluated it at each time point using
*CINeMA framework*
[2] through the dedicated software. Finally,
the information was synthesized at
**drug-level. **

We combined the efficacy, acceptability and confidence in the evidence at comparison-level in league tables, and at drug-level in 2 dimensional plots at each time point. These results demonstrate how evidence evolved in the past decades.

**The left side bar is not applicable to the illustration under this tab.

[2] CINeMA: Confidence in Network Meta-Analysis [Software]. Institute of Social and Preventive Medicine, University of Bern, 2017. Available from cinema.ispm.unibe.ch.

**Note: **
(1) Drugs in the upper right area are better balanced in both efficacy and acceptability.

(2) The size of the node is proportionate to the inverse of the width of 95%CI regarding efficacy. The bigger the node, the more precise the CI is.

(3) The colors in each pie chart is proportionate to the numbers of mixed estimates for the drug in question evaluated at four levels of GRADE certainty of evidence. The four levels are: green-high, blue-moderate, yellow-low, red-very low.

**Note: **
(1) The lower left half shows efficacy, and the upper right half shows acceptability.

(2) OR>1 indicates that the column-defined drug is favored.

(3) We colored each cell of the league table in terms of the overall confidence in evidence between two drugs: green indicated high, blue indicated moderate, yellow indicated low, and red indicated very low confidence in evidence.

Luo Y, Chaimani A, Furukawa TA, Kataoka Y, Ogawa Y, Cipriani A, Salanti G.
Visualizing the Evolution of Evidence: Cumulative Network Meta-Analyses of
New Generation Antidepressants in the Last 40 Years.
*Research Synthesis Methods*
2020 Apr;1–11. doi.org/ 10.1002/jrsm.1413.