Comparative Efficacy and Safety of Vortioxetine and Duloxetine for Major Depressive Disorder: A Systematic Review and Network Meta-Analysis of RCTs
Budget: $500 – $750 USD
Project Title
Comparative Efficacy and Safety of Vortioxetine and Duloxetine for Major Depressive Disorder: A Systematic Review and Network Meta-Analysis of RCTs.
Project Overview
This project is part of an academic thesis evaluating the comparative efficacy, tolerability, and safety of vortioxetine and duloxetine in adults with Major Depressive Disorder (MDD). Direct head-to-head randomized controlled trials (RCTs) between vortioxetine and duloxetine are limited, so an indirect comparison using network meta-analysis (NMA) is required.
The treatment network consists of vortioxetine, duloxetine, sertraline, desvenlafaxine, and placebo. Sertraline, desvenlafaxine, and placebo will be used as connecting comparators in the network, but the primary clinical focus and interpretation is on vortioxetine vs duloxetine.
The systematic search, title/abstract screening, and full-text review have already been completed. A final list of eligible RCTs and an initial data extraction template are available. The remaining work is data extraction, data preparation, and NMA (with SUCRA rankings) in R or STATA.
Objective
Direct head-to-head RCTs between vortioxetine and duloxetine are limited, making a network meta-analysis (NMA) necessary to combine both direct and indirect evidence. The analysis focuses on primary outcomes measured by the Montgomery–Åsberg Depression Rating Scale (MADRS) or the Hamilton Depression Rating Scale (HAM-D), to provide a comprehensive comparison of vortioxetine versus duloxetine and to establish SUCRA-based rankings for efficacy and tolerability within a network including sertraline, desvenlafaxine, and placebo.
Data Status (What Is Already Done)
- Comprehensive literature search across major databases is completed.
- Title and abstract screening are completed.
- Full-text screening is completed based on predefined inclusion/exclusion criteria.
- A final list of included RCTs (vortioxetine, duloxetine, sertraline, desvenlafaxine, placebo; adults with MDD; 2012 onwards) is available.
- A structured data extraction template (Excel) is already prepared, covering:
1. Study design and methodological details
2. Sample size and treatment arms
3. Drug, dose, and duration
4. Outcome scales (MADRS, HAM-D)
5. Statistical reporting format (mean ± SD, CI, etc.)
6. Dropout rates and analysis type (ITT/PP)
7. Risk of bias domains
I will provide:
- The list of included studies (with PDFs or links where possible)
- The current extraction template and instructions
- The inclusion/exclusion criteria and definitions for outcomes
Your Tasks
- Complete Data Extraction:
For each included RCT and each relevant arm (vortioxetine, duloxetine, sertraline, desvenlafaxine, placebo):
1. Primary efficacy (continuous)
a. Extract mean change from baseline in MADRS or HAM-D at the acute endpoint (e.g. 6–8 weeks) and corresponding SD of change.
b. If change scores are unavailable, extract endpoint means and SDs and clearly label them as such.
c. Record the time point (week) and scale used (MADRS or HAM-D) for each arm.
2. Binary efficacy outcomes (if reported)
a. Response (e.g. ≥50% reduction in MADRS/HAM-D, using each trial’s own definition).
b. Remission (e.g. MADRS ≤10 or HAM-D ≤7, using each trial’s own definition).
3. Safety and tolerability outcomes
a. All-cause discontinuation (any reason) per arm.
b. Discontinuation due to adverse events per arm.
c. Number of patients with at least one adverse event (if reported).
d. Key specific adverse events of interest (e.g. nausea, insomnia) where consistently reported.pubmed.
4. Dose and design information
a. Drug and dose (e.g. vortioxetine 10 mg, 20 mg; duloxetine 60 mg; sertraline dose; desvenlafaxine dose).
b. Duration of acute treatment phase (weeks).
c. Sample size per arm (randomized and analyzed).
d. Setting (e.g. outpatient vs inpatient), country/region, sponsorship.
5. Risk of bias
a. Domain-level judgments using standard Cochrane-style domains:
+ Random sequence generation
+ Allocation concealment
+ Blinding of participants/personnel/outcome assessors
+ Incomplete outcome data
+ Selective reporting
+ Other sources of bias
6. Documentation and assumptions
a. Clearly document any conversions (e.g. SE or CI to SD, p-value to SD) and formulas used.
b. Note any assumptions (e.g. converting medians/IQR to means/SD, if unavoidable and agreed).
c. Flag multi-arm trials, unusual designs, or deviations in a “Notes” column.
Deliverable: Updated and completed Excel/CSV extraction file(s) with all included trials and arms.
- Prepare Analysis-Ready Datasets
From the extraction file, prepare clean, analysis-ready datasets in study-arm format, with each row as one arm:
Core variables (minimum):
• study_id
• arm_id
• treatment (coded as VORT, DULO, SERT, DESV, PLAC)
• dose
• n (sample size)
• mean_change (or mean_endpoint, clearly labeled)
• sd_change (or sd_endpoint, clearly labeled)
• responders (number)
• remitters (number)
• discontinuation_any (number)
• discontinuation_ae (number)
• timepoint_week
• scale (MADRS/HAM-D)
Prepare separate CSV files for:
• Primary continuous efficacy dataset (acute endpoint).
• Binary efficacy (response).
• Binary efficacy (remission).
• Safety/acceptability (all-cause discontinuation).
• Tolerability (discontinuation due to adverse events).
• Risk of bias data.
Scales:
• Use MADRS where available; use HAM-D where MADRS is not reported, and clearly label the scale.
• Datasets should be suitable for computing standardized mean differences (SMDs) if MADRS and HAM-D must be combined.
Deliverable: Clean CSV files with a short data dictionary describing each variable.
- Conduct the Network Meta-Analysis
Use R or STATA for a random-effects network meta-analysis with the following treatments:
• Vortioxetine (VORT)
• Duloxetine (DULO)
• Sertraline (SERT)
• Desvenlafaxine (DESV)
• Placebo (PLAC)
The primary comparison of interest is vortioxetine vs duloxetine, with sertraline, desvenlafaxine, and placebo acting as network comparators.
Required analyses:
1. Primary efficacy (continuous)
a. Outcome: change in MADRS/HAM-D (SMD or MD; please specify and justify).
b. Model: random-effects NMA.
c. Outputs:
+ Effects vs placebo for all active drugs.
+ All pairwise comparisons (especially vortioxetine vs duloxetine).
+ Heterogeneity (τ2\tau^2τ2) and 95% CIs.
+ Any available consistency checks (e.g. design-by-treatment test, node-splitting).pmc.ncbi.nlm.nih+1
2. Binary efficacy
a. Outcomes: response and remission.
b. Effect measure: risk ratio or odds ratio (state which you use).
c. Same treatment set and NMA approach.
3. Safety/tolerability
a. Outcomes:
b. All-cause discontinuation.
c. Discontinuation due to adverse events.
d. Random-effects NMA, same treatment set.
4. Network description
a. Network plot showing nodes (VORT, DULO, SERT, DESV, PLAC) and edges.
b. Table listing number of trials per direct comparison and total patients per treatment.e-epih+1
Deliverable:
• Fully annotated R scripts or STATA .do files that import the data, run all models, and generate all outputs.
- SUCRA / Ranking
For each outcome:
• Primary continuous efficacy
• Response
• Remission
• All-cause discontinuation
• Discontinuation due to adverse events
Please:
• Compute SUCRA values (or P-scores in frequentist NMA) for all five treatments (VORT, DULO, SERT, DESV, PLAC).
• Provide:
a. A table of SUCRA/P-scores and ranks for each treatment and outcome.
b. Rankograms or bar plots if feasible.
Emphasis in interpretation should be on how vortioxetine ranks relative to duloxetine across efficacy and tolerability outcomes.
- Sensitivity Analyses (If Feasible)
If data and time permit:
• Exclude high risk-of-bias trials and repeat the primary NMA.
• Restrict to typical therapeutic doses (e.g. vortioxetine 10–20 mg; duloxetine 60 mg; standard doses for sertraline and desvenlafaxine).
• Consider a sensitivity analysis by the most common timepoint (e.g. week 6–8 only).
Clearly label which analyses are primary and which are sensitivity.
Final Deliverables
1. Data
a. Completed extraction file(s) (Excel/CSV).
b. Clean CSVs: continuous efficacy, binary response, binary remission, safety/acceptability, tolerability, risk of bias.
c. Short data dictionary.
2. Code
a. Fully commented R scripts or STATA .do files covering data import, NMA, SUCRA, and generation of tables/plots.
3. Tables and Figures
a. Network plot(s).
b. Forest plots for each outcome.
c. League tables with all pairwise comparisons and 95% CIs.
d. SUCRA/P-score tables and rankograms (if possible).
4. Methods & Results Summary (1–3 pages)
a. Description of data preparation and inclusion criteria.
b. Model details (random/fixed effects, effect measures, handling of multi-arm trials and different scales).
c. Main results, focusing on vortioxetine vs duloxetine for efficacy, acceptability, and tolerability.
d. Brief comments on heterogeneity and consistency.
Required Skills
• Proven experience with network meta-analysis and SUCRA/P-scores.
• Proficiency in R or STATA for NMA.
• Experience with RCT data and depression scales (MADRS, HAM-D) highly preferred.
• Strong documentation and reproducibility practices.
How to Apply
Please include:
• Your preferred software (R or STATA).
• An example (or description) of a previous NMA or meta-analysis you have worked on.
• Your estimated timeline to complete the project.
Communication and Stepwise Updates
I would like regular updates as you progress. Please agree to:
• Send a brief update and the intermediate files when each step is completed.pleted.
• Be available to clarify any methodological questions (e.g. scale choice, effect measure, handling of specific trials) before locking in the final analyses.
Comparative Efficacy and Safety of Vortioxetine and Duloxetine for Major Depressive Disorder: A Systematic Review and Network Meta-Analysis of RCTs.
Project Overview
This project is part of an academic thesis evaluating the comparative efficacy, tolerability, and safety of vortioxetine and duloxetine in adults with Major Depressive Disorder (MDD). Direct head-to-head randomized controlled trials (RCTs) between vortioxetine and duloxetine are limited, so an indirect comparison using network meta-analysis (NMA) is required.
The treatment network consists of vortioxetine, duloxetine, sertraline, desvenlafaxine, and placebo. Sertraline, desvenlafaxine, and placebo will be used as connecting comparators in the network, but the primary clinical focus and interpretation is on vortioxetine vs duloxetine.
The systematic search, title/abstract screening, and full-text review have already been completed. A final list of eligible RCTs and an initial data extraction template are available. The remaining work is data extraction, data preparation, and NMA (with SUCRA rankings) in R or STATA.
Objective
Direct head-to-head RCTs between vortioxetine and duloxetine are limited, making a network meta-analysis (NMA) necessary to combine both direct and indirect evidence. The analysis focuses on primary outcomes measured by the Montgomery–Åsberg Depression Rating Scale (MADRS) or the Hamilton Depression Rating Scale (HAM-D), to provide a comprehensive comparison of vortioxetine versus duloxetine and to establish SUCRA-based rankings for efficacy and tolerability within a network including sertraline, desvenlafaxine, and placebo.
Data Status (What Is Already Done)
- Comprehensive literature search across major databases is completed.
- Title and abstract screening are completed.
- Full-text screening is completed based on predefined inclusion/exclusion criteria.
- A final list of included RCTs (vortioxetine, duloxetine, sertraline, desvenlafaxine, placebo; adults with MDD; 2012 onwards) is available.
- A structured data extraction template (Excel) is already prepared, covering:
1. Study design and methodological details
2. Sample size and treatment arms
3. Drug, dose, and duration
4. Outcome scales (MADRS, HAM-D)
5. Statistical reporting format (mean ± SD, CI, etc.)
6. Dropout rates and analysis type (ITT/PP)
7. Risk of bias domains
I will provide:
- The list of included studies (with PDFs or links where possible)
- The current extraction template and instructions
- The inclusion/exclusion criteria and definitions for outcomes
Your Tasks
- Complete Data Extraction:
For each included RCT and each relevant arm (vortioxetine, duloxetine, sertraline, desvenlafaxine, placebo):
1. Primary efficacy (continuous)
a. Extract mean change from baseline in MADRS or HAM-D at the acute endpoint (e.g. 6–8 weeks) and corresponding SD of change.
b. If change scores are unavailable, extract endpoint means and SDs and clearly label them as such.
c. Record the time point (week) and scale used (MADRS or HAM-D) for each arm.
2. Binary efficacy outcomes (if reported)
a. Response (e.g. ≥50% reduction in MADRS/HAM-D, using each trial’s own definition).
b. Remission (e.g. MADRS ≤10 or HAM-D ≤7, using each trial’s own definition).
3. Safety and tolerability outcomes
a. All-cause discontinuation (any reason) per arm.
b. Discontinuation due to adverse events per arm.
c. Number of patients with at least one adverse event (if reported).
d. Key specific adverse events of interest (e.g. nausea, insomnia) where consistently reported.pubmed.
4. Dose and design information
a. Drug and dose (e.g. vortioxetine 10 mg, 20 mg; duloxetine 60 mg; sertraline dose; desvenlafaxine dose).
b. Duration of acute treatment phase (weeks).
c. Sample size per arm (randomized and analyzed).
d. Setting (e.g. outpatient vs inpatient), country/region, sponsorship.
5. Risk of bias
a. Domain-level judgments using standard Cochrane-style domains:
+ Random sequence generation
+ Allocation concealment
+ Blinding of participants/personnel/outcome assessors
+ Incomplete outcome data
+ Selective reporting
+ Other sources of bias
6. Documentation and assumptions
a. Clearly document any conversions (e.g. SE or CI to SD, p-value to SD) and formulas used.
b. Note any assumptions (e.g. converting medians/IQR to means/SD, if unavoidable and agreed).
c. Flag multi-arm trials, unusual designs, or deviations in a “Notes” column.
Deliverable: Updated and completed Excel/CSV extraction file(s) with all included trials and arms.
- Prepare Analysis-Ready Datasets
From the extraction file, prepare clean, analysis-ready datasets in study-arm format, with each row as one arm:
Core variables (minimum):
• study_id
• arm_id
• treatment (coded as VORT, DULO, SERT, DESV, PLAC)
• dose
• n (sample size)
• mean_change (or mean_endpoint, clearly labeled)
• sd_change (or sd_endpoint, clearly labeled)
• responders (number)
• remitters (number)
• discontinuation_any (number)
• discontinuation_ae (number)
• timepoint_week
• scale (MADRS/HAM-D)
Prepare separate CSV files for:
• Primary continuous efficacy dataset (acute endpoint).
• Binary efficacy (response).
• Binary efficacy (remission).
• Safety/acceptability (all-cause discontinuation).
• Tolerability (discontinuation due to adverse events).
• Risk of bias data.
Scales:
• Use MADRS where available; use HAM-D where MADRS is not reported, and clearly label the scale.
• Datasets should be suitable for computing standardized mean differences (SMDs) if MADRS and HAM-D must be combined.
Deliverable: Clean CSV files with a short data dictionary describing each variable.
- Conduct the Network Meta-Analysis
Use R or STATA for a random-effects network meta-analysis with the following treatments:
• Vortioxetine (VORT)
• Duloxetine (DULO)
• Sertraline (SERT)
• Desvenlafaxine (DESV)
• Placebo (PLAC)
The primary comparison of interest is vortioxetine vs duloxetine, with sertraline, desvenlafaxine, and placebo acting as network comparators.
Required analyses:
1. Primary efficacy (continuous)
a. Outcome: change in MADRS/HAM-D (SMD or MD; please specify and justify).
b. Model: random-effects NMA.
c. Outputs:
+ Effects vs placebo for all active drugs.
+ All pairwise comparisons (especially vortioxetine vs duloxetine).
+ Heterogeneity (τ2\tau^2τ2) and 95% CIs.
+ Any available consistency checks (e.g. design-by-treatment test, node-splitting).pmc.ncbi.nlm.nih+1
2. Binary efficacy
a. Outcomes: response and remission.
b. Effect measure: risk ratio or odds ratio (state which you use).
c. Same treatment set and NMA approach.
3. Safety/tolerability
a. Outcomes:
b. All-cause discontinuation.
c. Discontinuation due to adverse events.
d. Random-effects NMA, same treatment set.
4. Network description
a. Network plot showing nodes (VORT, DULO, SERT, DESV, PLAC) and edges.
b. Table listing number of trials per direct comparison and total patients per treatment.e-epih+1
Deliverable:
• Fully annotated R scripts or STATA .do files that import the data, run all models, and generate all outputs.
- SUCRA / Ranking
For each outcome:
• Primary continuous efficacy
• Response
• Remission
• All-cause discontinuation
• Discontinuation due to adverse events
Please:
• Compute SUCRA values (or P-scores in frequentist NMA) for all five treatments (VORT, DULO, SERT, DESV, PLAC).
• Provide:
a. A table of SUCRA/P-scores and ranks for each treatment and outcome.
b. Rankograms or bar plots if feasible.
Emphasis in interpretation should be on how vortioxetine ranks relative to duloxetine across efficacy and tolerability outcomes.
- Sensitivity Analyses (If Feasible)
If data and time permit:
• Exclude high risk-of-bias trials and repeat the primary NMA.
• Restrict to typical therapeutic doses (e.g. vortioxetine 10–20 mg; duloxetine 60 mg; standard doses for sertraline and desvenlafaxine).
• Consider a sensitivity analysis by the most common timepoint (e.g. week 6–8 only).
Clearly label which analyses are primary and which are sensitivity.
Final Deliverables
1. Data
a. Completed extraction file(s) (Excel/CSV).
b. Clean CSVs: continuous efficacy, binary response, binary remission, safety/acceptability, tolerability, risk of bias.
c. Short data dictionary.
2. Code
a. Fully commented R scripts or STATA .do files covering data import, NMA, SUCRA, and generation of tables/plots.
3. Tables and Figures
a. Network plot(s).
b. Forest plots for each outcome.
c. League tables with all pairwise comparisons and 95% CIs.
d. SUCRA/P-score tables and rankograms (if possible).
4. Methods & Results Summary (1–3 pages)
a. Description of data preparation and inclusion criteria.
b. Model details (random/fixed effects, effect measures, handling of multi-arm trials and different scales).
c. Main results, focusing on vortioxetine vs duloxetine for efficacy, acceptability, and tolerability.
d. Brief comments on heterogeneity and consistency.
Required Skills
• Proven experience with network meta-analysis and SUCRA/P-scores.
• Proficiency in R or STATA for NMA.
• Experience with RCT data and depression scales (MADRS, HAM-D) highly preferred.
• Strong documentation and reproducibility practices.
How to Apply
Please include:
• Your preferred software (R or STATA).
• An example (or description) of a previous NMA or meta-analysis you have worked on.
• Your estimated timeline to complete the project.
Communication and Stepwise Updates
I would like regular updates as you progress. Please agree to:
• Send a brief update and the intermediate files when each step is completed.pleted.
• Be available to clarify any methodological questions (e.g. scale choice, effect measure, handling of specific trials) before locking in the final analyses.