Skip to main content Skip to main content

LivingMetav1.0 — Self-Discovering Certified Meta-Analysis

1. Search 2. Screen 3. Extract 4. Synthesize 5. Bias 6. GRADE 7. Certify
📜 Protocol — PICO Framework
🎯 Primary & Secondary Outcomes
🔍 Search Strategy Preview
📚 Baseline Trials (Pre-loaded from Config)
📈 Published Benchmarks for Comparison
Trial Timeline
Data Quality Dashboard
Select a config to see data quality scoring and completeness analysis.
Add / Edit Trial
Import Trials from CSV

Paste CSV with columns: ID, Year, nTreatment, nControl, Events_Tx, Events_Ctrl, HR, HR_LCI, HR_UCI (header row required)

Cross-Config Comparison
Select two configs and click Compare to see side-by-side results.
Review Queue: 0
N/P = nav | I/E = propose | C = confirm
Select a study from the queue to review its details.
📊 Data Extraction — User Verification Required

Expand each trial to review extracted data. Source text is highlighted in yellow. Click values to override. Check ✔ to verify each field. ✓ 3 sources ⚠ 2 sources ✗ 1 source

Effect Size Converter
Outcome Harmonization Check
Run synthesis to check for composite endpoint mismatches across trials.
⚗ Evidence Synthesis
HKSJ
Bayes
Dual-Track Results
Track A: HR Pool (time-to-event)
Run analysis to see HR-based pooled estimates
Track B: OR Pool (dichotomous)
Run analysis to see OR-based pooled estimates
Forest Plot
Run analysis to generate forest plot
Heterogeneity Decomposition
Run synthesis first
Sensitivity Forest (Multi-Method Comparison)
Drapery Plot (P-value Function)
MCID Threshold Analysis
Run synthesis and set MCID to see how many studies show clinically meaningful effects.
Exclusion Sensitivity Matrix
Run synthesis first. Shows how excluding each study changes key metrics.
Study Contribution Waterfall
Confidence Interval Method Comparison
Run synthesis first to compare CI methods (Wald, HKSJ, Wald+t-critical, Profile Likelihood).
Study Weights Visualization
What-If Scenario (Add Hypothetical Trial)
Enter hypothetical trial parameters and click Run to preview impact on pooled estimate.
Method Comparison & Published Benchmarks
Run analysis to see method comparison table and benchmark comparison
Leave-One-Out & Cumulative Analysis
Leave-One-Out (Sensitivity)
Run analysis first
Cumulative Meta-Analysis (by year)
Run analysis first
Subgroup Analysis
Select a subgroup variable and run analysis to see stratified results with interaction test.
Subgroup forest plot will appear here
Trial Sequential Analysis (TSA)
TSA computes the Required Information Size and alpha-spending monitoring boundaries. Run synthesis first.
TSA monitoring boundary plot will appear here
Advanced Inference
Permutation Test (Exact)
Run synthesis first
Profile Likelihood for τ²
Run synthesis first
Fail-safe N
Run synthesis first
Robust Bayesian MA (RoBMA)
Run synthesis first
Fragility Index & NNT
Fragility Index
Run analysis first
Number Needed to Treat (NNT)
Run analysis first
E-value (Unmeasured Confounding)
Run synthesis first
Continuity Correction Sensitivity
Run synthesis first
Confidence Distribution Curve (Xie et al. 2011)
Run synthesis first to generate confidence distribution curve
Sequential Bayesian Updating
Run synthesis first to generate sequential Bayesian update plot
Deep Diagnostics
Sceptical P-value (Matthews 2018)
Run synthesis first
Q-profile CI for I²
Run synthesis first
Effective k & Entropy
Run synthesis first
Prior Sensitivity (3 Bayesian Priors)
Run synthesis first
Conditional Power & Predictive Probability
Run synthesis first
Bootstrap Prediction Interval (Parametric Bootstrap)
Run synthesis first
Replicability & Evidential Diagnostics
Intrinsic Credibility (Held 2019)
Run synthesis first
Harmonic Mean χ² (Held 2020)
Run synthesis first
Z-Curve ERR/EDR (Bartoš & Schimmack 2022)
Run synthesis first
Proportion of Clinically Meaningful Effects (Mathur & VanderWeele 2020)
Run synthesis first
Predictive Distribution of True Effects (Siemens, Borenstein, Schwarzer 2025)
Run synthesis first to generate predictive distribution plot
Bayesian Assurance (O'Hagan 2005)
Run synthesis first
Empirical Bayesian Heterogeneity Prior (Röver 2023)
Run synthesis first
Non-Collapsibility Sensitivity (OR/HR)
Run synthesis first
Firth Penalized Likelihood (Evrenoglou 2022)
Run synthesis first
Confidence Distribution Prediction Interval (Nagashima-Noma 2019)
Run synthesis first
Network Meta-Analysis (NMA)
NMA requires a config with 3+ drug classes (e.g., ATTR-CM). Run synthesis first.
Evidence Accrual Timeline
Run synthesis first to generate evidence accrual timeline
Sensitivity Analysis Summary
Run synthesis first
Bayesian Posterior Density (Prior Sensitivity)
Run synthesis first
NNT Curve Across Baseline Risk
Run synthesis first
Meta-Regression (Subgroup & Year)
Run synthesis first
Three-Level Meta-Analysis (Clustered Studies)
Run synthesis first. 3-level MA handles multi-arm trials and nested study structures.
Evidence Velocity & Maturity
Run synthesis first to see evidence accumulation rate and maturity assessment.
📈 Publication Bias Assessment
Funnel Plot
Run synthesis first, then bias tests will be available
Bias Test Results
Egger's test, Begg's test, trim-and-fill, and PET-PEESE results will appear here after synthesis.
Galbraith (Radial) Plot
Run synthesis first to generate Galbraith plot
Baujat Plot
Run synthesis first to generate Baujat plot
L'Abbé Plot
Run synthesis first to generate L'Abbé plot
GOSH Plot (All Subset Enumeration)
Run synthesis first to generate GOSH plot
Doi Plot & LFK Index
Run synthesis first
Run synthesis first to generate Doi plot
Harbord's Score Test (Harbord et al. 2006)
Run synthesis first
Influence Diagnostics Suite (Viechtbauer & Cheung 2010)
Run synthesis first
Trial Registration Quality
Run synthesis first. Scores each trial's CT.gov registration completeness.
Outlier Detection
Run synthesis first. Flags studies with extreme residuals, leverage, or effect sizes.
Cook's Distance & DFBETAS
Run synthesis first for detailed influence diagnostics.
Registry-Based Publication Bias Audit (ClinicalTrials.gov)
Run synthesis first
Enrollment-Weighted Selection Model (Sensitivity)
Run synthesis first
Copas Selection Model (Copas & Shi 2000)
Run synthesis first
CT.gov Evidence Delta (Registry/Publication Mismatch Audit)
Run synthesis first
Risk of Bias Traffic Light (Cochrane RoB 2 Style)
Run synthesis first to see RoB domain-level assessment across all trials.
★ GRADE — Summary of Findings
Summary of Findings table will be generated after synthesis and bias assessment. GRADE domains: Risk of Bias, Inconsistency, Indirectness, Imprecision, Publication Bias.
Evidence Convergence Dashboard
Run synthesis to see convergence across all methods and outcomes.
Auto-Generated Methods Paragraph
Run synthesis first to generate methods text.
Patient Risk Calculator
10%
Run synthesis first, then adjust baseline risk slider.
Summary Statistics (All Methods)
Run synthesis first, then click Generate to see all computed statistics in one table.
Multi-Outcome Heatmap
Click "Generate Heatmap" to run synthesis across all outcomes and display a cross-outcome comparison.
Power & Sample Size Calculator
Run synthesis first to see how many more patients/trials are needed for a definitive answer.
Decision Regret Analysis
Run synthesis first. Quantifies expected regret of acting vs waiting based on current evidence strength.
Full Manuscript Draft (Methods + Results)
Run synthesis for all outcomes, then click "Generate Full Draft" to produce complete Methods and Results sections suitable for journal submission.
🔒 TruthCert — Proof-Carrying Bundle
Bundle Status
Complete all previous steps (search, screen, extract, synthesize, bias, GRADE) to generate a TruthCert bundle.
Content Hash
Awaiting bundle generation...
Audit Trail
No audit events yet. Actions are logged as you use the app.
Share Analysis (URL)
Auto-Save Snapshots
Full Methodological Report (PRISMA 2020)
Run synthesis first. Then click the button below to generate a complete PRISMA 2020 compliant methods report.
PRISMA 2020 Flow Diagram
Run synthesis and click "PRISMA 2020 Flow (SVG)" to generate a PRISMA 2020 compliant flow diagram.
WebR: Validate with R in Your Browser
Not loaded
Gold-Standard Verification: This loads R + the metafor package directly in your browser using WebR (WebAssembly), then runs rma() — the same method used in Cochrane systematic reviews — and compares with LivingMeta.
First load: ~20-40 seconds (downloads R runtime + metafor package, cached after first load).
No data leaves your browser. All computation is local.
R

Click "Load R & Validate" to run an independent R meta-analysis on your current dataset

Requires: synthesis completed in the Synthesize tab. Uses metafor::rma() (DL + REML + HKSJ).

Alternative: Click "R Validation Script" above to download a standalone .R file, then run it in your local R/RStudio installation or paste into webr.r-wasm.org.