RWDExchange — Real-World Data Exchangeability Assessment
Demo data — your own assessments are stored separately and stay untouched.
BoyceLab · Clinical Informatics

RWDExchange

Evaluate real-world data exchangeability for use as external comparators in clinical trials.

v4.0

Step 1 of 4

Variable Assessment

Evaluate each candidate variable with six structured exchangeability questions. Add data quality metrics and confounders using the collapsible sections below.

Add New Variable
Exchangeability Questions
How these are scored
YesFull credit (1 point).
PartiallyHalf credit (0.5 point).
No / UnknownNo credit (0 points).
The six answers sum to a 0–6 feasibility score.
📊 Data Quality Metrics (optional)
🔗 Confounders & Analytical Approach (optional)
Assessed Variables
📋
No variables yet
Add your first variable using the form above to start assessing exchangeability. New here? The Demo tab loads a complete synthetic example you can explore.

Step 2 of 4

Pocock's Criteria

Evaluate the seven criteria for acceptable use of historical controls (Pocock, 1976).

Reference: Pocock SJ. The combination of randomized and historical controls in clinical trials. J Chronic Dis. 1976;29(3):175-88. doi: 10.1016/0021-9681(76)90044-8

Select Variable
🔬
Add variables first

Step 3 of 4

FDA Guidance Assessment

Evaluate alignment with the FDA 2023 guidance on externally controlled trials.

Reference: US FDA. "Considerations for the Design and Conduct of Externally Controlled Trials." February 2023. View guidance ↗

Select Variable
📄
Add variables first

Step 4 of 4

Gray et al. Framework

Eight methodological domains for evidence assessment in studies using external comparators from real-world data.

Reference: Gray CM, Grimson F, Layton D, Pocock S, Kim J. A Framework for Methodological Choice and Evidence Assessment for Studies Using External Comparators from Real-World Data. Drug Saf. 2020;43(7):623-633. doi: 10.1007/s40264-020-00944-1

Select Variable
📐
Add variables first

Study-Level

Confounder Inventory

Document each confounder once, at the study level: whether it is available in the RWD source, how well it is measured, and whether it enters the analytic model. Unavailable or poorly measured confounders are the central threat to a valid external comparator.

Add Confounder
Confounder Table
🔗
No confounders yet
Add the covariates you plan to balance between the trial and RWD arms. Availability and measurement quality drive residual confounding risk.

Export & Summary

Assessment Report

0
Variables
0
Strong Potential
0
Conditional
0
Limited Potential
0
Fully Assessed
Study Details

Used in the exported report header and the study-level summary.

Scoring Weights

Adjust how each framework contributes to the composite verdict. Weights normalize to 100% automatically, so verdicts stay transparent and defensible. All verdicts update live.

20%
27%
27%
26%
Study-Level Summary

Add variables to generate a summary.

Composite Verdicts
📊
No variables yet
Export

Download individual sections or the full combined report. JSON export preserves all data for re-import or sharing.

User Guide

How To Use RWDExchange

About This Tool

RWDExchange provides a structured, multi-framework assessment of whether real-world data can serve as a reliable external comparator in a clinical trial. It implements four complementary frameworks: a variable-level feasibility score, Pocock's seven historical control criteria, FDA 2023 guidance for externally controlled trials, and the Gray et al. (2020) methodological framework — synthesized into a single composite verdict per variable.

All data is stored in your browser via localStorage. Use Export JSON to back up and Share to generate a shareable URL. Open the Demo tab to explore the tool pre-loaded with a synthetic ALS external-comparator study — on a separate namespace, so it never touches your own data.

Steps
01

Variable Assessment

Add each variable. Answer six exchangeability questions for a 0–6 score. Optionally add data quality metrics (N, % missing, date range) and planned confounders.

02

Pocock's Criteria

Rate each of the seven Pocock criteria. The composite verdict ring updates live as you save.

03

FDA Guidance

Rate alignment with each of the eight 2023 FDA considerations.

04

Gray et al. Framework

Complete the eight methodological domain assessment covering research question fit, population representativeness, confounders, analytical approach, and reproducibility.

05

OMOP Mapping optional

Anchor each variable to OMOP standard concepts (domain, concept_id, vocabulary). Use the Athena link to look concepts up. Coverage appears in the variables table and the report.

06

Confounder Inventory

Document each confounder once: availability in the RWD source, measurement quality, and whether it enters the analytic model. This is the clearest signal of residual confounding risk.

07

Report & Export

Adjust framework weights (they normalize to 100%), review the auto-drafted study-level summary, then generate a print-ready PDF report or download CSV / JSON. Use Share to encode your full session in a URL.

Verdict Interpretation
🟢 STRONG (≥70%)

Variable is well-suited as an external comparator. Proceed to study design.

🟡 CONDITIONAL (45–69%)

Usable with documented caveats and sensitivity analyses.

🔴 LIMITED (<45%)

Significant limitations. Consider excluding or flagging as a major uncertainty.

⚪ PRELIMINARY

Variable score only. Complete all three frameworks for a full verdict.

Standardization

OMOP Concept Mapping

Anchor each assessed variable to OMOP CDM standard concepts. Documenting the domain, standard concept_id(s), and vocabulary makes your feasibility assessment reproducible and ready for federated / network characterization.

Tip: Use the Athena button to find standard concepts, then record the concept_id, name, vocabulary, and whether it is a Standard concept. Coverage is summarized per variable and in the report.

Select Variable
🔭
Add variables first
Map each assessed variable to OMOP standard concepts here.

Guided Example

Interactive Demo

Load a fully worked synthetic ALS external-comparator assessment to explore every feature. Demo data lives in a separate namespace — your real assessments are never touched, and you can exit anytime.