The Effective Statistician - in association with PSI Real-World Evidence vs. Randomized Trials: Can We Emulate Accuracy?
5 snips
Jun 16, 2025 Sebastian Schneeweiss, a Harvard Medical School professor specializing in medicine and epidemiology, shares insights on emulating randomized trials with real-world data. He discusses the motivation behind this approach, emphasizing its significance for regulatory-grade evidence. Topics include the challenges of data quality and adherence, use cases for real-world evidence in expanding treatment indications, and the innovative ‘Benchmark-Calibrate-Extrapolate’ strategy. This conversation is a must-listen for those in biostatistics, epidemiology, and healthcare innovation.
AI Snips
Chapters
Transcript
Episode notes
High-Quality Emulation Matches RCTs
- Emulating randomized controlled trials (RCTs) with real-world evidence (RWE) can produce similar findings if design and measurement quality are high.
- Variability in results arises from differences in populations, adherence, and measurement limitations rather than data quality alone.
CAROLINA Trial Predicted by RWE
- The CAROLINA trial's cardiovascular safety findings were predicted six months in advance by RWE emulation.
- The RWE study also anticipated benefits on hypoglycemic events, later confirmed by the trial.
RCT DUPLICATE Enhances RWE Validity
- FDA-funded RCT DUPLICATE project rigorously emulated 37 randomized trials using claims data.
- Successful emulation required pre-registered protocols, reproducible patient selection, and collaboration with FDA for regulatory-grade evidence.
