JAMAevidence JAMA Guide to Statistics and Methods Target Trial Emulation for Causal Inference From Observational Data With Dr Hernán
13 snips
May 2, 2024 Dr. Miguel A. Hernán, professor of epidemiology at Harvard T.H. Chan School of Public Health, discusses target trial emulation for causal inference from observational data with JAMA Statistical Editor. They explore the concept of target trial emulation, the importance of randomized clinical trials, estimating the effectiveness of Tocilizumab in ICU patients with COVID-19, and the complementary role of randomized trials and observational studies in generating evidence.
AI Snips
Chapters
Transcript
Episode notes
Hormone Therapy Example
- Observational studies on hormone therapy missed early harm due to delayed follow-up.
- Randomized trials avoid this by starting follow-up at treatment initiation.
Target Trial Emulation
- Target trial emulation helps address causal questions not feasible for randomized trials.
- Many questions are too expensive, untimely, or lack incentives for trials, like comparing existing vaccines.
Target Trial Emulation Steps
- Emulate a hypothetical randomized trial using observational data for causal inference.
- Define the causal question precisely, mirroring a pragmatic randomized trial.