Since a long-term, randomized trial for romiplostim is impractical, the proposed next step is to conduct large-scale, retrospective epidemiological studies. This approach would compare outcomes in matched populations of patients who did and did not receive the drug, representing a pragmatic shift in research strategy.

Establishing causation for a complex societal issue requires more than a single data set. The best approach is to build a "collage of evidence." This involves finding natural experiments—like states that enacted a policy before a national ruling—to test the hypothesis under different conditions and strengthen the causal claim.

Instead of the high-risk approach of replacing a trial's control arm with digital twins, Unlearn.ai adds counterfactual data to every participant. This method increases a trial's statistical power, allowing for smaller control arms or a higher chance of success, while satisfying regulatory constraints for pivotal trials.

The FDA's reversal on Unicure's Huntington's therapy re-validates using natural history data as a control in rare neuro diseases. This is critical for indications where placebo-controlled trials, especially those involving invasive surgery, are ethically and logistically challenging, providing a clearer path forward for similar programs.

The podcast highlights propensity score matching, a statistical method creating a comparable control group from large observational datasets like Enroll HD. This is an established, FDA-approved method for rare diseases where placebos are unethical or impractical, yet the agency rejected its pre-specified use in uniQure's case.

The medical field is so culturally ingrained with the belief that only randomized trials can prove causation that journals actively remove causal language (e.g., "X caused Y") from studies using observational data, even from rigorous natural experiments. This belief hinders the adoption of these valuable methods.

The key public health failure during the pandemic was not initial uncertainty, but the systemic inability to execute rapid experiments. Basic, knowable questions about transmission, masks, and safe distances went unanswered because of a failure to generate data through randomized trials.

To generate reliable findings from real-world data, researchers must avoid data dredging. The best practice is to simulate a 'target trial' by creating a formal protocol with pre-defined inclusion criteria and a statistical plan, mirroring the rigor of a prospective clinical trial. This approach is even guided by the FDA.

To de-risk its EMERALD trial for a poorly documented patient population, Resolution Therapeutics first ran a natural history study (OPOL). This provided crucial data to inform the trial protocol and, more importantly, allowed the creation of a matched external control arm, a clever and capital-efficient strategy.