CORE Seminar - October 19, 2026 - Dr. James Brophy

Value of the Information & Strength of the Argument

The talk provides significant value by illustrating how Bayesian methods can be applied to real-world clinical questions, offering a concrete alternative to conventional frequentist approaches. The speaker’s argumentation is strong, systematically building from a critique of a major trial to a reanalysis, a new trial design, and a network meta-analysis. He supports his points with visual examples and logical reasoning, addressing criticisms such as subjectivity while emphasizing transparency. The use of multiple priors adds robustness to his conclusions. The presentation is persuasive, but it is primarily a methodological demonstration rather than a fully peer-reviewed study, which slightly limits its immediate scientific impact.

Scientific Rigor, Source Quality, Title Accuracy

The rigor is high: the speaker references the PLATO trial, a well-known study on the immigration data (likely the ‘Many Analysts, One Data Set’ project), and his own unpublished trial. He does not provide formal citations or URLs, but the sources are identifiable in the medical literature. The lack of explicit references in the slides or description reduces the ability to verify claims directly. The title, while generic, is paired with a descriptive abstract that accurately reflects the content. The presentation adheres to the stated learning objectives, making the title-context alignment adequate.

The video title is generic (seminar series), but the description clarifies the actual topic of DAPT and Bayesian methods. The content fully matches the described subject, so the title is adequate.

8/10

The talk is given by a professor of medicine and epidemiology, with clear expertise in Bayesian methods. The content is methodologically sound, well-structured, and transparent about limitations. The speaker provides concrete examples and robust reasoning. However, it is an academic seminar, not peer-reviewed, and relies on personal analysis rather than published consensus.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction to PLATO trial: 18,000 patients, hazard ratio 0.84, p=0.001, but FDA did not approve due to regional differences.
• 06:00Illustration of model dependence: 73 teams analyzed identical data with different results; emphasizes that statistical models are as important as data.
• 12:00Bayesian hierarchical re-analysis of PLATO: pooling across regions expands credible intervals, reducing confidence in the overall benefit.
• 25:00Design of pragmatic cluster-randomized trial using administrative databases; cost per patient $300 versus conventional $33,000.
• 35:00Results with multiple priors: no robust evidence of ticagrelor superiority in North American patients.
• 45:00Bayesian network meta-analysis including prasugrel, demonstrating a broader synthesis of evidence.

• PLATO trial (ticagrelor vs clopidogrel) — Discussed as the primary trial with controversial results and regional heterogeneity.
• Many Analysts, One Data Set (immigration study) — Cited to illustrate that different analysts draw different conclusions from the same data.
• Randomized registry trial using administrative databases (Brophy et al., unpublished) — Presented as the author's own pragmatic trial using RAMQ data.