Episode 50, “The Fallacy of Ordinal Endpoints”
Episode 50 of "In the Interim..." is live. Dr. Scott Berry and Dr. Lindsay Berry tackle a critical question: What clinical weights are we really imposing when we analyze ordinal endpoints?
The debate centers on Modified Rankin Scale scores in stroke trials. Critics say utility-weighted analyses impose assumptions not all patients share. But here's the catch: proportional odds models impose weights too, they're just hidden behind the math, driven by prevalence rather than clinical judgment.
Lindsay demonstrates the mathematical equivalence, revealing that every ordinal analysis assigns relative values to clinical states. The question isn't whether to weight outcomes, it's how.
This matters for stroke trials, COVID platform trials, ALS functional scales, and win ratios. Statistical assumptions create clinical weights whether we acknowledge them or not.
Listen to explore the mathematics, the controversy, and why we might need to get more explicit about clinical value assignments. https://www.berryconsultants.com/resource/50-the-fallacy-of-ordinal-endpoints