News and Events
February 27, 2026
No items found.

Blog released on “Technical Realities of Ordinal Endpoint Analysis in Clinical Trials”

Play video
No items found.
A rigorous review of ordinal endpoint analyses, showing every approach—utility weighting, proportional odds, dichotomization, or non-parametric—inevitably assigns relative weights to outcome states. Berry Consultants’ mathematical demonstration reveals how proportional odds analysis embeds prevalence-based weights, underscoring the need for transparency and clinical input in trial design.

The fallacy of ordinal outcomes is the belief that analyses can only assume ordinality and be neutral to relative differences. Every approach to ordinal endpoints—proportional odds, dichotomization, nonparametric—imposes value judgments.

Common methods default to hidden weighting driven by data prevalence, not clinical meaning.

Some key points:

• All models impose weights—explicit or hidden.

• Statistical defaults, not clinical teams, often dictate impact.

• Transparent, clinically justified weights increase scientific integrity.

Have you examined the weights imposed in your own analyses? Read the full article to learn more: https://www.berryconsultants.com/resource/technical-realities-of-ordinal-endpoint-analysis-in-clinical-trials

Download PDF
View