Applying The Infamous Hazard Ratio

Research evidence excerpt

Applying The Infamous Hazard Ratio

Biotechnology

EQUITY RESEARCH Industry Report

July 15, 2026

Research Analysts:

Eric Schmidt In Part 1 of our hazard ratio tutorial (HERE), we deconstructed the Hazard

212-294-7724 Ratio (HR) from the ground up. We explained: Eric.Schmidt@cantor.com

Josh Schimmer ●What a hazard rate is

310-282-6513

Josh.Schimmer@cantor.com ●How a hazard rate is connected to the Kaplan-Meier curve

Yanni Souroutzidis

929-730-2656 ●How those hazard rates are converted into hazard ratios (HRs) through

Yanni.Souroutzidis@cantor.com Cox regression

●When HRs should or shouldn't be adjusted for co-variates

●How patient censoring can be properly and improperly applied to

influence time-to-event analyses

The goal of Part 1 was to make the HR less of a black box and provide

investors with a conceptual toolkit for understanding what lies beneath the

headline number.

In Part 2, we put that toolkit to work. We take the concepts from Part 1 and

apply them to a set of real-world situations that investors often encounter

when interpreting time-to-event data. We organize the report around four

specific questions:

●Is a clinical trial adequately powered to detect a given treatment

effect?

●Do late events carry more weight in the calculation of an HR?

●Should we expect a final HR to look similar to an interim HR?

●Why could a drug's treatment effect change over time and what should

investors do about it?

We begin with a discussion of clinical trial powering. Understanding the

statistical design of a study is critical to interpreting whether a trial is set

up for success.

●We explain how the minimal detectable HR represents the treatment

effect that a trial is designed to detect, while the threshold HR

represents the estimated weakest observed effect (highest HR)