Thursday 10/10

Non-Inferiority Theory - Day 2 Overview

  • Introduction

    • Focus on practicing and understanding non-inferiority concepts.

    • Outline of topics based on importance for the exam preparation.

    • Assurance that only discussed topics will be on the exam.

Non-Inferiority Margin

  • Definition and Importance

    • The non-inferiority margin represents the threshold for determining whether a new treatment is not much worse than the standard of care.

    • Emphasis on understanding delta, which reflects the minimum effect required to consider the new treatment as non-inferior.

  • Determining Delta

    • Delta calculation involves preserving 50% of the effect of standard care.

    • Example:

      • Risk of stroke without treatment: 3.5%

      • Risk with standard care: -3%

      • Difference: 6.7 percentage points in favor of standard care.

      • Lower bound confidence interval: 4.2 percentage points.

      • Delta (half of minimum effective treatment): 2.1 percentage points.

  • Rationale for Preserving Half

    • Using half is mathematically simpler than other fractional approaches.

FDA Guidelines on Non-Inferiority

  • Guidance on Margins

    • FDA definitions are less strict; they do not dictate a specific percentage for delta but emphasize clinical relevance and judgment.

    • M2: the threshold that new treatments should not cross regarding the standard care.

  • Non-Inferiority Assessment

    • Delta can vary; it should remain less than the calculated minimum (e.g., 4.2% in the stroke example).

    • Importance of ensuring delta is set thoughtfully, facilitating a reliable non-inferiority conclusion.

Comparing Treatments: Drug Examples

  • Assessment of Three Drugs

    • Presented three different drugs with specific estimates against the standard of care.

    • Findings:

      • Drug 3: Crosses M1, indicating no efficacy above no treatment; not considered non-inferior.

      • Drug 2: Falls within M2, potentially non-inferior if within acceptable thresholds.

      • Drug 1: Also non-inferior, does not cross M1; can reject the null hypothesis.

  • Clarifying Non-Inferiority

    • The focus is on showing a treatment is not much worse than the standard, not necessarily better.

    • Emphasis on statistical frameworks to dismiss or accept the hypothesis.

Reading Clinical Papers

  • Recommended Steps for Understanding Research

    • Engage critically with studies by asking specific questions about methodology, results, and implications for practice.