Analytics Deep Dive

Scale Down Model (SDM) for Bioprocess: Equivalence Acceptable Criteria (EAC)
/ Practical Significant Difference (PSD) & Sample Size Calculation

Determining the right equivalence acceptance criteria and statistically valid sample size is one of the most critical — and often underestimated — challenges in scale-down model qualification.

GT Graphtal Analytics Apr 18, 2026 9 min read

Determining the right equivalence acceptance criteria and statistically valid sample size is one of the most critical — and often underestimated — challenges in scale-down model qualification. This article explains how to calculate EAC/PSD from reference manufacturing data, how to interpret the sample size formula, and how Graphtal supports your biopharma team through this process.

Understanding EAC, PSD and Equivalence Margin

Before calculating, it's important to understand what each term means and how they relate to SDM qualification decisions.

EAC — Equivalence Acceptance Criteria

The statistically derived boundary within which the 90% confidence interval for the difference in means must fall for the SDM to be declared equivalent. EAC is derived from manufacturing batch data.

PSD — Practical Significant Difference

The minimum difference between scale-down and commercial scale that is considered practically meaningful. PSD defines the equivalence margin in terms of real-world process impact.

Equivalence Margin

Derived from reference (manufacturing scale) data. A wider margin means more tolerance for difference; a narrower margin demands closer alignment between SDM and commercial scale performance.

Reference Data

Manufacturing scale batch data forms the reference set. The number of available batches directly determines how many standard deviations (k) should be used — a critical input for sample size.

Calculating EAC / PSD from Manufacturing Scale Data

The sample size formula provides the mathematical foundation for determining how many reference batches are required to achieve a given statistical power at a specified confidence interval.

Sample Size Formula
N =
2 (Z1−α/2 + Z1−β)2
k2
Variable Definitions
  • α Level of confidence (Type I error rate) — typically 0.05 for 95% CI or 90% CI
  • 1−β Statistical power — probability of correctly detecting equivalence (e.g., 0.90 or 0.95)
  • N Sample size — number of reference manufacturing batches required
  • k Factor defining number of standard deviations used as equivalence margin (PSD)
  • Z Z-score corresponding to specified probability under a normal distribution

How to use this formula: Given your available manufacturing batches (N), solve for k to determine how many standard deviations you can use as equivalence margin. A larger k (wider margin) requires fewer batches; a smaller k (tighter margin) requires more batches.

Sample Size Calculator — Example Table

The table below illustrates calculated sample sizes across different combinations of confidence interval, power, and standard deviation factor (k).

CI α Power 1−(α/2) Z at 1−(α/2) Z at (1−β) Z Sum Z Sum² SD (k) Sample Size (N)
95% CI0.050.950.9751.961.643.612.96125.92
0.050.950.9751.961.643.612.9626.48
0.050.950.9751.961.643.612.9632.88
90% CI0.050.90.9751.961.283.2410.49120.99
0.050.90.9751.961.283.2410.4925.25
0.050.90.9751.961.283.2410.4932.33

How to Interpret the Calculator Results

The number of available manufacturing batches determines the appropriate standard deviation factor (k):

≥ 26

If you have 26+ batches of data, use k = 1 (1 Standard Deviation) — the tightest and most rigorous criterion.

≥ 6

If you have 6+ batches of data, use k = 2 (2 Standard Deviations) — a moderate criterion for mid-phase.

≥ 3

If you have 3+ batches of data, use k = 3 (3 Standard Deviations) — for early-stage or limited data scenarios.

Key insight: As k increases (wider margin), fewer reference batches are needed. This trade-off must be scientifically justified for regulatory submission.

Step-by-Step EAC Calculation Process

1

Compile manufacturing batch data

Gather all available at-scale manufacturing data. The quantity of batches determines your k factor selection.

2

Select confidence interval and power

Define α and 1−β. For SDM qualification, 90% CI (α = 0.05, power = 0.90) is standard TOST guidance.

3

Determine appropriate k

Match your batch count to k = 1 (≥26), k = 2 (≥6), or k = 3 (≥3).

4

Calculate EAC

Apply the formula to derive the sample size and define the EAC boundary based on manufacturing standard deviation × k.

5

Document Justification

Use derived EAC in TOST and provide clear scientific rationale for the choice of k and batch count.

Graphtal's Offering for Bioprocess SDM Qualification

Graphtal offers advanced analytics solutions specifically for mAb development and scale-down model qualification.

TOST & Equivalence Margin Setting

Expert support in deriving statistically robust EAC boundaries from manufacturing datasets for TOST validation.

MVDA Integration

Using MVDA to identify patterns and qualify models across multiple attributes simultaneously for improved accuracy.

Customised Calculations

Tailored sample size and EAC calculations adapted to your specific data constraints and regulatory context.

Ready to calculate your EAC?

Our bioprocess statisticians help you define statistically valid equivalence boundaries.

Get in Touch →
Go Back

Tell Us About Your Project

Our team of experienced professionals at Graphtal is ready to transform your project from idea to reality, ensuring alignment with your organisation goals.


Let's simplify your work

Want to make Insightful Analytics for Smarter Decisions?

Let's connect.