Introduction
A sponsor developing a generic modified-release product was mid-program when the FDA issued updated product-specific guidance that changed the rules of the study. What had been a standard bioequivalence study was no longer acceptable. The FDA now required a non-standard bridging approach built partly on historical data that wasn’t easily found in public documents.
The clock was the problem. Every week spent searching for the right statistical partner was a week lost on the submission timeline.
The Challenge
The FDA issued an updated product-specific guidance (PSG) to the sponsor mid-program while the sponsor was developing a generic modified-release product.
Why? Because the FDA determined that this study was not a “standard bioequivalence (BE) study.”
As such, the FDA required a non-standard bridging approach, one that required combining:
- Historical evidence from the original New Drug Application (NDA) program (publicly available clinical pharmacology data)
- A new relative bioavailability study comparing the test formulation to an intermediate reference
- A bioequivalence conclusion using PSG-defined formulas, including a variance homogeneity decision rule
Concerns
The Practical Problem
The PSG-required historical inputs (mean squared error and degrees of freedom for key endpoints) weren't explicitly tabulated in public documents. And the sponsor had no immediate access to historical subject-level data.
The Risk
Delay. Finding a statistical vendor both flexible enough to take on non-standard work and experienced enough to implement it correctly could take weeks, or months. The sponsor needed statistical analysis experts who could translate the guidance into executable statistics quickly, without rework cycles.
Identifying The Issues
Sample size and methodology uncertainty
Given the new PSG requirements for the bridging study, would the standard two-one-sided test (TOST) method still work with historical data from the original drug? What assumptions could be made? What sample size would ensure adequate power under the new bridging framework?
Historical parameter reconstruction
The required historical MSE and degrees of freedom weren’t tabulated in public NDA documents. A transparent method to reconstruct these from published aggregate data was needed.
Audit trail requirements
The final bioequivalence conclusion needed full traceability from input data to final result, with defensible documentation for regulatory review.
Solution
The scope of work included biostatistics and statistical consulting focused on PSG bridging methodology, statistical simulation for sample size determination under new guidance, mini SAP authoring for guidance-specific computations and analysis planning and documentation for FDA submission
Timeline: Under 2 weeks for FDA correspondence and statistical support to protocol drafting (statistical method and sample size determination)
Statistical Simulation for Sample Size Determination
- Conducted statistical simulations to determine the correct sample size given the historical data constraints and new variance assumptions
- Evaluated whether standard TOST methodology remained valid under the new PSG bridging requirements
- Documented the assumptions and sensitivity analyses for regulatory correspondence
Reconstructable Historical Parameters
- Developed a methodology to obtain historical parameters using the standard log-scale relationship between mean squared error and intra-subject coefficient of variation
- Documented this approach for FDA controlled correspondence with FDA’s confirmation
Mini SAP for PSG-Specific Computations
- Created a focused statistical analysis plan (mini SAP) documenting the data and analysis needed to fill the gap from the new PSG requirements
- Specified the exact inputs required, calculations to be performed, and outputs to be generated
White-Box Transparency Throughout
- Created documentation with formula traceability mapping inputs to derived quantities
- Produced analysis and documentation ready for FDA submission
Results
Figures summarized at a level appropriate for a public case study to protect sponsor confidentiality.
- Sponsor maintained submission timeline: Complete methodology delivered in under two weeks, before the new study completed, allowing the sponsor to stay on schedule despite the mid-program guidance change
- Sponsor proceeded with regulatory confidence: Controlled correspondence with FDA confirmed the proposed approach, including reconstruction of historical parameters from aggregate data, before study start
- Study designed with validated sample size: Statistical simulation confirmed the appropriate sample size under the new bridging framework, avoiding underpowered results or wasted enrollment costs
- Submission package ready without rework: Audit-ready documentation with full traceability suitable for inclusion in the submission dossier
Lesson
When FDA guidance changes mid-program, the bottleneck isn’t usually the science, it’s finding a partner who can translate new requirements into audit-ready execution without a lengthy onboarding cycle.
Free Consultation
- Statistical consulting and simulation for study design and sample size determination
- SAP development with audit-ready documentation
- Regulatory strategy support for complex or non-standard submissions
- Transparent, defensible analyses that teams can explain and defend
