The Impact of Unblinding on Statistical Analysis Plans in Clinical Trials

Introduction: Why Unblinding Affects the SAP

The Statistical Analysis Plan (SAP) is a regulatory document that pre-defines how trial data will be analyzed. Its credibility relies on the principle of blinding, which prevents bias in decision-making. When unblinding occurs—either at the patient level during emergencies or at the trial level during interim analyses—it can have significant implications for the SAP. Regulators such as the FDA, EMA, and ICH E9 (R1) emphasize that sponsors must carefully manage unblinding’s impact to preserve trial validity and regulatory acceptability.

This tutorial examines how unblinding affects SAPs, what regulators expect, and best practices for integrating unblinding safeguards into trial planning.

How Patient-Level Unblinding Impacts the SAP

Patient-level unblinding occurs in emergencies to guide individual treatment decisions. While necessary for safety, it can affect the SAP in the following ways:

• Data censoring: Analyses may require censoring unblinded subjects from blinded efficacy endpoints.
• Bias risk: Patient knowledge of treatment may influence reporting of subjective outcomes.
• Documentation: SAP must specify how unblinded patient data will be handled in efficacy and safety analyses.
• Regulatory reporting: Each unblinding event must be described in the Clinical Study Report (CSR) and may affect final analyses.

Example: In an oncology trial, emergency unblinding of patients with infusion reactions required exclusion of certain safety outcomes from blinded analysis, as pre-specified in the SAP.

How Trial-Level Unblinding Impacts the SAP

Trial-level unblinding during interim analyses or at final database lock can significantly alter the SAP:

• Interim modifications: SAPs must specify when interim looks occur and how unblinded data is used.
• Adaptations: Changes such as dose arm dropping, futility decisions, or sample size adjustments must be outlined in advance.
• Independent oversight: DSMBs typically access unblinded data, while sponsors remain blinded.
• Error control: SAP must include statistical safeguards to preserve Type I error across interim looks.

Illustration: In a vaccine trial, the SAP defined Bayesian predictive probabilities for interim unblinded data review, with final modifications documented in the CSR.

Regulatory Expectations on SAP and Unblinding

Agencies require SAPs to be explicit about unblinding:

• FDA: SAPs must define how unblinded data will be incorporated, censored, or adjusted in analyses.
• EMA: Requires SAPs to include charters and SOP references for trial-level unblinding oversight.
• ICH E9 (R1): Emphasizes estimand strategies that account for unblinded events.
• MHRA: Inspects TMFs for SAP amendments and unblinding justifications.

Example: EMA required revisions to a cardiovascular trial SAP after interim unblinding raised concerns about multiplicity control.

Case Studies of Unblinding Impact on SAP

Case Study 1 – Oncology Trial: Emergency unblinding of multiple patients for toxicity management required SAP adjustments to exclude affected efficacy endpoints. FDA inspectors confirmed compliance.

Case Study 2 – Vaccine Development: Interim unblinding for dose selection required DSMB oversight. SAP simulations were adjusted to maintain Type I error control, which EMA validated during inspection.

Case Study 3 – Rare Disease Therapy: MHRA identified gaps in SAP handling of unblinded data. CAPAs were required, including SOP revisions and SAP amendments.

Challenges in Managing SAP Unblinding Impacts

Sponsors face challenges in ensuring SAPs remain robust despite unblinding:

• Complexity: Adaptive designs introduce multiple interim unblinding points requiring simulations.
• Documentation burden: SAPs must integrate with SOPs, DSMB charters, and TMF entries.
• Regulatory variability: FDA, EMA, and PMDA differ in their expectations for SAP handling of unblinded data.
• Bias mitigation: Ensuring investigators remain blinded while statisticians access unblinded data is operationally difficult.

Illustration: In a multi-regional cardiovascular trial, inconsistent SAP documentation of unblinded data handling led to EMA requiring additional simulations before approval.

Best Practices for Sponsors

To align with regulatory expectations, sponsors should:

• Pre-specify unblinding handling strategies within SAPs.
• Ensure SAPs are version-controlled and integrated into TMFs.
• Embed statistical simulations demonstrating Type I error preservation under unblinding conditions.
• Coordinate SAP updates with DSMB charters and SOPs.
• Train statisticians and data managers on SAP unblinding procedures.

One oncology sponsor embedded an “unblinding appendix” in their SAP, clarifying how subject-level and trial-level events would be managed. Regulators praised the transparency during inspections.

Ethical and Regulatory Consequences of Weak SAP Integration

Improper handling of unblinding within SAPs can lead to:

• Regulatory findings: FDA or EMA may issue critical observations for vague or missing unblinding provisions.
• Data integrity risks: Biased analyses may undermine trial conclusions.
• Ethical issues: Patient safety may be compromised if unblinding events are not appropriately analyzed.
• Reputational damage: Scientific credibility may suffer if results are questioned.

Key Takeaways

The SAP is deeply affected by unblinding events. Sponsors must:

• Pre-specify both patient-level and trial-level unblinding management strategies.
• Ensure SAPs integrate with DSMB charters, SOPs, and TMFs.
• Document and archive all unblinding-related SAP changes for regulatory inspection.
• Conduct simulations and sensitivity analyses to safeguard statistical validity.

By following these steps, sponsors can ensure SAPs remain credible and regulatory-compliant, even when unblinding occurs during trial execution.

Integrating DSM Plans with Statistical Analysis Plans in Clinical Trials

Introduction: Why Integration Matters

In clinical trials, interim analyses are governed by two critical documents: the Data and Safety Monitoring (DSM) plan and the Statistical Analysis Plan (SAP). While the DSM plan focuses on oversight, safety, and operational procedures, the SAP details statistical methodologies, including stopping thresholds for efficacy, futility, and safety. If these documents are not harmonized, inconsistencies can create confusion for Data Monitoring Committees (DMCs), undermine trial integrity, and trigger regulatory findings. Agencies such as the FDA, EMA, and ICH E9 stress the importance of aligning DSM and SAP documents to ensure transparency, error control, and ethical oversight.

This tutorial explains how DSM plans should be integrated with SAPs, providing step-by-step guidance, examples, and case studies from oncology, cardiovascular, and vaccine trials.

Regulatory Requirements for Integration

Regulators expect clear linkage between DSM and SAP documents:

• FDA: Requires DSM plans to reference SAP-defined stopping rules and document how DMCs apply them.
• EMA: Expects DSM plans, SAPs, and DMC charters to be consistent; discrepancies may be cited during inspections.
• ICH E9: Emphasizes that interim analyses must be pre-specified and documented in both operational and statistical frameworks.
• WHO: Advises harmonization of monitoring and statistical oversight, especially in multi-country vaccine trials.

For example, during an EMA inspection, one oncology sponsor was cited for inconsistent futility definitions between the DSM plan and SAP, requiring corrective action.

Key Components of a DSM Plan

The DSM plan typically includes:

• Roles and responsibilities: Defines DMC membership, independence, and scope of oversight.
• Meeting frequency: Specifies how often interim reviews occur.
• Safety reporting: Describes how adverse events and safety signals are monitored.
• Stopping rule framework: References thresholds that trigger DMC consideration.
• Communication pathways: Details how recommendations are relayed to sponsors and sites.

The SAP, in contrast, provides the statistical details of boundaries, error spending, and conditional power calculations.

How to Align DSM and SAP Documents

Integration requires cross-referencing and consistent terminology:

1. Cross-reference stopping rules: DSM plan should cite SAP-defined boundaries (e.g., O’Brien–Fleming thresholds).
2. Synchronize timing: Both documents should use identical information fractions and interim analysis points.
3. Align language: Terminology for efficacy, futility, and safety rules must match across documents.
4. Document communication: DSM plan should explain how SAP results are shared with the DMC.
5. Archive consistency: All versions should be filed in the Trial Master File (TMF) with cross-referenced version control.

Illustration: A vaccine program ensured alignment by appending SAP stopping rules to the DSM plan, which regulators praised for transparency.

Case Studies in DSM-SAP Integration

Case Study 1 – Oncology Trial: A futility rule was described in the SAP as conditional power <15%, but the DSM plan cited <20%. Regulators flagged this as inconsistent, requiring immediate harmonization.

Case Study 2 – Cardiovascular Program: The DSM plan referenced O’Brien–Fleming rules, while the SAP specified Lan-DeMets spending. FDA reviewers questioned the discrepancy, delaying approval until corrected.

Case Study 3 – Vaccine Trial: SAP and DSM plan were fully harmonized, with appendices showing simulations. This alignment allowed rapid FDA and EMA acceptance of interim stopping decisions during a pandemic.

Challenges in Integration

Common challenges include:

• Multiple authorship: DSM plans and SAPs are often written by different teams, leading to misalignment.
• Frequent amendments: Adaptive trials may require updates to both documents simultaneously.
• Regulatory differences: FDA and EMA may have different expectations for level of detail.
• Operational timing: DSM plans may reference meeting schedules that don’t align with SAP event-driven looks.

For example, in a global cardiovascular outcomes trial, amendments to the SAP were not reflected in the DSM plan, creating confusion for DMC members during review.

Best Practices for Sponsors

To avoid inconsistencies and regulatory findings, sponsors should:

• Draft DSM and SAP documents collaboratively, with cross-functional teams.
• Use consistent statistical thresholds and terminology across both plans.
• Maintain version control logs to track updates across documents.
• Append SAP excerpts directly into DSM plans where possible.
• Ensure DMC training includes review of both documents side by side.

One sponsor implemented an integrated SAP-DSM master document that combined statistical and operational oversight. Regulators cited this as a model of best practice.

Regulatory and Ethical Consequences of Misalignment

If DSM plans and SAPs are not aligned, sponsors risk:

• Regulatory citations: FDA or EMA may classify inconsistencies as major findings.
• Trial delays: Misaligned documents can confuse DMCs and delay interim decisions.
• Ethical risks: Participants may face harm if safety stopping rules are misinterpreted.
• Loss of credibility: Sponsors may appear disorganized or noncompliant during audits.

Key Takeaways

Integrating DSM plans with SAPs is essential for consistent and transparent trial monitoring. To ensure success, sponsors should:

• Cross-reference and harmonize stopping rules in both documents.
• Align timing, language, and thresholds across SAPs and DSM plans.
• Document and archive integration in the TMF for inspection readiness.
• Adopt collaborative drafting and training approaches for teams and DMCs.

By embedding these practices, sponsors can ensure that interim analyses are scientifically rigorous, ethically sound, and regulatorily compliant.

Statistical Analysis Plan (SAP) Considerations for Interim Analysis in Clinical Trials

The Statistical Analysis Plan (SAP) is a foundational document in clinical trials, outlining all statistical methodologies, endpoints, and data handling rules. When an interim analysis is planned, the SAP must provide specific, regulatory-compliant guidance on how these analyses are conducted, interpreted, and used to make decisions. The integrity of the trial and its acceptability by regulatory agencies like the USFDA or EMA often hinges on how well interim analyses are pre-specified in the SAP.

This article provides a detailed tutorial for pharma and clinical trial professionals on structuring SAP content for interim analysis, covering statistical methodology, firewalls, data access, adaptation, and documentation strategies.

Why the SAP Is Critical for Interim Analysis

Interim analysis involves reviewing accumulating data while the trial is ongoing. Without a predefined plan, such reviews can introduce bias, inflate Type I error, or violate ethical and regulatory standards.

Including detailed interim analysis strategies in the SAP ensures:

• Prevention of operational bias
• Protection of statistical integrity
• Clear decision-making rules for DMCs
• Transparency with regulatory bodies

Key Elements of Interim Analysis in the SAP

The SAP must address several key areas when interim analyses are planned:

1. Timing and Number of Interim Analyses

• Specify the number and timing of planned interim looks (e.g., after 50% of events)
• Define event triggers or calendar-based schedules
• Ensure consistency with protocol and GMP SOP documentation

2. Purpose and Type of Interim Analyses

• Is the goal safety monitoring, futility assessment, efficacy determination, or adaptive design modifications?
• State whether the analysis is blinded or unblinded
• Clarify whether the analysis is binding or non-binding

3. Statistical Methods and Boundaries

• Describe alpha-spending functions (e.g., O’Brien-Fleming, Pocock)
• State efficacy and futility thresholds
• Include conditional or predictive power calculations
• Mention simulation assumptions to justify boundary selection

4. Data Handling Procedures

• Explain data cut-off procedures for interim analysis
• Define derived variables, imputation strategies, and analysis sets (e.g., ITT, PP)
• Clarify treatment of missing or censored data

5. Firewalls and Blinding

• Specify who will conduct the interim analysis (typically a firewall statistician)
• Ensure operational teams remain blinded to treatment assignments
• State how interim data will be protected using access controls and firewall SOPs
• Detail the format of DMC communications (e.g., blinded vs unblinded summaries)

6. Decision-Making Criteria

• Clearly state under what conditions the trial will be stopped or modified
• Differentiate between DMC recommendations and sponsor actions
• Link interim decisions to predefined adaptive rules if applicable

7. Documentation and Version Control

• Maintain a dated version history of the SAP
• Document any SAP updates with justification and approval logs
• Include the SAP in the Trial Master File (TMF)

Special Considerations for Adaptive Trial SAPs

For adaptive designs, the SAP must also include:

• Pre-specified adaptation strategies (e.g., sample size re-estimation)
• Modeling and simulation reports showing error control
• Independent decision rules triggered by interim data
• Clear description of how operational bias will be minimized

Tools such as EAST, ADDPLAN, or R packages like gsDesign are commonly referenced for simulation validation.

FDA and EMA Expectations for Interim SAPs

FDA:

• Expects the SAP to be finalized before database lock or interim data unblinding
• May request simulation reports as part of IND or NDA submissions
• Requires justification for any protocol-SAP inconsistencies

EMA:

• Stresses pre-specification of interim boundaries and stopping logic
• Encourages inclusion of the DMC charter and SAP in submission dossiers
• Reviews SAP updates in the context of trial integrity

Failing to meet these expectations may delay approvals or require resubmission with additional justification.

Case Study: Interim SAP in an Oncology Trial

In a Phase III breast cancer trial, the SAP outlined a single interim analysis after 60% of PFS events. The SAP included O’Brien-Fleming boundaries, a detailed DMC communication flowchart, and firewalled team responsibilities. Conditional power and simulation outputs were attached as appendices. During NDA review, the FDA found the SAP acceptable and approved the data cut-off strategy and interim analysis results.

Best Practices for Interim SAP Development

1. Start SAP development early, aligned with protocol design
2. Engage statisticians experienced in adaptive and interim analysis
3. Include DMC charter elements as reference
4. Perform trial simulations to validate operating characteristics
5. Ensure cross-functional review (medical, regulatory, QA)
6. Maintain version control and transparent change logs
7. Submit SAP with protocol to regulatory bodies if required

Conclusion: Interim SAP Planning Is Crucial to Trial Success

A well-crafted SAP not only guides sound statistical analysis but also builds credibility with regulators. When interim analyses are involved, the SAP becomes a critical safeguard against bias and misinterpretation. By including clear methods, decision criteria, firewall processes, and regulatory documentation, sponsors can ensure that interim analyses contribute meaningfully to trial oversight while maintaining full compliance.

Essential Sections in a Clinical Trial Protocol: A Step-by-Step Writing Guide

A well-written clinical trial protocol is the foundation for ethical, regulatory-compliant, and scientifically robust studies. It outlines every aspect of a clinical trial, ensuring that all stakeholders—from investigators and sponsors to regulators—are aligned. This tutorial explains each key section of a clinical trial protocol, providing practical writing guidance for professionals engaged in drug development and research documentation.

Understanding how to structure and draft the protocol not only satisfies regulatory agencies like the EMA but also ensures operational efficiency, risk mitigation, and subject protection.

Introduction and General Information:

Every clinical trial protocol should start with a clear title page and introductory section. This area typically includes:

• Protocol Title: Full, descriptive name of the study including study number and investigational product name.
• Protocol Number and Version: Ensure version control is properly tracked.
• Sponsor Details: Organization name, address, and primary contact.
• Confidentiality Statement: Optional legal language asserting proprietary content.

This section sets the tone and provides traceability throughout the trial lifecycle. As per GMP documentation principles, maintaining consistency in protocol identification is critical during audits and inspections.

Background and Rationale:

This section outlines the scientific and medical basis of the study. Include:

• Current disease burden and unmet need
• Mechanism of action of the investigational product
• Summary of preclinical and clinical data
• Justification for dose selection, route, and regimen

This section must logically lead to the objectives and design. Ensure that references to prior studies or Stability Studies are cited when relevant to justify safety or formulation assumptions.

Study Objectives and Endpoints:

Clearly define:

• Primary Objective: The main scientific question being answered
• Secondary Objectives: Supporting outcomes that provide context or safety data
• Exploratory Objectives: Optional biomarker or pharmacogenomic insights

List endpoints directly tied to these objectives. For instance, if your primary objective is to evaluate efficacy, the primary endpoint may be a change from baseline in a validated clinical scale.

Study Design:

This is a critical section describing how the trial is conducted. It should include:

• Type of study (randomized, blinded, parallel, crossover)
• Randomization methods and stratification criteria
• Blinding techniques (single, double, open-label)
• Control arms (placebo, active comparator, none)
• Estimated study duration
• Trial flow diagram (SPIRIT-compliant)

Design should align with your validation master plan and regulatory requirements to ensure scientific rigor and ethical acceptability.

Eligibility Criteria:

Eligibility defines who can and cannot participate:

• Inclusion Criteria: Clearly defined patient attributes
• Exclusion Criteria: Risk minimization for safety or confounding

Ensure each criterion is justified and feasible to screen within your chosen clinical setting.

Study Treatments and Administration:

This section details investigational product usage:

• Product name, dosage form, strength, and route
• Dosing schedule and titration rules
• Packaging, labeling, and accountability
• Storage and stability (include reference to SOP validation in pharma)

Include rescue medications and prohibited drugs if applicable.

Assessment Schedule:

Use a standardized Schedule of Assessments (SoA) table. It should include:

• Visit windows
• Timing of assessments
• Lab tests, imaging, ECG, and other procedures

Ensure all assessments align with endpoint definitions and regulatory expectations.

Safety and Adverse Event Monitoring:

Clearly describe:

• AE/SAE definitions and reporting windows
• Role of investigators in causality assessment
• Stopping rules and safety review committees

This section is critical for drug regulatory compliance and must be harmonized with your global safety strategy.

Statistical Considerations:

• Sample size calculation with assumptions
• Statistical hypothesis and test methods
• Interim analyses and stopping boundaries
• Analysis populations (ITT, PP, Safety)
• Missing data handling

The SAP (Statistical Analysis Plan) is typically a standalone document but should be summarized here.

Data Management and Record Keeping:

• Use of EDC or paper CRFs
• Data query processes
• Audit trails and version control
• Archival timelines

Comply with GMP quality control and ALCOA+ principles.

Monitoring, Audits, and Protocol Deviations:

This section defines how quality oversight is maintained:

• Monitoring plans and CRA responsibilities
• Audit preparedness and escalation pathways
• Deviation management and reporting

Ensure alignment with your broader Stability testing or product lifecycle monitoring strategy if applicable.

Ethical Considerations and Informed Consent:

Describe the consent process, including:

• Timing and documentation
• Languages and literacy levels
• Witness requirements for vulnerable subjects
• IRB/IEC submission and renewal timelines

Publication and Data Disclosure:

Define who owns the data, how results will be disseminated, and how trial registration and transparency are ensured (e.g., ClinicalTrials.gov).

Conclusion:

Writing a clinical trial protocol requires attention to detail, regulatory knowledge, and clear scientific articulation. This structured guide ensures that you include all essential elements, minimizing ambiguity and facilitating compliance, quality, and reproducibility. By following best practices, you enable all stakeholders—from site investigators to regulatory reviewers—to operate with clarity and confidence.

Essential Components of a Statistical Analysis Plan (SAP) for Clinical Trials

The Statistical Analysis Plan (SAP) is a cornerstone document in any clinical trial. It outlines the methodology and statistical approaches that will be used to analyze trial data, and serves as the blueprint for transforming raw data into clinical evidence. A well-written SAP ensures transparency, reproducibility, and regulatory compliance.

This guide offers a step-by-step breakdown of what should be included in an SAP, why each component matters, and how to align it with protocol objectives and regulatory expectations.

What Is a Statistical Analysis Plan (SAP)?

An SAP is a detailed, standalone document that supplements the clinical trial protocol. It defines the statistical techniques, models, and outputs that will be used to analyze primary and secondary endpoints, safety data, and exploratory objectives. According to USFDA and ICH E9 guidelines, the SAP should be finalized before database lock and unblinding of data.

It is essential for regulatory submissions, Clinical Study Reports (CSRs), and publication of trial results.

Why a Comprehensive SAP Matters

• Ensures consistent and objective analysis of data
• Prevents post-hoc manipulation or data dredging
• Facilitates regulatory review and approval processes
• Supports reproducibility of findings
• Serves as a roadmap for biostatistical programming and validation

A clear SAP also aligns biostatistics teams, sponsors, and regulatory bodies, making it indispensable in evidence generation.

Core Sections of a Statistical Analysis Plan

While formats may vary, these key sections are generally expected in any SAP:

1. Title Page and Document History

• Study title, protocol number, version, and dates
• Sponsor and CRO contact details
• Document revision history and approvals

2. Introduction and Study Objectives

• Brief background of the trial
• Primary, secondary, and exploratory objectives

This section connects the SAP to the protocol and Clinical Development Plan (CDP).

3. Study Design Overview

• Type of trial (e.g., randomized, double-blind)
• Treatment arms, duration, and study flow diagram

4. Analysis Populations

• Definitions of ITT, per-protocol, safety, and modified ITT populations
• Inclusion/exclusion rules for each population

5. Endpoints and Variables

• Clearly defined primary, secondary, and exploratory endpoints
• Derived variables, scoring algorithms, and coding dictionaries (e.g., MedDRA, WHO Drug)

6. Statistical Hypotheses

• Null and alternative hypotheses for each endpoint
• Superiority, non-inferiority, or equivalence assumptions

7. Sample Size Justification

• Power calculations and assumptions
• Effect size, alpha level, dropout rate
• References to sample size simulations or literature

8. Randomization and Blinding

• Randomization method (e.g., stratified block)
• Unblinding procedures and roles involved

This aligns with data integrity expectations in clinical data management.

9. General Statistical Methods

• Types of statistical tests (e.g., ANCOVA, logistic regression)
• Handling of missing data (e.g., LOCF, multiple imputation)
• Adjustments for multiplicity

10. Interim Analysis and Stopping Rules

• Timing, scope, and methodology of interim analysis
• Data Monitoring Committee (DMC) responsibilities
• Statistical boundaries (e.g., O’Brien-Fleming)

11. Subgroup and Sensitivity Analyses

• Predefined subgroup analyses (e.g., age, gender)
• Sensitivity checks for model robustness

12. Safety and Tolerability Analysis

• Adverse events (AEs) and serious adverse events (SAEs)
• Laboratory, ECG, vital signs, and physical exams
• Incidence, severity, and relatedness summaries

13. Statistical Software and Validation

• List of statistical software and versions used (e.g., SAS, R)
• Details of programming validation and code review

Documenting tools ensures compliance with computer system validation standards.

14. Mock Tables, Listings, and Figures (TLFs)

• Annotated mock outputs for key endpoints
• Layout, structure, and footnotes for each TLF

15. References and Appendices

• Citations to published methods, previous trials, or regulatory guidance
• Appendices for SAP templates, derivation rules, or shell displays

Best Practices for Writing a Statistical Analysis Plan

1. Involve Biostatisticians Early: Collaborate during protocol development
2. Use SAP Templates: Standardize across studies for quality and efficiency
3. Document Assumptions: Clearly state all statistical assumptions and rationale
4. Maintain Version Control: Track changes and approvals systematically
5. Ensure Review by All Stakeholders: Clinical, data management, regulatory, and QA teams

Regulatory Guidance for SAPs

Key guidelines that shape SAP development include:

• ICH E9 – Statistical Principles for Clinical Trials
• FDA’s Guidance for Industry on Statistical Aspects of Clinical Trials
• EMA Guideline on Adjustment for Covariates in Clinical Trials

Aligning your SAP with these ensures smoother regulatory review and approval.

Common SAP Pitfalls to Avoid

• Inadequate detail on derived variables
• Vague endpoint definitions
• Absence of handling instructions for missing data
• No documentation of interim analyses
• No version control or stakeholder review history

Each of these can lead to regulatory queries or delays in clinical development timelines.

Conclusion: The SAP Is the Bridge Between Data and Decisions

A robust Statistical Analysis Plan not only satisfies regulatory requirements but also provides a transparent, reproducible path for transforming raw trial data into evidence that supports labeling claims, peer-reviewed publications, and regulatory submissions. By including the right components and adhering to best practices, pharma professionals and clinical teams ensure both compliance and scientific credibility.

Further Learning Resources

• Pharma SOP checklist
• Stability testing