As a project biostatistician in a clinical trial, a “decision based on statistical support” refers to:

Professional judgments and determinations made by the project biostatistician—based on statistical principles, trial design, and study data—regarding how analyses are conducted, how data are handled, and how scientific and regulatory integrity is ensured, in accordance with the protocol, SAP, and regulations.
These decisions support project and medical decision-making but do not replace final medical or managerial decisions.

---

I. What Constitutes “Decisions Based on Statistical Support” for a Project Biostatistician

1️⃣ Decisions on Analysis Methods and Strategy

• Confirming or recommending:
• Analysis methods for primary and secondary endpoints
• Statistical models, tests, and multiplicity adjustment strategies
• Assessing whether the chosen methods:
• Align with the trial design
• Meet regulatory expectations
• Support robust and interpretable conclusions

Example:

Determining whether MMRM is more appropriate than ANCOVA given the data structure.

---

2️⃣ Decisions on Data Handling Rules

• Statistical judgment on:
• Missing data handling approaches
• Strategies for intercurrent events
• Definition of analysis populations (ITT, PP, Safety)
• Principles for handling outliers

Example:

Deciding whether subjects who discontinue early without follow-up data should be included in the primary analysis set.

---

3️⃣ Statistical Judgment Following QC Findings

• Assessing whether QC-identified issues:
• Affect primary or key secondary endpoints
• Impact statistical conclusions
• Determining whether:
• Re-analysis is required
• SAP clarification or amendment is necessary

Example:

Evaluating whether omission of a covariate from a programmed model constitutes a material deviation.

---

4️⃣ Judgments on Result Plausibility and Robustness

• Assessing whether results:
• Are consistent with prespecified assumptions
• Are sensitive to reasonable methodological changes
• Deciding whether additional:
• Sensitivity analyses
• Supportive analyses
  are warranted

Example:

Determining whether inconsistent results between primary and sensitivity analyses require additional analyses or explanation.

---

5️⃣ Decisions Related to Compliance and Disclosure

• Determining whether:
• An SAP amendment is required
• Protocol clarification is needed
• Additional statistical disclosure or limitations should be included in the CSR

Example:

Deciding whether a subgroup analysis insufficiently specified in the SAP should be clearly described as exploratory in the CSR.

---

II. What Does Not Constitute Statistical Support Decisions

❌ Medical judgments

• Whether treatment effects are clinically meaningful
• Whether dosing regimens should be modified

❌ Management or strategic decisions

• Whether to terminate a trial early
• Whether to proceed with regulatory submission

❌ Independent DMC decisions

• Interim determinations on continuation, modification, or termination of the trial

The project biostatistician provides statistical judgment and evidence, but does not make final medical or operational decisions.

---

III. One-Sentence Summary (Project Biostatistician Perspective)

As a project biostatistician, decisions based on statistical support are professional statistical judgments regarding analysis methods, data handling, result robustness, and compliance pathways, made to support—but not replace—medical and project decision-making.