Role of the Project Statistician in the Estimand Framework

Overall Role

Within the ICH E9(R1) estimand framework, the Project Statistician serves as the primary architect and steward of the treatment effect definition, ensuring that the clinical question is translated into a clear, estimable, interpretable, and regulatorily acceptable estimand.

---

Key Responsibilities

1. Define the Estimand Based on the Clinical Question

The Project Statistician works closely with clinical and regulatory colleagues to define the estimand by specifying:

• Population
• Treatment and comparator
• Outcome (variable/endpoint)
• Intercurrent event handling strategy
• Summary measure

The goal is to ensure the estimand is complete, unambiguous, and aligned with the study objective.

---

2. Lead Intercurrent Event Strategy Selection

The Project Statistician:

• Identifies key intercurrent events that may affect interpretation of the treatment effect
• Selects appropriate strategies (e.g., treatment policy, composite, hypothetical)
• Balances scientific validity, operational feasibility, and regulatory expectations

---

3. Ensure Alignment With Trial Design and Conduct

The Project Statistician ensures that:

• The randomization scheme and follow-up schedule support the target estimand
• Interim analyses or adaptive features do not alter the meaning of the estimand
• Estimands are consistent across study stages (e.g., Phase II/III seamless designs)

---

4. Maintain Separation Between Estimand and Estimator

A critical role of the Project Statistician is to:

• Clearly distinguish what is being estimated (estimand) from how it is estimated (estimator)
• Define the estimand independently of the statistical analysis method
• Select appropriate estimators later in the SAP to estimate the pre-defined estimand

---

5. Ensure Consistency Across Documents and Regulatory Interactions

The Project Statistician:

• Authors or reviews estimand sections in the protocol and SAP
• Ensures consistency across protocol, SAP, CSR, and submission documents
• Explains and defends the estimand strategy in regulatory meetings (e.g., FDA Type B or Type C meetings)

---

Common Pitfalls the Project Statistician Prevents

• Mixing estimands with estimators
• Post hoc changes to intercurrent event handling
• Inconsistent estimand definitions across documents
• Using analysis methods to redefine the scientific question

---

---

One-Sentence Summary

The Project Statistician ensures that the study answers the right scientific question and that this question remains unchanged from protocol design through regulatory submission.

An intercurrent event (IE) refers to:

An event that occurs after randomization but before the endpoint assessment,

and which may affect the observation of the endpoint or the interpretation of the treatment effect.

e.g. How should we use the data collected after discontinuing the medication?

Intercurrent Event Strategies – Summary Table

Strategy	Data After Treatment Discontinuation	Core Concept
Treatment policy	Included	Ignore intercurrent events; reflects an ITT / real-world treatment strategy
Hypothetical	Modeled or imputed	Assume the intercurrent event did not occur (counterfactual scenario)
Composite	Not used	The intercurrent event is counted as part of the endpoint (e.g., PFS)
While-on-treatment	Not used	Consider outcomes only while patients are on treatment; PP-like, commonly used for safety
Principal stratum	Restricted to a subgroup	Focus on a hypothetical subgroup defined by post-randomization behavior

Endpoint	FDA IE strategies
OS	Treatment policy
PFS	Composite
ORR	While-on-treatment / Composite
DoR	Treatment policy
EFS	Composite
Safety	While-on-treatment