It’s been about seven years since the ICH E9 (R1) addendum was adopted, but it is still often a source of confusion and frustration for our study teams. For the statisticians, most of us understand what an estimand is; we’ve had the training, read the guidance and papers and attended the webinars. However, when it comes to writing the estimands sections in our protocols, we often fall short of the cross-functional collaboration that the addendum intended. The estimand decisions are seen as a statistical responsibility and documented without the detailed discussions that are needed with the clinical team, or decisions are deferred to later in the clinical trial life cycle. This means that, whilst we might have an estimand section in the protocol, it is likely that there will be issues with ambiguity and interpretation further down the line.

So, this isn’t another blog about how to define estimands or specific thoughts about which intercurrent event strategy we should use in a given situation. It’s about something that is more fundamental:

How we communicate them.

With this in mind, here are some top tips for statisticians communicating estimands in terms our study teams care about: patients and clinical outcomes.

8 practical ways to improve estimand conversations

1. Start with the clinical question

Focus on the decisions or clinical questions the study is meant to support before introducing the estimands framework 

“At the end of this study, what do we want to be able to say about this treatment?“

2. Reduce jargon 

Translate concepts into everyday language wherever possible. For example, instead of naming a strategy, describe the real-world scenario: What happens if a patient stops treatment early?  

3. Ask questions before proposing solutions 

Rather than presenting a fully formed estimand from the start, ask clinicians what matters most to them. Use open questions to let the clinical team articulate what they care about first, and their answers will often effectively define the estimand.  

4. Anchor to clinical thinking 

Use concepts clinicians already apply. Clinical teams already reason through complex patient journeys every day. Position estimands as a formalization of that thinking, instead of a new construct. 

5. Treat intercurrent events as clinical reality 

Intercurrent events are often the part most likely to derail the conversation. I suggest using them as conversation starters. Frame them as real-world scenarios to plan for: What happens if a patient switches treatment? Misses doses? Discontinues early?  

6. Make consequences explicit 

Estimand decisions affect study design, data collection, analysis, and interpretation. It’s important to ensure clarity around how early decisions impact these aspects: If we don’t agree on this now, how will we interpret the results later? 

Think about ways to support this discussion:  

• Use patient-level examples: walk through a specific scenario, showing how different approaches would treat that patient 
• Visualize scenarios to support abstract decisions 
• Simulate outcomes under different strategies to highlight trade-offs 

7. Focus on one decision at a time 

Trying to define the full estimand in one conversation can overwhelm even experienced teams. Break it down. Build agreement across discussions rather than aiming for completeness in a single meeting. 

8. Make it genuinely collaborative 

Make it clear that estimands shouldn’t sit with statistics alone. They reflect shared decisions across the team. Shared ownership leads to stronger alignment and more meaningful study outcomes. 

By focusing on clinical meaning, rather than the estimand framework, we make estimands not just understandable, but genuinely useful. 

Better estimands don’t come from more detailed frameworks. 

They come from better conversations that happen earlier and are clearer and more collaborative.