Tools to help patients and other stakeholders’ input into choice of intercurrent event strategy for estimands in randomised trials Authors: Joanna Hindley, Charlotte Hartley, Jennifer Hellier, Kate Sturgeon, Sophie Greenwood, Ian Newsome, Katherine Barrett, Debs Smith, Tra My Pham, Dongquan Bi, Beatriz Goulao, Suzie Cro, Brennan C. Kahan Published: 2026 ABSTRACT BackgroundEstimands can help to clarify the research questions being addressed in randomised trials. Because the choice of estimand can affect how relevant trial results are to patients and other stakeholders, such as clinicians or policymakers, it is important for them to be involved in these decisions. However, there are barriers to having these conversations. For instance, discussions around how intercurrent events (post-randomisation events which affect the interpretation or existence of the outcome) should be addressed in the estimand definition typically involve complex concepts as well as technical language. We aimed to provide tools that could facilitate conversations between researchers and patients and other stakeholders about the choice of intercurrent event strategy for estimands.MethodsWe developed three tools: (i) a video explaining the concept of an estimand and the five different ways that intercurrent events can be incorporated into the estimand definition; (ii) an infographic outlining these five strategies; and (iii) an editable PowerPoint slide which can be completed with trial-specific details to facilitate conversations around choice of estimand for a particular trial. Each tool was produced through collaboration between researchers and public partners. This involved (i) an initial meeting between researchers and public partners to discuss the aims of the tool; (ii) a draft of the tool being prepared by the research team; (iii) public partners providing feedback; and (iv) the research team updating and finalising the tool.ResultsThese resources can help to start conversations between the trial team and patients and other stakeholders about the best choice of estimand and intercurrent event strategies for a randomised trial. The video and infographic—which explain estimands and intercurrent events with reference to imagined examples—can be sent to stakeholders in advance of a consultation, or presented in the meeting itself. It is important that a member of the trial team is available to answer questions or clarify concepts following this. The editable slide can be completed by the trial team with the specific details of their trial, and then shown to patients or other stakeholders during the meeting to facilitate discussion around which intercurrent event strategy is most relevant for the trial. An example of a completed editable slide is also provided for an example weight loss trial.ConclusionsWe developed three tools to help researchers to have conversations with patients and other stakeholders about estimands, and how intercurrent events should be incorporated into the target estimand for a randomised trial. Further work to evaluate the tools in real-world settings across different stakeholder groups could help to validate the tools and reveal any further refinements necessary to improve their utility. Background Randomisedtrials are used to answer questions about how safe and effective treatments are, but for any given trial there are usually many different questions that could be addressed. For instance, a trial could evaluate the effect of a treatment if it were taken exactly as intended, or alternatively, regardless of any treatment non-adherence. In order to clarify which question a trial is addressing, the use of estimands (which provide a precise description of the treatment effect to be estimated) is recommended [1–9]. An important part of choosing an estimand is specifying which strategy will be used to handle intercurrent events. Intercurrent events are post-randomisation events that affect the interpretation or existence of outcome data. Common examples of intercurrent events include not receiving the assigned treatment, stopping the assigned treatment early, or receiving a different treatment to the one intended [2, 5, 7]. There are five different strategies that can be used to handle intercurrent events in the estimand definition: treatment policy, composite, while-on-treatment, hypothetical, and principal stratum. Details of these can be found elsewhere [1, 4]. The choice of intercurrent event strategy can impact the relevance of trial results for patients and other stakeholders, such as clinicians or policymakers [4]. For instance, an estimand which uses a hypothetical strategy to evaluate the effect of an intervention under perfect adherence may be less relevant for patients who are unlikely to be able to fully adhere. It is therefore important for patients and other stakeholders to be involved in these decisions for trials in which the results may inform future treatment choices [10]. Previous work has indicated that patients value being part of decisions around numerical aspects of trials [11] and in choices around estimands specifically [10]. However, there are currently substantial barriers to involving patients and other stakeholders in discussions around the choice of estimand. The concepts around the different intercurrent event strategies can be challenging to understand, even to those with a statistical background, and much of the terminology and language is highly technical and may be inaccessible to many. Literature suggests that accessible language and communication tools can help to facilitate patient involvement in the more technical aspects of randomised trials, such as statistical aspects [12]. Cro et al. have previously co-developed tools with public partners to introduce the concept of estimands [10] and explain the five attributes [13]. However, these do not introduce the strategies to handle intercurrent events, which are essential to the process of choosing an estimand. We therefore co-developed tools with public partners that aim to facilitate discussions between members of the trial team and patients and other stakeholders about the choice of intercurrent event strategy to be used in the estimand definition. This paper outlines how these tools were developed, and how they can be used to facilitate discussions between researchers and stakeholders around the choice of estimand. This paper is primarily aimed at researchers, with the intent of describing how the tools were developed and how they could be employed within their own trials. Methods In conjunction with public partners, we co-developed three tools to help facilitate discussions between researchers and patients and other stakeholders around the choice of estimand and intercurrent event strategies, and to enable these stakeholders to better contribute to these decisions. The three tools are:i.A video explaining estimands and intercurrent events, and covering the five strategies that can be used to address intercurrent events in the estimand definition;ii.An infographic which describes each of these five intercurrent event strategies; andiii.An editable PowerPoint slide which can be completed with trial-specific details to facilitate conversations around choice of estimand for a particular trial. i.A video explaining estimands and intercurrent events, and covering the five strategies that can be used to address intercurrent events in the estimand definition; A video explaining estimands and intercurrent events, and covering the five strategies that can be used to address intercurrent events in the estimand definition; ii.An infographic which describes each of these five intercurrent event strategies; and An infographic which describes each of these five intercurrent event strategies; and iii.An editable PowerPoint slide which can be completed with trial-specific details to facilitate conversations around choice of estimand for a particular trial. An editable PowerPoint slide which can be completed with trial-specific details to facilitate conversations around choice of estimand for a particular trial. We chose to develop the video and infographic to provide a lay overview of key concepts to interested stakeholders. We further developed the editable PowerPoint slide as we felt that an adaptable tool that was specific to the particular trial of interest was essential to facilitate discussion between researchers and stakeholders. All public partners had previous experience inputting into trials and/or statistical methodology projects, but none had prior experience with estimands. Video development The video focussed on describing estimands, intercurrent events, and the different intercurrent event strategies, using a fictional mental health trial as a motivating example. We advertised through the Patient and Public Involvement team affiliated with the National Institute for Health and Care (NIHR) Maudsley Biomedical Research Centre (BRC) at King’s College London for public partners to help us co-develop the video. Two public partners (KB, DS) were recruited. An initial meeting was held between several researchers and the public partners, to discuss the aims of the project and answer any questions. An outline of the video was prepared by a team of researchers (CH, SG, TMP, JH, DB, KS, and BCK). Using this outline, a draft script was written by a science communications officer (CH). Feedback on the draft was sought from the research team and the public partners, and this was incorporated into a finalised script. The animation was developed by the science communications officer and feedback was given by the research team (but not public partners). Adjustments were made based on this feedback and the video was finalised. Infographic and editable slide The infographic focussed on describing intercurrent events and the different intercurrent event strategies. The editable slide focussed on enabling researchers to facilitate discussions with patients and other stakeholders around choice of strategy for a particular intercurrent event in a particular trial. We advertised through the Patient and Public Involvement team affiliated with the MRC Clinical Trials Unit at UCL for public partners to help us co-develop these tools. One public partner (IN) was recruited. An initial meeting was held between one researcher (BCK) and the public partner, to discuss the aims of the project and answer any questions. First drafts of the infographic, the editable PowerPoint slide, and a completed example of the editable slide were developed by members of a research team (BCK, SC) and feedback was solicited from the public partner. This was used to update and finalise the infographic, editable slide, and a completed example of the editable slide. Results Video The video uses an imagined interaction between a doctor and a patient who is looking for a treatment for stress. The doctor looks at the results of a recently published randomised trial evaluating the use of cognitive behavioural therapy (CBT) for stress. A narrator explains how outcomes in this trial might depend on whether patients started taking anti-anxiety medication at some point during their course of CBT, and explains that this is an example of an intercurrent event. The narrator then explains that different research questions—or estimands—can be targeted depending on how this intercurrent event is addressed, and outlines the five different strategies for handling intercurrent events using the example trial. It is explained that the estimand should be chosen when the trial is being planned, and that not using an estimand may make the results of the trial difficult to interpret for patients and doctors. The video is available online at: https://www.youtube.com/watch?v=UX8MmPo27rc. Infographic The infographic is shown in Fig. 1, and is available online at https://osf.io/sg4h7/, and with the supplementary material for this article.Fig. 1An infographic explaining different ways that intercurrent events can be addressed in the estimand definition An infographic explaining different ways that intercurrent events can be addressed in the estimand definition Using an example of an imagined trial to evaluate a weight loss drug, it explains what estimands and intercurrent events are. Then, using an example of stopping treatment early as an intercurrent event, the five strategies for addressing this in the estimand definition are outlined for this trial, and it is emphasised that there is no ‘right’ estimand. Editable PowerPoint slide The editable slide is shown in Fig. 2, and is available online at https://osf.io/sg4h7/, and with the supplementary material for this article.Fig. 2An editable slide that can be used to illustrate the different approaches to addressing an intercurrent event in the estimand definition for a particular trial. This figure shows a generic version of the slide that can be edited for a particular trial An editable slide that can be used to illustrate the different approaches to addressing an intercurrent event in the estimand definition for a particular trial. This figure shows a generic version of the slide that can be edited for a particular trial The highlighted fields show where researchers should update the slide with details of the randomised trial that they are planning. When completed, the slide will outline the different options that can be used to handle a specific intercurrent event in the planned trial. An example of how this may be completed is shown in Fig. 3, and is available online at https://osf.io/sg4h7/, and with the supplementary material for this article.Fig. 3An illustration of how the editable slide can be completed for an imagined weight loss trial. This figure shows an example of how the slide from Fig. 2 can be completed for an imagined weight loss trial An illustration of how the editable slide can be completed for an imagined weight loss trial. This figure shows an example of how the slide from Fig. 2 can be completed for an imagined weight loss trial This completed example uses a weight loss trial with the intercurrent event of stopping treatment early to show an example of how the editable slide could be completed. Recommendations for using these resources Each of these tools takes a different approach to introduce intercurrent events and outline how these might be addressed in the estimand definition. The tools can be used individually or in combination with each other to facilitate conversations between a trial team and patient partners and other stakeholders. Some suggestions for using these tools are shown in Table 1, though we advocate the tools should be used in whatever way trial teams find most useful. Table 1Suggestions for using these tools to facilitate discussions around choice of intercurrent event strategies with stakeholders. Although this table provides suggestions for using these tools, it should be noted that the tools could be used in other ways too (e.g. the infographic/video could be shown during a meeting, etc)Before a consultation meeting • Send infographic and video link to stakeholders in advance of meeting • Modify the editable slide to be aligned to the trial of interest (with separate slides for each intercurrent event). • If applicable, inappropriate intercurrent event strategies should be removed from the editable slides at this stage (e.g. treatment policy when the intercurrent event is death)During a consultation meeting • At the start of the meeting, the research team should: ◦ Provide all relevant background information (e.g. summarise the trial, the meeting aims, etc) ◦ Set expectations for how the final choice of estimand(s) will be made, e.g. to clarify whether the trial team will make the final decision after soliciting views from all stakeholders ◦ Communicate to stakeholders what information would be most useful to help inform the final choice of estimand(s) (e.g. whether researchers want to know all strategies the stakeholders would find acceptable, how they would rank strategies in order of relevance, or only what their preferred option is) • Then, they should present the editable slide for each intercurrent event strategy, making sure to clarify any estimation challenges for specific strategies if applicable (e.g. if a strategy requires strong, untestable assumptions to estimate robustly). • Solicit stakeholder views for each intercurrent event. Table 1Suggestions for using these tools to facilitate discussions around choice of intercurrent event strategies with stakeholders. Although this table provides suggestions for using these tools, it should be noted that the tools could be used in other ways too (e.g. the infographic/video could be shown during a meeting, etc)Before a consultation meeting • Send infographic and video link to stakeholders in advance of meeting • Modify the editable slide to be aligned to the trial of interest (with separate slides for each intercurrent event). • If applicable, inappropriate intercurrent event strategies should be removed from the editable slides at this stage (e.g. treatment policy when the intercurrent event is death)During a consultation meeting • At the start of the meeting, the research team should: ◦ Provide all relevant background information (e.g. summarise the trial, the meeting aims, etc) ◦ Set expectations for how the final choice of estimand(s) will be made, e.g. to clarify whether the trial team will make the final decision after soliciting views from all stakeholders ◦ Communicate to stakeholders what information would be most useful to help inform the final choice of estimand(s) (e.g. whether researchers want to know all strategies the stakeholders would find acceptable, how they would rank strategies in order of relevance, or only what their preferred option is) • Then, they should present the editable slide for each intercurrent event strategy, making sure to clarify any estimation challenges for specific strategies if applicable (e.g. if a strategy requires strong, untestable assumptions to estimate robustly). • Solicit stakeholder views for each intercurrent event. Suggestions for using these tools to facilitate discussions around choice of intercurrent event strategies with stakeholders. Although this table provides suggestions for using these tools, it should be noted that the tools could be used in other ways too (e.g. the infographic/video could be shown during a meeting, etc) While these tools can provide a starting point for these discussions, they do not give a comprehensive overview of the topic, and our aim is for them to form part of a wider consultation between members of the trial team and stakeholders around the choice of estimand for a trial. At these consultation meetings, at least one member of the trial team should have an understanding of estimands and the different intercurrent event strategies, so that they might answer any questions that arise. The video and infographic provide a general introduction to estimands and intercurrent events. They can be presented during a meeting with stakeholders but can also be shared with them in advance to familiarise them with the topics to be discussed. Any questions from stakeholders can then be addressed at the meeting. The editable slide shows more specifically how each approach to addressing intercurrent events in an estimand can look in the context of a planned clinical trial. It is intended to be used in a discussion aimed at gathering feedback on choice of estimand for a trial, involving members of a trial team with an understanding of estimands. It is therefore not intended for the slide to be shared with stakeholders in advance. When preparing the slide with the details of their planned trial, the trial team can consult our example shown in Fig. 2—this imagined example is not itself intended for presentation. Of note, the editable slide assumes the broad framework for the other estimand attributes (population, treatment conditions, endpoint, summary measure) is already in place (while recognising that the choice of intercurrent event strategy might affect these). Other tools are available to help facilitate discussions with stakeholders about these attributes [10, 13], and it would be useful to summarise these attributes to stakeholders at the beginning of any meeting to discuss intercurrent events. All of these tools examine approaches to including a single intercurrent event in an estimand definition. In reality, it is likely that multiple intercurrent events will occur in a trial, and each of these will require a different set of considerations when choosing an appropriate strategy. This is a point that can be emphasised in discussions, and it could be useful to prepare multiple versions of the editable slide, one for each anticipated intercurrent event. The aim of these tools is to facilitate discussion between researchers and stakeholders around the choice of intercurrent event strategy. However, it is possible that different stakeholders will have different preferences on the choice of intercurrent event strategy, and these tools are not intended to address these differences. Ultimately, it will be down to the trial team to decide on the estimand(s) to use, taking into account the views of different stakeholders so far as possible. Therefore, it will be useful for researchers to think through what information from stakeholders would be most helpful in making this decision (Table 1). For instance, whether they want to know all the strategies that stakeholders would find acceptable, regardless of their order of preference, or they only want to know the stakeholders’ top-ranked preference. Discussion It is vital that patients and other stakeholders, such as clinicians or policymakers, are given the opportunity to contribute to the choice of estimand to ensure that trials are addressing questions relevant to those using the results. However, this has so far been limited by the complex nature of some concepts and the language used to describe these, together with the lack of tools to facilitate these conversations [10, 13]. We developed three tools to address this gap, specifically focussing on how intercurrent events could be included in the estimand definition. These tools offer a general introduction to estimands and intercurrent events, and provide specific guidance on what different intercurrent event strategies look like in the context of the trial under consideration. Collectively and individually these tools can be used to initiate informed dialogue between the trial team and stakeholders. These resources have the strength of being developed with input from public partners, who have given feedback on how they might be received in practice. They have not, however, been tested in real-world examples, or across different stakeholder groups, and doing so might reveal further refinements or further need for Patient and Public Involvement. While the video is fixed in its format, the infographic and editable slide can be updated directly by researchers for their planned trials. Furthermore, the tools were developed with a relatively low number of public partners. Further work to evaluate the tools with a broader group of public partners would be useful. Another potential limitation of the tools is that we have prioritised brevity in each. While this suits the purpose of these as a precursor to—or part of—meetings with a trial team, it may also mean that some concepts are not immediately clear for all users. Additionally, while these tools can facilitate discussion between researchers and stakeholders, they do not help researchers to resolve differences in views across different stakeholders (such as if one stakeholder group prefers a treatment policy strategy while another prefers a hypothetical strategy). Further work to understand how to choose estimands in the presence of differing views would be useful. Finally, although clinical relevance is essential when choosing an estimand, it is not the only consideration. For instance, the ability to robustly estimate the chosen estimand is also important, and this should be considered alongside clinical relevance to inform the choice of estimand [14]. As these tools do not address the fact that some intercurrent event strategies may be challenging to robustly estimate, further work around how these issues should be raised with stakeholders during discussions would be valuable. Conclusions We have developed three tools that can be used by research teams to facilitate discussions with patients and other stakeholders around how intercurrent events should be addressed in the estimand.