What to write
Specific objectives related to benefits and harms
Example
“To evaluate whether a structured exercise programme improved functional and health related quality of life outcomes compared with usual care for women at high risk of upper limb disability after breast cancer surgery.”1
Explanation
Objectives are the questions that the trial was designed to answer. Adequate reporting of the research question is essential to allow readers to appraise and interpret the trial results. The PICO framework, which requires defining the patient population (P); the experimental intervention (I); the comparator intervention or condition (C); and the outcome or outcomes (O) of interest, has been proposed to help define the research question. PICO is sometimes styled as PICOTS, to include T (the timeframe) and/or S (the setting).2
Treatment decisions require an evaluation of the balance between benefit and harm; however, information about harms is frequently omitted or incompletely reported in published reports of trial results.3–6 Trials whose primary objective is to evaluate benefits of an intervention may not be powered to detect harms, but authors should still report whether they considered harms outcomes when planning the trial.7
Authors should clarify whether the aim is to establish superiority of the experimental intervention, or non-inferiority or equivalence, as compared with the comparator intervention.8 Authors should also report whether the trial is intended to provide preliminary data (a pilot or feasibility trial),9 explore pharmacokinetic properties, or generate confirmatory results.
For multi-arm trials, authors should clarify which treatment group comparisons are of interest (eg, A v B; A v C). If authors planned to readjust the objective during the trial (eg, in some platform trials or basket trials10), this should be reported. Finally, trials can be designed to study the effect of the experimental intervention under different conditions, often described on a spectrum from ideal conditions (explanatory trial) to standard clinical care conditions (pragmatic trial).11
The objectives should be phrased using neutral wording (eg, “to compare the effect of treatment A versus treatment B on outcome X for persons with condition Y”) rather than in terms of a particular direction of effect.12 The trial objectives should align with what was specified in the trial registry and protocol; any changes to the trial objectives after it commenced should be reported with reasons (item 10).
Recently, some trials have been designed using the estimands framework to define the research question and trial objectives. While the terminology surrounding estimands may be new to some investigators, it is expected that the use of this framework will become more widespread. Box 1 provides more information about the estimands framework and how it is being used.
Estimands
Concerns have been raised that the precise research questions that randomised trials set out to answer are often unclear.13 In particular, there is often ambiguity around how events occurring after randomisation (termed intercurrent events) are handled. Specifying the research question using an estimands framework is increasingly used to improve clarity. Despite calls for estimands to be included in the CONSORT 2025 statement,13,14 their inclusion did not reach consensus. However, we provide a brief overview of estimands and introduce terminology, so they can be applied and reported if used. A more detailed primer on the estimand framework which provides practical guidance on estimands in studies of healthcare interventions can be found elsewhere.14
ICH E9(R1) defines an estimand as “a precise description of the treatment effect reflecting the clinical question posed by a given clinical trial objective.”15 The estimands framework provides a structured description of the objectives in an attempt to bring clarity in specifying the research question, which can be used to guide the study design, data collection, and statistical analysis methods. In brief, an estimand comprises five key attributes: population, treatment groups, endpoint, summary measure, and handling of intercurrent events (see Table 1). A separate estimand should be defined for each study outcome, and for some outcomes, more than one estimand may be defined.
| Attribute | Definition |
|---|---|
| Population | Patients for whom researchers want to estimate the treatment effect |
| Treatment groups | Different intervention strategies being compared in the treatment effect definition |
| Endpoint | Outcome for each participant that is used in the treatment effect definition |
| Summary measure | Method used to summarise and compare the endpoint between treatment conditions (eg, risk ratio, odds ratio) |
| Handling of intercurrent events | Strategies used to handle each intercurrent event (see below) in the treatment effect definition; different strategies could be used for different types of intercurrent events |
Intercurrent events are post-baseline events (or post-randomisation events in randomised trials) that affect the interpretation or existence of outcome data. These events frequently affect receipt of treatment (eg, treatment switching or treatment discontinuation) or preclude existence of the outcome (eg, death, if it is not defined as part of the outcome).
The ICH E9(R1) outlines five strategies for handling intercurrent events, which are at the core of the estimand framework (see Table 2).
| Strategy | Description |
|---|---|
| Treatment policy | The occurrence of the intercurrent event is considered irrelevant in defining the treatment effect of interest: the value for the outcome of interest is used regardless of whether the intercurrent event occurs |
| Hypothetical | The treatment effect in a scenario where the intercurrent event did not occur is of interest |
| Composite | The intercurrent event is incorporated into the outcome definition |
| While on treatment | The outcome before the occurrence of the intercurrent event is of interest |
| Principal stratum | The outcome in a subpopulation of patients who would not (or would) experience the intercurrent event is of interest. |
Although the terminology surrounding estimands may be new to some investigators, it is expected that defining research questions using the estimands framework will become more widespread. A number of existing reporting guidelines have recently included estimands with the reporting recommendations.16–18 If the estimands framework has been used to design the trial or the data collection or to inform the statistical analysis (by guiding choice of appropriate methods), then this should be made clear in the manuscript and the methods and results should be reported within the framework.
Training
The UK EQUATOR Centre runs training on how to write using reporting guidelines.
Discuss this item
Visit this items’ discussion page to ask questions and give feedback.
