An international , Delphi consensus study to identify priorities for methodological research in behavioural trials : A study protocol

Background: Effective behaviour change interventions are needed to impact important health outcomes, including morbidity and mortality. However, the uptake and impact of behavioural interventions have been limited by methodological challenges. The International Behavioural Trials Network (IBTN) was established in 2013 to facilitate global improvement in methodological quality of behavioural trials. There has been no formal process, within the network or in the broader literature, to define the most important research priorities to achieve this aim. In this project, we will conduct an international, Delphi consensus study to identify and achieve consensus on priorities for methodological research in behavioural trials among IBTN members. Methods: Fifteen core members of IBTN, who are experts in the field of behavioural intervention research, will be invited to generate a list of all items they consider priority areas for methodological research in trials of behavioural interventions. The IBTN Research Prioritisation team (the authors) will review all items generated, removing duplicates and merging similar topics, and generate a ‘long-list’ of items. This long-list will be sent to the 15 IBTN core members for approval. We will then administer two online Delphi surveys to all IBTN members. In the first survey, respondents will be asked to rate the importance of each item on a nine-point scale and rank their top five priorities. In the second survey, respondents will receive feedback on others’ responses and a reminder of their own responses in survey 1, and will be asked to re-rate items and re-select their ‘top five’. Discussion: Findings from the project will be used to inform the research agenda of the IBTN and to make recommendations for Open Peer Review


Introduction
There is now compelling evidence that behaviour change is central to global reduction of chronic disease prevalence, mortality, and burden of disease 1 . However, despite the significant potential for interventions that target behaviour to improve health and clinical outcomes, the reach and impact of behavioural interventions remains limited 2 . Behavioural interventions are "interventions that require the active participation of a target group (e.g., the patient/individual, health professional, health care systems) in a program delivered by a trained interventionist with the proximal or ultimate goal of changing health-related behaviour" 3 . Behavioural interventions may be delivered in person or digitally, employing digital technologies such as the Internet, telephones, and mobile and environmental sensors 4 . Experts in the area have concluded that the uptake and impact of behavioural interventions have been limited by methodological challenges specific to the design and conduct of behavioural trials 3 and a lack of investment from funders to conduct research to tackle these issues 5 .
To address these challenges, the International Behavioural Trial Network (IBTN) was established by a team of behavioural researchers in June 2013 with three main goals: first, to facilitate the global improvement of the quality of behavioural trials; second, to create networks and capacity to undertake more and higher quality trials; and third, to develop a repository of resources of existing recommendations, tools, and methodology papers on behavioural trials and intervention development.
Improving the quality and potential of behavioural trials requires methodological issues in this area to be identified and research conducted with the specific aim of addressing these issues. Previously identified methodological issues include intervention development and piloting, intervention reporting, identifying suitable comparison groups, selection of appropriate outcome measures and intervention fidelity 3 . A formal process to identify and specify methodological priorities is now needed to facilitate the development of an international and cohesive behavioural trials research agenda.
Research prioritisation provides a process for key stakeholders to generate ideas and reach consensus on important research topics. The prioritisation process has been used to identify priorities across conditions and populations 6,7 . In the area of trials, research prioritisation has been conducted with Directors of UK Clinical Research Collaboration Clinical Trials Units to inform the broader trials methodological research agenda 8 . In addition, a priority setting exercise has recently been reported to inform the global health trials methodology research agenda 9 .
Our aim is to identify priorities for methodology research specific to trials of behavioural interventions and to inform and guide the direction of the behavioural trials research agenda internationally. We will use a Delphi priority setting consensus approach, inviting all members of the IBTN to participate.

Methods
The Delphi process An electronic Delphi (e-Delphi), with online administration of questionnaires, will be used for this research prioritisation to facilitate international participation 10 . The Delphi process is a structured group facilitation technique to obtain consensus among anonymous respondents through iterative rounds with feedback 11 . The Delphi approach has been widely used in health research 12 . The features of the Delphi process that make it suitable for gaining consensus include: anonymity to facilitate balanced participation and iterative rounds to allow participants to change their opinion in response to controlled feedback where participants are provided with information on the distribution of overall group responses from previous rounds 12 . The Delphi process has been identified as valuable for large groups or when participants are separated geographically 12 .

Participants
Three levels of participants will be included in the process. • Core IBTN members (n = 15), selected by the research team as individuals within the Network with extensive

Amendments from Version 1
Based on reviewers' reports on version 1, we made the following changes to our manuscript for version 2: 1. Some words have been changed, e.g. in the abstract, 'brainstorm' was changed to 'generate'; in the abstract, 'develop' was changed to 'inform'.
2. We added a referenced sentence to the first paragraph of the introduction, to acknowledge that behavioural interventions may also be digitally delivered.
3. In the first paragraph of the methods section, we added a reference to back up our claim that the Delphi approach has been widely used in health research.
4. We added the following sentence at the end of the description in the methods section of the Delphi, phase 1: 'The final list of items will be approved and agreed by the IBTN Research Prioritisation Team.' 5. We provided a reference for Limesurvey software in the methods section.
6. We added the following sentence to clarify what was asked of participants within the e-Delphi survey: 'Participants are asked to rate the perceived importance, from their own individual point of view, of the items for methodological research in trials of behavioural interventions.' 7. We added new text to the methods section to clarify the procedure of including new items in survey 2 of items generated in the free-text comment box in survey 1.
8. In the dissemination section, we removed 'at the IBTN Conference 2018' and added detail about how data will be presented as follows: 'Items will be reported according to importance ratings for individual research items in surveys 1 and 2. The numbers and percentages of participants who rated each item as their top priority in surveys 1 and 2 will also be reported. The findings will be written up following recent reporting guidelines.'

See referee reports
REVISED expertise in behavioural trials, will be invited to participate in the expert topic generation process.
• Current IBTN members (n = 295) will be invited to take part in the two-stage e-Delphi survey.
Delphi phases Delphi Phase 1: Expert topic generation. A member of the research team (MB) will email core IBTN members and experts in behavioural trials (n = 15) to invite them to generate a list of all possible topics or research questions that they think are important for behavioural trials methodology research. Demographic details of participants including job title and years of experience will also be requested.
Two members of the research team (MB and JMS) will review generated items initially and propose a 'long-list' of items. These will be reviewed, discussed and approved by the full ITBN Research Prioritisation Team. This list will be sent by email to the full list of core IBTN members for comment and approval in advance of Phase 2. The final list of items will be approved and agreed by the IBTN Research Prioritisation Team.

Delphi Phase 2: E-Delphi survey.
All members of the IBTN will be invited by email to participate in two online surveys, using LimeSurvey online survey software 13 . In survey 1, participants rate their agreement with the importance of the items identified in Phase 1. Participants are asked to rate the perceived importance, from their own individual point of view, of the items for methodological research in trials of behavioural interventions. They will be asked to assign a score between 1 and 9 to indicate the priority they place on each topic. The cut-off scales proposed by the RAND group and used in similar research prioritisation exercises 8 will be used, with scores of 1-3 indicating a topic is not important, scores of 4 to 6 indicating that the topic is important but not critical, and scores of 7 to 9 indicating that the topic is critical. In addition, participants will be asked to select and rank their 'top-five' overall priorities. Participants will also be invited to add any items that they believe are important and missing from the list in a free text comment box at the end of the survey. Participants will be asked to provide demographic information including: sex, current professional position, country of residence, age group and number of years of experience of working in the area of trials of behavioural interventions.
In the second survey (administered 3 weeks after the closing of survey 1), participants will receive information about how others rated and ranked the items in survey 1 (a bar chart plotting total ratings for each item, the mean rating score for each item and the percentage of all respondents placing each item in their 'top-five' priority selection). They will also receive a reminder of how they rated each item and what items they selected and ranked as their 'top-five'. Any additional items proposed in the free-text comment box in survey 1 will be discussed by the research team and included for rating in survey 2 if the majority of team members agrees that the item is a unique, novel, previously excluded item. Respondents will be asked to re-rate each item and re-select and rank items to indicate their 'top-five' priorities. New items added to survey 2 will be rated only once. Ratings and levels of consensus from survey 1 and 2, as well as endorsement of items as 'top-five', will be collated and presented.

Ethical approval
Ethical approval was granted by the National University of Ireland Galway Research Ethics Committee (reference: 17-Jun-13). Participation in the surveys is taken to indicate participant consent; this is explicitly stated on the introductory page of the surveys.

Dissemination of research
Findings from this research prioritisation project will be shared with members of the IBTN to guide the development of a research agenda for the Network. All data will be stored anonymously on password protected computers to which only members of the research team have access. Items will be reported according to importance ratings for individual research items in surveys 1 and 2. The numbers and percentages of participants who rated each item as their top priority in surveys 1 and 2 will also be reported. The findings will be written up following recent reporting guidelines 14 for publication in a high quality, open-access international journal to promote visibility of the findings among behavioural trialists internationally, enabling them to focus research on high priority methodological questions, which will improve the quality of behavioural trials into the future.

Current study status
The first e-Delphi survey was launched in late January 2018.

Data availability
No raw data are associated with this article.

Competing interests
No competing interests were disclosed. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
the word 'inform' rather than 'develop': 'Findings from the project will be used to inform the research agenda of the IBTN and to make recommendations for future research.' Introduction The methodological issues examples given in the 3 rd paragraph are common issues for any complex interventions. To highlight the issue of methodological problems within the behaviour change field, and the importance of this Delphi consensus study, I would like to see more evidence of specific issues with behaviour change interventions that we already know about, within disease areas or types of interventions. For example fidelity in weight loss interventions.

1.
Authors' response: We acknowledge that the terms 'behavioural trials' has significant overlap with the term 'complex intervention trials'. Indeed, it could be argued that the terms could be used interchangeably, that all behavioural trials are complex, and that all complex intervention trials involve behaviour change. In this section of the protocol, we reference a paper specifically discussing challenges to 'behaviour change interventions', so we have used this terminology here. In the final reporting of this study, we will incorporate a more elaborate discussion around the definition of behaviour change intervention trial, and how it relates to complex interventions. However, we consider such a discussion to be beyond the scope of this protocol paper. Methods 5 th paragraph -does LimeSurvey software need to be cited with LTD or a web address? 1.
reported.' Is the rationale for, and objectives of, the study clearly described?Yes ○

Is the study design appropriate for the research question?Yes
○ I wonder if the authors might want to collect the same demographic information between Phase 1 and Phase 2 to allow comparison.
It might be useful to add a sentence on how additional comments from the 15 experts at the end of Phase 1 will be handled.
Similarly, how will new items integrated into Survey 2 following Survey 1 be handled. Will these be rated only once in Survey 2?
When the participants are asked to rate the importance, it might be useful to specify what importance relates to. Someone who is interested in theory development might rate importance different to someone primarily interested in clinical implementation of evidence. The exact phrasing of the question is thus crucial to avoid introducing unwanted variation in the provided ratings and to be specific as to what the IBTN members are asked to judge. This is a well written protocol of a timely study, the authors are conducting important and vital work to further advance behaviour change science. I have only minor comments for consideration by the authors. Abstract: The authors use the term brainstorm in the abstract, but not the methods. It might be good to use the same terms for the proposed activities. Authors' response: The sentence in the abstract has been changed to: 'Fifteen core members of IBTN, who are experts in the field of behavioural intervention research, will be invited to generate a list of all items they consider priority areas for methodological research in trials of behavioural interventions.' The definition of behavioural intervention on p. 3 is very useful ("interventions that require the active participation of a target group (e.g., the patient/individual, health professional, health care systems) in a program delivered by a trained interventionist with the proximal or ultimate goal of changing health-related behaviour"). However, might this definition omit interventions which are delivered remotely and electronically, such as ehealth and mhealth interventions which are typically delivered automatically and not by a trained interventionist? I wonder if the authors might want to collect the same demographic information between Phase 1 and Phase 2 to allow comparison. Authors' response: The following demographic data will be collected from participants at time 1: sex, current professional position, country of residence, age group and number of years of experience of working in the area of trials of behavioural interventions. As these are not going to change between survey time 1 and time 2, we do not plan to collect this data again for survey 2.
It might be useful to add a sentence on how additional comments from the 15 experts at the end of Phase 1 will be handled. Authors' response: The following sentence has been added at the of the methods section Phase 1: 'The final list of items will be approved and agreed by the IBTN Research Prioritisation Team.'