The aim is to develop a COS for cranioplasty research following TBI or stroke by defining a standardized and agreed set of outcomes that should be reported as a minimum [13] while adhering to the COS-STAD (Core Outcomes Set-Standards for Development) recommendations [14] for COS development (Textboxes 1 and 2).

This study has 4 objectives:

This is a mixed methods study and will be divided into 2 main phases (Figure 2):

Advocates from multiple stakeholder groups were a part of the COS development [13,15] by COS methodologists. Adhering to this, the COAST study has 4 stakeholder groups:

These 4 stakeholder groups represent the key personnel in the acute and long-term management and rehabilitation of patients who have undergone or will undergo a cranioplasty. The groups will provide a wealth of experience and views from all aspects of the patient’s care pathway.

An international steering committee has been formed to oversee the development of the COS. The committee represents a broad range of disciplines and regions of the world, all with clinical interest in and involvement with cranioplasty from a research perspective. A project team oversees the COS development process’ day-to-day management, which includes 3 steering committee members and a coordinator. In addition, the project team will establish the methodology and address the key aspects of the project. Key definitions and concepts used to develop this COS have stemmed from the OMERACT initiative (Textbox 2).

Which outcome domains are used within the published literature for patients undergoing a cranioplasty following TBI or stroke and their recovery and rehabilitation?

Participants should be adult patients aged >16 years who have undergone a decompressive craniectomy.

The intervention being studied will be cranioplasty.

A search of three databases (PubMed, Embase, and Web of Science) from 1990 to the present analysis will be undertaken to identify either prospective cohort studies with >10 cranioplasty patients or retrospective cohort studies with >20 cranioplasty patients, and two reviewers (AMCL and HM) will independently assess the abstracts of the studies resulting from the search.

Two reviewers (ACL and HM) will independently review the full texts of the screened abstracts, with any disagreements for inclusion being resolved through discussion. If required, a third reviewer (AK) will be consulted.

Data will be extracted independently by two reviewers (AMCL and HM). The data will then be reviewed to ensure all outcomes have been identified and no unresolved issues arise. Any disagreements will be resolved through discussion with a third reviewer (AK) available for consultation if required.

The included studies’ outcomes will be categorized into domains under one of the four core areas of the OMERACT 2.0 filter framework [12]. Outcomes will be classified as primary when explicitly identified by the study authors, and all other outcomes will be classified as secondary. Each core area and subsequent domain evaluation of the number and frequency of the different outcomes will be analyzed.

To ensure a comprehensive list of outcomes, a qualitative study involving participants in the 4 stakeholder groups will be run, exploring cranioplasty experiences and services and discussing which outcomes are essential to the individual or group. These will then be combined with the outcomes from WP1 to develop the Delphi survey questionnaire.

The qualitative study will include both interviews and focus groups. Participants will be sampled purposively, and the sample size will be guided by data saturation, with 25 and 40 participants across the stakeholder groups likely to be required to reach saturation. A thematic analysis [16] will be used using deductive and inductive coding to analyze known topics of interest and generate new information as the study proceeds. Disconfirming evidence and outlying data will be searched to enhance the reliability and rigor of the analysis process and findings [17]. Data will be compared across and within groups. Qualitative data analysis software (ATLAS.ti) [18] will facilitate data management and analysis.

The survey will be online using the Delphi Manager software developed by COMET at the University of Liverpool. Participants will be asked to register and assign themselves to one of the four stakeholder groups. Registration will be implied consent.

This will be drawn from outcomes collated in phase 1 (WP1 and 2) and reviewed and signed off by the steering committee before use.

This will involve centrally coordinating with phase 1 participants invited to participate in further recruitment by post, email, or in person. Participants will also be advised of the importance of completing the Delphi survey, and email reminders will be generated as required to help with compliance.

There are no specific guidelines for the optimal number of participants in a Delphi survey, with previous COS studies having reported 140-150 participants in total [19,20]. Therefore, a pragmatic approach will be taken with the number of participants recruited for each stakeholder group. No new recruitment will occur after the completion of round 1.

The survey is an iterative process, with a minimum of two rounds but occasionally requiring a third or fourth. All participants score all outcomes using a 9-point Likert scale [21] (Table 1). There will be an “unable to score” option, allowing participants to abstain from scoring a particular outcome if they feel they do not have the expertise to do so; this has been demonstrated to work well in previous COS development projects [22]. A comments box will accompany each outcome to allow feedback as required.

Between rounds, scores are analyzed, summarized, and fed back to all participants in a histogram displaying the distribution of the scores of each outcome, with their scores highlighted. In addition, if there are any descriptive comments for a particular outcome, these will be summarized and presented.

All outcomes from round 1 will be carried forward to round 2, with any additional recommended outcomes also included in round 2. Participants who completed round 1 in the time allocated will be asked to participate in round 2. Each participant will be shown all the outcomes and scores, the number of respondents, the distribution of respondents' scores, and their scores from round 1. The participant will then be asked to rescore each outcome measure with the feedback from round 1 but categorized by stakeholder group. Brookes et al [21] have shown that this feedback method may improve overall agreement in scoring because it enables reflection based on expert views in other groups. Following round 2, no new outcome measures will be considered, and the consensus definition (see WP4) will be applied.

The steering committee will be informed of scores and analysis between rounds, with any issues or conflicts being addressed.

In line with a previously proposed definition by Williamson et al [13], our consensus definition has been discussed and agreed upon by the steering committee and will be applied in round 2 of the Delphi study.

The consensus criterion needs to be met for each outcome across the 4 stakeholder groups separately for inclusion or exclusion in the final COS; otherwise, the outcome will be classed as “no consensus” and taken forward for discussion.

The results of the Delphi survey will be discussed at a “final consensus meeting” involving invited participants from the 4 stakeholder groups. Outcomes categorized as “consensus in” across the stakeholder groups during the Delphi will be included in the final COS, and those outcomes categorized as “consensus out” will be excluded. Other outcomes that either have no consensus or partial in or out consensus across the stakeholder groups will be discussed within the consensus meeting, including a round of voting on all debated outcomes. If a final COS cannot be decided by the end of this consensus meeting, further consensus meetings may need to be planned to reach the final consensus.