Population.

There are no restrictions on the population.

Concept.

The review will consider literature on the conduct of PPIE. We will use the NIHR description of public involvement in research as “research being carried out ‘with’ or ‘by’ members of the public rather than ‘to’, ‘about’ or ‘for’ them” [19]. We will also include studies which consider engagement in the context of PPI, which the NIHR define as ‘activity where information and knowledge about research is provided and disseminated’ [20]. However, terminology around PPIE is still evolving and is not internationally standardised so other phrases that encompass public involvement such as “co-production”, “community participation” or “patient engagement” will also be eligible for inclusion [21].

Context.

We will define statistical methodology research as the development, evaluation or comparison of methods for designing studies, analysing data, and presenting results. Here we are specifically interested in statistical methodology research which aims to improve the conduct of applied health and care research studies.

We will be broad and inclusive in our definition of statistical methodology research for applied healthcare and consider the following (non-definitive) list of areas of quantitative healthcare methodological research to be part of that definition: observational studies, clinical trials, health technology assessment, health data science, mathematical modelling and simulation studies.

Types of sources.

We will include both published peer reviewed articles and grey literature. We anticipate, from our expertise in the discipline, that there will be relevant sources of information which are not published in peer reviewed journals (e.g., websites, blogs or webinars).

We will restrict articles to those published from 1996 onwards as this was when the Department of Health in England established a Standing Advisory Group on ‘Consumer Involvement in the NHS Research and Development Programme’, which represents the beginning of what we now call PPIE [22].

Search strategy.

A three-step search strategy will be conducted following the JBI recommendations with the aim to locate both published and unpublished literature. The search strategy for the first and second searches were developed in collaboration with an information specialist.

In the development of this protocol, the first step of the search strategy involved an initial search of MEDLINE and Scopus to identify articles on the topic. A validated filter was used to identify PPIE articles [23]. We made some adaptations to the PPIE search filter to include the following additional relevant terms: carer, co-production, co-development and co-creation. Keywords in the titles and abstracts and the MeSH terms of the relevant articles identified in this first step were used to adapt and refine the search for the second step (S2 Appendix).

The finalised search filter will be used in the second step of the search strategy, with the search filter adapted for selected databases judged to be most relevant to the topic (MEDLINE (Ovid), CINAHL (EBSCOHost) and APA PsycInfo (Ovid)). Given the lack of consistent terminology for both PPIE and statistical methodology for database searching, we anticipate additional search methods will be important to find relevant literature.

We will hand search the following journals and websites that we have identified as being highly relevant to this review:

• BMC Research Involvement and Engagement
• Journal of Participatory Medicine
• Health Expectations
• Research for All
• https://www.learningforinvolvement.org.uk/

To identify unpublished or other hard-to-find literature, Google, DuckDuckGo and Mednar will be searched in incognito mode to remove the influence of past search histories and, in the case of DuckDuckGo, with location set to ‘all regions’. We will conduct our searches using combinations of terms for both PPIE and statistical methodology, e.g., “Co-production” AND “statistical methodology”, “PPIE” AND “Data Science” (S2 Appendix). This will result in many searches, therefore, we will restrict our findings to the first 20 results to maximise search efficiency. We will contact our network of statisticians involved in PPIE (PPIE for Statisticians Community of Practice) to identify additional relevant literature or organisational websites. All types of grey literature (including non-textual resources such as videos) will be considered as long as there is sufficient detail about the PPIE to allow for meaningful data extraction.

The third step of the search will be to screen the reference lists of all studies selected for data extraction to ensure no relevant studies have been missed.

Evidence selection.

Following the search, all identified sources will be imported into Covidence [24] and duplicates removed. Pilot testing of both title and abstract screening and full text screening will take place using 25 randomly selected articles. The team will screen these using the eligibility criteria and then meet to discuss discrepancies and make any modifications as required.

Following pilot testing, title and abstract screening and full-text screening will both be performed by two reviewers independently with a third reviewer deciding when there are any disagreements. A PRISMA flow diagram of the process will be produced outlining the number of articles at each stage as well as the reasons for exclusions [25].

Data extraction.

Data extraction will be conducted using both Covidence and NVivo [26]. In Covidence, structured data will be extracted by two independent reviewers with a third reviewer resolving any disagreements. During the development of the data extraction form it will be pilot tested on three articles and modifications will be made on an iterative basis as needed as the data extraction proceeds. The final data extraction form will be reported in the scoping review.

The data extraction form will capture key information including: year of publication, country of origin, study funding, category of medical statistics methodology research, PPIE methodology, number of PPIE participants, participant remuneration, participant acknowledgement and diversity of PPIE members. Study-level risk of bias for published papers will be assessed using the CASP Qualitative Studies Checklist [27].

Using NVivo, we will extract narrative data on the objectives, approaches, benefits, challenges, evaluation, impacts, resources, key findings, recommendations and evidence gaps for PPIE for statistical methodology research.

Patient and public involvement.

This review is the first part of a larger programme of work which aims to develop guidance for conducting PPIE to inform statistical methodology research. Our PPIE co-applicant, who is part of the research team and has reviewed this manuscript, will chair a PPIE advisory group. This advisory group will meet three times a year (regular monthly updates in the form of ‘PPIE postcards’ will be sent in between meetings) and will advise on all aspects of the project, including this review. Our PPIE input has informed this scoping review, suggesting that both the positive and negative experiences of PPIE in this area should be collected and the importance of highlighting how the research, either directly or indirectly, will ultimately make a difference to people. Our PPI co-applicant and Patient Advisory Group will be involved with the co-development of a visual evidence map, attending the externally facilitated meeting.

Ethics and dissemination.

Ethical approval is not required for this review as it will use existing published aggregate data only.

The results of the review will be published in a peer reviewed journal and, in addition, will be shared through our social media accounts on LinkedIn and Bluesky.