Mapping the role of patient and public involvement during the different stages of healthcare innovation: A scoping review

Patient and public involvement (PPI) has become increasingly important in the development, delivery and improvement of healthcare. PPI is used in healthcare innovation; yet, how it is used has been under‐reported. The aim of this scoping review is to identify and map the current available empirical evidence on the role of PPI during different stages of healthcare innovation.


| BACKGROUND
Patient and public involvement (PPI) is premised on the principle that healthcare should be patient-centred, involving patients and the public in the design, conduct and dissemination of research and improvement work. Since the mid-1990s, PPI has become increasingly integrated into research, delivery and improvement in health and social care systems across the globe 1 to the extent that it is now widely considered best practice. 2 This paper focuses on PPI in healthcare innovation. While many studies have focused on reporting the benefits and challenges of PPI, these studies have tended to focus on earlier stages of the research process, such as research design and conduct. In their systematic review of 66 studies reporting the impact of PPI on health and social care research, for example, Brett et al. 3 found little regarding the impact of PPI upon the implementation of innovations. These findings mirror other studies 4,5 that appear to have generated substantial evidence regarding PPI involvement in the early stages of research, but where less is known about PPI is the later stages such as implementation, spread and scale-up. 6 Although there are many reviews on the role of PPI, they are often targeted within particular health services, including cancer care 7 and mental health, 8 or address a specific part of the research or innovation journey. Two previous scoping reviews have focused upon PPI in health research (the early stage of innovation) 9 and PPI in health policy-making. 10 To date, no reviews have been carried out to provide a holistic picture of PPI across all stages of innovation and across different areas of healthcare. In response to this evidence gap, this paper reports the findings of a scoping review to identify and map the currently available empirical evidence regarding the role of PPI during the different stages of healthcare innovation.
Before the scoping review is presented, we first outline our use of the terms PPI and innovation. Both are terms that lack a universal definition and so require clarification when used. Regarding PPI, we adapted the definition that the UK-based Na- into practice, including prototyping, piloting and evaluating safety and effectiveness; and (4) diffusion, the wider uptake of the ideas, services or products into use across the whole organisation/s. Also following Gabriel et al., 13 we focused on three types of innovation: technological and clinical innovations, which include new drugs, diagnostic tests, medical devices, software and surgical techniques; process and service innovations, including new institutions, business models, service models, clinical pathways, roles, education and training; and systems innovations, including policy innovations and systems reforms. Such mapping sought to identify good practice as well as evidence gaps (if any) to inform future research objectives.

| METHODS
We conducted the scoping review in accordance with Tricco et al. 14 A protocol was developed, following Peters et al., 15 and was revised by members of the research team and two experts in the field of healthcare innovation. The protocol is available upon request from the corresponding author.

| Types of participants
We included studies that considered the involvement of the public or patients across healthcare innovations, often referred to as service users or expert patients. CLULEY ET AL. | 841 2.1.2 | Concept/phenomenon of interest Our primary focus was to identify and map the involvement of the public or patients across any stage of healthcare innovation. As mentioned previously, our interest related to any practice innovation of the delivery of healthcare, including new medication, medical devices, care models, treatment programmes and service or quality improvement initiatives. To reflect our focus on healthcare delivery, we excluded innovations relating to healthcare financial management or governance, following the Cochrane Effective Practice and Organisation of Care taxonomy of health systems interventions. 16 Papers related to commissioning, educational and/or workforce development or those only focusing on evidence or knowledge utilisation were also excluded. While our definition of systems innovation includes policy innovation, we chose to exclude papers detailing policy to avoid duplicating a previous review. 10 This is discussed in further detail in the discussion and limitations sections.

| Context
The context for this review included peer-reviewed studies within any healthcare setting or context including the primary, secondary, acute or community setting, in any country.

| Types of studies
We included any empirical study type that reported on PPI in healthcare innovation within our inclusion criteria, including quantitative, qualitative and mixed-method studies. Theoretical frameworks, conceptual, scientific or grey literature such as case reports, evaluations, guidelines on how PPI should be conducted or service reviews as well as discussion and opinion papers were excluded.
There were no restrictions on language; the internal research team had the capacity to translate papers in Greek, Spanish, Italian, German, Dutch and Chinese. We aimed to use Google Translate for studies in other languages, but this was not required. Studies were limited from 2004, to reflect the publication of the definition of innovation used to shape the review. 12 2.1.5 | Types of outcomes Specific outcomes were not applicable in this study because our intention was to identify and map the empirical evidence in relation to PPI involvement across all stages of innovation-successful or unsuccessful.

| Search strategy
The search strategy followed Peters et al.'s 15 three-step process.

| Step 1
An initial search of Medline was carried out to identify the key words required for Step 2. However, due to the excessive number of hits (>70,000), a number of revised stages were carried out within the research team, to ensure a balance between adequate searching and unnecessary hits (see Appendix S1 for the revised version). As such, a second database was not searched and analysed to provide search terms. An example of the revised search terms included Delivery of health care; Patient or public or citizen or service user or lay or people; Involve* or participa* or co-crea* or co-design or co-produ*; Implement*or innovat* or spread or diffusion of innovation or evidence-based practice or quality improvement or adopt or translational medical research.

| Step 2
Using the key words and revised search strategy, the following

| Step 3
Citation checking was carried out for all included texts generated from the database sources to identify any additional papers that fulfilled our inclusion/exclusion criteria.

| Review screening process and results
All searches were carried out by C. F. For the bibliographic database search results, the citations were uploaded to Mendeley, deduplicated and uploaded to Rayyan, a systematic review screening software. A. Z. and C. S. carried out the initial screening, where titles and abstract were screened against the inclusion criteria. For papers that fulfilled the criteria, C. F. obtained the full texts and uploaded them to Rayyan. Full-text screening was carried out in a blinded manner by A. Z. and C. S. They screened the papers independently from each other and then met to discuss and resolve potential conflicts. Following this, C. F. carried out citation checking of all the included papers that captured additional papers. A. Z. and C. S. doublescreened the papers from citation checking. C. F. carried out a Google Scholar search to cross-check against our included papers, which were double-screened by A. Z. and C. S. Five papers were found in incidental searches during the initial phase, when C. F., who ran the searches, familiarized herself with the project. These papers were double-screened by A. Z. and C. S., and all were included in the final review.

| Data charting process
A data extraction tool was adapted from the Joanna Briggs Institute and refined with input from the research team to meet the specific needs of this review. C. F. charted the data from the main searches, and C. S. updated and refined the form. V. C. and E. O. conducted a blinded double-check of the data extraction. Any discrepancies were resolved between C. S., V. C. and E. O.
A critical appraisal of individual papers was beyond the scope of this study. 17

| Synthesis of data
We first summarized the studies by the type and stage of innovation.
Studies were then grouped by the innovation stage that they focused on, and narrative synthesis was used to identify the main findings that they presented.

| RESULTS
The PRISMA diagram below (Figure 1), developed following the updated guidelines for reporting systematic reviews, 18 provides an overview of our search and screening process.

| Characteristics of the included studies
The key characteristics of the included studies are summarized in Table 1. Of the 87 studies included in this review, 48 (55%) used qualitative methods, 2 (2%) used quantitative methods and 14 (16%) used mixed methods. Eighteen (21%) were case studies and five (6%) were randomized-controlled trials.
The studies considered a wide range of conditions. The conditions that were seen more frequently among the 87 studies  Table 1. Only five studies examined systems innovation (6%). One study focused on all types of innovation (1%). The studies detailing process and service innovation-focused predominantly on quality improvement. Of the seven studies on technological and clinical innovations, two looked at digital health, 19,20 one at clinical platforms 21 and one at patient-recorded outcomes. 22

| Stages of innovation
Mirroring research on PPI in general, it is evident from this review that PPI is more common in the earlier stages of innovation, particularly problem identification and invention, in comparison to adoption and diffusion.
Five studies focused on problem identification in healthcare research. [23][24][25][26][27] Fourteen studies focused on the invention. One involved consent processes; 28 five focused on quality improvement using 'living labs', 29 patient surveys, 30 patient forums 31 or a combination of different methods. 32,33 Twenty-seven studies looked at both problem identification and invention, the majority of which were quality improvement initiatives. Twenty-seven studies considered adoption, while only six analysed PPI in wider diffusion or scaling up of innovations. Finally, only five studies looked at all stages of the innovation process.
A summary of studies by stage and type of innovation is presented in Table 2 below. The data are also presented using an infographic developed by a design company (Design Science: bring knowledge to life. https://design-science.org.uk/) to facilitate interpretation by different audiences (Figure 2).

| Methods of PPI engagement
The majority of the studies included in the review (n = 32, 37%) described public and patient involvement and engagement in quality improvement initiatives.
The review included 24 studies (28%) involving PPI in codesign or codevelopment of health services. The methods used included patient journey mapping, surveys and workshops, 20,45,55,100,102 expert panels of patients and carers, 56,81 a user board, 57 PPI in the format of a conference, 36 Delphi methods to reach consensus, 28 living laboratories for technology innovation, 29 stakeholders activities 22,35,101 and interviews with patients. 58 Two studies presented codevelopment of clinical guidelines using qualitative methods. 40,82 Other studies used various methods. 19,37,49,[83][84][85] Participation in research activities was described in 14 studies (16%). [24][25][26][27]34,[59][60][61][62][63][64][65][66]83 Patients and the public participated in various ways, through advisory and face-to-face discussion groups, 34,62,63 virtual steering groups, 59 online surveys and workshops 26 and network activities. 65 Some innovative initiatives included video-reflexive ethnography that allowed clinicians to explore the needs of patients and their families in end-of-life care. 60 Eleven studies described the direct involvement of patients and the public in health service delivery either via peer-led activities or volunteering. 38,39,52,67,[86][87][88][89][90][91] Peer-service delivery was predominantly described in mental health services and considered eating disorders, 67,89 maternal depression in South Asia, 87  Drawing on the studies reviewed and the literature in the field, we next discuss why these trends might be prevalent.

| PPI is most frequently used in service improvement innovation
As outlined, the majority of the studies reviewed focused on service innovation (n = 62, 71%). Armstrong et al., 68  Traditional methods such as interviews, 58 focus groups 70 and surveys 45 were frequently used; however, more innovative and participatory methods were also found. Hacket et al., 55 for example, used experienced-based codesign to include young people in the improvement of youth mental health services. Collier and Wyer 60 used reflexive video ethnography to improve end-of-life care.
One reason for this trend could be that process and service innovation is an aspect of healthcare innovation that has long been associated with patient inclusion. Coproduction, experienced-based codesign and other patient-centred methods for improvement have an evidencebased background for effective healthcare improvement work. 104 PPI was found to be used least in technology (n = 7, 8%) and systems (n = 5, 6%) innovation. Various studies addressing the use of PPI in technological innovation have also identified this trend. Caution and ambivalence towards the use of PPI in health technology development have long been noted. 106 Recently, it was identified that the use of PPI is increasing in this area of innovation; however, challenges continue, including lack of public knowledge/awareness and lack of guidance on how to use PPI in health technology innovation. 103 It has been suggested that the slow take-up of PPI in this field of innovation is due to the dual roles of the health technology community to contribute to both research and policy-making. 107 Where PPI is well established and accepted in health research, studies have shown that the lack of a universal and rigorous approach to PPI is perceived to be unreliable by policy-makers. 106,107 While PPI was found in fewer studies detailing technology innovation, interestingly, the use of PPI in these studies tended to be grounded in participatory approaches rather than traditional research methods with a participatory angle such as interviews and surveys.
Codesign was a particularly common approach to PPI in this area of innovation. Codesign techniques were used to develop the use of technology in nursing homes, 29 to develop diabetes technology 49 and to develop an online mental health support platform. 20 This is perhaps a reflection of the design element involved in these methods and the parallels found in the design of technology. Additionally, such methods tend to focus on the earlier stages of innovation, perhaps going some way towards explaining the trend to utilize PPI at the beginning of the innovation journey rather than the end.

| PPI is most frequently used at the beginning of the innovation journey
In the introduction, we set out the innovation journey to include four stages: problem identification, invention, adoption and effectiveness and diffusion. 13 The majority of studies concentrated on one or two of these stages and these tended to focus on the earlier stages of problem identification, invention and adoption. As outlined, only six studies focused specifically on the use of PPI at the final diffusion stage. 38,52,82,84,92,101 Significantly, only five studies addressed PPI at all stages of the innovation journey. 37,44,50,83,102 PPI is intended to be a wholeprocess approach, meaning that PPI is included at all stages of the research or innovation journey. The 'nothing about us, without us' slogan borrowed from disability studies and often used to describe the ethos of PPI summarizes this aim well. Despite this, PPI work is often criticized for its limited use. Other studies have identified that PPI work tends to take place at the beginning of a project to aid planning and agenda-setting and then dwindle off. 3,9,10 A similar trend has been observed in technological innovation in healthcare, 108 whereby, it has been identified that PPI is more likely to be found at planning and implementation stages in comparison to monitoring and dissemination stages. This has led to PPI being criticized as a tick-box exercise to fulfil funding requirements, 109 and as virtue signalling. 110 It has recently been suggested that 'what gets done is what can be measured'. 111 This claim is based on the increasing call to measure the impact of PPI and how this is prescribed by funding bodies.
Russel et al. 111 use the NIHR as an example and show that reporting the impact of PPI is typically focused on things that can be measured quantitatively, such as the number of events and participants. Here, we suggest that this observation goes some way towards explaining why PPI tends to be used at earlier stages in the innovation process.
Related to this point, research and innovation tend to be lengthy processes. To include PPI throughout requires commitment and support, time and resources from both researchers and patient and public participants. 112 There are numerous evidence-based strategies to support effective PPI and retention of participants; however, as illustrated in this review and others, PPI is seldom used in the later stages of research. 113,114 Another reason for this trend is that PPI work is notoriously underfunded and often delegated to junior staff members, who are more likely to move onto other roles within the course of a project. 115 Those tasked with facilitating PPI work often do so in addition to existing roles, which can lead to PPI activity falling by the way side when other work has to be prioritized. 109 In response to these limiting issues, Boylan et al. 109

| Limitations
The studies found reflected our search terms and inclusion and exclusion criteria. These terms and criteria limited the study as follows. 3. A whole-system approach to PPI is adopted in all forms of healthcare innovation work to ensure that PPI is used throughout the innovation journey rather than just at the early stages. In a recent systematic review of barriers and enablers of PPI, 118 it was found that a whole-system approach will increase the buy-in and partnership working necessary to support successful PPI work.
Based on our study limitations, we make the following observations and recommendations. While the benefits of PPI in healthcare research and delivery are widely accepted, PPI has also been subject to wide-ranging critique. PPI activity is significantly under-evaluated, resulting in a poor evidence base and limited understanding of how PPI can improve research processes and outcomes. 119 Where the impact of PPI is reported, the attention tends to focus on improvement and delivery outcomes. The impact on the participants themselves is underexplored. 120 As outlined, we did not comment on this based on our research purpose to identify and map PPI in healthcare innovation. In addition to this, PPI work is increasingly being criticized for its lack of inclusivity and diversity. 121 The typical PPI participant has been described as white, middle class and male. 122 Based on these criticisms and the limitations of our search, we recommend that further research is needed to explore the impact of PPI in healthcare innovation, particularly focusing on patient and public experience. We also recommend that further research is needed to explore and promote diversity in PPI work in healthcare innovation.
In addition to this, as observed in our review findings and limitations, the majority of the studies found were conducted in highincome countries. With the exception of a few published papers, 123,124 there has been limited exploration of why PPI is seldom used in low-and middle-income countries nor how PPI could be used in these countries. Building on this, we recommend that further research is needed to explore the use of PPI or similar engagement strategies in healthcare innovations in low-and middle-income countries.

| CONCLUSION
At the time of conducting this scoping review, no reviews had been carried out to provide a holistic picture of PPI across all stages of innovation and across different areas of healthcare. This scoping CLULEY ET AL.
| 851 review set out to address this evidence gap by identifying and mapping the use of PPI work in healthcare innovation. In doing this, we have highlighted two main findings that are generally consistent with the PPI and healthcare innovation literature: first, PPI is used most frequently in service improvement innovations and the least in system and technology innovations and, second, PPI is most used in the early stages of innovation.
At present, PPI in healthcare innovation runs the risk of being described as a tick-box exercise or virtue signalling. As outlined above, we have set out a range of recommendations in response to this. What is most important, however, is that if PPI is accepted to be as beneficial as it is often reported to be, that there are systems of support in place to guide its use and ensure its accessibility and inclusivity across the whole innovation journey. While PPI should not be a one-size-fits-all approach, 114 it has been identified that the lack of measurable variables and the lack of a universal definition or approach render evidence gained from PPI work unpalatable to certain and influential audiences such as policy-makers. As suggested earlier, the UK Standards for Public Involvement could be a step in the right direction. 116 To support this conclusion, as outlined in our recommendations, we advocate for a whole-system approach to PPI to ensure that PPI is used across the whole innovation journey, not just at the earlier stages. Stronger PPI in later stages is likely to support the adoption and diffusion of innovation.

AUTHOR CONTRIBUTIONS
Claire Feeley carried out all the searches. Claire Feeley, Alexandra Ziemann, Victoria Cluley, Ellinor K. Olander and Charitini Stavropoulou were involved in the data extraction and analysis. All authors were involved in the interpretation of the results, contributed to the writing up of the manuscript and approved the submitted version.

DATA AVAILABILITY STATEMENT
All data generated or analysed during this study were included in this published article and/or its Supporting Information Materials.