Patient and public involvement (PPI) in research is defined as “research being carried out ‘with’ or ‘by’ members of the public rather than ‘to,’ ‘about,’ or ‘for’ them” [1]. PPI means that patients or the public are actively involved in the research process rather than being included only as participants. Public engagement is the 2-way process of engagement activities that benefits both the researcher and the public [2].

Research that involves patients and the public can reduce the mismatch between what matters to patients and what is actually being done in the research [3]. A waste of research resources can be generated when the needs of those likely to use the research results, such as patients and caregivers, are not taken into account [4]. PPI can contribute to the identification and selection of high-priority research questions, planning and performing of more focused research, and improvement of participants’ enrollment in clinical trials [5]. Ultimately, this can result in a higher societal benefit through better use of resources for research. PPI improves the quality of the study and makes research more relevant [6].

Involve, a UK-funded program, aims to improve the quality of research through the integration of PPI throughout the research cycle (identification and prioritization, commissioning, design and management, implementation, dissemination, implementation, and evaluation). Involve has published a report with guidelines to help researchers start new projects when they intend to include PPI in their projects [7].

Digital epidemiology has the same objectives as epidemiology, which are the observation of disease patterns, their evolution, and the causes of these patterns to improve population health and prevent diseases, but digital epidemiology uses digital data [8]. A digital or e-cohort study can integrate data that were not generated for the research (social media and registries) or were generated with digital tools (wearables, sensors, smartphone technologies, and e-questionnaires through web platforms) [9]. Before starting any epidemiological, clinical, or population-based study, researchers need to choose the best methodology to incorporate PPI throughout the project [10]. However, although there are approaches to integrate PPI in a research project, we think that there are no clear recommendations of which methods are the most appropriate, in particular with respect to the launch of cohort studies with digital sources of data.

Although PPI is recommended in research projects, this involvement is often not described or is incompletely reported [11]. A reason for underreporting may be to avoid describing an unsuccessful PPI attempt or that there was no involvement [12]. In addition, there is some evidence that PPI is seldom used in many countries [13]. In the case of cohort studies and, in particular, e-cohorts, we believe that this insufficient involvement may be due to a lack of knowledge of the methods, barriers, or facilitators to apply PPI. In addition, we think that there is a need to have concrete and clear examples for applying PPI in this type of study.

The aims of this study are to (1) review methodologies used to include PPI in research, (2) provide the viewpoint of a final user of research results, and (3) provide practical guidelines and recommendations about how to initiate and run an e-cohort study with PPI based on the review and the point of view of a patient advocate.

Data for this review were identified by searches of PubMed, Google Scholar, targeted websites about PPI, reports, and existing PPI guidelines, as well as Google Search and references from relevant studies. We used the following search terms: “patient and public participation,” “patient engagement,” “patient involvement,” “consumer involvement,” “community involvement,” “participatory health research,” “community based research,” “research ideas,” “co-writing,” “coproduction,” “co-design,” “cohort study,” “e-cohort,” and “longitudinal study.” We included original studies describing PPI methods using the Involve definition with at least one type of PPI in the study. We included information on studies published in English from 2000 to the present.

We used relevant definitions from the European Patients’ Academy on Therapeutic Innovation for patients. We defined patients as individuals with personal experience of the disease, caregivers as individuals supporting a patient, and patient advocates as individuals representing large numbers of patients with a specific disease [14].

We invited a patient advocate to present her perspective and expectations regarding PPI in the context of diabetes research and the use of digital tools and, in particular, the use of social media.

We integrated the results of the review and the patient advocate’s viewpoint to identify practical guidelines on how to increase PPI in future epidemiological studies. We used the implementation of a digital cohort study as an exemplary case for testing and illustrating established guidelines for PPI [15]. In addition, we integrated in our recommendations the revised version of Guidance for Reporting Involvement of Patients and the Public (GRIPP2) checklist as an instrument to improve the quality of PPI reporting [16].

This study was based on a collaboration of patients and researchers. A patient advocate (RS), patient researchers (TS and TB), and researchers (GAA, CG, AF, VTT, PR, LH, and GF) were involved in the preparation of the manuscript (cowriting, editing, and critical review).

We analyzed data from 97 studies published from 2000 to 2020. The studies were performed in 9 countries: the United Kingdom (58/97, 60%), Canada (13/97, 13%), the United States (8/97, 8%), Australia (6/97, 6%), Ireland (6/97, 6%), Denmark (2/97, 2%), France (1/97, 1%), Portugal (1/97, 1%), Indonesia (1/97, 1%), and the United Kingdom and Australia (1/97, 1%). The most frequent methods of PPI were workshops (47/97, 48%); surveys (33/97, 34%); meetings, events, or conferences (28/97, 29%); focus groups (25/97, 26%); interviews (23/97, 24%); consensus techniques (8/97, 8%); James Lind Alliance consensus technique (7/97, 7%); social media analysis (6/97, 6%); and experience-based co-design (3/97, 3%). Of the 97 studies, 34 (35%) used only 1 method, whereas 30 (31%), 22 (23%), 8 (8%), and 3 (3%) used 2, 3, 4, and 5 methods, respectively. The use of ≥3 methods together was observed from 2017 onwards (Figure 1 and Multimedia Appendix 1 [17-114]).

Digital methods can be enablers of PPI, and the public can help modulate and develop digital technology [127]. Dedicated websites can enhance PPI not only by providing information, but also by organizing PPI itself. An example is the use of a webpage to organize a patient-led research hub. This is an initiative aimed at making patients and the public leaders of their own research projects. The researchers assist and support the patients and the public who proposed projects [37].

Crowdsourcing is a method by which many people are engaged on the web for a common goal such as obtaining new ideas and analyzing data. The main advantage is that it is possible to have a large number of contributors in a very short time [128]. However, because crowdsourcing participants are generally younger than participants from traditional nonweb-based procedures, these participants are not always representative of the target population [129].

Social media analysis is the use of digital data from social networks for epidemiological purposes. This information increases knowledge about epidemiological trends and may be very different from information obtained by traditional methods [130]. Social network data sources can be used, for example, to obtain information about the patient’s research priorities [38]. This often generates large quantities of data and can offer the opportunity to use special techniques such as natural language processing for the analysis [131].

Web-based platforms are internet services where the public and innovators meet and can be created for allowing coworking [132]. Vasilica et al [39] reported creating a web-based network to co-design a social media–based platform with the aim of generating information to improve disease outcomes [39]. In addition, web-based platforms through a web-based voting system have been used for establishing research priorities [133].

We found that many studies (60/97, 62%) involved patients and the public for generation of new ideas. Patients or the public contributed also in coproduction, co-design, or study scope (42/97, 43%) and in establishing research priorities (31/97, 32%). Other forms of PPI included participation in data analysis (25/97, 26%), as coauthors of a scientific article (17/97, 18%), as members of a steering committee or advisory group (16/97, 16%), reviewing or writing protocols (14/97, 14%), in the interpretation of results (12/97, 12%), in the dissemination of results and advocacy (11/97, 11%), in the data collection (9/97, 9%), in the development of the recruitment strategy (7/97, 7%), as a project manager (4/97, 4%), and being coinvestigator or having patient- or public-led projects (2/97, 2%). We found only 1 study where patients were involved in the co-design of mobile health tools (1/97, 1%). Finally, recent studies reported an early PPI in at least three different stages or fields of involvement and throughout the life of the research project (Table 1 and Figure 1).

To develop a successful PPI project, patients or the public and researchers must have, or develop, certain skills. For example, researchers need to become familiar with PPI as a research approach, know how to manage a PPI project, and how to deal with conflict. As for the patients and the public, they must understand the research process and develop capacities for management and conflict management. However, it is not mandatory that patients have specific vocational or educational training [136].

Web-based trials are more and more frequently described in the literature. Price et al [137] performed a systematic review of web-based trials and found that PPI was only reported in 24% (10/41) of the trials included in the review. Face-to-face meetings and email contact were the most common ways of interaction [137].

Taylor et al [97] performed a cohort study in patients with cancer that involved the patients in the creation and choice of a brand for the cohort. With a 1-day workshop, patients and researchers co-designed the brand. The results showed higher acceptance and retention of the study than expected. An ongoing cohort study used social media (Facebook) for PPI by creating a closed group of patients and families to bring new ideas to the project [80]. Meetings with a family advisory committee were organized regularly.

Morris et al [116], in the context of an epidemiological study, investigated PPI with surveys and postevent interviews and wrote recommendations. Before an event, they suggest sending a detailed document with the topics to be discussed. During the event, they suggest having enough space between tables to allow all participants to be heard, providing materials to facilitate note taking, taking a whole day for the meeting, and arranging a facilitator for each table. Finally, after the event, they recommend a follow-up by sending the notes to the participants [116].

Domecq et al [138] in a systematic review described some barriers to PPI. They highlighted 2 barriers: the excessive time taken for training activities and attendance and the risk of a tokenistic involvement. Another barrier reported in participants who were frail was frustration because of discontinuity in the involvement [139]. Barriers reported by researchers were concerns about the quality of research, ethical issues, lack of funding, failure of the PPI in the past, and not being convinced of the real need for PPI in the cohort [140]. Maccarthy et al [113] described communication issues as a key barrier to PPI. Researchers fear not being able to explain the project and not being able to engage patients and the public in the project; they also feel discomfort speaking about their experiences with patients and the public and fear having misunderstandings.

Creating a safe and welcoming environment where each contributor feels empowered and confident facilitates PPI [94]. The coproduction process can give participants the self-confidence to take responsibility for the entire duration of a project [78]. In addition, an iterative process of PPI evaluation during the entire research cycle has been proposed to ensure success in PPI [113]. Mathie et al [141] reported that feedback for patients, when provided, motivated them to continue their collaboration with researchers. Concerns for well-being, trust, mutual respect, and flexibility in time and methods were facilitators of PPI [139]. Chambers et al [142] found key areas that may facilitate or hinder the development of PPI. These key areas were the following: good role definition, recognition of difficulties, integration through organizations, training, developing networking, considering different perspectives, improving communication, and recognizing the relevance of emotional impact. Finally, concerning digital interventions and PPI, O’Connor et al [143] recommend investing in raising awareness of the usefulness of digital tools, improving health literacy, and using optimal tool design.

When writing this review, it was natural to allow a potential representative of study participants to express how they see PPI so far in research and what they are expecting from researchers to increase the participation of people with diabetes in research (Textbox 1, written by RS).

Digital cohort studies are longitudinal studies in which the data come either totally (e-cohort) or partially (hybrid: e-cohort and traditional cohort) from digital sources. Modern cohort studies increasingly incorporate digital tools such as data generated on the web and connected devices that allow much wider use of data generated for multiple projects [9].

We elaborated a strategy of PPI for a digital cohort study. Tables 2 and 3 show recommendations for PPI at all stages of research. This participation was defined in 2 categories: Recommended participation activities, in which patients and the public participate more passively, helping to generate research ideas and prioritize those ideas by participating in surveys, and Recommended involvement activities, in which patients actively participate in collaborative work with researchers on an equal footing. Examples of these activities are events, meetings, workshops, and focus groups. The chosen strategies for the digital cohort study are detailed in 32 actions in total for participation as well as participation and involvement, corresponding to the different steps of the research process, and are based on the current recommendations of Involve [7], our review results, and the viewpoint of a patient advocate (Textbox 1). In addition, for the realization of certain activities, we suggest a recommended duration of the activity based on our review of the literature and the point of view of the patient advocate.

The advantage of a digital cohort study is that digital tools can be used to promote PPI at each stage as a primary method or as a complementary or alternative method. A digital cohort can facilitate PPI, allowing participation from remote locations, using a smartphone app with web-based questionnaires, organizing most of the meetings through videoconferences, and using web-based tools for coworking. However, we think that face-to-face activities are also recommended and these 2 approaches may be complementary.

The recommendations are as follows:

Figure 2 shows our vision of integrating PPI in a digital cohort study in the whole research cycle.

We observed that the most popular methods for PPI are workshops, focus groups, interviews, and surveys. The appropriate method must be carefully chosen to fit the research objective. For example, workshops and focus groups can be a rich source of information in the prestudy phase; they can be adapted to the context of the participants; and they allow direct observation of, and interaction among, participants. The downside is that they are highly dependent on the capacity of the moderators.

We have also described digital tools to collect information from the patients and the public for research, such as social media listening and a particular web-based crowdsourcing survey with the magic wand–like question [24]. We think that digital tools may not only facilitate patient engagement in research, but can also stimulate a continuous and long-term participation and involvement of patients and the public.

Overall, none of these methods can address all relevant questions for scientists and patients or the public related to a research project. Therefore, we recommend a mix of methods to obtain optimal and meaningful information for the research project at stake. Different methods can be applied at different times of the project depending on the objective. Regarding the timing of introducing PPI, we recommend that PPI be included from the beginning (eg, co-design and research plan) to the end (publication and dissemination), followed by a PPI assessment.

We suggest pragmatic steps to integrate PPI in future epidemiological studies. These recommendations are as generic as possible but may not be applicable in all cases. They are based on current guidelines for PPI with concrete examples for a digital cohort study [14,15]. These recommendations should be understood as modular, meaning that they must be adapted to the study design, population, available budget, duration of the project, and local context. For PPI report and assessment, we recommend using the GRIPP2 checklist, a dedicated PPI reporting checklist [16].

Domecq et al [138], in a systematic review, identified barriers to, and facilitators of, PPI. They found no evidence indicating which method was best for PPI. In our review, we found that a combination of methods is more common with greater involvement (patients included in many areas or across the research cycle). We integrated the barriers and facilitators described by Domecq et al [138] in our recommendations.

Liabo et al [145] performed a systematic review of good PPI practices. They compared the results of the systematic review with 3 involvement groups and found that the priorities were similar. However, the involvement groups found additional values that were not described in the literature, such as the enthusiasm of the participants and the choice of welcoming venues for the meetings. We integrated these reflections in our recommendations.

Nunn et al [146] reviewed reports of 96 human genomic research projects and found that only 33% (32/96) declared PPI. From these, most of the PPI activities were organized in formal groups (20/32, 63%), with 22% (7/32) using web-based tools (website, social media, and web-based communities). We found similar results with social media reporting in 6% (6/97) of the studies with PPI. We think that there is room for more PPI using digital tools.

Miah et al [147] conducted a scoping review of PPI in dementia research and found 19 studies from the United Kingdom and 1 from the Netherlands. Biddle et al [13] found an uneven distribution of PPI in Europe. They attributed this to a lack of infrastructure, support, and guidance. An example of support is that research-funding institutions in the United Kingdom require PPI in project applications. However, funding agencies in many countries do not have this requirement. We found similar results with most of the studies on PPI from the United Kingdom. We believe that PPI is still underreported or not performed in many countries.

Few epidemiological or clinical studies report PPI in the research process [11,146]. Studies that include patients and the public most often involve them only at the stage of idea generation, but not in the whole research cycle. This suggests that PPI is symbolically added or very limited.

Individual interviews are useful for tackling sensitive questions because participants do not feel dominated or influenced by the opinion of other participants. Surveys (web-based, telephone-based, or paper-based) are a pragmatic method of obtaining large-scale information from many people quickly. Differently, a survey can be very useful for generating new ideas and for refining and improving them when a new survey containing the generated ideas is launched for another sample of people.

PPI is a project in the project, creating its own challenges such as completing appropriate regulatory tasks and obtaining approval from an ethics committee. In the informed consent, the nature of participation must be specified, with clear and fair terms and conditions. This includes whether remuneration, cost coverage for the patient organization, and travel costs of patients or the public are provided and whether there are other benefits for the patients, such as nonfinancial compensation for the time allocated to project participation [14].

Several reasons for a lack of, or delay in, PPI have been mentioned in the literature. Some researchers do not include patients or the public in the research, arguing that the patient or public point of view is too subjective [148]. Furthermore, researchers may consider PPI only as a requirement of funding agencies for the project to be approved; therefore, PPI arrives late in the process and is treated as an afterthought [149]. We believe that obstacles can be overcome when they are identified and taken into account in the plan and recommendations.

When PPI is considered, there is the risk of selective PPI, which means that only those within the community who agreed with the research objectives are included. In addition, this may have the risk of hearing only the opinion of the most active patients or public, which may not reflect the opinion of most of the other patients and can therefore lead to research designs that still do not research the questions that interest most patients. We recommend nonselective PPI, meaning that the aim should be to involve a diverse selection of participants in the PPI process, including patient organizations, patient advocates who are legitimately speaking on behalf of a patient community, and individual patients, to ensure that the opinions and views of the participants are representative of most patients with the same condition [47].

Researchers should make a greater effort to minimize the burden of participation for patients and maximize the benefits for participants at the same time. A lack of participation is more likely to occur when there is poor or 1-way communication, resulting in poorly organized protocols and demanding follow-ups that lead to noninformed, noninvolved, and nonmotivated participants. Nevertheless, when done properly, patients can associate their participation with feelings such as usefulness, empowerment, and consideration [150].

This study includes certain limitations. Our search was limited to studies in English. In addition, a part of what is happening in PPI is described in the gray literature and has therefore been excluded from the review. Our review may give the impression that early involvement in setting research priorities is the norm, but this may not be true because of a potential publication bias. For example, the results of studies with PPI are often not intended to show improvements in efficacy in clinical studies; therefore, they are less likely to be accepted by publishers for publication or publication may be delayed because of negative or statistically insignificant results.

There are, and rightly so, many expectations on the part of patients and the public to be actively involved in research and not only by providing data, but also as research partners. PPI can contribute to patient empowerment by increasing disease awareness and according recognition as actors of their own condition [151]. However, PPI is uneven among countries and research institutions, and even now many patients and the public are not yet involved in research [13] and ignore or do not have access to research protocols [152]. Digital tools such as websites, social media, and connected devices have been increasingly incorporated into cohort studies and could be leveraged to increase PPI [153]. Digital tools can facilitate PPI by providing an opportunity for remote access and therefore easier participation. In addition, digital tools can facilitate PPI by enabling feedback and interaction between researchers and patient collaborators [154].

PPI can be a powerful approach to increase the relevance of research projects. We have shown that PPI must be planned in the initial phases of the development of a new epidemiological study and then be considered throughout the life of the research project. Combining different approaches of PPI seems to be the most effective strategy for improving the quality of research.

Some techniques such as persona-scenario are very powerful for idea generation and can be combined with digital tools. In addition, web-based surveys are easy to implement and allow involving many participants (crowdsourcing).

Digital methods such as social media listening or web-based magic wand–like questions can also offer useful complementary channels of interaction and help to identify key information such as research gaps at large scale with a limited cost directly from the target population. As such, we recommend that these methods be also integrated in the PPI process. With the example of a new digital cohort study, we offer practical guidelines to implement and run a patient- or public-centered research study. We therefore encourage investigators to rely on our practical recommendations to increase PPI in future epidemiological studies.