Co-creation and community engagement in implementation research with vulnerable populations: a co-creation process in China

Background

Effective community engagement is linked to improved health outcomes across a variety of settings.^1,2 Community engagement is defined by the World Health Organization as a process of developing relationships that enable stakeholders to work together to address health-related issues and promote well-being to achieve positive health impact and outcomes.³ Researchers have an ethical responsibility to engage a diverse range of stakeholders, especially those who are most affected by a program or intervention.⁴ Research funding agencies increasingly require that public health researchers and implementation scientists partner with community members or program beneficiaries.⁵

There are various approaches to include community members in research, each with a different balance of role, power and influence among key stakeholders. McCloskey et al. argue that community engagement can be regarded as a continuum of community involvement,¹ ranging from an outreach level to a shared leadership with a strong bidirectional relationship.¹ Researchers in public health have used a variety of community-engaged methods to engage diverse stakeholders. For instance, community advisory boards consist of community representatives that provide community feedback to the research project or initiative. Community advisory boards engage the community in a less active way, where community advisors mainly offer feedback to the project at researchers’ discretion.⁶ In contrast, co-creation aims to create a shared leadership role that involves the community in the development and design of programs from the beginning. Co-creators are empowered in the decision-making process for key implementation contexts, intervention components and implementation strategies, and/or fleshing out detailed steps to better meet their needs.

Despite an increasing use of co-creation and other community-engaged methods, there is a lack of clarity in defining, implementing and evaluating them.⁷ Researchers use terms, such as co-creation, co-design and co-production, interchangeably without distinctions or consensus definitions. Some examples of purported co-creation may not meet the shared leadership goal of co-creation and, in reality, result in only tokenistic engagement.^7,8 In this paper, we adapt the definition and conceptual framework proposed by Vargas et al.⁹ on how to use these terms in public health research. We defined co-creation as ‘the collaborative approach of creative problem solving between diverse stakeholders at all stages of an initiative, from the problem identification and solution generation through to implementation and evaluation’.⁹ Further, program beneficiaries are regarded as experts through their lived experience, and serve with researchers as equal members of the design team.^10,11

Co-creation has been extensively used in sexual health research, particularly in STD prevention.¹² Designing contextually appropriate interventions in diverse cultural and geographical settings is key to addressing sexual health challenges. However, very few studies have used co-creation among key populations, such as men who have sex with men (MSM). Moreover, there is no consensus on how co-creation should be carried out.¹³ Co-creation evaluations are also sparse, and it is unclear whether co-creation processes actually result in effective, high-level community engagement,⁷ especially for sensitive or stigmatized topics when stakeholders may be less willing to participate.¹⁴

The goal of this paper was to describe and evaluate a co-creation process to design and implement a sexual health intervention for MSM in China. Our earlier trials showed the preliminary effectiveness of this intervention for increasing gonorrhea and chlamydia test uptake.^15,16 We adopted the co-creation approach to adapt and implement our gonorrhea and chlamydia testing service across multiple public STD clinics and MSM community-based organizations (CBOs) in Guangdong Province.

Methods

Our co-creation model was guided by the principles and recommendations for co-creation of public health interventions derived by Leask et al.¹⁷ The adapted four stages included: (1) needs assessment (literature review, trial experiences and community feedback), (2) planning (co-creation leadership, group structure, roles and participants, and potential topics for group discussion), (3) conducting (co-creation group help with project design, develop project names, logos, recruitment materials and strategies), and (4) evaluating (co-creator surveys, comments and researcher reflections). Stages (2–4) can iteratively feed back towards stage (1) and dynamically inform problems/tasks to be solved through co-creation. Fig. 1 illustrates how we used the four steps of the co-creation process for our study. Ethical approval was obtained through The University of Carolina at Chapel Hill. All participants gave written informed consent and agreed to participate in the study. We analyzed our co-creation process though a Likert scale post-discussion survey and a rapid qualitative analysis methods based on observation notes and open-ended feedback.

Fig. 1.

Co-creation process overview.

Planning

Planning for co-creation involved establishing leadership, recruiting co-creators, sampling decisions and deciding on co-creation group structure.

Establishing roles

We began by identifying roles for the co-creation process. Co-creation is coordinated by facilitators, who orient team members to the co-creation process, balance roles, address potential conflicts and clarify how participants’ input will be used.⁷ To develop an atmosphere that enables the equal contribution of all stakeholders, facilitators must ensure that sessions are relaxed, informal and free of jargon.⁸ Facilitators often utilize design tools – such as games, visualizations and probes tailored for use in the target context – to engage participants and spur creativity.^18,19

We then mapped out key stakeholders to invite as co-creators, and the co-creation group structure, roles and responsibilities (Table 1). Key stakeholders included provider-side representatives (health professionals and leaders from community clinics) and community members. Additionally, two researchers from the project team would attend co-creation meetings and observe the group dynamics, reflect on the whole process, and iteratively used group feedback to optimize the co-created interventions and strategies.

Table 1.Roles and responsibilities of co-creation team members (n = 17).

Roles	Backgrounds	Responsibilities in the co-creation process
Lead (n = 1)	Project co-investigator with prior experiences of implementing the public health intervention	To oversee the process of the co-creation activities and ensure that activities meet the overall project need/purpose. Be observative, reflective, but less engaged during the activities. Jointly decide with co-chairs on key issues to be discussed in the process.
Facilitators (n = 2)	One project research assistant with research training; one gay community influencer who has prior experiences of providing STD testing services to local gay communities	The project assistant coordinated the whole process, came up with draft lists of topics relevant to the project, discussed with the Lead and the other co-chair issues to be solved in the upcoming co-creation activity. The gay influencer reviewed all relevant materials and topics to be discussed, provided constructive feedback, and jointly decided with the Lead and research assistant co-chair on issues to be discussed with the group. Both served as moderators of the co-creation activities throughout the process.
Researcher implementers and observers (n = 2)	Two research fellows who have been intensively engaged in implementing the project	To observe the group dynamics, reflect on the whole process and iteratively use group feedback to optimize the intervention delivery during a pilot study. Presented and observed the discussions. Communicate with other project staff on co-creation activities and contribute to the interactively developed discussion topics.
Provider side representatives (n = 4)	Two health professionals from public clinics; two lay health workers (gay men) from community-led clinics	To provide inputs on contextual information about clinical settings in both public and community-led clinics; advise on optimal user pathways for gay men, especially closeted ones, to learn about and navigate our intervention.
Community members (n = 8)	Eight men who have sex with men including closeted gay men	To provide user perceptions on demand and needs, STD testing services provision in public and community-led clinics, co-produce intervention components (e.g. messages, postcards, videos) and engagement strategies.

Recruiting co-creators and sampling considerations

We invited potential co-creators through social network recruitment and online open recruitment. Health workers were recruited through our professional networks. Open recruitment for MSM community members was advertised on the WeChat public official account (a Chinese multifunctional social networking app similar to Facebook) of Social Entrepreneurship to Spur Health. Interested men filled in the recruitment survey and were assessed for eligibility (i.e. biologically assigned as male at birth, age >18 years and ever had sex with another man).

We followed the principles discussed by Leask et al. with respect to sampling, which are based on prior recommendations to recruit 6–12 participants for focus group studies.¹⁷ We aimed to recruit ~10–12 total co-creators to also allow for division into smaller groups. Initially, we used convenience sampling to recruit co-creators who seemed likely to remain engaged. We used stratified sampling across age, geographic location and occupation to ensure some diversity in our co-creators. However, we also later used opportunistic sampling to recruit additional MSM as our ongoing needs assessment suggested that more input from a wider range of MSM would be valuable.

Results

Conducting

Between 31 August 2021 and 22 November 2022, we conducted a total of eight focus group discussions with a total of 17 participants. The main outputs of the co-creation sessions were as follows:

Strategies to decrease stigma and tailor interventions to MSM at public STD clinics

The co-creation group noted that there was still substantial stigma associated with STDs, and that many MSM patients were afraid of disclosing their sexual orientation to other people, including healthcare providers. As our intervention involves public clinic sites, this could cause challenges in identifying and recruiting MSM. Our co-creators generated icons, symbols, images or slogans that were subtle, but still indicated affinity with the gay community, to help signpost potential participants to MSM-friendly clinics in the public sector. Co-creators also suggested increasing the presence of MSM staff at project sites to help gain trust, reduce stigma and improve acceptability in public clinics.

Strategies to integrate STD testing services into existing community-led clinics

The community-led human immunodeficiency virus (HIV) testing clinics are trusted by the MSM community, but typically only provide rapid HIV, syphilis and hepatitis C blood testing, and lack infrastructure to provide gonorrhea or chlamydia polymerase chain reaction swabs. Furthermore, MSM co-creators suggested that they had a good understanding of HIV, but not on other STDs, such as gonorrhea and chlamydia. Thus, co-creators developed easy-to-understand educational pamphlet about gonorrhea and chlamydia to be disseminated among MSM to encourage testing (Fig. 2). Co-creators also helped with generating new workflows (Fig. 3) to integrate gonorrhea and chlamydia testing into existing workflows at clinics and CBOs.

Fig. 2.

Educational pamphlet.

Fig. 3.

Intervention delivery workflow.

Intervention components to enhance acceptability and community engagement

To make the intervention more acceptable and engaging, co-creators brainstormed additional online and offline components. Co-creators suggested an online engagement platform (Fig. 4: WeChat Mini-app main page) to engage participants and streamline their testing experience. The online platform is being developed as a WeChat mini-app that digitalizes key decision-making steps in participating pay-it-forward gonorrhea testing and donation behaviors. Co-creators also suggested more onsite activities involving the MSM community, including souvenirs or small gifts, to socially recognize participants’ inputs and engagement. Co-creators also helped design many of the intervention’s key promotional materials, including postcards to be used during the intervention (Fig. 5), re-conceptualized Chinese names for pay-it-forward, a project introductory video and project recruitment flyers (Fig. 6).

Fig. 4.

WeChat Mini-app.

Fig. 5.

Panel of co-created postcard and messages.

Fig. 6.

Panel of recruitment flyers.

Discussion

We present our experience using co-creation to adapt and implement a gonorrhea and chlamydia testing service for Chinese men. Our paper extends the literature by describing our four-stage, iterative, structured approach to co-creation and evaluating the co-creation experience. Our results highlight three principal findings: (1) the importance of community leadership in strengthening ownership and inclusivity, (2) community members’ enthusiasm and motivation for co-creative tasks, and (3) strategies to improve co-creation with marginalized communities in low- and middle-income settings.

First, designating appropriate community leadership can help create ownership and promote inclusivity during the co-creation process. Selecting a well-respected MSM as one of our facilitators greatly contributed to creating a safe, welcoming atmosphere, and seemed to indicate to many co-creators that we were sincere about their ownership, so community members were willing to openly share their sexual health issues and needs. This suggests the importance of the needs assessment and preparation phases of co-creation in understanding the community’s preferences and selecting an appropriate leader before even beginning to hold meetings. Although co-creation has long been recognized as a collaboration with researchers and community, previous literature showed concern that this process can sometimes be dominated by academics, which may weaken the community’s role and participation, and fail to generate community-led outcomes.^21,22 We successfully involved community leadership from the beginning to mitigate these issues and strengthen community ownership.

Second, our co-creators demonstrated their enthusiasm, diligence and engagement throughout the process, often demanding increased accountability. Increasingly, experts in public health call for the voice of community in participating in public health decision-making.^3,23,24 Our co-creators requested the project team to regularly report to the co-creation group about how their inputs influence the intervention and implementation, and many reported interest in future co-creation and even implementing this intervention. Our evaluation data also show co-creators’ high satisfaction and their willingness to adapt the intervention. Most co-creators’ participations were active, and their performance in taking initiative and demanding a stronger role showed a positive response to the shared leadership described in the community engagement continuum.¹ Such partnerships have potential for long-term, sustainable, public health interventions led by communities.

Finally, this paper presents our strategies for working with vulnerable and marginalized populations in resource-limited settings. Previous co-creation and community research has shown the challenges of working with vulnerable groups as many are structurally invisible²⁵ and fear exposing themselves to researchers, making it difficult for researchers to build rapport with vulnerable groups.^26,27 Our co-creation therefore developed effective approaches to mitigate some of the above issues. In terms of recruitment strategy, we started with an open recruitment process, inviting co-creators to participate, as well as a convenience sampling in our immediate MSM community, targeting those who we knew were highly engaged in their communities to ensure adequate participation in co-creation. Moreover, our hybrid format of co-creation process allowed both MSM who were more comfortable about being out to others, and men who preferred to stay closeted. This is primarily because closeted MSM may have concerns or feel anxious about outing to strangers, and such distress can be eliminated via anonymous online participation.²⁸ Recruiting MSM who had not widely disclosed their sexual orientation was a priority, because they are often underserved, and may be at higher risk for HIV and other STDs.²⁹

This study had several limitations. First, we reported data from one co-creation project convened in a single location focusing on one community, and may not necessarily be applicable to co-creation processes in other locations on other topics, with other communities. However, we collected rich data regarding the Chinese MSM community’s sexual health needs through eight co-creation meetings. We also conducted co-creation evaluation through surveys and open-ended questions to increase our data’s richness, which provides significant detail and summary on our co-creation process. Second, as previously discussed, there was selection bias in choosing our co-creators; in general, our co-creators tended to be well-educated and young, which may have biased our co-creation process. Despite an open recruitment for MSM end users, our participant recruitment strategies still risked insufficient representation of the community.

Third, we observed some power imbalance among co-creators during the co-creation process. Despite our efforts to maintain that all co-creators were equal, this resulted in dominance of the conversation by physicians and MSM community leaders over others. We later mitigated this by holding co-creation meetings with only MSM community, and emphasized this issue during the later sessions. Fourth, we admit that not all co-created ideas and contents were included in the intervention, and some needed to compromise due to the limited availability of resources in the public clinic and CBO settings. However, we documented these ideas, so that they may be incorporated in the future.

Overall, this report of our experience can help inform future co-creation processes for public health interventions and implementation strategies. A previous review³⁰ suggested that many so called ‘co-creation’ activities generated limited community engagement at a tokenistic level and there is a lack of practical consensus to organize co-creation in a structured way. Our description detailed the organization process and provided an empirical example of co-creating with a vulnerable population on a sensitive topic in low- and middle-income settings. Our co-creation experiences may also have practical implications for public health practitioners and campaign designers interested in deepening community ownership and engagement.

Conclusion

We successfully adapted a four-stage co-creation process to adapt and implement a sexual health testing service among MSM in China for community-led and public clinics. This approach effectively enhanced community leadership and fostered a sense of ownership, and resulted in practical, useful adaptations and implementation strategies. Our co-creation experience may prove beneficial for engaging other underserved and marginalized populations in low- and middle-income settings.