Vaccine hesitancy is one of the many factors impeding public health efforts to control the COVID-19 pandemic [1,2]. High vaccine uptake, especially in vulnerable and high-risk groups, is crucial to reduce severe disease hospitalization and death; however, vaccination rates differ markedly across countries (eg, 50% in Russia, 77% in Germany, and 91% in Singapore [3-5]). The World Health Organization (WHO) defines vaccine hesitancy as “a motivational state of being conflicted about, or opposed to, getting vaccinated; this includes intentions and willingness” [6], and it is a dynamic state that can evolve over the course of a pandemic. Reasons for vaccine hesitancy include a misunderstanding regarding COVID-19, a general mistrust of vaccination, safety and effectiveness concerns, concerns about the rapid development and testing of the vaccine, lack of awareness of the social benefits of vaccination, and logistical or opportunity-related factors [2,7-11]. These concerns have been exacerbated by the COVID-19 infodemic, in which “too much information including false or misleading information in digital and physical environments during a disease outbreak,” can lead to confusion and loss of trust in health authorities and public health responses [12]. COVID-19 misinformation has not only undermined public trust in vaccination [13] but has also led to a high social cost where close social relationships have experienced conflict or disagreements about pandemic control efforts [14-16]. In response, the WHO and others have called for greater application of behavioral science interventions targeting behavioral change and to support individuals in dealing with the psychological burdens associated with pandemic prevention and response [17-19]. For instance, evidence-based interventions that can help strengthen the resilience of individuals and communities to misinformation and other factors affecting trust in information will be a valuable tool for infodemic management [20].

The aim of this paper is to describe the theory behind the development of a new digital behavioral science intervention—The Good Talk!—designed to target vaccine-hesitant individuals through their close contacts (eg, family, friends, and colleagues) and to describe the methodology of a research study to evaluate its efficacy. The Good Talk! uses a serious game—a game that does not have “entertainment, enjoyment or fun as their primary purpose” [21]—to boost the competences of vaccine advocates to have open conversations regarding COVID-19 with their close contacts who are vaccine hesitant. Serious games have been used as educational tools to build skills, including teaching perspective taking to help students improve their open conversation skills [22] and inoculating people against misinformation (eg, the Bad News game [23]). Peer-led initiatives have been used to address mistrust in other domains, and communicating through relatable and trusted sources can lead to more acceptable and personal information exchanges, promoting prosocial behaviors when common identities are shared [1,19]. Furthermore, by focusing on boosting open conversation skills rather than addressing a specific cause of vaccine hesitancy, the approach has the potential to address a variety of reasons why an individual is vaccine hesitant [1,24] as well as build transferrable skills that can be applied across a variety of situations [25]. The Good Talk! is publicly available worldwide and will be evaluated in a preregistered randomized trial. Evaluating the impact of the intervention will be the subject of a future publication.

In 2019, the WHO identified vaccine hesitancy as one of the 10 threats to public health [26]. Despite strong and sustained evidence of the efficacy of vaccinations, perceptions of the safety and effectiveness of some vaccines have varied over the past years [27,28], affecting vaccine uptake [29] and contributing to disease outbreaks [24]. The influence of vaccine hesitancy on public health efforts to manage disease outbreaks has been accentuated during the COVID-19 pandemic, in which vaccine uptake in some countries has been low or lagging despite high vaccine access [3,4,30,31]. Consequently, longer durations of public health and social measures to reduce COVID-19 transmission are required, resulting in a variety of economic, health, and social costs with substantial psychological burden on individuals and communities [19,32].

The social costs of vaccine hesitancy have also been fueled by the COVID-19 infodemic, in which an overabundance of information has made it more difficult for the public to differentiate between correct and misinformation. When the public’s questions and concerns are left unanswered, information voids occur, which can lead to widespread false and misleading information. For instance, searches involving antivaccination rhetoric increased during the COVID-19 pandemic as people sought information about vaccine safety [33]. The infodemic has diluted trust in institutions, leading to confusion and disagreement about vaccination and safety measures [34,35], and has resulted in an increasingly polarized environment in which people avoid engaging with those who hold opposing or divergent views, negatively affecting interpersonal relationships [36].

For instance, 56% of participants in a nationwide UK-based study reported having arguments, feeling angry, or falling out with others because of discussions surrounding COVID-19 [16]. In addition, people report relationship strain, interpersonal conflict, and increased isolation associated with sharing differing opinions or information about COVID-19 with family members or friends or avoiding conversations altogether to prevent conflict [14,15]. Thus, interventions to help people develop skills to discuss potentially polarizing topics and maintain supportive relationships with those who hold different opinions are needed to reduce the negative social costs of the pandemic [15]. Furthermore, considering the evolving nature of COVID-19—both in terms of the emergence of new variants and cyclical gravity because of seasons—governments have imposed, lifted, and changed regulations frequently. As such, the topic of COVID-19 and vaccinations will hold relevance for many years to come [37].

We sought to develop an evidence-based behavioral science intervention that could address diverse drivers of vaccine hesitancy while also supporting individuals in managing the psychological burden of the COVID-19 pandemic. Vaccine hesitancy is context specific and falls along a dynamic continuum, with motivation toward vaccination existing on a spectrum of those who are fully, partially, or not yet vaccinated [6,38,39]. As such, one-size-fits-all public health communication strategies are unlikely to reach all such individuals and will require a tailored, targeted, and multifaceted approach [1,24].

Reviews of a broad range of intervention approaches for addressing vaccine hesitancy report a lack of evidence or mixed findings on the effectiveness of many interventions [6,40-42], with dialogue-based approaches being among the most effective strategies [40]. Dialogue-based approaches aim to facilitate and encourage open communication to understand a specific audience and their underlying concerns [40]. Open communication–based approaches have the potential to address the negative effects of the COVID-19 infodemic on polarization and interpersonal relationships. For instance, exposure to divergent views through interpersonal discussion can weaken the association between selective exposure—exposure to content and groups that reinforce one’s opinions—and polarization [43]. Open communication is important for showing respect and empathy toward a communication partner, which can reduce defensiveness, increase receptiveness, and pave the way for open-minded conversations in which speakers are willing to share their views, consider new information, and potentially change their attitudes [44].

In addition, personalized communication can increase COVID-19 vaccine uptake, and these interventions are most effective when tailored to and delivered by a trusted peer or community member [6,42,45]. Advocating normative or prosocial behavior is more effective when a member of one’s social network models the behavior and there is an expectation of social approval [19]. Social approval is important, as people have a strong need to belong [46] and people listen to others in their social environment [47]. Accordingly, vaccine-hesitant individuals may be more likely to listen to and trust their close ties, as public trust in health institutions among vaccine-hesitant individuals is exceedingly low [19,48], and peer-led initiatives have shown positive outcomes when addressing mistrust in other domains [1]. Peer-led initiatives capitalize on social networks to communicate and can influence social norms to promote behavior change [41]. Furthermore, communicating through trusted and relatable sources can facilitate more acceptable and personal information exchange [1].

Thus, the target group for the intervention is the close ties of vaccine-hesitant individuals who are vaccinated or are supportive of COVID-19 vaccination. For the purposes of this study, we refer to these individuals as vaccine advocates. As shown in the Theory of Change for the intervention (Figure 1), we aim to target vaccine advocates who are motivated to have a conversation about vaccination but lack the self-efficacy, skills, or knowledge to start a conversation or who fear resistance from their conversation partner. Motivation is an important driver of behavior change, and behavioral interventions are most effective when an individual is sufficiently motivated to learn new skills or competences [49,50]. In addition, targeting vaccine advocates increases the ecological validity and potential scope of the intervention, as they are likely to have more direct access and frequent encounters with vaccine hesitants and can reach broad and diverse audiences.

Accordingly, the goal of The Good Talk! intervention is to increase open conversations about COVID-19 vaccination by encouraging vaccine advocates to start a conversation with an individual they know or encounter who holds divergent or opposing views on vaccination or public health responses. We focus on starting an open conversation for several reasons. First, encouraging people to initiate a conversation is a first step to counter people’s motivations to avoid conversations about difficult topics (eg, owing to conflict concerns) [51,52]. Second, starting an open conversation is a concrete behavior as opposed to the more ill-defined concept of having a conversation and does not focus on the outcome of the conversation as an indicator of success. Although open conversations are important for building trust over the longer term, they may not lead to immediate changes in vaccination behavior or in clearly identifiable positive or negative outcomes following each conversation instance [53]. Thus, the aim of the intervention is to develop a participant’s competences, skills, and self-efficacy to approach rather than avoid conversations about COVID-19 vaccination. The intervention targets vaccine hesitancy indirectly by fostering more supportive interpersonal relationships between vaccine hesitants and vaccine advocates that encourage exposure to and consideration of divergent views. Ultimately, this should lead to continued exposure to divergent views between close ties and ongoing dialogue, which may eventually result in greater openness to the positions of others and remain open to evidence.

We applied the theoretical framework of boosting to guide the development of our intervention. Boosting is an evidence-based behavioral science approach that focuses on building people’s skills and competences rather than targeting a specific behavior [25,49]. The aim of boosting approaches is to foster people’s existing skills and competences or to develop new ones, thus helping individuals build their skills, knowledge, and self-efficacy to shift their motivations and translate their intentions or preferences into behavior [25,49]. A core feature of boosts is that they preserve people’s agency and enable them to exercise that agency to make their own choices [49]. Boosts are recommended for behavioral interventions where there is uncertainty or heterogeneity in people’s goals, when generalizability and longevity of a behavior is desired, and boosts are most effective when people are motivated to acquire new skills or competences [49,54]. Accordingly, a boosting approach is well suited to the design of an intervention that seeks to develop open conversation skills that are generalizable and flexible to addressing diverse causes of vaccine hesitancy, and for interventions that enable individuals to make their own choices on how to apply the skills to achieve their interpersonal relationship goals.

The focus of boosting approaches on fostering competences aligns with the objectives and key drivers of the intervention target behavior. Competence in interpersonal communication skills is associated with positive conversational outcomes, such as personal conversational satisfaction and more positive evaluations from conversation partners [55,56]. Demonstrating open conversation skills also leads to reduced defensiveness, greater awareness of attitude contradictions, decreased attitude extremity, and more open mindedness in conversation partners [57-59]. Furthermore, interventions designed to teach open conversation skills (eg, high-quality listening) or simple receptive conversation strategies have been found to signal positive collaborative intentions and receptiveness to alternative views while reducing negative conversational outcomes, such as defensiveness or threat reactions [57,59].

Developing open conversation skills can also build confidence and self-efficacy in conversations [60,61], and perceptions of self-efficacy are associated with conversational satisfaction [55]. Boosts foster self-efficacy by helping people learn and apply skills and knowledge through the development of competencies, rather than by providing information per se [25]. For instance, rather than teaching universal skills (eg, declarative knowledge), boosts use techniques such as teaching procedural rules, simple heuristics, or trainings (eg, learning how to harness implementation intentions) that can be generalized across different contexts or environments. The approach is conceptually similar to the inoculation approach used in misinformation detection interventions (eg, the Bad News game [23]), in which participants are exposed to the techniques that underlie misinformation and encouraged to generate their own antibodies rather than learn passively through being shown examples of manipulative strategies [62]. Learning through feedback on the effectiveness of skills can help build self-efficacy and confidence in understanding where and when to apply the skills to achieve one’s objectives [54].

Thus, our objective was to review the literature on open conversation skills to identify a set of simple strategies that could be integrated into a short behavioral intervention in the form of boosts. We aimed to design an intervention that would help individuals learn the effectiveness of these strategies on conversational outcomes through guidance and feedback, thus fostering participants’ self-efficacy in applying these strategies to initiate open conversations. In the following section, we will provide an overview of the evidence basis for the open conversation skills. The Methods section describes how we designed the intervention to provide participants with the opportunity to apply the strategies through an engaging and motivational serious game: The Good Talk!

We sought to deliver the intervention through a medium that had a strong evidence basis for promoting behavior change, was realistic and had strong ecological validity in terms of interpersonal communications during the COVID-19 pandemic, would motivate people to engage with the intervention, and could reach broad and diverse audiences. Gamified behavioral interventions have been found to promote positive behavior change across a variety of health-related areas, including asthma control, medication adherence, and diet [89], and were found to be more effective than inactive controls or nongamified interventions in improving physical activity behavior [90]. Games have also been used as educational tools across a variety of fields [91,92], such as teaching students role-playing and perspective taking to improve open conversation skills [22], and have been found to be effective at inoculating people against disinformation [93,94], specifically COVID-19 disinformation [95].

Serious games are games that do not have “entertainment, enjoyment or fun as their primary purpose” [21]. The learning benefits of serious games align with the goals of boosting interventions, in which the aim is to boost competences and in which motivation is key to intervention efficacy [49]. Serious games increase knowledge, build skills, and increase self-efficacy through design features within the game play that help players to select relevant information and integrate and organize the information with prior knowledge [96]. For instance, a first-person perspective can engage players in the learning experience, narrative and choice instill a sense of autonomy, interactions with game characters and personal reflection provoke empathy and relationship or community building, and mistakes and feedback processes can foster persistence [92,96]. Furthermore, serious games not only enhance motivation to learn but also have a strong motivational appeal [92,96].

Serious web-based games have the potential to reach a broad and diverse audience, can be accessed worldwide on mobile and desktop devices, and can be shared with others via web-based media. Edutainment mobile games aimed at educating remote or low-resource communities on issues such as agriculture and farming practices, health, livelihoods, and financial literacy have been disseminated in over 20 countries, with COVID-19–focused games reaching >2.7 million people across 11 countries in 2020 [97]. Furthermore, as social distancing and safety measures during COVID-19 have limited close interpersonal contact and people have increasingly turned to web-based communication to stay connected [98], mimicking widely used web-based communication applications (eg, using a messenger-type app) [99] has strong ecological validity [100]. Another goal of using this format was to facilitate adaptation, translation, and implementation into existing public health channels that may be fighting misinformation via chatbots (eg, WHO’s Health Alert Services [101-103]), including in low-resource settings [104]. Therefore, we designed a serious game based on a messenger app to build self-efficacy and open conversation skills by mimicking real-world conversations between vaccine advocates and their close contacts. Further details on the specific design features of the serious game and how the open conversations skills were implemented are provided in the Methods section.

In the previous sections, we have detailed the theoretical development of The Good Talk! intervention. Our next step is to conduct a research study to evaluate the efficacy of The Good Talk! serious game to improve people’s skills and competences to have open conversations about COVID-19 with vaccine-hesitant individuals. In the following sections, we present a detailed research protocol describing the planned evaluation study to be conducted along with the implementation of the game. Once completed, the results of the evaluation study will be reported in a separate publication.

The efficacy of the intervention game will be evaluated in a preregistered randomized controlled trial. Specifically, we aim to answer the following research question: “How well does The Good Talk! game improve people’s open conversation skills, perceptions of self-efficacy, and behavioral intentions to have open conversations with vaccine-hesitant individuals who are close contacts?” We will compare a group of participants who play the intervention game with a control group that will play the widely known noneducational game Tetris. Tetris was selected as the game for the control condition as it is in the public domain, and most people know how to play it without the need for practice or instruction. Tetris involves a similar amount of cognitive effort to play as the intervention game, it contains no content related to the study, and it has been used as a control game to validate other digital game-based learning interventions [62,94,105]. Tetris is an inactive control group (receiving no educational material), which is an appropriate design for assessing the absolute efficacy of the intervention [106].

Our primary hypothesis is as follows:

We also hypothesize the following:

Participants will be recruited to play the game via video or push notifications promoting The Good Talk! game on social media platforms (eg, WHO chatbot channels) and advertised via other WHO informational campaigns. Recruitment will be open to all interested individuals worldwide but will be limited to English language speakers. Participants will self-subscribe to the study by accessing the WHO game information web page via the public link or when shared by members in their social network. On the game information page, participants will be asked if, in addition to playing the game, they would like to participate in the research element (a similar approach was used to recruit participants to research elements in other serious web-based games [93,95]). Participants who are not interested in participating in the study will be sent directly to play The Good Talk! game; otherwise, they will be asked to opt-in to the study after giving informed consent. Participation in the research element of the game will be restricted to individuals aged ≥18 years. Participation in the study is voluntary, and no remuneration will be offered to complete the study.

Previous studies evaluating the effect of similar serious games on the development of competences report small-to-medium effect sizes averaging around f=0.18 [95,100,105]. A power calculation with a small effect size (f=0.18), power of 0.90, and α=.05 for an analysis of covariance on posttest open conversation skills suggests a total sample size of n=327. We will aim to collect 450 responses (225 per condition) to account for potential technical issues or possible poor-quality data. Data collection will continue until the prespecified sample size for each condition is reached. Preregistration of this study can be found on the Open Science Framework [107].

The intervention game was developed in collaboration with the game design company “Tiltstudio,” that has experience designing serious games including “Bad News,” “Go Viral,” and “Harmony Square,” targeting misinformation, COVID-19 misinformation, and political misinformation tactics, respectively. The game was designed to simulate messenger-type conversations between participants and close contacts in their social network. Messenger-type technologies are dyadic, informationally rich, synchronous, and mimic interpersonal communication and are therefore better suited for establishing a sense of connection and generating empathy [19]. At the same time, they represent a realistic communication platform familiar to many users worldwide [99], which increases the ecological validity of the intervention [100]. The game design implements instructional techniques found in a meta-analysis to be effective at improving learning and motivation in serious games, such as feedback (providing feedback or guidance on whether an answer is correct), reflection (participants are asked to reflect on their responses), level of realism (the game uses basic and cartoon-like representations), and narrative elements [96,108].

The game consists of 5 chapters in which the participants have conversations with characters representing their close contacts. In the first chapter, participants are introduced to the objective of the game, that is, to have as many people attend an event that they are organizing and to earn points for the good talk skills. The participant is asked to select an avatar to represent themselves and can choose between 1 of 4 social events they would like to organize (eg, birthday party, dinner, dance party, or housewarming).

The remaining 4 chapters see the participants encounter characters within their close ties who are motivated by reactance (negative affective reaction when one feels personal freedoms are being threatened), skepticism (doubting knowledge or claims), and inertia (opposing change or action) when discussing COVID-19 measures for the social event. As the conversation develops, the player must select from ≥2 conversational responses that affect the participant’s pathway toward the game objective and the points they earn on a Good Talk Meter. The Good Talk Meter increases with response choices that represent good open conversation skills and decreases for those that represent bad open conversation skills. In each chapter, participants are guided by a Good Talk Guide who offers advice and reinforces or educates participants about the strategies based on their selections. A brief description of the objective of each chapter is provided in Table 1, and a screenshot of the game interface is shown in Figure 2.

Participants who enter the game website and agree to participate in the research study will be randomized to the experimental or control group (between subjects) and complete both a pre- and postgame survey (within subjects). The experimental group will play The Good Talk! game, and the control group will play the noneducational game Tetris.

Participants will first answer the questions regarding their support for COVID-19 vaccination and their past experience speaking with vaccine-hesitant individuals. Participants will then complete a pretest survey containing the 8 open communication skills items and the self-efficacy and behavioral intention items. The experimental group will then play The Good Talk! game, which will take approximately 10 minutes to complete. The control group will play Tetris for a similar amount of time.

Following the game, participants will respond to a posttest survey containing the open communication skills, self-efficacy, and behavioral intention questions. Participants will also complete the implementation intention questions and a short demographic survey. At completion of the study, participants will be thanked for their time and invited to share the study with their friends. Participants in the control group will be given the option to play The Good Talk!. Figure 3 presents the flowchart of the study design and procedure.

Ethics approval for the study has been submitted and is awaiting approval from the WHO Ethics Review Committee (application ID: CERC.0177).

We hypothesize that The Good Talk! game will lead to a greater increase in open conversation skills, self-efficacy, and intentions to initiate conversations about vaccination between vaccine advocates and their vaccine-hesitant close ties relative to the control game. We hope that the game will help people to talk to one another so that supportive relationships can be maintained, and over time, opportunities to share with consideration of divergent views may help combat some of the negative social costs of the pandemic. Furthermore, we hope that this study will highlight how serious games can be a promising forum for delivering evidence-based health behavior interventions to the public.

The Good Talk! game represents an evidence-based behavioral science application based on the theoretical framework of boosting that aims to boost people’s skills and competences rather than target a specific cause of vaccine hesitancy. As such, this approach has the potential to build transferrable skills that can be applied across a variety of situations. Delivering the intervention via a worldwide, publicly available serious game means that it has the potential to reach a broad and diverse audience while using an ecologically valid medium that mimics real-world web-based communications.

Nevertheless, there are potential limitations to the insights that can be drawn from this study. First, as participants will access the study via a public link that is associated with the promotion of the game, it is not possible to blind participants completely from their assigned experimental condition. Although participants will be blinded to their assigned conditions at the initial pretest survey stage, it is not possible to avoid dropouts or potential self-selection to their assigned condition. To limit this effect, all participants will be told that both games are very important for the research and that everyone will get the opportunity to play The Good Talk! game. We will also collect data on the number of dropouts and partially complete and completed surveys per condition to evaluate their impact on study completion. Second, the game may attract a diverse set of participants from various backgrounds worldwide, including different levels of English proficiency, experience with COVID-19 vaccination, and social and political contexts. However, as the intention for the game intervention is to reach a broad and diverse audience and be applicable to individuals from a wide range of cultural, economic, and educational backgrounds, the evaluation will also contribute data on the effectiveness of the intervention in a real-world context.

We envisage that the positive results of the study can lead to concrete practical applications, for instance, as a valuable resource for health care professionals whose patients request advice on how to have vaccination conversations with those close to them. Furthermore, health care systems can refer to The Good Talk! game to boost the competencies of their patients to have open conversations about other difficult topics with their close contacts. More broadly, we hope that this study can motivate health care policies to consider serious games such as ours as an additional and efficacious method for promoting behavior change in public health communications.

The Good Talk! intervention represents an application of boosting theory to the field of behavior change for public health purposes. By describing the methods through which the intervention was designed, we hope to enable others to replicate and innovate our findings. We hope our approach will encourage more health ministers to engage in their mission to reach their citizens directly with digital health solutions and to consider such interventions as an important tool in infodemic management.