Social media has been extensively used worldwide to communicate health-related information. For example, in China, one-third of the users of the social media platform WeChat—which is widely used for instant messaging and social networking [1]—receive and read health information through the platform [2]. In the United States, 32% of social media users post messages about friends and family members’ health experiences on social media [3]. Health care professionals use social media to provide health information and answer medical questions [4], and patients and caregivers use social media for self-care and health literacy [5]. In health care research, social media platforms such as Twitter, Facebook, and YouTube are increasingly used for participant recruitment, intervention implementation, data mining and collection, and the sharing of research findings [6].

Social media, with its free access, interactive features, and widespread reach, has become increasingly used by individual researchers and professional organizations who wish to translate research evidence into health care practice. For example, the Joanna Briggs Institute (JBI) at Fudan University in China has been using WeChat to disseminate nursing evidence since 2014. In the first 2 years, their WeChat account reached 22,369 followers from 34 provinces in China [7,8]. The Cochrane Child Health groups in Canada and Portugal used social media strategies to disseminate child health evidence to health care providers, and within 6 months of initiating the strategy, their blog received 2555 visitors and 3967 page views, and their Twitter account gained 469 new followers from a geographically diverse population [9]. A social media initiative called It Doesn’t Have to Hurt, led by health care researchers in Canada, developed a short YouTube video on evidence-based strategies, such as distraction and using topical anesthetics for reducing procedural pain in children. Their video received 237,132 unique views from 182 countries 5 years after its launch, with patients and health care providers reporting strong acceptance and high intention to use the strategies [10]. The number of parents reporting the use of topical anesthetic creams to reduce pain increased from 18% to 63% after watching the video [11].

There has also been a surge in social media initiatives during the COVID-19 pandemic, which are aimed at helping health care professionals, patients, and the public better understand the coronavirus and cope with its impacts. Global evidence synthesis networks such as Cochrane, JBI, and Campbell Collaboration use social media to disseminate rapid review findings related to COVID-19. In China, the Beijing University of Chinese Medicine (BUCM) Cochrane Center, together with 20 evidence-based health care research teams and organizations, launched the Fighting COVID-19 with Evidence initiative. They collect urgent clinical questions about COVID-19 diagnosis, treatment, and nursing care through WeChat and share recommendations after a rapid search and synthesis of research evidence [12]. In England, the Center for Evidence-Based Medicine at Oxford University uses Twitter (@CebmOxford) to share COVID-19 relevant recommendations to a global audience. In Canada, the COVID-19 Evidence Network to support Decision-making initiative (@COVID_E_N_D) collects the best available evidence related to COVID-19 and shares this information on Twitter to support decision-making.

Despite its popularity, many researchers and organizational decision-makers upload research findings onto social media platforms without deliberately planning how to facilitate its use by recipients in policies, programs, or practices. In their systematic review, Webb et al [13] concluded that theory-based internet interventions had greater impacts on health behaviors than non–theory-based interventions, with interventions based on the theory of planned behavior having larger effects than those based on the transtheoretical model or social cognitive theory. However, despite these benefits, theoretical frameworks are rarely used to guide the development of social media interventions aimed at facilitating research use. In their systematic review, Arguel et al [14] only identified 15 experimental studies published between 2005 and 2016 that applied theoretical approaches to guide the development of social media interventions.

Ngai et al [15,16] classified 31 theories used in social media studies into three categories: personal behavior theories, social behavior theories, and mass communication theories. Personal behavior theories (eg, the theory of planned behavior and technology acceptance model) focus on personal factors that affect user behavior on social media. Social behavior theories (eg, social capital theory and social cognitive theory) identify key social factors that stimulate individuals to participate in collective actions on social media. Mass communication theories (eg, parasocial interaction theory) reveal the distinct characteristics of social communications that can assist in the use of social media for communication and marketing [15,16]. These theories provide valuable insights into social media’s role in behavior change; however, the following two limitations exist in fully understanding the research use process:

These 2 limitations were partially addressed by Ritterband et al [17], who developed a behavior change model for internet interventions, which posited that website use was influenced by support, characteristics of the websites and users, and environmental factors. Behavior change from information on websites is then influenced by various mechanisms (eg, knowledge and motivation). This model has been used to guide the development and evaluation of internet interventions in health care [18,19]. Although not exactly the same, websites that allow for multiway interaction are normally considered to be social media [20,21], and the Ritterband et al [17] model has been used in the social media context [22]. It addresses the limitations of the aforementioned social media theories, as it considers the platform—which in this case is the website—and accounts for the multilayered contexts in shaping behavior, such as personal, professional, and community contexts, as well as the health care system [17]. However, the Ritterband et al model [17] does not make mechanisms of change explicit and presents a linear process for using the internet to change behavior when real-world practice is often complex [17].

Despite its extensive use for disseminating health care research evidence, social media is rarely used in a well-planned way with end users in mind, which largely limits its potential to bridge evidence–practice gaps and contribute to health care practices. Studies on the use of research evidence through social media are sparse [14]. Large theoretical gaps exist in understanding how social media interventions affect health care practices and decision-making. Unpacking the process by which social media works as a knowledge translation strategy is important to not only advance science but also inform interventions for improving health care practices and patient outcomes.

The purpose of this paper is to propose a conceptual framework to understand how social media works as a knowledge translation strategy for health care providers, policy makers, and patients to inform their health care decision-making.

Through a review of social media initiatives (n=5), theories (n=5), and empirical studies (n=58), including papers (15/58, 26%) relevant to the 5 social media initiatives [7,8,10,24,26,27,35-43] and papers (43/58, 74%) [9,13,15,44-83] from citation tracking, snowballing, or consultation, we developed the SMILE framework (Figure 1). Table 1 summarizes the key constructs and their supporting evidence. The SMILE framework provides a preliminary understanding of how social media can be used as a knowledge translation strategy to inform health care practices and decision-making. It has six key constructs: (1) developers, (2) messages and delivery strategies, (3) recipients, (4) context, (5) triggers, and (6) outcomes. For social media to enable recipients to use research evidence in their practice or decision-making, the framework proposes that developers respond to the needs and context of target recipients to develop relevant messages and appropriate delivery strategies. Recipients’ use of social media messages is influenced by the virtual–technical, individual, organizational, and system contexts and can be activated by different types of triggers, described as sparks, facilitators, and signals. Next, we describe the constructs of the SMILE framework.

Developers are individuals, groups, and organizations responsible for the management of social media contents. Developer activities may include designing and periodic uploading of information, monitoring operations, collecting data on impact, and answering questions or comments from viewers. Developers can be health care researchers who produce research evidence and share it directly via social media for public access. Barton [78] proposed a new research-to-practice continuum where researchers not only disseminate research findings through traditional journal publications but also create multimedia messages and disseminate them to the public. Developers can also be intermediaries who serve as a link between research producers and end users by translating research evidence into user-friendly messages for dissemination on social media.

Although it might be simple for individuals to develop and upload research findings to social media, a fast, frugal, and hope-the-change-happens approach has limitations. One of the propositions embedded in the framework is that the composition of the development team, availability of resources, scope of topics, and vision of impact influence the development of relevant and appropriate social media interventions, thus affecting recipients’ engagement with and use of the messages.

We suggest bringing together a multidisciplinary collaborative team of health care professionals, target users, social media experts, and audiovisual technicians (eg, camera operators and video editors) to best support the development of social media interventions [9]. Health care professionals can assist with the identification of different types of evidence resources; target users can strengthen the relevance and accessibility of messages; social media experts can contribute to the operation of the platform; and audiovisual technicians can provide support when the team wants to deliver messages using videos or animations. For example, the It Doesn’t Have to Hurt initiative has built a large interdisciplinary collaborative team of researchers, trainees, patients, and other stakeholders to facilitate the stable operation of their social media program [10,26,85]. Similarly, the ECHO research program has created various videos, animations, and posters on child health with a multidisciplinary team [28]. With different knowledge, skills, and perspectives, the team can generate high-quality and influential social media products. The long-term collaborative approach can additionally promote the sustainability of these initiatives.

The availability of resources to develop and manage social media initiatives, such as time and budget, must be taken into consideration when planning it. In the It Doesn’t Have to Hurt initiative, it cost the team Can $15,000 (US $11,802) and considerable efforts to develop and promote their YouTube video, and the developers stated that financial and time costs could be a hindrance for individual researchers to undertake the work [10]. In their social media initiative to disseminate Cochrane Child Health evidence, Dyson et al [9] also found that the team invested enormous time and human resources in managing the platform. Therefore, we suggest that adequate time and budget be allocated to social media initiatives before their commencement.

The scope of topics covered is closely linked to the amount of time and resources invested. Some initiatives, such as the Fudan JBI Initiative [7,8,24] and BUCM Cochrane Initiative [25], have broad scopes that are open to a range of topics in nursing and medicine. Some initiatives focus only on specific topics; for example, the It Doesn’t Have to Hurt [10,26] and Be Sweet to Babies initiatives [27] target reducing procedural pain for children and infants, respectively. Other initiatives center on a certain field, such as the ECHO initiative, which covers common childhood conditions. The topics covered should be balanced with the consideration of practical issues. Dyson et al [9] suggested that starting from a specific content area and engaging with a stable social media community was more effective for developing a social media network.

It is also essential that the development team builds a shared vision of the impact they are looking to achieve and tracks the performance of their social media initiatives [51]. Building and sustaining a social media initiative is demanding work that requires collaboration and investment. An explicit team vision of the impact of social media can motivate the team to work toward a common goal. For example, since 2016, the Fudan JBI Initiative has openly shared its social media vision in its annual center report and at conferences [84]. Gates et al [79] also emphasized the importance of setting goals and tracking achievements after the evaluation of their social media initiative.

In the SMILE framework, we propose that the engagement of a multidisciplinary team, time, and resource investments are essential for developing relevant and appropriate social media interventions to influence research use. Developers should balance the topics covered with practical considerations and create a shared vision of the goals of their social media initiatives.

Recipients are the target audience of social media messages and have the potential to direct, influence, or be affected by messages. In our framework, we consider health care providers, policy makers, and health care consumers as recipients. We also propose that using social media messages in health care decision-making involves two distinct, interconnected layers: using the social media and then using the message. It is a prerequisite for recipients to first accept and use the social media before they can engage with the messages. We distinguish between these 2 layers and consider the factors that influence each layer separately. We contend that the characteristics of recipients and the virtual–technical context are the two main domains that influence people’s use of social media, and the individual, organizational, and system contextual domains shape the message use.

Together with frameworks from the social media and technology research field [15,48,55,65], the i-PARIHS framework [29], behavior change model for internet interventions [17], COM-B model [30], and Fogg behavior model [31] have provided valuable insights into the characteristics of recipients that influence social media use. On the Basis of their theoretical constructs, four aspects of recipients’ characteristics were incorporated into our framework: demographics, personal traits, motivation, and capability.

Demographics include age, gender, geography, socioeconomic status, ethnicity, and lifestyles [15,17,55]. Large quantities of research data from Twitter and Facebook revealed differences in social media use by gender, ethnicity, and geography [45,53]. Personal traits of openness, conscientiousness, extraversion, agreeableness, and neuroticism—rooted in genetics—are perceived as one of the fundamental theories that explain personal behavior [15]. They are closely associated with social media use [64]. In a national survey in the United States, Correa et al [46] found that although extraversion and openness were positively related to social media use, emotional stability—a central measure of neuroticism—was a negative predictor. These findings differed by gender and age [46].

Motivation and capability are 2 summative characteristics of social media recipients that the SMILE framework identifies as affecting social media use. These characteristics are based on the Fogg behavior model [31] and the COM-B model [30]. Within motivation, perceived needs [65], attitude [50], intention [48,50,55], self-efficacy [17], and goals [55] are factors motivating individuals to use social media. Within capability [48,50], knowledge and skills [17,31] enable individuals to use social media. Together, all four characteristics of recipients (demographics, personal traits, motivation, and capability) are determinants affecting social media use in the SMILE framework.

In the SMILE framework, we specify the knowledge translation outcome as research use, which is a multidimensional concept that involves conceptual, instrumental, and persuasive use of research findings [33,34]. Conceptual research use refers to using research evidence to change the levels of knowledge, understanding, or attitude of a recipient. Both the It Doesn’t Have to Hurt initiative [10,26] and Be Sweet to Babies initiative [27,35-39] have demonstrated that when recipients receive relevant and appropriate messages on social media that respond to their needs and context, they are highly likely to improve conceptual research use. Instrumental research use involves the direct application of research evidence in practice to change behavior. Modanloo et al [41] and Tunnecliff et al [76] have shown, in their randomized controlled trials, that different types of triggers, such as sparks or facilitators, are essential for the active uptake of research evidence and behavior change by recipients. Persuasive research use refers to using research evidence as a political or persuasive tool to justify an action, attain power, or achieve goals [33,34]. One of the most typical examples is the #WearingMasks social media campaign during COVID-19, which has made a huge impact on public behavior and government policy making.

In this paper, we present the SMILE framework, which is based on a review of 5 social media initiatives, 5 theories, and 58 empirical studies. The framework provides a preliminary understanding of how social media works as a knowledge translation strategy for health care providers, policy makers, and patients to inform their health care decision-making.

The SMILE framework has implications for research by offering a heuristic device for the development of social media interventions to promote evidence use. We suggest that it be used in combination with process frameworks, which provide step-by-step guidance on implementing web-based knowledge translation interventions [86] or evaluation frameworks to evaluate the multilevel outcomes and impacts of social media interventions [54].

On the basis of this framework, we offer several suggestions for researchers and organizations who intend to use social media to promote research use. First, in the preparation stage, it is important for developers to assess their readiness to start a social media initiative. Some probing questions may be considered during this stage, such as is there an explicit topic to be covered? Does the team have enough time, resources, and expertise to develop the intervention and monitor the operation?

Once the infrastructure has been built for the social media initiative, the team begins developing a message and delivery strategy. Developers should recognize target users’ needs and their context and, if possible, engage them in the development process. The six attributes of messages (ie, relevance, aesthetics, readability, findability, credibility, and usability) and three dimensions of delivery strategies (ie, social media platform, mode of delivery, and specific parameters) need to be taken into account when creating the social media interventions.

The team can then start the activation stage, where they make efforts to embed triggers into the social media delivery mechanisms for recipients to use the messages. Developers may interact with multilevel stakeholders and investigate the enablers of and barriers to recipients’ use of the messages. By tailoring behavior change techniques to identified barriers and enablers, the development team can develop a social media strategy that has the greatest potential to affect message use in practice.

We fully acknowledge the complexity of developing and implementing social media interventions and incorporate the notion of complexity within this framework in several ways. As information and communication sciences are fast-growing fields, new features and functions for social media platforms are continually emerging. Consequently, the approaches to developing messages and delivery strategies may become more diversified as technology advances. The dynamic interactions between constructs within the SMILE framework, such as the interaction between developers, recipients, and their situated contexts, make it challenging to undertake firm predictions [87,88]. Developers should immerse themselves in the human–social media system and capture underlining interactive patterns to inform the development of the most relevant and targeted activating techniques.

We also acknowledge the nonlinear aspect of the social media implementation processes, in which different levels of context influence and shape behavior. Promoting research use through social media is not a linear, straightforward process, and constant adaptations should be expected and embraced to optimize interventions. Finally, as each construct within this framework does not have a fixed and predetermined effect, and the interactions between constructs are dynamic and complex, we recognize that the framework has not been empirically validated and may not reveal all of the mechanisms at play for social media to influence research use. Nevertheless, the framework is based on current empirical evidence and well-recognized theories to provide plausible explanations for the successes and failures of social media interventions. Overall, the framework explicates the complexities of using social media in real-world practice and elucidates the key domains that developers, recipients, and researchers should attend to when developing or evaluating social media interventions.

As Maloney et al [72] suggested, “rather than looking at whether or not social media is effective for health professional education, it may be time to look at how various modalities can be optimized, both in terms of how the messages are delivered and how learners can be supported to engage.” Using social media to disseminate research evidence has become such an inexorable global trend that researchers should go beyond the investigation of the effectiveness of social media interventions and delve into the theoretical field on how to make it effective. The next stage of our project will be to test and refine the SMILE framework through a realist methodology that unpacks the mechanisms of how and under what circumstances social media works as a knowledge translation strategy for health care professionals to improve the delivery of research-based care.

The SMILE framework and its development process have 2 limitations. First, because of the multiple interactive components involved in developing and using social media for knowledge translation, as well as the massive amount of literature available from the relevant disciplinary fields, it was challenging to retrieve all pertinent theories and studies using a full systematic review approach. Instead, we used a targeted and flexible approach to select studies that allowed us to prioritize articles based on our framework’s development needs. It is possible that we missed some research and embedded our own values into the propositions by using this approach; thus, our next step is to test and refine the framework. Second, as the use of social media for knowledge translation in real-world practice is still in its infancy, we could not locate studies that captured all the SMILE framework’s propositions. More empirical studies of social media initiatives are needed to test the propositions of this framework.

In this paper, we propose the SMILE framework based on a review of social media initiatives, theories, and empirical studies as a preliminary understanding of how social media works as a knowledge translation strategy in health care decision-making. We provide a detailed description of each construct in the framework and offer suggestions for researchers and developers who intend to develop social media initiatives and interventions. For social media to be effective in enabling recipients to use research evidence in their practice decision-making, the SMILE framework purports that developers respond to target recipients’ needs and context to develop relevant social media messages and appropriate delivery strategies. Recipients’ use of messages is influenced by the virtual–technical, individual, organizational, and system contexts and can be activated by three types of triggers: sparks, facilitators, and signals. The SMILE framework maps the factors that are hypothesized to influence recipients’ social media message use and offers a heuristic device for social media developers and researchers to develop social media interventions. More empirical studies and social media initiatives are needed to test the propositions of the SMILE framework.