Web 2.0 for Health Promotion: Reviewing the Current Evidence

Abstract

As Web 2.0 and social media make the communication landscape increasingly participatory, empirical evidence is needed regarding their impact on and utility for health promotion. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we searched 4 medical and social science databases for literature (2004–present) on the intersection of Web 2.0 and health.

A total of 514 unique publications matched our criteria. We classified references as commentaries and reviews (n = 267), descriptive studies (n = 213), and pilot intervention studies (n = 34).

The scarcity of empirical evidence points to the need for more interventions with participatory and user-generated features. Innovative study designs and measurement methods are needed to understand the communication landscape and to critically assess intervention effectiveness. To address health disparities, interventions must consider accessibility for vulnerable populations.

WEB 2.0 TECHNOLOGIES HAVE significantly changed the health communication landscape in recent years. Leaving the health impact aside, we have witnessed a shift in the online environment, from unidirectional and “read-only” (whereby information is “pushed” on passive audiences) to multidirectional communication characterized by participation, collaboration, and openness.^1,2 Through social media sites such as Facebook, YouTube, blogs, and forums, individuals obtain information and contribute to online content in an interactive, networked environment. For instance, the online community PatientsLikeMe helps individuals connect with others with the same illness and share information about symptoms and treatment options.³ If one considers public health, Web 2.0 media are challenging traditional health promotion models and prompting the advancement of innovative health promotion and communication methods with rigorous impact assessment.^2,4–6

In response to the growing participatory communication environment, public health practitioners and researchers have begun undertaking descriptive and intervention studies to assess how Web 2.0 and social media shape health-related knowledge, attitudes, and behavior. In addition, a number of Web 2.0–based interventions have recently been tested to evaluate feasibility, usability, and effectiveness for different populations. Health outcome impact assessments for a diverse range of health topics are only beginning to take shape given the rapid developments in Web 2.0 technologies.

A key research question facing the changing communication landscape centers on the issue of the digital divide. Inequalities in access to and use of technologies have been observed, whereby those of lower socioeconomic status, minority race/ethnicity, older age, poorer health, and living in geographically isolated locations are less likely to have adequate access.^7–13 A 2009 Federal Communication Commission survey indicated that the main dividing lines for broadband access are along socioeconomic dimensions such as income and education.¹⁴ Furthermore, research on health information technology adoption has found disparities in access and utilization patterns across socioeconomic groups and health literacy levels.¹⁵ Indeed, the digital divide parallels patterns of health disparities in many ways.

On the other hand, there is growing evidence that the digital divide may be narrowing. Nationally representative data indicate that, when Internet access is controlled for, race/ethnicity, income, and education do not dictate social media participation.¹⁶ With equal likelihood of accessing social media, populations suffering disproportionately from health disparities are afforded unprecedented opportunities for online information and services. Moreover, mobile technologies are increasingly accessible regardless of demographics. In 2010, 40% of American adults used a mobile phone to access the Internet and text messages, up from 32% in 2009¹⁷; of interest, ethnic minorities were more likely than Whites to own a cell phone and to utilize a wider variety of its data-based functions. In a similar way, mobile phone use has begun to diminish the impact of income on Internet access; a recent survey found that 27% of adolescents with mobile phones used them to access the Internet, and, for adolescents from lower-income households, the rate jumped to 41%.¹⁸ This is evidence that mobile phones are lowering the access barrier to Web 2.0 technologies.

Empirical evidence is gradually emerging to inform participatory and “bottom-up” health interventions for various population segments. Although the characteristics of social media are identified in some recent publications, a number of key questions remain to be answered through a critical literature review: How is the current Web 2.0 environment affecting health attitudes and beliefs? What are the outcomes of social media–based health interventions? How have they been assessed? What proportion of interventions target populations that face health disparities and what are their implications, if any, for alleviating health disparities?

To investigate these questions, we conducted a comprehensive review of published studies on the impact of Web 2.0 on health. Although we are mindful of the increasing diversity of platforms in which scientific information is shared beyond traditional peer-reviewed venues (e.g., personal blogs, articles available through general Web-based searches), to delineate the scope of the study and make it more replicable, we focused on published references searchable through major bibliographic databases. We sought to (1) summarize existing descriptive studies of health-related Web 2.0 activities and efforts; (2) critically review a range of intervention studies, including feasibility, usability, pilot, and randomized clinical trial projects; and (3) discuss the implications of the current results for health promotion and health communication, particularly with marginalized communities. Note that this review adheres strictly to the defining characteristic of Web 2.0, namely its participatory nature; hence, it will not cover Web-based interventions without a user-generated component, or unidirectional mobile health programs. Finally, although numerous Web 2.0 programs provide information and social support to diverse populations, only a small portion have been empirically evaluated and reported in peer-reviewed journals; this review excludes unpublished and untested programs.

METHODS

Our search process conformed to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for reporting systematic reviews.¹⁹ PRISMA facilitated our design of a screening protocol for study inclusion, and also provided useful guidance in our qualitative and quantitative syntheses of culled studies.

Literature Search and Review Protocol

Three researchers from the team (W. C., A. P., and C. L.) collectively established the coding protocol following an integrative (qualitative and quantitative) paradigm. We determined databases, search terms, and inclusion and exclusion criteria through an iterative, mixed methods approach.^20,21 That is, each member separately conducted a qualitative review of selected Web 2.0–based publications to explore themes, approaches, measures, and populations covered in the works; we then convened to create a review protocol.

After initial qualitative explorations, all authors convened to identify electronic bibliographic databases and search terms through a consensus process. We selected 4 of the largest databases in medicine, health, and social sciences: Web of Science, Scopus, PubMed, and PsychINFO. Members each brought a list of relevant search terms, and we collectively decided on the final set. Terms used in search through titles or abstracts included the following: social media, new media, participatory media, user-generated content, Facebook, MySpace, Twitter, YouTube, Second Life, LinkedIn, wiki*, blog*, Web 2.0, online social network, and social networking. Informed by the qualitative exploration, we decided not to limit studies further by methodology; we agreed on being inclusive of study design, participant, and setting, because of the nascent state of Web 2.0–related health research. In particular, we chose to include a wide variety of platforms, including blogs and microblogging technologies, social networking sites, video sharing programs, and mobile health applications.

Moreover, seeing many nonhealth studies in our initial exploration, we decided to only include articles with an identifiable health outcome (e.g., improving or promoting health knowledge, attitudes, or behaviors in disease prevention and management). To this end, our search only included references in which 1 or more of the terms appeared together with the word “health” in the title or abstract. We further limited the search to English-language studies published since 2004, when the term “Web 2.0” was coined to describe the shift to a more participatory online landscape.⁵

We divided up the search tasks, with each member of the team searching 1 or 2 of the bibliographic databases. All searches were performed on December 20, 2011, and resulted in 1258 total entries from Web of Science (n = 507), Scopus (n = 482), PubMed (n = 267), and PsychINFO (n = 2). After rendering the initial search results, we manually deleted duplicate articles across databases (total n = 542). The team then identified a manual coding process, beginning with the exclusion criteria. Although we erred on the side of inclusivity in the automatic searches, to truly represent the defining characteristics of Web 2.0, we included in the protocol a manual exclusion process, whereby we excluded studies without a participatory component (i.e., some level of participation or interaction by the target audience) and those without any substantial connection to health (that is, even if the word “health” were mentioned in the title or abstract). We also manually excluded publications that were not in English or lacked sufficient information (e.g., missing author names) for this review.

The manual exclusion was done by dividing up the remaining search results, and 3 research team members (W. C., A. P., and C. L.) independently reviewed their respective sections—reading abstracts or full articles—to identify potential articles to exclude on the basis of these criteria. Decisions on the final sample were based on team consensus using these inclusion and exclusion criteria.

Data Analysis Protocol

Being mindful of our inclusive approach in selecting publications, we felt it was important to reflect the type of work represented by categorizing them in terms of overall approach and study types. We reached a consensus to code all references into 3 general types: commentaries and reviews, descriptive or observational studies, and intervention studies. These 3 categories posed minimal overlap, and the categorization process was straightforward for all coders.

For descriptive or intervention studies, the coding protocol included identifying the race/ethnicity of the target population to determine whether a study targeted minorities or underserved populations. This coding decision was predicated on our hypothesis that the majority of Web 2.0 health studies did not account for vulnerable populations suffering from health disparities. Moreover, among intervention studies, we developed further codes, including study design (feasibility studies, pilot interventions, and full-scale interventions), age of target population (youth vs adult), and targeted health outcomes (health promotion, professional education, and others). These are distinguishing features noted by team members and proposed during protocol development.

In addition to the study sample, we coded the excluded studies for reason of exclusion in an effort to inform future systematic reviews related to Web 2.0 communication. The development of these categories was iterative and a posteriori. Figure 1 summarizes the search and review process.

FIGURE 1—

Literature search and review process regarding Web 2.0 for health promotion.

In addition to these quantitative findings, the review relied on qualitative synthesis of the articles. To do so, the team collectively identified emerging themes in the review and collated common themes through an iterative discussion. We decided to present a narrative synthesis of the literature by illustrating specific examples, maximizing the range of studies to optimally characterize the full spectrum of work.

RESULTS

First, we coded excluded articles a posteriori into 5 general categories. In total, the research team manually excluded 202 miscategorized publications. Among these were

1. 154 publications that did not include a participatory component (e.g., unidirectional Web site physical activity intervention),

2. 36 publications that did not have a health focus (e.g., assessing information accuracy on Wikipedia without a specific focus on health topics),

3. 6 publications wherein a lab’s social media presence (e.g., Facebook page or Twitter handle) was the sole mention of Web 2.0 technology (e.g., “Each (bovine) gene page is linked to a wiki page.”),

4. 5 non-English publications, and

5. 1 publication where the author was anonymous and reference to the research project setting was missing.

Among the final study sample of 514 unique citations, we identified 3 main types of studies: (1) commentaries and reviews (n = 267), (2) descriptive or observational studies (n = 213), and (3) intervention studies including feasibility or usability and pilot interventions (n = 34). Key results are synthesized in the next paragraphs. For the intervention studies, the team selected examples, listed in Tables 1 and 2, to illustrate the nature and scope of the projects.

TABLE 1—

Selected Feasibility Studies Regarding Web 2.0 for Health Promotion

Reference	Health Topic	Platform	Study Goal	Key Findings
O’Dea and Campbell22	Mental health	SNS	Can SNS promote peer support for mental health support seekers?	Of participants in need of mental health, 53% turned to the Internet, 82% used SNS regularly, 47% believed SNS could address mental health problems by linking people with shared conditions.
Ralph et al.23	Sexual health for lower-income adolescents of diverse ethnic backgrounds	SNS and mobile health (mHealth)	What are adolescents’ views on Web 2.0 and texting for health education?	Minority adolescents are less likely to access Internet from home; adolescents like anonymity and convenience, and are concerned about information accuracy; moderate acceptance of MySpace friend request from clinics. Clinical staff use SNS to reach and build relationships with adolescents and reinforce messages, but are concerned about hacking, victimization, time consumption, and constant changes and transience of SM platforms.
Rozental24	Patient–physician communication	SNS	Are SNS viable channels for patient–provider communication?	Age, education, computer ownership predicted SNS use. Facebook, MySpace, and Twitter are most common platforms. Only 31% interested in using SNS to communicate with physicians.
Rushing and Stephens25	Health and fitness for American Indian youths	SNS	Is SM a feasible way to reach youths with health messages?	Youths want information on fitness, drug and alcohol use, nutrition, and stress management. Specific information and design features targeting American Indians desired (e.g., traditional healing methods, spiritual beliefs, culture, stories, history, symbols, and design).
Silenzio et al.26	Suicide prevention for GLB young adults	SNS	Can SNS be used for suicide prevention?	Social network analysis and simulation predicts that a peer-driven prevention intervention could have a wide reach; dissemination potential quantified.

Note. GLB = gay, lesbian, and bisexual; SM = social media; SNS = social networking site.

TABLE 2—

Usability and Pilot Studies on Selected Social Media–Based Interventions Regarding Web 2.0 for Health Promotion

Reference	Health Topic	SM Platform	Study Goal	Intervention Design	Primary Findings
Chiu et al.27	Water intake	Mobile app	Tested mobile app of multiuser game to increase water intake	Playful Bottle: mHealth app to track water intake; single-player game with system reminders; multiuser game with reminders from peer players	Both games improved water intake; social reminders were more effective than system reminders.
Foster et al.28	Physical activity	Facebook app	Tested whether competitive interaction motivates exercise	StepMatron: Facebook app to provide social and competitive context increase exercise	Ability to view, compare, and comment on physical activity of other users led to slight increase in steps.
Franklin et al.29	Pediatric type I diabetes	Mobile app	Tested app to enhance self-efficacy, insulin therapy uptake, and glycemic control in patients	Sweet Talk: mHealth app with text messages for diabetes self-management and peer support; direct link to care team	Blood glucose levels improved for those who used app with intensive therapy. App improved self-efficacy, adherence; seen as useful and sustainable self-management tool.
Gay et al.30	Emotional well-being, social support	Mobile app	Tested app to enhance evaluation, awareness, and sharing of emotions	Aurora: mHealth social emotion recording and sharing app to encourage reflection, awareness, and sharing of emotions	Users reported increased comfort in socially expressing and sharing emotions. App may have encouraged social support.
Linehan et al.31	Diet and nutrition	Web app	Tested if a food diary app motivated uploading of meal photos; explored user tags in USDA database	Tagliatelle: Web-based food diary where users upload photos of their meals to be tagged for nutritional content	Participants liked tagging images; 65% of tags in USDA database, but most lacked specificity to facilitate derivation of nutritional information.
Marciel32	Management of cystic fibrosis in adolescents	Mobile app	Tested whether app improved knowledge, social support, and treatment adherence of patients with cystic fibrosis	CFFONE: Web-enabled phone with information and social support to boost treatment adherence among patients with cystic fibrosis; included care management calendar, chat, text reminders, social networking, games, direct link to care team	CFFONE was considered likely to improve knowledge, social support, and adherence and facilitated social support; found acceptable, useful, and feasible.
Moreno et al.33	Sexual health education for adolescents	MySpace	Tested whether physician intervention limits adolescents’ displays of risky behavior and encouraged STD testing	Physician e-mail suggesting that adolescents moderate displays of risky behaviors on profiles and offering information on STDs, STI testing	Intervention decreased public mentions in profiles of sexual behavior but not substance use and no changes in privacy settings. One of the 3 changes was implemented by 42% of intervention group versus 30% of control group.
Stanforth34	Physical activity	YouTube	Tested YouTube intervention to increase exercise self-efficacy and behavior	YouTube videos of peers providing exercise tips, stories, and encouragement	More than 50% who visited site watched the videos; no differences in exercise behavior, self-efficacy, or social support between those who watched videos and those who did not.
Yellowlees and Cook35	Mental health	Second Life	Tested a SIM environment as an educational tool about psychotic hallucinations	Second Life SIM of psychiatric inpatient ward, including simulations of auditory and visual hallucinations	SIM improved understanding of schizophrenics’ auditory and visual hallucinations. Almost half rated SIM “disturbing”; most would recommend to others.
Zheng et al.36	End stage renal disease and young adults	Custom SNS	Tested Web site to provide social support and sense of community	Ktalk.org: SNS to help young adults with end stage renal failure develop new normal lives, restore social identities, and regain confidence; features video interviews of peer mentors, forum, blogs, SNS profile, useful sites and information	Participants indicated a strong intention to continue use and recommend to others. Perceived usefulness and ease of use predicted intentions to use Ktalk.

Note. SIM = simulation; SM = social media; SNS = social networking site; STD = sexually transmitted disease; STI = sexually transmitted infection; USDA = United States Department of Agriculture.

Commentaries and Reviews

The commentaries and reviews heralded social media (or “participative Internet”) as a powerful tool for interactive health promotion.^4,8,17,37 According to these manuscripts, the key advantages of these platforms included improved reach, increased interactivity, low cost, and the capacity to communicate personalized and tailored messages quickly.³⁷ Compared with other health intervention approaches and channels, social media approaches were advocated for their ability to improve intervention reach because of their ubiquitous nature.^4,37 In particular, scholars suggested that social media may have the capacity to engage traditionally “hard-to-reach” populations, including adolescents³⁸ and ethnic minorities.³⁹ Second, the interactivity of Web 2.0 was argued to increase audience participation, allowing them to coconstruct health promotion messages and interventions.^2,4 Third, compared with traditional media, social media–based health information dissemination efforts were said to be low in cost.⁴⁰ Finally, social media represents a hybrid channel for intervention delivery with the benefits of both mass media (e.g., wide reach) and interpersonal communication (e.g., personalized messages to motivate behavior change).⁴¹ As a consequence, interventions may “take advantage of the synergistic contributions of mass and interpersonal media needed to effect change on individual, institutional, and social levels.”^2(p333)

The commentaries also spoke to the limitations of social media for health communication. Several articles alluded to the possibility of differential effectiveness across populations of different socioeconomic status or health literacy levels.⁴ Moreover, authors acknowledged limitations in outcome measurement of social media–based health interventions, pointing to the ubiquity of the Internet, difficulties in tracking engagement, and limitations in fielding online surveys to assess participants’ feedback and perceptions toward Web 2.0 for health communication.

Descriptive Studies

There were ample descriptive studies that made observations on social media participants’ conversations about and attitudes toward health topics. The most common method was content analysis of blogs and social networking sites focusing on health promotion or disease prevention or management (e.g., smoking cessation and pediatric cancer support groups^42,43). Some analyses evaluated the validity and credibility of discussions compared with scientific evidence and clinical guidelines (e.g., prostate cancer information on YouTube vs clinical guidelines⁴⁴). Some studies surveyed social media participants (e.g., Caringbridge or PatientsLikeMe users³) on needs, preferences, and participation features on health-related social media sites. The results, which varied by forum, generally pointed to potential for improving health through these platforms. Included in many descriptive studies were discussions of future interventions, from preliminary ideas to detailed research designs. These forward-looking projects were categorized as descriptive because they lacked empirical data or an evaluation component. Finally, of note, among the 213 descriptive studies, only 12 were coded to focus on minorities or underserved groups.

Intervention Studies

Social media–based health intervention studies were categorized into 2 subtypes: feasibility or usability studies and pilot interventions. Selected studies are listed in Tables 1 and 2, where we have taken care to maximize the range of publications to represent various disease outcomes, platforms, and research questions. Feasibility studies addressed the question of whether social media is an appropriate and “feasible” venue to deliver health communication interventions. Key questions included the acceptability of a particular channel to stimulate conversation among a target population around a specific health topic, target users’ preference for content, Web 2.0 features, and mode of communication. Many publications did not include the design and implementation of an actual intervention, but merely assessed the likelihood that participants will embrace and benefit from the planned interventions. Table 1 highlights representative feasibility studies.

Beyond feasibility assessment, pilot interventions (n = 10) have been undertaken to determine preliminary outcomes. The range of target endpoints included awareness of a health topic^33,35; engagement³¹; changes in attitude, affect, or behavior^27,28,30,34; and health outcomes.^29,32 Table 2 lists exemplary studies in each domain for illustration.

Social comparison and network connection were most commonly harnessed characteristics of social media. Several pilot interventions used social competition to encourage health behaviors; for instance, 1 mobile app featured social games to increase water intake, and a separate Twitter intervention promoted user competition in physical activity.^27,45 Some interventions leveraged network connections by enabling individuals to broadcast emotional, informational, or instrumental support needs so their social networks could respond accordingly.³⁰

In terms of health outcomes, the 2 most common intervention objectives were health promotion and professional education. Among the 34 intervention studies, 17 targeted health promotion, including nutrition, sunscreen use, and physical activity. In addition, 9 trials were directed at the education of allied health professionals, pointing to the potential of social media to train and support health care personnel. In terms of representation across populations, only 3 of the 34 studies targeted racial minorities most affected by health disparities.

Finally, the pilot interventions were generally conducted with a small sample size. For example, only 16 participants were included in an initial test of an mHealth application for water consumption,²⁷ and another pilot test of a Twitter-based intervention to increase physical activity included only 4 adolescent female participants.⁴⁵ Although there is value in these initial tests, larger and more diverse samples are necessary to strengthen the validity of the conclusions.

DISCUSSION

Our findings highlight several key themes and suggest future directions for research and practice related to Web 2.0.

Emerging Critical Themes

Need to harness the participatory nature of social media.

Although commentaries and reviews focus on the unique opportunities afforded by social media, namely the utility of participatory, user-generated content in enhancing health interventions, the majority of the interventions do not optimally leverage this participatory nature. To illustrate, a recent study tested the efficacy of various modes of advocacy for a screening service for sexually transmitted infections.⁴⁶ The authors analyzed social media channels (e.g., Facebook) together with a range of traditional promotional avenues (e.g., newspaper ads, Web pages, and posters) to measure the number of generated referrals to the service, and as such, missed an opportunity to evaluate the unique facets of social media for health message diffusion.

Furthermore, user-generated content and audience feedback on an intervention or message can serve as valuable evaluative data toward the design or refinement of future health promotion efforts. For example, an experimental study confirmed the influence of viewer comments on other audience members’ perceptions of health-related YouTube clips,⁴⁷ demonstrating the power of the participatory nature of the Web to enhance (or in some cases, detract from) the efficacy of health promotion efforts. Moreover, content analyses of YouTube clips revealed high prevalence of fat or obesity stigmatization. In fact, compared with traditional media, social media and YouTube contained a far higher proportion of weight-based teasing and stigmatization, and comments on YouTube clips about obesity were similarly stigmatizing and negative.^48,49 To this end, we urgently need more observational research with similar aims to document and evaluate the role of user-generated content uncontrolled by preplanned interventions.

Information source and accuracy.

The observational studies we reviewed generally showed user-generated health content on social media (e.g., YouTube, Twitter) to be inconsistent with clinical guidelines or scientific evidence, raising concern over the credibility of health information in this forum.⁴⁴ Although these concerns are valid, the transparency of social media presents an opportunity to monitor the spread of misinformation and intervene with credible information. Furthermore, descriptive studies of social media health information sharing may illuminate alternative strategies in designing social media–based health campaigns. For instance, Prochaska et al. found that although only a small portion of Tweets on smoking cessation adhered to clinical guidelines, those who were Tweeting were sharing tips about how they quit smoking (i.e., advice drawn from personal experience).⁵⁰ This type of peer-to-peer health exchange parallels a main finding in online support group studies where sharing personal advice dominates over sharing knowledge and guidelines.^51,52

On a broader level, studies demonstrate that individuals seek different types of information from online support groups and social media than from various online informational tools for guiding health decisions (e.g., mayoclinic.com for cancer treatment options). Future interventions may incorporate a holistic approach to disseminating guidelines or evidence-based health information by combining system-generated content (e.g., clinical guidelines) and user- or peer-generated content related to personal experience.

Implications for the digital divide.

Although commentaries repeatedly point to the potential for social media health interventions to reach underserved and marginalized populations, consequently decreasing the digital divide and information disparities, our review demonstrates that this stated potential has not been realized in empirical studies. Limited empirical work presents the most apparent challenge: as evidenced by the relatively small number of studies reviewed here targeting underserved populations, more work is urgently needed to examine how populations affected most by health disparities use and can best benefit from health-related social media. Moreover, health literacy research has demonstrated differential engagement levels and outcomes of technology-based communication interventions across different SES groups and health literacy levels.^15,53 Interventions need to account for factors such as literacy demand, message comprehension, relevancy, and trust of information source, and determine intervention effectiveness for audiences with limited literacy. These important questions will help us understand social media’s promised potential for alleviating health disparities.

Despite much excitement in this area, presently we do not have sufficient evidence that social media can reduce disparities or ameliorate problems associated with the digital divide. We have more convincing data on the overall growing accessibility and lowering of access barriers: national survey data suggest the disappearance of differential social media and mobile technologies access among racial/ethnic minorities. This suggests that the access barrier is diminishing, at least for nonrural populations. Enabling equitable Internet access for poor and rural populations remains a high priority to continue to lower the access barrier for everyone.

Suggestions for Future Research

Considering intervention feasibility and usability.

Although commentaries generally concluded that social media can expand the reach of health communication messages to wider audiences, current evidence only tentatively supports this claim. For instance, one feasibility study found that fewer than half of adolescents would accept a friend request on MySpace from a sexual health clinic.²³ On another health topic, a study reported that only 31% of patients wanted to use social networking sites to communicate with their doctor.²⁴ Depending on the intended goals and message content, those already well-immersed in regular social media participation may not automatically be open to using it for particular health-related functions—“If you build it, they might not come.” It continues to be a high priority to use descriptive and pilot studies to ascertain feasibility (including content, platform choice, and design features) to sustain and engage target groups.

Innovative assessment methods.

This review highlights growing opportunities for health interventions to utilize an array of innovative measurement tools for assessing dissemination, exposure, engagement, and effectiveness. In fact, as experience with Web 2.0 grows, methods and approaches are becoming more sophisticated. For instance, on a YouTube health campaign, number of views, average time spent viewing, comments generated, and links shared can be easily assessed in real time. These metrics can illuminate the reach and impact of a particular message, as well as user-generated content in response to the message. Moreover, to better understand the complex interactions across users and platforms, techniques such as social network analysis are now commonly adopted in health research—for instance, to study large cohorts as well as online smoking cessation communities.^54,55 In addition, data mining and cloud-based computing techniques are increasingly recognized and utilized to facilitate the acquisition and analysis of large volumes of user-generated social media content. Specialized platforms can now be used to monitor as well as quantitatively and qualitatively analyze social media content around particular key words or phrases.⁵⁶ Validated machine learning models can be built to assess patterns of social media content as well.⁵⁷ These services are complemented by computational data analytic methods, such as corpus linguistics or natural language processing approaches to synthesize the content, sentiment, and associated interactions surrounding a particular health topic.

Intervention design and implementation.

The field is in need of more dynamic intervention designs to accommodate the rapidly evolving technologies and communication landscape. The references we reviewed repeatedly suggested that traditional study designs (e.g., randomized controlled trials and longitudinal cohort studies)—though still the gold standards in some areas of health sciences—are simply unfeasible in this constantly evolving communication environment. Fields outside health sciences (including engineering, computer science, education, sociology, and anthropology) offer valuable complementary approaches. For example, in rapid prototyping, research and development are conducted in tandem to create a prototype that is evaluated in advance of the final product. This paradigm could be useful to allow intervention components to be tested in real time as they are designed and built, thus helping evidence-based health promotion to keep pace with the rapid development of technology.^58,59

The incentive structure also needs to evolve to facilitate innovative intervention design and evaluation; this includes knowledge dissemination or publication venues, multidisciplinary review criteria, and support mechanisms. Although our review focused on published academic journals searchable through scientific databases, the field is beginning to more broadly consider open-access journals, rapid and open peer-review models, conference proceedings, and expert blogs to complement the traditional publication model. The advantages of being more inclusive are the lowered access barrier and faster pace of knowledge dissemination. Yet, challenges also accompany the democratization of scientific information and diminished gate-keeping functions of peer-reviewed publication venues, resulting in potentially mixed quality, decreased rigor in evaluation, limited replicability, and less use of common measures. Finally, funding mechanisms and review criteria need to be modified to be technologically nimble and reward innovative measurement and rapid study designs such as the ones outlined here.

Toward a Model for Health Promotion 2.0

The themes emerging from the review are poised to enhance existing health promotion or eHealth models. To illustrate, we chose Skinner, Maley, and Norman’s Spiral Technology Action Research (STAR) model⁶⁰ as an example because of its theory-driven model and dynamic, cyclic structure. The STAR model was developed to facilitate eHealth intervention and stresses engagement in ongoing dialogue with the target users of an information communication technology system. This model comprises 5 cycles of development:

1. identifying and understanding users’ needs,

2. planning ways that technology can meet these needs,

3. implementing these plans in system design,

4. reviewing the system and adjusting the design based on user feedback, and

5. implementing the system.

In this way, the model is analogous to social media–based health promotion efforts, whereby communication is seen as a conversation rather than unidirectional message transmission. The capacity of Web 2.0 technology components and capabilities to enhance each cycle of the STAR model is discussed in the next paragraphs.

First, one key finding from this review is the value of descriptive studies to understanding naturally occurring user-generated content. In enacting the cycles 1 and 2 of the STAR model, programs can benefit from “listening” to health dialogue on social media to identify users’ (information, behavior, support) needs. Considerations must also be given to target user communities’ demographics, psychographics, health literacy, and digital competence. In cycle 3, health promoters and community members may cocreate systems that offer digital spaces for connecting with others and exchanging information and support. These systems may be created from the bottom up, such as disease-specific social networking sites,³⁶ or they may be built on existing platforms, such as Facebook, forums, or blogs.^33,61,62 Another design consideration is striking a balance between system-generated and user-generated content. Is the intervention a space where health messages are “pushed” out for the audience’s response and commentary, and, if so, how is audience feedback incorporated into ongoing message design? Alternatively, does the intervention involve the audience as cocreators of content? If so, how might health promoters moderate or “vet” this content to ensure accuracy and appropriateness?

In cycles 4 and 5, the metrics associated with social media provide opportunities for comprehensive outcome evaluations ranging from developmental assessment (process outcomes) to changes in health-related knowledge, attitude, and behavior (health outcomes). Process outcomes can include satisfaction, engagement, usability, diffusion, and intervention “stickiness.” Assessment of health outcomes can determine intervention efficacy and effectiveness. As per the STAR model, development and program planning is sequential as well as iterative, allowing evaluation to inform in real-time intervention design and health message delivery.

Study Limitations

There are limitations to the present review. First, despite a systematic search strategy with carefully chosen search terms, some publications on the impact of Web 2.0 on health may not have been included in this review because of the variety of additional terms that may be used for this area (e.g., mHealth was not used). Second, we chose to focus on publications indexed in main medical and social science databases to maintain replicability of the review. As a consequence, certain gray literature and knowledge products in nontraditional dissemination venues such as blogs and organizational Web sites may not have been adequately included. In particular, significant strides in the use of social media have been made in private and not-for-profit sectors (e.g., health information firms, HMO health plans), but only a small proportion of the successes have been reported in the published literature. Toward this end, we see tremendous value in research examining unpublished documents and reports on this topic.

Third, a meta-analysis with definitive quantitative results was not possible because there were limited published feasibility and pilot studies, which contained a variety of data collection methods and targeted outcome measures, making it impossible to compare or harmonize outcome data across studies. Finally, whether improvements in outcomes (e.g., knowledge, physical activities) were the direct result of engagement with social media or associated mediators and moderators (e.g., increased peer support, information delivery) remains difficult to assess. Future research can employ rigorous assessment on several different levels and the results will shed light on new interventions.

Conclusions

Although the current review finds limited empirical evidence to support the efficacy and effectiveness of social media health promotion interventions, innovative design and measurement, coupled with the continuing growth of Web 2.0 penetration, place us at the cusp of changes in public health research and practice. Although rigorous intervention studies remain to be implemented, the field is already witnessing rapid changes. To date, the bulk of empirical evidence has been generated from descriptive research. For instance, there are abundant interesting discoveries stemming from content analyses of smoking cessation Tweets and YouTube videos about cancer screening.^44,50,63 The observational studies, as well as the feasibility and usability evaluation studies, provide important insights for best practices.

This review offers findings that can cohere toward developing an integrative health promotion model that maximizes the participatory nature of social media. We believe that rather than constructing a brand-new framework, it is better to avoid “reinventing the wheel” by adapting to and enhancing existing frameworks. As illustrated through the STAR model, additional Web 2.0–related elements can easily enhance and facilitate the use of existing models.

The present review summarizes the current state of research and practice in this growing area. Looking ahead, we anticipate a rapid growth of intervention studies, with evidence becoming available across diverse knowledge dissemination venues. As more evidence continues to be generated, our conceptual framework will evolve and our understanding of the utility of social media for a variety of public health objectives will improve.