This review adhered to the guidelines outlined in the Cochrane Handbook for Systematic Reviews of Interventions, version 6.3, 2022 [12]. The reporting of the systematic review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Multimedia Appendix 1) [13]. The review protocol was registered in PROSPERO on March 14, 2022, with the ID CRD42022304229 [14]. Searches were conducted from the year 1946 to June 29, 2023.

We conducted searches for publications of randomized controlled trials (RCTs) using keywords related to social media and vaccination campaigns. Journal articles were sought in databases including MEDLINE, Embase, PsycINFO, CINAHL (Cumulative Index to Nursing and Allied Health), and CENTRAL (Cochrane Central Register of Controlled Trials). Additionally, Google Scholar was searched, and the first 200 hits were assessed for eligibility criteria. Searches were also conducted in the International Clinical Trials Registry Platform (ICTRP) and ClinicalTrials.gov for ongoing studies. The reference lists of systematic reviews and other relevant publications were checked for studies that might not have been identified in the initial database searches. Gray literature was consulted by conducting searches on Google (Alphabet Inc.) using terms such as social media, vaccine, vaccination, and randomized trial. For the complete search strategy, please refer to Multimedia Appendix 2.

The identified references were uploaded to EndNote (Clarivate Plc.) version 20.3 [15] and Rayyan (Rayyan Systems Inc.) [16]. Two reviewers (RKH and EG) independently participated in the selection of studies. Any disagreements were resolved through discussion with a third reviewer (NB).

Inclusion and exclusion criteria were based on PICO (population, intervention/exposures, comparison, outcomes) elements and are listed in Textbox 1.

Two reviewers (RKH and EG) independently assessed the risk of bias using the Cochrane Risk-of-Bias Tool for RCTs [18], and the certainty of evidence was evaluated using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) assessment [19]. Any disagreements were resolved through discussion with a third reviewer (NB).

A single reviewer (RKH) conducted the data extraction, and a second reviewer (EG) verified the accuracy and completeness of the extracted data. The data extraction focused on the categorization of studies and how to potentially pool the results. The extracted information included the following: (1) bibliographic information (authors, date, title, and country); (2) study characteristics (duration of the study, study setting, study design, loss to follow-up, and type of vaccines); (3) population (average/mean age, gender, sample size, ethnicity, and socioeconomic status); (4) intervention (type of social media); (5) comparison or control group; (6) outcome/measurement (days unvaccinated, vaccine acceptance, and uptake of vaccines compared with baseline); and (7) outcomes (vaccine rate and the authors’ conclusions from their studies).

To visualize the effect of the intervention, an effect direction plot was created following the guidance outlined in the Cochrane Handbook on alternative synthesis methods. P values were calculated using GraphPad (GraphPad Software Inc.) [20] to test the probability of the null hypothesis. The number of positive and negative effect direction arrows was counted for each outcome domain. All inconsistent effect directions were excluded from this sign test, given the calculation of a 1-tailed P value for each outcome domain.

A total of 469 hits were identified, and 126 duplicate articles were removed. The remaining 343 articles were assessed for relevance, with 289 being excluded due to incorrect outcomes, study design, or other reasons. Subsequently, 54 articles were eligible for full-text screening, of which 42 articles were excluded (refer to Multimedia Appendix 3 for an overview of the rejected articles and the reasons for their rejection; also see [21-62]). Ultimately, the result of these screenings resulted in the inclusion of 12 articles in this systematic review. Notably, 3 of these articles originated from the same protocol and were considered as 1 study. Refer to Figure 1 for a visual representation of the screening process [13].

This review encompasses 10 distinct studies presented across 12 articles. Notably, 3 articles [63-65] explored 3 different outcomes from the same study protocol [66]. The studies included in this review span from 2012 [67] to the most recent in 2022 [68]. Table 1 presents an overview of the included articles.

The outcomes of the 12 included articles were categorized according to the prespecified outcomes. Six studies [65,67,69,71,73,75] reported on the uptake of vaccines compared with baseline, 5 [63,68,70,72,74] assessed vaccine acceptance, and 1 [64] analyzed the days unvaccinated.

Four studies [65,67,69,71] of good or fair quality examined the uptake of vaccines compared with baseline. One of the studies [69] reported no difference between the 2 study groups. The other 3 studies reported a positive effect from the intervention: coaching through Facebook live sessions was found to be effective in reducing COVID-19 vaccine hesitancy [71]; personal health management system had a small but significant effect on influenza vaccination rates [67]; and web-based interventions, with and without social media components, showed higher uptake rates of the influenza vaccine in pregnant women receiving the intervention [65]. The 2 studies [73,75] of poor quality reported no differences in vaccination outcomes between groups.

Five studies investigated vaccine acceptance as the outcome. Four of them [63,68,70,72], of good and fair quality, reported that the intervention had a positive impact on vaccine acceptance. The fifth study [74], of poor quality, reported that the Facebook intervention had a positive effect on vaccine knowledge and acceptance. For a visualization of the effect direction, see Table 2.

Because of variations in outcome measurement and reporting among the included studies, pooling the data across studies to generate a single-effect estimate was not possible. However, to provide a systematic and transparent assessment of the certainty of evidence, we performed a GRADE assessment based on 5 GRADE domains to judge our certainty in the studies. The certainty of evidence was influenced by methodological limitations or risk of bias, indirectness, imprecision, inconsistency, and the likelihood of publication bias within the domains [19]. The grading results indicate that there is a reason to have less confidence in the effect estimate. The GRADE assessment for the outcomes is presented in Table 3.

Considering a health care system under pressure, it is crucial to explore how existing information channels can be leveraged to optimize the available resources within the health care system. In our review, we identified 12 randomized controlled trials (RCTs) that utilized social media to enhance vaccination rates. Although each of the studies contributes value by demonstrating positive results using various forms of social media to increase vaccine rates, the overall evidence remains limited.

As this review demonstrates, social media interventions have the potential to enhance knowledge about vaccines and increase the willingness to get vaccinated for oneself and one’s children. Previous research on vaccine hesitancy and behavior change theory–based social media interventions has also indicated this positive effect [5].

Robichaud et al [21] stated that there is an opportunity for public health organizations to actively engage in promoting factual and useful health messages regarding the benefits of vaccination using social media. Even 8 years later, D’Souza et al [77] investigated YouTube as a source of medical information on the COVID-19 virus disease. The authors advocated for information materials from official health agencies to disseminate valid and informative information to the public [77]. They also suggest that social media should be monitored by established health care personnel to maintain the platforms with fact-based knowledge on health issues and ensure that misleading and harmful information is not spread [77].

Previous research indicates that direct communication between health care personnel and the public is a factor that reduces vaccine concerns and might improve vaccine uptake [78]. Therefore, the use of social media platforms by health care personnel to enhance meaningful dialog regarding vaccine acceptance is encouraged [78]. This statement aligns with the findings in our review, where 3 studies [67,69,71] underscored the importance of active involvement from health care personnel in settings where health issues are communicated via social media platforms. It appears to be of great importance that health professionals, assuming roles such as informants, moderators, and effective discussion partners, play a role in distributing accurate and fact-based information on social media platforms. Mothers whose daughters have completed the vaccine program are considered effective representatives in influencing vaccine programs, as assessed by Buller et al [22]. Liao et al [69] stated that online information effectively promotes mothers’ self-efficacy to vaccinate their children against seasonal influenza. Nevertheless, the authors highlight that the active involvement of health professionals in online discussions is important in shaping positive discussions about vaccinations. Abdel-Qader et al [71] concluded that coaching by pharmacists and physicians through Facebook groups is effective in reducing rates of COVID-19 vaccine resistance and hesitancy. Ugarte et al [68] concluded that colleague guidance in the form of peer-led online Facebook groups can be useful for disseminating health information to help combat COVID-19 vaccine hesitation among essential health care workers.

The WHO estimates a projected shortfall of 10 million health workers by 2030, with the majority occurring in low- and lower-middle-income countries [79]. Health workforce shortages and the changing health needs of the public are contexts where digital transformation can offer unique opportunities [80]. Given the existing shortage of health personnel and the increasing burden on those who remain, it will be crucial to enhance the efficiency of those who will carry out health work in the future.

It is imperative to conduct further research on the mechanisms at play on social media with the aim of intervening in health issues such as vaccination. More systematic studies are needed to investigate how commercial social media platforms can effectively influence vaccination rates, allowing the results to be generalized to other settings and, potentially, to address other health issues. Digging deeper into specific issues, such as populations’ vaccine attitudes, would be significant for implementing timely interventions aimed at averting adverse public health consequences [81].

According to this review, there is a need to further explore which populations are most receptive to this type of intervention. Additionally, it is important to uncover the main features and characteristics of the most effective social media campaigns for vaccination.

When creating a social media intervention, establishing trust between the target population and the authorities and health care personnel is crucial [23,82]. Additionally, several other factors merit consideration: providing information on both risks and benefits, and acknowledging the concerns of the audience are essential components. Avoiding scientific jargon is imperative, and it is crucial to be transparent about funding sources. Referencing all sources of health information is equally important, along with providing quick responses and tailored personalized information [23]. It is essential to recognize that vaccine hesitancy is a complex phenomenon, not solely rooted in a deficit of comprehension. Vaccine hesitancy encompasses multifaceted considerations, including religious beliefs, safety concerns, low confidence in governments, and a range of other factors [83-85]. Recognizing this diversity of perspectives is crucial when formulating effective strategies to address and mitigate vaccine hesitancy within communities [82].

Previous research has shown a high prevalence of vaccine-related misinformation on social media [86], leading to vaccine hesitancy [83]. It is suggested that including fact-checking labels on posts containing misinformation can make viewers more favorable toward vaccines [70]. Designing, building, and evaluating theory-driven social media platforms aimed at making intervention recipients feel more comfortable about vaccines are suggested in the literature [23]. Additionally, monitoring by experts such as nurses, doctors, and other health care providers is recommended.

To influence the vaccine decision-making process, key factors include the source delivering the information, the network structure, and the framing of the information [87]. Similar findings are evident in other studies on this topic. For instance, a study revealed that vaccine-critical websites and blogs negatively impact the intention to vaccinate [78]. Moreover, even brief exposure to vaccine-critical websites increases beliefs in vaccine risk and hesitancy [88]. The question of framing the content of vaccine information becomes crucial in the construction of social media interventions. Lee et al [89] explored media design and choice for promoting HPV vaccination online, highlighting that the content itself plays a vital role in promoting health. They reference previous studies that describe messages emphasizing the negative consequences of neglecting recommended behavior, known as loss-framed messages, as more effective than the opposite kind of messages, namely, gain-framed messages [89].

We acknowledge the presence of several limitations in this review. Language limitations were a factor, and as a result, we may have overlooked relevant studies published in other languages. It is possible that we did not identify all studies eligible for this review, but the likelihood of this is considered minor. Additionally, there may be studies published after the conclusion of the searches for this review.

Several limitations are associated with the included studies in this review. The level of heterogeneity was notably high, which limits the potential for quantitative comparisons in a meta-analysis and the ability to conduct subgroup analyses. This heterogeneity arises from divergent data across study populations, varied data collection methods, differences in exposures and outcomes assessed, and diverse applied methodologies. The utilization of social media platforms also varied significantly, with some studies describing platforms that cannot be directly generalized to other conditions. Furthermore, some studies were assigned a high risk of bias by the review members. The included studies also vary significantly in size, ranging from the smallest with 58 randomized participants to the largest with 21,592 participants. Summarizing the material available through this review, considering both the benefits and limitations of using social media as a means of communication for distributing vaccination information, is challenging due to these reasons. Additionally, the studies included are context specific, further complicating a comprehensive summary.

The distribution of the studied population is skewed, with a notable focus on women in a large portion of the studies. It is conceivable that a more balanced gender distribution might have yielded different results in some of the studies. Exploring how gender influences the receptivity of information distributed through social media is of significant importance and interest. Understanding whether there are differences in the way messages should be adapted to different gender categories could provide valuable insights. It is noteworthy that studies focusing on women were exclusively conducted in developed countries, and therefore, the results may not be readily generalized beyond these settings. In these environments, where there is a high probability that women’s decisions carry weight within a family considering the advantages and disadvantages of vaccination, the findings may be context specific. In developing countries, the situation can differ, and men’s voices may hold more influence. In cultures where gender roles strongly shape knowledge and acceptance of vaccination, it is crucial to consider these dynamics when planning how to effectively reach participants in a vaccination program.

Given that blinding was not applicable in the included studies, the domains related to this aspect were not taken into account during the GRADE assessment. Despite this, all 3 outcomes were graded, including the outcome assessed by only 1 study. Confidence in the evidence was primarily downgraded due to heterogeneity between the studies and concerns related to study designs. Two of the outcomes were downgraded due to the low quality of the studies as assessed in the risk-of-bias assessment, imprecision stemming from nonsignificant confidence intervals, and inconsistency in the varying forms of reporting results across different studies. Overall, the body of evidence is graded as low, indicating that the results must be interpreted with caution.

This review underscores the substantial and untapped potential associated with using social media as a communication channel for health issues. With a strategic understanding of how to harness these mechanisms effectively, social media has the potential to reach a wide audience rapidly and in a cost-effective manner. Social media, when used as a supplementary promotional channel, can serve as an instrument for transmitting information that has the potential to increase vaccination rates in a population. However, the effectiveness of these tools relies on authorized personnel closely monitoring and moderating discussions. Numerous studies have explored how social media contributes to increased vaccine resistance. However, there is a pressing need for more knowledge on how social media can be optimally utilized to enhance vaccination rates in a population.