Social media may assist in helping health services engage new audiences in the process of QI [41,42,44,51,52,56,59,62,63,73,80,85,86,88,91], but success was mixed with many studies reporting underutilisation by target audiences [42,45,47,51,52,56,59,60,63,64,65,66,70,74,75,76,78,79,83,84,86,89,90,96,97], particularly by people viewed as ‘hard-to-reach’ [98]. Based on their experience as stakeholders involved in health design and QI, the advisory committee believed that the most practical ways to overcome this limitation were educating potential users on how to use social media [43,50,51,52,61,69,73,81] and using social media in addition to other, more traditional methods of engagement (e.g., face to face) rather than relying on social media alone [39,51,52,60,69,71,73,74,80,85,88].

The risk of negative, false or malicious messaging was highlighted as an important finding, even though this risk was only mentioned in thirteen studies [41,42,47,49,51,52,53,64,74,84,86,90,97]. The discrepancy between the relatively small number of included studies which viewed negative, false or malicious messaging as a risk, and the importance placed on this risk by the advisory committee members may reflect issues around delays between conducting and publishing academic research in the context of a rapidly changing digital environment. For example, issues such as the strategic use of misinformation and bots to influence policy debate have only emerged in some of the more recent studies [47,60,64,90]. Other recent social media issues, such as social media algorithms creating highly polarised filter bubbles [99], and selling of user data to third parties [100], were not investigated in the studies included. It is likely that these issues will emerge in future research, particular given the complex advocacy, social support, information- and misinformation-dissemination role that social media has played in the COVID-19 pandemic [101]. Future research should continue to explore emergent risks and barriers to social media use and how they can be managed or overcome.

The advisory committee felt that the strategies of monitoring [49,55,65,69,71,72,73,84], moderation [39,65,69,93] and group rules [39,69] reported within the theme of ‘making discussions safe’ were the most important and relevant ways to manage the risk of negative, false or malicious messaging for stakeholders engaged in health service design or QI activities. Given the perceived importance of these strategies in making social media spaces safer, researchers would have expected they would be more frequently reported as an enabler for using social media to engage stakeholders in design and QI. This potential under-reporting of moderation, monitoring and group rules may reflect the types of platforms and their access features used in the included studies. Monitoring and moderation are likely easier to undertake and report on in registration-required or closed social media environments, where monitoring and moderation of small online groups is likely done by group members, organisers or paid content moderators [102]. The potential under-reporting of monitoring and moderation may be because the majority of included studies occurred in public social media spaces, where monitoring and moderation may be done through often opaque artificial intelligence mechanisms or anonymous reporting of problematic content [103], and is therefore difficult to report on, or not identified as an enabler in research articles.

Whatever the reason for the potential under-reporting of monitoring, moderation and group rules, the fact that the advisory committee highlighted the importance of these strategies indicates that they may be a key area for future research, in order to better understand the impacts of monitoring and moderation practices on stakeholder engagement in health service and system design and QI.

Social media can have positive impacts on trust and the quality of relationships between consumers and their providers or health organisations. The potential for social media use to make organisations more transparent [51,52,66,69,74,84], and help consumers better understand how organisations work [51,52], are possible ways to overcome the distrust that some consumers may feel towards health services [104]. Related to this could be the potential of social media to create more collaborative relationships, through both the levelling of hierarchies [63,71,81,92] and facilitating interactions between people and groups who rarely interact [37,43,56,63,67,69,80,82,86,89,91], creating new opportunities for innovation.

However, the findings of the review also demonstrate that this experience of improved relationships is not universal. Some of the included studies identify risks of negative experiences for users through social media messaging or online behaviour [42,51,52,53,74,86,90,97], or just an inability to reach key user groups for relationship building [42,47,51,52,56,59,60,63,64,66,70,75,76,83,84,86,90,97]. These risks and limitations would need to be managed if relationship building through social media were to be successful.

There was some discrepancy between consumers wanting some ability to maintain privacy and anonymity in social media spaces [71,91], and providers wanting to restrict the ability for users to be fully anonymous or use pseudonyms [72], although this was only reported in a small number of studies. This discrepancy may be reflective of the type of activities being conducted and the participants involved. For example, both of the studies which advocated for anonymity in social media spaces examined the experiences of consumers who shared experiences of care to try and influence change at the health service [71] or public health policy [91] levels. Moorley et al. (2014) found value in undisguised interactions between members of a health professional group using social media for networking and seeking professional advice [72]. Discrepant findings are expected given the range of audiences, populations and issues reviewed and the complexity of modes of social media use [9]. Future efforts by services could be informed by models and examples that are relevant to their needs and aims [9] and by planning social media use in health service design and QI activities in partnership with advisory groups of clinical and consumer stakeholders [105].