TABLE 4 |.

Agenda for future research on communicating about HIV disparities.

Research activity	Rationale
Experimentally compare messages about HIV-related disparities across the continuum of care using progress, impact, and social comparison frames.	Social comparison-framed messages about health disparities do not always increase risk perceptions in the target audience (Bigman, 2014; Landrine and Corral, 2015; Dunham et al., 2016) and can have adverse effects, including stigmatization and stereotyping, negative emotional reactions, and distrust of the information (Landrine and Corral, 2015; Lee et al., 2017). Additionally, only one previous study compared framing strategies in the context of HIV (Dunham et al., 2016). Thus, it would be beneficial to examine whether frames other than a social comparison frame can increase risk perceptions without having adverse effects.
Experimentally test message strategies that acknowledge the multiple factors (i.e., social determinants) that contribute to HIV risk and may be outside of an individual’s control, while also acknowledging the role of the individual in reducing risk.	Challenges that emerge from the literature on health disparities messages are that (1) messages that discuss the contribution of social determinants of health may not address individual behavior and thus may not motivate health behavior change (Lundell et al., 2013), and (2) messages that address individual health behavior alone may be perceived as stigmatizing and can perpetuate misunderstanding about the cause of a disparity (Matthews et al., 2016). Thus, it would be useful to investigate whether messages that discuss both social/societal and individual factors reduce negative responses to messages while also motivating behavior change. This is a type of mixed or competitive frame that tends to be overlooked in research on message framing effects (Guenther et al., 2020).
Assess how images and exemplars can be incorporated in HIV disparities messages to increase personal relevance, attention, and persuasiveness without reinforcing disparities related stereotypes. Studies can evaluate different combinations of images, exemplars, and text to assess emotional response and effects on risk perceptions and other outcomes.	Exemplars and images can prime stereotypes and bias judgment (Guttman and Salmon, 2004; Coleman and Hatley Major, 2014; Arpan et al., 2017). Visual images within health messages tend to be more influential than text; although they are often used to increase message relevance, they can be also perceived as stereotypical and offensive by members of the target audience (Uhrig et al., 2017; Drumhiller et al., 2018).
Examine how anti-stigma communication approaches found to be effective in reducing mental health stigma may be used in HIV communication interventions.	Meta-analyses of mental health anti-stigma communication research found that approaches that facilitate interpersonal or “mediated” contact successfully reduced stigma associated with mental illness (Corrigan et al., 2012, 2015). Creative approaches are needed to develop opportunities for “contact” with people with HIV and to evaluate effects on audiences. One way of mediating contact with stigmatized groups is by using photovoice, an approach used to counter stereotypes, external stigma, and internal stigma (Wang et al., 2000; Russinova et al., 2014; Centers for Disease Control and Prevention (CDC), 2019). Although these studies suggest some promise, this approach has not been tested in combination with framing or in the context of HIV
Examine whether messages about HIV disparities designed to elicit positive emotions such as encouragement and hope—similar to progress-framed messages—are effective in motivating positive behavior change.	According to functional theories of emotion in psychology, emotions are elicited in response to our environment and motivate action in ways that are consistent with personal goals (Lazarus, 1991). This perspective on emotion suggests that the response to shame, which is associated with stigma and the perception that one is being stereotyped, is to hide and avoid facing what may be perceived by oneself or others as personal failure (Lazarus, 1991). As this is not the desired response to health risk messages, other approaches need to be investigated. Two studies that tested responses to skin cancer prevention messages found that hope was positively associated with self-efficacy perceptions and that hope and self-efficacy predicted intentions to engage in skin cancer prevention behaviors (Nabi and Myrick, 2019).
Investigate whether integrating self-affirmation with health risk messages about HIV disparities will be effective in promoting positive behavior change.	Messages that present information about health disparities can be perceived as threatening to one’s social identity (Tajfel, 1979). Previous research found that engaging in a self-affirmation exercise before exposure to a threatening health message can reaffirm one’s self-concept, thus increasing message acceptance and positive health behavior change (Epton and Harris, 2008). A recent study found that a health risk message that incorporated self-affirming text in the message produced greater intentions to reduce risky behaviors (Arpan et al., 2017).
Examine whether messages focused on fostering individual and community empowerment will increase trust and be more effective in generating positive individual- and community-level responses to reduce HIV disparities.	Messages about health disparities can be perceived as blaming individuals for poor health outcomes, which can produce a multitude of adverse effects and make messages ineffective or, worse, harmful. Incorporating community members and their feedback into message development and campaign implementation can increase the likelihood that these interventions will be effective (Earnshaw et al., 2013). It can also increase the capacity of messages and communication campaigns to empower individuals and communities to engage in positive behaviors and actions that reduce health disparities (Douglas et al., 2016; Thompson et al., 2016).