Innovative perception analysis of HIV prevention messaging for black women in college: a proof of concept study

Project iMPPACS [12] implemented by Sznitman and colleagues [9] was a multi-site mass media campaign using television and radio advertisements containing culturally and developmentally sensitive HIV prevention information for Black adolescents of low socioeconomic status (SES). The mass media campaign aimed to increase HIV transmission perception among Black adolescents with outcomes that included increasing condom usage and reducing HIV incidence-related behaviors [9, 13]. Project iMPPACS demonstrated influence in risk reduction behavior change, and change patterns, among the priority population [9, 13]. However, it is unclear if the messaging and images used in this intervention would resonate specifically with Black women in college.

HIV prevention initiatives typically target communities of low socioeconomic status and, while this work is necessary given their risk, Black women in college and of higher socioeconomic statuses are often not considered in the development of such interventions despite experiencing similar high HIV risks. Given the lack of attention placed on this population, Black women in college may not relate to media messages targeting lower income Black populations due to differences in lived experiences [14]. Partnering with Black women who attend college to conceptualize and design Sexual/HIV risk reduction communication might render the greatest degree of relatability among Black college women. The purpose of this study was to examine appropriateness, acceptability and relevance of Project iMPPACS advertisements and other publicly available audio-only media (Public Service Announcements [PSAs] among Black college women. Our aim was to have participants identify specific prevention-related content within the HIV prevention laden advertisements they found to be appealing, unpleasant, or critically important in efforts to inform and develop tailored messaging. Perception analyzers were used to record in real-time participant attitudes and reactions to media messages, and the feasibility of using this novel technology was explored. Explicit identification of relevant and health-motivating content may inform future message design, production, and dissemination for this population.

This study was guided by the Social Norms Theory developed by Perkins and Berkowitz (1986) [15]. Social norms have remained a fundamental part of creative communications strategies particularly in health and entertainment education. This theory is widely used in promoting positive health-related behaviors and operates under the premise that individuals misperceive their peers’ behaviors and attitudes, and under-and-over estimate behaviors and peer approval [16]. This theory consists of six constructs: 1) descriptive norms (beliefs about what other people do) 2)injunctive norms (beliefs about what others approve of/think people should do) 3)reference group (a group of people that a person feels a connection or identity with) 4)sanctions and punishments (beliefs about the perceived sanctions/punishments) 5)benefits and rewards (beliefs about the perceived benefits/rewards) 6) collective norms (actual prevalence of a behavior) [17]. In addition, certain advertisements involved Black men and women in the community discussing the importance of safe sex and reducing HIV risk. Undertaking these approaches were largely aimed to normalize safe sex, address misperceptions regarding sex in the community, and promote condom use to reduce HIV risk within the Black community. Social norms theory has been used to guide social marketing campaigns targeted at the group level, with the use of mass media to disseminate social normative feedback [16]. However, literature on social norms in increasing health promotion behaviors has been relatively mixed,with certain studies demonstrating significance in behavior change and misperceived norms and other studies not demonstrating any significance [16]. While this theory has been widely used in health behavior interventions, literature has been limited on how social norms theory can be used to guide health promotion interventions that use digital technology and digital media. Furthermore, there is a need to understand how this theory can be used and adapted to guide studies within communities such as Historically Black Colleges and Universities (HBCUs) in which the majority of the population are young adults who identify as Black. This study focused on the constructs of descriptive norms, injunctive norms, and reference groups. The Project iMPACCS vignettes and PSAs were developed for Black youth to encourage HIV and STI (sexually transmitted infection) risk reduction through the use of strategies like message delivery by influential members of the Black community (e.g. popular artists and actors within the Black community). It was unclear whether the content in Project iMPACCS or the selected publicly available PSAs would be effective among Black women in college.

Black women who were enrolled as either full or part-time students at a southern HBCU were invited to participate. All participants were required to self-report their race as Black, have no auditory or visual impairment, and speak and understand English. Additionally, participants had to be 18–24 years old, previously or currently sexually active within 6 months, and had to have reported media consumption (e.g., radio, television, social media, other audio/visual media) for at least one hour per day.

A convenience sample was obtained through face-to-face, campus (e.g., dining halls), and electronic newsletter distributions of a study flyer. This research opportunity was also presented to various academic colleges (e.g., Media & Technology) and student organizations within the university. The creatively-designed advertisement included the pre-determined focus group schedule, participant incentive, and study contact information. A pre-screening script was used to guide determine study eligibility. If all criteria were met, participants were given the opportunity to reserve attendance for an open session. This schedule was updated daily during active enrollment. Participants who contacted study staff by phone or email received a return reply within one business day. The pre-screening script was also used to facilitate enrollment by phone. Students who expressed interest after the focus group was closed to enrollment were asked to provide their name and contact information in the event of cancellations. This information was destroyed upon study completion. Each participant received weekly telephone and email reminders of their study appointment (i.e., location and date/time). Plans for attendance were confirmed one day prior to their focus group session. Participants were also notified that, upon arrival, they would be required to sign a consent form which explains the study and its potential risks/benefits, confidentiality, and voluntary withdrawal. A copy of this consent was emailed to each participant for advanced review.

The primary outcome measure of this study was to evaluate the acceptability of the Project iMPPACS and publicly available audio-only HIV prevention messages among Black college women, and whether students believed the content would be beneficial in promoting safe sexual behaviors. The secondary outcome was assessing the feasibility of using perception analyzers in recording participant attitudes and reactions to Project iMPPACS and other media messages in real-time.

The study was conducted in a lecture hall on a southern HBCU campus. Participants were introduced to the study staff and the Dialsmith administrator, followed by a brief overview of the study. The informed consent document was then reviewed; participants were given the opportunity to ask questions, summarize key principals in their own words, or exit the perception analysis session if they chose not to participate. All remaining participants then signed the informed consent document; these were collected and verified prior to the start of the focus group discussion. The Dialsmith administrator then distributed a perception analyzer tool to each participant. Each tool was labeled with a unique number that allowed study staff to identify each individual (along with self-reported classification and sexual activity) as well as their synchronous feedback. Prior to showing the Project iMPPACS media content or playing the PSAs, participants were asked questions pertaining to demographic data along with media intake. Participants were also asked a set of warm-up questions to test the perception analyzer for accuracy.

The perception analyzer responses from the focus groups were coded and exported as a Microsoft Excel spreadsheet. The baseline demographics of the sample were characterized using measures of central tendency. Matched pre-and-post-test responses were obtained from the perception analyzers and used for data analysis. Descriptive statistics and t-tests were used to characterize the data, with Type 1 error rate for statistical tests being set at 0.05 for significance and 95% confidence intervals provided where appropriate.

A total of 10 students participated in the study. The mean age of study participants was 19.5 with ages ranging from 18–22 years old. Participants largely held Freshman classification (n = 6, 60%) with the majority of students reporting that they were currently sexually active (n = 6, 60%). Among those who were sexually active, 90% (n = 9) reported being sexually active with men see Table 2.

Participants reported watching or listening to social media more often over radio and print media (77.8±17.01). When asked about whether participants watched or listened to television or radio more frequently, participants reported watching television more frequently over listening to the radio (23.1±16.51; 95% CI 11.29,34.91). Participants were more likely to state that HIV topics pertaining to risk, exposure, testing, and disclosure were more likely to be discussed in their current sexual situations as opposed to STI topics; and that condoms and other family planning options presented in media would have the most influence on their immediate decision to have protected sex (78.9±17.96; 95% CI 66.05,91.75) (Table 3).

The degree to which participants identified with the vignettes’ content ranged from slightly to moderately well [1.25±1.39; 95% CI 0.09, 2.41] for the first vignette, and moderately well for both the second [2±1.70; 95% CI 0.78,3.22] and the third vignette [2.1±1.53; 95% CI 1.01, 3.19]. Participants reported greater and more intense personal connections with the fourth vignette compared to the other vignettes [3.88±0.36; 95% CI 3.58, 4.17]. Participants reported that the content they remembered the most in the fourth vignette were main discussion points (n = 6, 75%). Additional content that participants remembered the most within the other vignettes included profound statements (n = 5; 50%) followed by people and actors (n = 3; 30%) and main discussion points (n = 3, 30). Participants had higher average responses for the fourth vignette regarding their immediate interest in learning more about the topics discussed, with participants wanting to learn more about the fourth vignette compared to the other vignettes [2.34±1.38; 95% CI 1.38, 2.27].

Among the vignettes, participants were more likely to recommend the fourth vignette to their close friends and family [3.5±0,76; 95% CI 2.87,4.13] and believed that the video would be somewhat effective to very effective if shown on campus [2.62±1.19; 95% CI, 1.632, 3.618] and if broadcasted on their preferred viewing stations [3.63±0.74; 95% CI 3.00, 4.25]. For each of the vignettes, the large majority of participants responded that if they were to watch a similar video advertisement before watching sexually-charged visual media, then it would make them more aware of their sexual practices and any potential risks (n = 3, 30% for 1^st video message; n = 6, 60% for 2^nd video message; n = 6, 75% for 3^rd video message; n = 5, 62.5% for 4^th video message). Students were asked what content was missing from the vignettes that could influence them to engage in safer sex. In over half of the vignettes (75%) participants reported that content representing students’ current society, culture, or interests was missing. Additional content that students reported to be missing included “catch” phrases (e.g., hook lines and punch lines), profound statements, and songs, jingles, and music.

Responses largely varied in regard to how participants personally identified with audio message content, with average responses ranging from slightly well to very well across the individual audio messages. Participants were more likely to recommend the audio messages that they strongly identified to their friends and family (2.5±1.69; 95% CI 1.09,3.91) in comparison to audio messages that they did not identify with strongly. Additionally, participants reported that the messages they personally identified with more would be between somewhat effective to very effective if they were to be broadcasted on campus and on their preferred listening stations. The majority of participants (n = 5, 62.5% for 1^st audio message; n = 6, 75% for 2^nd audio message) reported that the people and actors in the audio messages were the most memorable aspect of the advertisement, followed by profound statements (n = 3, 37.5%) and main discussion points (n = 3, 37.5%). Participants were less likely to be interested in learning more about the topics discussed in the audio message that they did not personally identify with strongly (1.5±0.76, 95% CI 0.87,2.13).

When asked about content that was missing from the audio messages that could influence participants to engage in safer sex, among all three of the audio messages participants reported content that represented their current society, culture, or interests was missing. Additional content that was reported to be missing were catch phrases and profound statements. In two of the audio messages, at least half of participants reported that if they were to listen to the audio message content before listening to sexually charged radio/songs, they would be more likely to participate in safer-sex activities (n = 5, 62.5% for 1^st audio message; n = 4, 50% for 2^nd audio message). Interestingly, in the third audio message which happened to be the message participants were least likely to identify with, half of participants responded that they would feel annoyed if they were to listen to the audio message prior to listening to sexually-charged radio/songs with only two (25%) of participants stating that the audio message would make them more aware of their sexual practices and risks.

Perception Analyzers were feasible to use in this setting and allowed for students to establish an objective and independent assessment of each HIV prevention advertisement, which was followed by more descriptive discussions of real-time outcomes from the perception analyzer data. Students appreciated the ability to give distinct feedback about varies content within the advertisement. For instance, the participant may like specific content or visuals in one portion of the advertisement and not in others. Using this method, we were able to get nuanced, objective data about each element of the advertisement (Additional file 1: Fig. 1).

Future research should assess the specific components of media advertisements that can result in greater uptake of safe sex practices among this population. There is a need for future interventions focusing on the development of mass media campaigns for Black women in college, thus necessitating program developers to ascertain the features and content that Black women in college want to see included in HIV media messaging. Several participants in our study noted that many of the advertisements lacked content pertaining to their society, culture and interests. Additionally, a number of participants reported not strongly identifying with some of the advertisements. Despite this, participants were more likely to consider participating in safer-sex activities after listening to or viewing the advertisements. It is unclear as to whether the consideration of participating in safer-sex activities after watching or listening to advertisements results in the actual uptake of behaviors, and whether this effect can be sustained over the long term. It is also unclear as to whether there is an association between personally identifying with advertisement content and the uptake of safe sexual behavior practices. Thus, future research interventions should evaluate this. While we focused our study within an HBCU setting, there is a need for future research to evaluate similar interventions among Black female college students attending PWIs. Finally, many participants in our study reported watching or listening to social media more than radio advertisements. With the increased popularity and visibility of social media, there is a need for the development of mass media campaigns that target Black women in college through social media platforms.

Our study is not without limitations. First, non-probability sampling was used and our sample size was small consisting of students attending a southern HBCU. Thus, our results cannot be generalized to all Black female college students.

An essential element for future studies employing this method is to fully assess the relevance of demographic information related to socio-economic status, religiosity, social factors that may impact media preference and availability, previous exposure to HIV or treatment for an STI, and the emotional states of the participants before, during and after media exposure. In the context of a proof-of-concept, we a confident that this approach will yield useful information for effective messages.

In addition, given our sampling framework, there was potential for the presence of selection bias. Also, although perception analyzers were used to obtain an objective measure for participant responses and to limit response bias, self-reported behaviors and perceptions regarding HIV and sexual health is a highly sensitive topic. Thus, there was potential for reporting bias; perception analyzers were used to mitigate this. It can be inferred that higher perception values correspond to greater satisfaction and appeal of message content and that lower values correspond with decreased satisfaction of message content. However due to the study being a proof of concept, we cannot directly extrapolate impact. Furthermore, Project iMPPACS’s messages and PSAs were not specifically developed for Black female college students, and originally targeted Black teens in urban communities. However, this media content was chosen given that it was intended for Black youth and contained aspects that could be applicable to Black female college students [23]. Thus, we determined that this was an appropriate model to use for our study.