This study utilized data collected over a 6-week period (February and March 2015) for a cross-sectional health survey conducted on the Web. Participants were recruited through a social marketing campaign, which included Web recruitment via social media, flyers posted in community settings, and snowball sampling. Web-based venues used in Web recruitment included Facebook, Twitter, LinkedIn, and Craigslist. Flyers were posted in community venues that African American women frequent (eg, churches, hair salon, health centers, and colleges in the New York metropolitan area). As this is a historically difficult-to-reach population, snowball sampling was used to recruit African American women [42]. Study eligibility was as follows: (1) reported female identity, (2) self-identifying as black or African American, (3) aged 18 years or older, (4) ability to read English, and (5) willingness to participate. Those who clicked on the recruitment banner were routed to a screening survey on a secure study website. A total of 384 individuals consented to participate in the study. Of those, 34.3% (132/384) individuals were ineligible after not meeting the inclusion criteria. An additional 10.9% (42/384) of the surveys had missing data and, therefore, were also ineligible to be included in the sample. Furthermore, 7 eligible surveys were identified as duplicate responses based on identical IP addresses and were excluded from the sample (n=116). Moreover, per our original inclusion criteria, participants who did not watch the entire video were excluded from the regression model to ensure that the participants in the analysis received the same exposure during the study. As such, the final sample size was N=91.

Those who were deemed ineligible were thanked for their time and exited the site. Eligible participants were directed to the Web-based consent form, which outlined the study’s purpose, risks, and benefits followed by a prevideo survey, a 9-min eHealth video (PEP and PrEP for Women), and a brief postvideo survey. Those who completed the study were entered into a random drawing to receive an Amazon gift card (US $100, US $200, or US $300) and asked to forward the link to potential eligible participants in their network. The institutional review board at Teachers College Columbia University reviewed and approved all study procedures.

We assessed a comprehensive range of sociodemographic variables, including age, employment status (unemployed vs employed), annual income level (from “no income” to “more than $50,000” in increments of $20,000), education level (high school/GED or less vs some college vs college or more), current student (yes/no), relationship status (single vs partnered), and gender of current sexual partners (male, female, or both). We also assessed previous knowledge of both PEP and PrEP. Additional information related to sexual risk behaviors including drug/alcohol use, drug/alcohol use during sex, HIV testing history, relationship dynamics [43], condom use, self-efficacy for condom use [44], history of adult and child sexual abuse, and sexual assault within their social network was collected in the prevideo survey and entered into the regression model but not reported in this paper. After the completion of the prevideo survey, participants were then directed to the 9-min PEP and PrEP for Women eHealth video.

The 2-part PEP and PrEP for Women eHealth video was created using Goanimate Technology [45]. Goanimate is a software program used to create professional, animated, avatar-led videos with an entertainment-education style [45]. Research supports the premise of technology-based approaches [45] to facilitate learning and decision making.

Part 1 of the video addresses a storyline that warrants the initiation and use of PEP. In scene 1, the main character, Tania, informs her friend, Michelle, that she has been sexually assaulted (Multimedia Appendix 1). In scene 2, Tania is in the emergency room speaking with her physician after being examined. The physician explains how PEP could be used by women who may have been exposed to HIV. Tania decides to begin PEP treatment to prevent HIV from her possible exposure. In scene 3, Tania returns for a follow-up visit 28 days after she started her PEP regimen at the hospital clinic. This concluded part 1 of the eHealth video.

Part 2 addressed a storyline that warrants the use of PrEP to prevent HIV transmission. In scene 4, we introduce 2 new characters for this narrative, Kia and Jackie. The main character Kia has a discussion with her friend, Jackie, about her boyfriend Mike (Multimedia Appendix 2). Mike has not been monogamous in his relationship with Kia, and she is concerned. Jackie informs Kia about PrEP and suggests that she talks to a health care provider about initiating PrEP to protect her from possibly getting HIV. In scene 5, Kia’s physician explains the history of PrEP and how PrEP can be used by women who are in a relationship with an HIV-positive partner or a nonmonogamous partner who may be at high risk of obtaining and transmitting HIV, such as in Kia’s current situation. In addition, Kia’s physician discusses some common concerns about using PrEP, such as cost, side effects, and the routine laboratory tests that are required as a part of treatment. Finally, Kia’s physician empowers her with this information and lets her decide what is best for her—either starting prophylactic treatment now or considering its future use. This concluded part 2 of the eHealth video.

After the participants watched the video, we conducted a brief survey to measure the participant’s perceived acceptability of the video on a 5-point scale. First, we accessed their dose of exposure to the video to identify if the participants watch all, most, some, or none of the video. Second, we asked participants to rate the video with the following response anchors: poor, fair, good, very good, or excellent. Next, participants were asked if they would recommend the video to other women. A dialogue box was provided to ascertain why or why not they would share this video and information with other women. Finally, they were asked about their intentions to use PrEP and PEP in the future, if needed, and if they would recommend PrEP and PEP to other women.

Quantitative data were downloaded from SurveyMonkey, transferred to SPSS, and analyzed using SPSS 21 [46]. Descriptive statistics, including frequencies and proportions, were computed for all sociodemographic variables. Tabulations were performed for all responses. Bivariate correlations were calculated among the study variables using the Pearson coefficient of correlation (Pearson r). The outcome variable of interest was the participant’s ratings of the PEP and PrEP for Women eHealth video. To refine our model to exclude insignificant or highly correlated variables, backward stepwise regression was used. The rationale for using this approach was based on the work by Mantel who explained that backward selection serves to reduce the degrees of freedom, has joint predictor capability, and removes noise caused by including unrelated variables or variables that may be highly correlated with each other [47]. This was conducted by including the full group of 22 predictor variables that were all entered into the equation regardless of their worth for the creation of the regression model. The backward stepwise regression analyses included 22 independent variables as potential predictors: (1) age, (2) level of education, (3) annual household income, (4) has a current partner, (5) is a student, (6) employed, (7) relationship status, (8) lifetime HIV testing, (9) tested positive for an STI, (10) current drug use, (11) have other sex partners, (12) had any unprotected sex with main/steady and any other partner(s) in the past 3 months, (13) heard of PEP before watching the video, (14) heard of PrEP before watching the video or not, (15) higher percentage of time had sex while using alcohol/drugs, (16) prevalence of exposure to sexual assault/rape before 18 years old, (17) prevalence of exposure to sexual assault/rape since 18 years old, (18) prevalence of sexual assault experiences, (19) prevalence of sexual assault in their social network, (20) self-efficacy for condom use [44], (21) better partner characteristics [43], and (22) more self-control in a relationship [43]. Next, the backward stepwise method involved the least significant variable (one with the largest P value) being removed when the model was refitted. Then, a new model is built in the absence of that 1 independent variable and the evaluation process is repeated again, removing the least significant variable. This removal process and the equation-reconstruction process were continued until only significant independent variables (P<.05) remained—as the final model reported for the backward stepwise regression.

Analyses of the open-ended questions were thematic, focusing on dominant themes that emerged and were organized by the question asked; attention was paid to the level of endorsement of a theme across the sample. For the qualitative analysis, a grounded theoretical approach was used, permitting the use of categorizing strategies to code and analyze the data [48,49]. First, the transcripts were summarized in a digest that identified the major themes of the responses. Using QSR International’s NVivo 9 qualitative software, the coding of the data was conducted in 3 steps. First, based on the transcripts of the responses, a list of analytic areas represented in the data was composed and given a code (ie, a closed code) derived from relevant literature pertaining to entertainment-education narratives [36]. Second, the primary analyst (principal investigator) reread the transcripts and identified blocks of text to be given a descriptive label (ie, either a label from the closed code list or an original one, termed an open code). Next, the open-coded data were organized under and integrated into the closed code list. Third, the data under each thematic code (eg, empowering, educational / informative, usability, suitability, and entertainment) were reread, and, if needed, recoded into subcategories to refine the analytic categories used. As the themes of the responses emerged, special attention was paid to data that did not confirm emerging themes, noting these for exploration in future research. In addition to using these methods of analysis, we also ensured analytic rigor by engaging several peer reviews of early analytic claims [50]. For this analysis, we will focus on the questions related to recommendations for use of the video.

In our sample (N=116), the mean age was 34 years. Most women were born in the United States (109/116, 94%), identified as black or African American (115, 99%), were in a relationship (71, 61%), and were intimate with male partners only (110, 95%). Approximately one-third (30, 27%) of them had a high school education, and a similar proportion (38, 33%) of them were current students. The majority (83, 72%) of them reported either current full-time or part-time employment. Almost half (52, 47%) of them reported a household income of US $50,000 or more. Before watching the video, only 18% (21) of the participants reported that they heard of PrEP and 22% (26) reported that they heard of PEP (Table 1).

Overall, 78% (91/116) of the eligible participants watched all of the video (Table 2). In addition, 11% (13) of them watched most of the video and 3% (3) watched some of the video. Furthermore, 8% (9) of the eligible participants did not watch any of the video (Table 2). Analyses of the postvideo survey and regression model only included those who reported watching the entire eHealth video.

Overall, 89% (81/91) of the participants rated the video as good or higher; 29% (26) of them rated the video as excellent, 32% (29) rated the video as very good, and 29% (26) rated the video as good. Only 10% (9) of participants rated the video as fair and 1% (1) rated the video as poor. The mean rating for the video was very good (mean 4.77, SD 1.01). Among the sample, 91% (83) of participants reported that they would recommend the video to others, whereas 9% (8) of them reported that they would not recommend the video.

After watching the video, 97% (88/91) of participants reported that they would seek out PEP if they felt that they might have been exposed to HIV, including taking the recommended oral medication for 28 days. In addition, 93% (85) of them reported that they would recommend to another woman to seek out PEP if they knew the woman might have been exposed to HIV. Regarding the use of PrEP, after watching the video, 76% (69) of them reported that they would seek out PrEP. In addition, 91% (83) of participants reported that they would recommend to another woman to seek out PrEP (Table 3).

Regression analysis identified the best predictors for giving the PEP and PrEP video a higher rating (ie, good, very good, and excellent). The backward stepwise regression analysis showed that the 5 variables that best predicted a high rating of the video were not engaging in any current drug use (beta=−.814; P=.004), using condoms all the time in the past 3 months (beta=−.488; P=.03), a higher household income (beta=−.149; P=.03), lower education level (beta=−.267; P=.005), and less experience of sexual abuse as an adult (beta=−.313; P=.004). For this model, the adjusted R square was .224 (F=6.067; P<.001; df=5), meaning that 22.4% of the variance was explained by this regression model (Table 4).

Thematic analysis was used to analyze the open-ended questions to identify reasons for recommending the video to other women. Overall, 91% (83/91) of the participants expressed that they would recommend the video to other women. Our analysis yielded 3 themes about the eHealth video, which were (1) educational, (2) entertaining, and (3) empowering (Table 5).

Under the first theme, educational, subthemes covered how the video provided awareness or knowledge of PEP and PrEP and how it was adaptable. The participants reported that they were not aware of PEP and PrEP before watching the video and that the video was informative and easy to follow. For the second theme, entertaining, subthemes covered how the video had characters that were relatable, and the information was provided via a video with a storyline and characters. The participants reported that they would recommend the video because the at-risk scenarios were relatable to women’s current relationship situations, and it may encourage them to change their behaviors. Other participants commented that the video was nice to watch and entertaining. For the third theme, empowering, subthemes covered 3 areas: (a) female-controlled HIV prevention methods, (b) culturally relevant, and (c) heightened perception of risk. The participants reported that they would recommend the video because it provided information on an HIV prevention method that was female controlled. In addition, the video was considered to be culturally relevant based on the scenarios depicted in the video, and it was not judgmental. Finally, the participants reported that the video provides scenarios that would heighten women’s awareness of their own risk because of their partner’s behavior in sexual relationships.

In addition, 8 participants reported that they would not recommend the video to other women (information was excluded from the table). There were 3 emergent themes, including subthemes. First, participants expressed criticism of the script—with the subtheme of their identification with all of the characters. One participant wrote regarding the characters in part 2:

Participants stated that they would have preferred for the health information to not be delivered by the doctor character in the video. Second, participants reported problems maintaining attention because of the length of the video:

Third, participants expressed a preference for other modes of communication—with subthemes of age appropriateness and video format. The participants reported that the animated format was too cartoonish and would be more appropriate for younger women. As expressed by 1 of the participants:

This intervention aimed to increase the knowledge and initiation of PEP and PrEP among African American women by modeling situations and discussions that would be relevant to the target population. The goal was to also make the video entertaining (ie, edutainment) and engaging, while having diverse characters suited for connecting to African American women [51]. More specifically, the intervention included personal narratives that were culturally tailored to the experience of HIV epidemic among African American women and provide information on PEP and PrEP [52]. For example, in PEP and PrEP for Women, the viewer learns through a conversation between 2 friends (Kia and Jackie) that the main character, Kia, does not have to rely on her male partner to use condoms to prevent HIV. In this scene, Kia learns about PrEP from her friend, and she is referred to a clinic that provides PrEP. The video does not attempt to answer the question of personal responsibility or criticize complex relationship dynamics that influence HIV risk, but rather it increased the viewer’s knowledge and awareness of an alternative HIV prevention option via involvement with the story characters [39]. Furthermore, the viewers may identify with the protagonists of the stories and thus might be better able to relate the scenario to their own lives while lessening the resistance to the health messages [40].

The participants reported a positive perception of the video, with 89% of those who watched all of the video reporting a video rating of good or higher. In addition to rating the video, 91% of them reported that they would recommend the video to other women. Moreover, 4 salient factors facilitated the recommendation of the video by participants. The high acceptability of the video could be attributed to its focus on providing educational messages in an entertainment content to raise awareness and increase knowledge of 2 novel HIV prevention strategies, PEP and PrEP. First, the information provided on PEP and PrEP increased their knowledge of novel prevention strategies [53]. Second, the format of the video was inclusive of phenotypically relatable characters [54]. Third, the animations and subject matter were entertaining using visual storytelling and the information was valuable [55]. Finally, the subject matter was culturally relevant [52] and focused on a topic about female-controlled HIV prevention methods [15].

Results from regression analysis showed that the major predictors of giving the video a high rating included not having unprotected sex in the past 3 months, lower education, higher income, less sexual abuse as an adult, and no current use of drugs. This could be interpreted that women who would be not considered at high risk for HIV or potential candidates with PEP or PrEP were more receptive to the video. Previous research has shown that people who are already receptive to or substantially aware of the issues were more inclined to adhere to the health message and evaluate it favorably [56].

It is important to acknowledge that African American women are disproportionately affected by HIV not only because of behavior but also because of vulnerabilities created by unequal cultural, social, and economic status [57]. Our sample represents an atypical Web-based sample of African American women with high socioeconomic status, including high levels of education, employment, and income [58]. The higher socioeconomic status is perhaps reflective of attributes of those who have access to computers and the internet or the impact of the snowball data collection method, resulting in a network of women with high incomes participating in the study. However, the high socioeconomic status of the sample may have important implications that are reflected in the body of data obtained. Very few studies have focused on HIV/AIDS risk among African American women with higher socioeconomic statuses and the challenges that they may face in their environment that are similar to women of color with lower socioeconomic statuses [59,60].

The content addressed in the video, information on HIV prevention methods that were female controlled, was provided as one of the reasons that the participants would be willing to recommend the video to other women. Among the sample, there was low awareness of both PEP and PrEP before participating in the study. This mode of eHealth videos that incorporate entertainment-education dramas can persuade individuals because they depict characters who changed their behavior by initiating PEP and PrEP to improve their lives [61]. Drama incorporates a component of emotional response to the informational content, and the combination of emotion and information works together to promote individual intentions to engage in prevention activities [61,62]. The participants reported that the characters were relatable and they had personal knowledge of similar situations experienced by the characters in the PEP and PrEP for Women video. After viewing the video, 88% and 69% of the participants reported that they would seek out PEP and PrEP, respectively, for themselves. This is reflective of the impact of entertainment-education used as a tool to change health behavior [63,64]. Bandura recommended that characteristics of the models should be similar to the viewers to increase the impact of educational modeling [64]. Perhaps the most important lesson is that although a Web-based intervention relies on audiovisual content, it is also important to tailor the intervention to the target population.

A generally positive response to eHealth as an innovation for disseminating HIV prevention was found in this study, but 22% of the participants reported that they did not watch all of the video. There are several possible reasons for why the individuals did not completely watch the eHealth video. Participants who were not able to watch the entire video expressed that they experienced technical problems with the video, which included not being able to watch the video on specific handheld devices. This issue could frustrate participants and cause them to skip over watching the video to complete the survey. An additional reason for not completing the video could be participant’s fatigue because of the length of the video. In this study, some of the anticipated issues related to fatigue were addressed by creating 1 long video instead of 2 different videos for PEP and PrEP, but the length of the video was cited as a reason to not recommend the video. Participants reported that the video was informative, but the length of the video may make it difficult for viewers to maintain their attention. Other reasons cited for not recommending the video included the educator/health care provide character and the use of animation. Participants felt that the PEP and PrEP educator character should have been a peer and not a health care provider. Previous research reported that groups that have traditionally lacked power, such as African American women, may be more sensitive to an interventionist’s physical characteristics [65]. Despite the fact that the doctor in the PEP and PrEP for Women video is an African American woman, the participants felt disengaged from the character. This may be reflective of the lack of cultural competence among medical providers, institutional racism, and personally mediated racism and sexism that have negatively affected the sexual and reproductive health of African American women [57]. Participants also stated that the animated format was not appropriate for all age groups. They reported that it was more suitable for teenagers or younger adults.

Although some consumers may criticize the limitations of the technology (eg, age appropriateness), the vast majority found the avatar video product to be informative, accessible, and engaging. Those individuals who have a positive response to the avatar video were willing to adopt the innovation and engage in the diffusion of it within their social networks, which includes disseminating on social media networking sites. Studies have shown that HIV prevention programs grounded in the diffusion of innovation theory can be useful for increasing HIV/AIDS risk reduction knowledge among early adopters and diffusing information throughout social networks; therefore, preventing HIV among vulnerable populations [66]. This study specifically demonstrated how programs may need to strive to incorporate social media technologies to disseminate information. In addition, prevention programs should also take relationship status into consideration during planning and implementation of effective programs to reduce women’s HIV/AIDS risk.

Finally, this study offers recommendations and suggestions for Web-based research as well as Web-based recruitment using modern-day social networking. Although the sample may not be representative of all African American women in the United States with respect to socioeconomic status, this strategy allowed for rapid recruitment of the sample population with minimum resources, such as staff and funding. Furthermore, this strategy potentially provides opportunities not only for recruitment purposes but for eHealth intervention purposes. Researchers and practitioners should consider the benefits of not only Web-based survey research but also of Web-based health education, health promotion, and disease prevention to reach more diverse samples of subpopulations affected by the HIV epidemic. eHealth intervention has proven to be extremely useful in educating African American women on a number of topics related to HIV/AIDS or STI prevention. Web-based health education can be beneficial for African American women if they use the internet to access and share information with their peers. In addition, as previous research studies have shown, using eHealth interventions may help reach a subgroup of African American women who are not usually addressed in the literature pertaining to HIV risk and prevention [67].

This study had some limitations. First, this was a small, cross-sectional, Web-based sample of African American women, thus making it difficult to reach generalizations from this data source. Second, internet access familiarity with Web-based technology may have influenced the decision to participate in the study. Women who use a different type of electronic device (eg, cell phones and tablets) to access the internet may be able to navigate websites easily and may encounter challenges with technical issues. Third, the video and script were not previously tested with different groups of African American women. Having focus groups or concept-mapping sessions before the launch of the study would have enhanced the script of the video for tailoring of the message. Future research should involve conducting focus groups with different age groups, socioeconomic statuses with regard to the scripting, video display, and avatar characters to reach a broader audience of African American women. Fourth, given that this is a feasibility and acceptability study, it is not possible to make any conclusion regarding the efficacy of the video. In addition, participants’ responses to PEP and PrEP adoption intentions were based on a hypothetical situation rather than their actual experience. Increasing knowledge does not necessarily correlate with motivating behavioral changes. Finally, the results could have been influenced by social desirability bias [68]. To minimize bias, participants were able to anonymously take the survey on their 1 device. Despite these limitations, the results of this study have important implications for interventions targeting HIV risk behavior among African American women. In addition, it also has important implications for utilizing technology to disseminate HIV prevention information to this population [69].

This study has demonstrated that the PEP and PrEP for Women video using eHealth and entertainment-education health message narratives for a targeted, culturally centered, HIV/AIDS preventative education intervention for African American women can be readily adopted and diffused as an innovation, while still using low-cost technology in the public domain. Findings from this eHealth video study demonstrated that (1) a culturally tailored eHealth video could be used to increase awareness of HIV prevention methods and (2) it is feasible to recruit and engage African American women into an eHealth intervention that is delivered via the internet and social media. The PEP and PrEP for Women video was viewed as culturally centered Web-based content by the participants and capitalized on the widespread use of the internet to reach a population that is not only underserved in eHealth interventions but also in traditional HIV research targeting African American women, therefore, adding to the limited body of research on computer-mediated HIV interventions targeting African American women [1,67]. The eHealth video is an acceptable mode of education to address HIV prevention methods. It follows the CDC’s recommendation to create health messages that are relevant, useful, and interesting to encourage the audience to interact and be engaged [52].

Given the disproportionate impact of the HIV epidemic on African American women, the lack of awareness of both PEP and PrEP [3], and poor participation rates of minority populations in prevention interventions [33], it is essential to develop strategies to reach underserved groups and reduce some of the challenges associated with face-to-face preventive interventions. eHealth technologies have the potential to reduce the gap between what is known about PEP and PrEP and its utilization among high-risk populations by increasing individuals’ sexual health knowledge. Future research is warranted on improving the tailoring and design of the eHealth video by including African American women and other underrepresented groups, such as transgender individuals, to participate in the design of an updated PEP and PrEP for HIV video. This can be explored through formative qualitative research to first assess the cultural and structural factors that may influence uptake of both PEP and PrEP.