Feasibility of Delivering Evidence-Based HIV/STI Prevention Programming to A Community Sample of African-American Teen Girls via the Internet

Abstract

The current study examined the feasibility of an HIV/STI prevention intervention for African American female adolescents. The intervention, SiHLEWeb, is a web-based adaptation of the evidence-based intervention, Sistas, Informing, Healing, Living, and Empowering (SiHLE). Participants were 41 African-American girls aged 13 to 18 years, recruited in collaboration with community partners (local high schools, Department of Juvenile Justice, child advocacy center, medical university). Results support the feasibility of recruitment, screening, and follow-up retention methods. The majority (63.4%) of recruited participants completed the intervention, taking an average of 4.5 site visits. Completers of SiHLEWeb demonstrated increases in knowledge regarding HIV/STI risks and risk reduction behavior, as well as significant increases in condom use self-efficacy. Findings provide preliminary support for the large-scale, randomized-controlled trial of the efficacy of SiHLEWeb to reduce high-risk sexual behavior among female African-American adolescents.

Introduction

HIV is a costly epidemic that has disproportionate effects on young African American women (Chen et al., 2012) and is currently the third leading cause of death African American females aged 25 to 34. According to predictions by the Centers for Disease Control and Prevention, approximately 1 in 32 African American women will be diagnosed with HIV at some point in their lifetime (CDC, 2011; 2012). If current trends continue, the vast majority of these women will be infected in their youth and young adulthood through heterosexual contact (CDC, 2011). Within the past five years (2007), African Americans and teen girls represented 68% and 40%, respectively, of HIV cases among 13 to 19 year olds (CDC, 2008).

Culturally-tailored prevention programming targeting African American female adolescents and young adults has been developed and evaluated and has demonstrated efficacious impact on key outcomes – including HIV/STI (Sexually Transmitted Infection) diagnosis, number of sexual partners, condom use, condom use efficacy, and assertive communication with partners (DiClemente et al., 2004; DiClemente, Wingood, Rose, Sales, & Crosby, 2010; DiClemente et al., 2009; Downs et al., 2004; Jemmott, Jemmott, Fong, & Morales, 2010). Successful interventions possess common characteristics – they tend to be skills-based, peer-based, and interactive rather than purely didactic in nature (Darbes, Crepaz, Lyles, Kennedy, & Rutherford, 2008; Hemmige, McFadden, Cook, Tang, & Schneider, 2012; Johnson et al., 2009). Currently, the vast majority of culturally-tailored HIV/STI prevention interventions are delivered either in entirety or in part by health workers, resulting in a notably high financial and time cost of dissemination (Hemmige et al., 2012).

Computer-based adaptations of evidence-based prevention programming have demonstrated efficacy in reducing HIV/STI risk behavior among a variety of populations, including African-American teen girls and young women (Hightow-Weidman et al., 2011; Klein & Card, 2011; Lightfoot, Comulada, & Stover, 2007; Noar, Black, & Pierce, 2009; Noar et al., 2011; Wingood et al., 2011). Use of multi-media and technological delivery modalities overcomes several barriers to access of traditional services – facilitator time, cost, potential stigma, pacing (Noar et al., 2009). The past decade has seen a dramatic growth in the reach of the Internet. Recent data from Pew Internet and American Life Project (2012) indicate that 95% of American teens (aged 12 to 17) have regular Internet access. Further, no significant racial or ethnic differences in access rates exist, although living in a household with an annual income below $20,000 remains a negative predictor of access (Zickuhr & Smith, 2012). Advances in technology and access make the Internet a promising tool for extending the reach and public health impact of culturally-tailored HIV/STI prevention interventions to African American teen girls with the potential to overcome traditional barriers like transportation, stigma, and cost (Chillag et al., 2002). However, to date, there is a remarkable dearth of empirical evaluation of the feasibility e-health delivery of evidence-based HIV prevention interventions tailored to African American adolescent girls, including among a community-dwelling samples.

To this end, our group recently adapted the Sistas Informing, Healing, Living, and Empowering (SiHLE) HIV prevention intervention for delivery via a web-based platform. SIHLE is a culturally-tailored HIV/STI prevention program targeting African-American adolescents that has demonstrated efficacy with respect to increasing consistent condom use and decreasing the incidence of unprotected vaginal sex, number of new sexual partners, and number of new chlamydia infections among participants (DiClemente et al., 2004). Traditionally delivered, SiHLE consists of four 4-hour small group sessions consisting largely of interactive skills training activities delivered to groups of 10 to 12 girls by a trained Health Educator and “Near Peer.” SiHLEWeb was created in consultation with SiHLE developers (DiClemente and Wingood) and consists of four 1-hour sessions (modules) that simulate the experience of live group participation by using an interactive, video-based design to present Health Educator/Near Peer content, as well as to follow five characters lives and development as they progress through the SiHLE program. As they progress through SiHLEWeb, users have the opportunity to complete interactive activities and receive real-time feedback on their responses from their video peers, Health Educator, and Near Peer. In contrast to a prior computer-based adaptation of SiHLE (Multimedia SiHLE) (Card et al., 2011; Klein & Card, 2011), the SiHLEWeb intervention was designed with the potential to be a stand-alone, multi-session intervention that, due to capitalizing on a web-based delivery platform (rather than a single-session, computer-based intervention), could be completed by African American teen girls in a setting and timeframe of their choosing.

Whereas prior SIHLE dissemination efforts have focused on public health department/clinical populations (Card et al., 2011; DiClemente et al., 2004; DiClemente et al., 2009), the current study evaluates the feasibility of web-based delivery of this evidence-based HIV/STI prevention programming to a community sample of traditionally underserved African American teen girls residing in the southeast. Specifically, it was predicted: (1) that community-based recruitment of at-risk African American adolescent girls (i.e., girls engaging in risky sexual behavior) would be feasible; and (2) that the majority of recruited African American teen girls would complete the SiHLE-Web intervention independently within a one-month timeframe. Further, exploratory analyses examined pre- to 3-months post-intervention changes in HIV/STI risk-reduction knowledge and efficacy among SiHLE-Web completers.

Methods

Participants

Participants were 41 African-American girls aged 13 to 18 years (M= 15.85, SD= 1.42) recruited from the local community (large Southeastern city) in collaboration with community partners (local high schools, Department of Juvenile Justice, child advocacy center, medical university) through the use of flyers, postings, word-of-mouth, and bulletin advertisements. Participants were compensated $20 for completion of the baseline interview, $15 per module completed of SiHLEWeb, and $20 for completion of the 3-month follow-up interview.

Procedures

Participants were informed about all study procedures and IRB-approved written informed consent and informed assent were obtained from a parental guardians and adolescents, respectively, prior to participation in any study procedure. Participants were screened for study eligibility via phone. Eligibility criteria included meeting all of the following: (a) identifying as African American, (b) being between 12 and 19 years, (c) being female, and (d) currently being/having been in a serious dating relationship or contemplating being in a serious dating relationship in the coming year. Baseline assessments were completed by the adolescent in-person via paper-and-pencil questionnaires. Upon completion of baseline assessment, girls were provided the website address for SiHLEWeb and given a unique code to allow them access to the site. Participant baseline and follow-up data were connected to web-based data via this unique access code identifier. Girls were told that they would have one month (30 days) to complete the site and that they may go through the site at their own pace and any location (or variety of locations) with high-speed internet access that is convenient for them.

Participants were sent weekly reminders via email, phone call, or text message (preference indicated by participant at baseline assessment), and a study coordinator was available to respond to technical queries or assist participants in the case of lost/forgotten log-in information (i.e., web address, access code) during the one-month timeframe allotted for SiHLEWeb completion. There were no self-initiated technical assistance queries; however, 7 participants requested this information during weekly reminder phone calls. Approximately 3 months (average of 88 days) post-baseline, participants were re-contacted and asked to complete a follow-up assessment. Of participants completing baseline assessments, 83% also completed 3-month follow-up assessment via paper-and-pencil questionnaires.

Measures

Demographics assessed at screening, prior to baseline, included gender, age, ethnic identity, and presence of either a dating history or intent to date in the coming year. Knowledge change was measured via pre-module and post-module delivered questions (4 questions per module), embedded in the website, related to the key themes and core elements of the content presented by each SiHLEWeb module.

HIV/STI related outcomes were assessed using psychometrically-sound standardized measures utilized by the developers of the SiHLE intervention in prior evaluations of traditional (in-person) delivery of HIV prevention curriculum, including SIHLE (e.g., (DiClemente et al., 2004). These measures were administered at baseline and at the 3-month follow-up. Outcomes assessed included:

1. Sexual Behavior and Condom Use, assessed by 39 items querying girls’ sexual behaviors (including vaginal, oral, and anal sex behaviors) over their lifetime (baseline only), past 90 days, and past week (DiClemente et al., 2004);

2. Condom Self-efficacy was assessed by 9 items asking girls to indicate how “much of a problem” – a Likert scale ranging from “none (0)” to “a lot (4)” – they would have completing behaviors involved in correct use of a condom (Wingood & DiClemente, 1998). The summed scale ranges from 0 to 36, with higher values reflecting greater efficacy;

3. Partner Communication History was assessed by an 11 item query of “how hard” – a Likert scale ranging from very easy to very hard – it would be for girls to engage in various conversations with their partner related to protected sex (e.g., ask how many sex partners he has had, demand that he use a condom, ask if he has an STI), as well as how often girls engaged in conversations related to safer sex practices with their partner in the previous 90 days (Wingood & DiClemente, 1998);

4. Self-esteem was assessed using the Rosenberg Self-Esteem Scale – a 10-item scale, measured global self-esteem (Rosenberg, 1979). Possible scores range from 10 to 40, with higher scores indicating higher levels of self-esteem; and

5. Ethnic Pride was assessed by a 9 item, modified version of the Multidimensional Inventory of Black Identity (Sellers, Rowley, Chavous, Shelton, & Smith, 1997), that queries girls regarding the degree of their positive associations with Black identity. The scale ranges from 9 to 45, with higher values representing stronger ethnic pride.

Results

Recruitment Feasibility: Sample Descriptive Information

Girls were recruited from four main sources: local area high schools (n=11), Department of Juvenile Justice (n=10), community advertisements (n=10), and other means (n=10). A subsample of the girls had either previously participated (n=10) or were currently participating (n=6) in an in-person SIHLE group. The majority (n=23; 56.10%) reported having previously engaged in consensual vaginal intercourse at least once in their lifetime. Sexually active girls reported a modal age of sexual debut of 15 years of age and a mean of 2.52 (SD=1.16) lifetime sexual partners. More than one-quarter of sexually active girls (26.1%) reported that they did not use a condom during their most recent sexual encounter. Additional information regarding baseline levels of sexual risk behavior reported by participants is presented in Table 1.

Table 1.

Descriptive information regarding reported sexual risk behaviors at baseline (N=41).

Sexual Risk Behavior	N	% of full sample	% of active sample
Ever had vaginal sex	23	56.1	----
Used a condom during last sexual encounter	17	73.9	----
Had vaginal sex in past week	5	12.2	21.7
Condom used each time had sex in past week	2	-----	40.0a
Had vaginal sex in past 90 days	17	41.5	73.9
Condom used each time had sex past 90 days	12	-----	70.6b
Had vaginal sex in past 6 months	18	43.9	78.3
Condom used each time had sex past 6 months	10	-----	55.6c
Ever had anal sex	2	4.9	8.7
Ever had oral sex	2	4.9	8.7

^aDenominator is number of girls sexually active in past week (n=5)

^bDenominator is number of girls sexually active in past 90 days (n=17)

^cDenominator is number of girls sexually active in past 6 months (n=18)

Feasibility of SiHLEWeb Completion

Completion data represent whether a person logging into the site with unique participant identifiers viewed all pages of the SIHLEWeb intervention. Among the full recruited sample (N=41), the mean number of modules completed was 3.28 (of 4.4 total; SD = 1.63;), in a mean of 4.00 (SD=3.71; Range of 0 to 18) log-ins. Nearly two-thirds of the recruited sample (63.40%; n=26) completed the full website, in an average of 4.96 (SD=4.10) log-ins. Five girls (12.20% of the full sample) never accessed the site. Having an unspecified recruitment source (X² [1; N=41] = 6.37, p<.05) was associated with a greater likelihood of non-completion, whereas lifetime history of sexual activity (X² [1; N=41] = 0.07, ns), Age (13 to 15 v. 16-18; X² [1; N=41] = 0.85, ns), and history of participation in traditional SiHLE programming (X² [1; N=41] = 1.52, ns) were not significantly associated with likelihood of SiHLEWeb completion. Among the girls who accessed the site at least once (N=36), the mean number of modules completed was 3.74 (of 4.4 total; SD=1.13), the mean number of log-ins was 4.56 (SD=3.63), and a large majority (72.2%; n=26) completed the site in its entirety. Reasons for non-completion were assessed at the follow-up assessment and are presented in Table 2.

Table 2.

Reasons for not completing the SiHLEWeb site in its entirety (n=14^*).

Reason	N	% of non-completers
Too busy	1	7.1
Site not helpful	1	7.1
Security concerns	0	0.0
Privacy concerns	0	0.0
Information not relevant	0	0.0
Had a hard time using the site	0	0.0
Forgot to/how to access the site	4	28.6
Took too much time to complete	0	0.0
Did not have access to the Internet	1	7.7

^*Although 15 participants did not complete the site, only 14 responded to questions regarding reasons for non-completion.

Preliminary Investigation of Pre-to-Post Changes

Among completers of the full website, knowledge – assessed at the outset of each module (pre) and at the conclusion of each module (post) – significantly improved, t (18)=4.74, p<.001. With respect to individual modules, knowledge significantly improved for Module 2 (STD education and condom demonstration), t (25)=3.46, p<.01. Knowledge did not significantly improve for content in Module 1 (ethnic and gender pride) [t (33)=1.98, ns], Module 3 (communication and assertiveness) [t (22)=1.55, ns], or Module 4 (healthy relationships and self-care) t (24)=1.00, ns]. Among completers of the full website, condom use self-efficacy – assessed prior to intervention completion and 3-months following baseline assessment – significantly increased, t (16)=2.41, p = .03. Partner communication [t (19)=0.37, ns], ethnic pride [t (22)=0.81, ns], and self-esteem [t (20)=1.25, ns] did not significantly improve from baseline to follow-up assessment.

Discussion

The current study was an open pilot trial to determine the feasibility of screening, recruitment, and retention efforts, as well as establish preliminary rates of treatment adherence among female African-American teen participants in a web-adapted, evidence-based HIV/STI prevention program (Leon, Davis, & Kraemer, 2011). Community-based recruitment methods have been used in numerous trials of traditionally delivered, culturally tailored, HIV prevention interventions that have been designated as “Best Evidence Group Level Interventions” by the Centers for Disease Control (DiClemente et al., 2004; Jemmott, Jemmott, Braverman, & Fong, 2005; Jemmott et al., 2010; St Lawrence et al., 1995; Stanton et al., 1996). Unlike the community-based recruitment efforts of the majority of the aforementioned trials, the current study did not recruit primarily from community health clinics serving adolescents; however, the community-based recruitment efforts for the current study were successful in recruiting at-risk African American female participants for this web-based intervention trial. More than three-quarters of study participants were recruited through solicitation of interested referrals from partnering community-based organizations (e.g., juvenile justice agency, children’s advocacy center, high schools) and/or advertising in local media outlets (e.g., university broadcast messaging, community flyers). Further, telephone-based screening, including a brief assessment of demographic and relationship status, was an efficient and low-cost procedure that resulted in an “at-risk” sample of adolescents. Over half of recruited girls reported being sexually active and one-quarter of sexually active girls indicated that they did not use a condom during their most recent sexual encounter.

SiHLEWeb completion rates were remarkably high and provide preliminary support for the feasibility of the web-based delivery of an evidence-based, culturally tailored, HIV/STI prevention intervention. Nearly two-thirds (63.4%) of all girls recruited into the study completed the intervention, taking an average of 4.5 visits to complete the content. Among those logging into the site at least once, the average girl completed well over three-quarters (3.72 out of 4) of the intervention. Consistent with prior research on access and attrition rates in e-health interventions, the current data suggest that the greatest barrier to site completion was overcoming the initial barrier of accessing the site (Price, Gros, McCauley, Gros, & Ruggiero, 2012) and the most frequently endorsed reasons for non-completion included forgetting (how) to access the site, being too busy, or finding the site content irrelevant (Christensen, Griffiths, & Farrer, 2009; Melville, Casey, & Kavanagh, 2010). Also consistent with recent national indications of the decreasing digital divide (Zickuhr & Smith, 2012), only one person identified lack of Internet access as a barrier to site completion. Rather, access and completion of the intervention may be enhanced in future efficacy and effectiveness trials through enhancing the frequency of and channels through which study reminders are sent to participants (e.g., text-message reminders), as well as addressing potential gaps in health literacy that may contribute to differential access of e-health interventions (Lustria, Smith, & Hinnant, 2011).

Retention rates at follow-up were notable, with 37 of 41 participants (90.2%) completing assessments at three-months post-recruitment. Exploratory analyses examining pre-intervention to 3-months post-intervention changes in HIV/STI risk-reduction knowledge and self-efficacy among SiHLEWeb completers demonstrated overall increases in knowledge of key content from pre-to-post module assessments that were specifically driven by significant increases in knowledge regarding STI education and condom use demonstration content. However, knowledge is widely considered a necessary but insufficient component of behavior change (Darbes et al., 2008; Romero, Galbraith, Wilson-Williams, & Gloppen, 2011). Notably, completers of SiHLEWeb also demonstrated significant increases in their condom use self-efficacy. Condom use self-efficacy has been shown to be a predictor of actual condom use behaviors in prior research among African-American teen girls (DiIorio et al., 2001; Reitman et al., 1996). No significant improvements were detected in partner communication, self-esteem, or ethnic pride at follow-up. Lack of significant improvement in the aforementioned domains may be a result of insufficient power to detect all outcome effects of interest (Leon et al., 2011), attributable to differential effectiveness of activity translation for web-based delivery, or may be indicative of potential limitations of web-based delivery more broadly. For example, partner communication skills – practiced in traditional SiHLE groups through role-play exercises – may be one partial mediator of risk-reducing behaviors that does not fully translate to completion of an individually based intervention (Sales et al., 2012; Wingood & DiClemente, 1998). The subsequent efficacy trial should be adequately powered to detect significant changes in both behavioral and behavior-mediating outcomes.

Consistent with prior search (DiClemente et al., 2004), validated self-reported measures were used to assess primary outcomes – including sexual self-efficacy, partner communication, and sexual risk behavior. In addition, the current study included a follow-up timeframe of 90 days for the purposes of establishing the feasibility of follow-up methods, as well as assessing sustained changes in condom use self-efficacy, partner communication, and ethnic pride. However, for the purposes of subsequent efficacy evaluation, follow-up assessment should extend to include 6- and 12-month post-intervention assessment and should incorporate biological assessment of HIV/STI status alongside self-report measures, consistent with efficacy evaluations of traditionally delivered, group level HIV/STI prevention interventions identified by the CDC as meeting “best evidence” criteria (DiClemente et al., 2004; Jemmott et al., 2005; St Lawrence et al., 1995; Stanton et al., 1996). The current study cannot assure that the designated participant, as opposed to a proxy, completed the site; as noted earlier, completion data represent whether a person logging into the site with unique participant identifiers viewed all pages of the intervention. Similarly, because participants were not monitored during their completion of the site, we cannot speak to the extent of engagement with site content. For the purposes of subsequent efficacy evaluation, monitored completion of the site with integration of assessments of engagement would be preferred (Klein & Card, 2011). In line with the state purposes of open pilot trials (Leon et al., 2011), participants were reimbursed for their time in completing the site as a component of evaluating the feasibility of methods; however, participant remuneration limits the ability of the current pilot to speak to the potential effectiveness of recruiting adolescents to access and complete the intervention in the absence of incentives.

Findings from this open pilot trial of the SiHLEWeb intervention demonstrate the feasibility of methods of recruitment, screening, and retention. Further, results indicate that e-health delivery of HIV/STI prevention interventions may be a low-cost, viable method for expanding the reach, and therefore the impact, of evidence-based HIV/STI prevention programming among at-risk adolescent populations. Subsequent research should establish the efficacy of SiHLEWeb in producing sustained changes in HIV/STI risk behavior over time through randomized-controlled-trial methodology, as well as explore the viability and effectiveness of efforts to increase the reach of prevention programming through the dissemination SiHLEWeb to community-based agencies that may not have the capacity to deliver more cost-intensive traditional prevention programming.