Sexual health disparities exist for women of color as the burden of sexually transmitted infections (STIs), including HIV, falls on ethnic racial minorities (Centers for Disease Control and Prevention, 2016). Sexually transmitted infections can result in serious health issues for women as they increase the risk for HIV, infertility, complications during pregnancy, and death (Centers for Disease Control and Prevention, 2016). The STI epidemic continues to disproportionately affect women of color. For example, compared with White women, chlamydia is contracted twice as frequently in Hispanic women, four times more often in Native women, and five times more frequently in African American women (Centers for Disease Control and Prevention, 2016). Comparatively, Asian American women have lower STI rates but are an understudied population regarding sexual health outcomes (Centers for Disease Control and Prevention, 2016). Numerous behavioral interventions have been designed to reduce STIs in women of color, yet, in spite of these interventions, STIs such as chlamydia, are increasing at alarming rates (Centers for Disease Control and Prevention, 2015).

In addition to women having anatomical differences making them more susceptible to STIs, women of color experience sociocultural conditions that increase their vulnerability to STIs (National Women's Health Information Center, 2016). For the purposes of this review, "women of color" was defined as women who self-identify or belong to a population made up of persons "of mixed racial descent or of other nonwhite descent" (https://www.thefreedictionary.com/Women+of+color, p. 3). Women of color share experiences as a result of membership in marginalized groups and, as a result, endure social, economic, and cultural barriers (i.e., environment, discrimination, poverty, and access to care) to preventive health information that would reduce their STI/HIV risks (Short & Williams, 2014). The majority of STI/HIV behavioral interventions developed for women of color have (a) specifically targeted African American and Latina women, (b) incorporated culturally specific elements, and (c) focused on modifying risky sexual behaviors, such as increasing condom use and reducing number of sexual partners (Crepaz et al., 2009). Additionally, these interventions focused on increasing STI/HIV knowledge, skill building for negotiating safer sex, and counseling about behavior change as the primary goal of reducing the spread of STIs (Crepaz et al., 2009; Wetmore, Manhart, & Wasserheit, 2010). Many researchers have suggested that grounding STI/HIV prevention in a target group's culture would make it more effective, but few have explicitly described how they have integrated culture into their interventions (Crepaz et al., 2009; Logan et al., 2002; Wilson & Miller, 2003). Our intent was to fill this gap by reviewing all STI/HIV preventive interventions published between 2008 and 2018 that stated they had incorporated culture into their interventions.

What Is Culture?

Culture is a complex, multidimensional construct (Barrera, Castro, Strycker, & Toobert, 2013; Wilson & Miller, 2003). Fiske (2002) captured the constructs of culture:

A culture is a socially transmitted or socially constructed constellation consisting of such things as practices, competencies, ideas, schemas, symbols, values, norms, institutions, goals, constitutive rules, artifacts, and modifications of the physical environment. (p. 85)

Culture is a way of life and can be passed down through generations (Wilson & Miller, 2003). Health is influenced by culture and behavior, which can be used to explain disease processes (Barrera et al., 2013). For women of color, cultural aspects such as beliefs, values, traditions, and norms may influence the risk for STIs (Wilson & Miller, 2003). For example, some Hispanic/Latino cultures discourage discussion about topics such as STIs and sex because of machismo (cultural expectation to respect and be submissive to males) and Catholic beliefs (Gipson & Frasier, 2003). Additionally, the historical context of African American women's sexual experiences (i.e., slavery) can cause internalized oppression and discrimination, which may be a barrier to accessing sexual and reproductive health care (Short & Williams, 2014). Sexualized and stereotyped messaging is also known to influence African American female sexuality, sexual attitudes, perceptions of self, and self-esteem, which can increase risky sexual behaviors (Rosenthal & Lobel, 2016; Townsend, 2008). Discomfort in discussing sexuality and sexual behavior due to cultural norms makes it difficult for women of color to negotiate safer sex practices and set limitations with sexual partners, which may increase the risk for STIs.

Infusing Culture Into Intervention Research

There is variation in what is considered to be the gold standard for incorporating culture into intervention research. Further, multiple terms have been used interchangeably to describe the idea of incorporating culture into research and intervention (e.g., culturally tailored, culturally sensitive, culturally adapted). Barrera et al. (2013) noted many distinctions between these terms, but consistency has been lacking in qualification for a culturally grounded intervention. Resnicow, Soler, Braithwaite, Ahluwalia, and Butler (2000) defined culturally sensitive interventions as, "ethnic cultural characteristics, experiences, norms, values, behavioral patterns, and beliefs of a priority population as well as relevant historical, environmental and social forces[horizontal ellipsis]in the design, delivery and evaluation" (p. 272). Barrera et al. (2013) found the term "culturally tailoring" as often being synonymous with "cultural adaption." The CDC defined adaption as, "the process of modifying key characteristics of an intervention, recommended activities and delivery methods, without competing with or contradicting the core elements, theory, and internal logic of the intervention thought most likely to produce the intervention's main effects" (McKleroy et al., 2006). Kreuter and Skinner (2000) proposed that "tailored" be defined as, "any combination of information or change strategies intent to reach one specific person based on characteristics that are unique to that person, related to the outcome of interest, and have been derived from an individual assessment" (p. 1).

Culturally grounded interventions have the potential to improve client engagement, increase retention, develop a trusting relationship between participants and researchers, and address culturally relevant social norms in the population of interest (Marsiglia & Booth, 2015). Therefore, we focused our review on STI/HIV behavioral interventions for women of color that included some element of culture in their interventions. There are discrepancies in how to incorporate culture into interventions as well as in the consideration of culturally adapted theories. Race and gender disparities in US STI/HIV rates have persisted despite efforts of "culturally grounded" behavioral interventions aimed at women of color (CDC, 2016). We aimed to examine the state of STI/HIV behavioral interventions for women of color, evaluate how culture was incorporated into interventions, and identify gaps in the literature.

Methods

Eligibility Criteria

We included published, peer-reviewed empirical studies addressing behavioral interventions to prevent STIs/HIV in women of color. We used the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines to guide the review (Liberati et al., 2009). The PRISMA guidelines helped the authors create criteria in order to determine which articles should be included in the analysis. Articles were reviewed for inclusion based on the following criteria: (a) included women of color, (b) focused on STI/HIV prevention, (c) infused culture into the intervention, (d) used empirical methods such as randomized control trials and evidence-based interventions (EBIs), and (e) were written in English. Articles were excluded based on the following criteria: (a) studies included men in the sample; (b) interventions did not incorporate elements of culture into the intervention; (c) interventions targeted at sex workers, churches, parents, women living with HIV, incarcerated females, or health care providers; (d) interventions using vaccinations as the sole prevention; or (e) published abstracts.

Literature Search Strategy

An electronic search of articles published from January 2008 to January 2018 was performed using PubMed and Cumulative Index to Nursing and Allied Health Literature (CINAHL) with the following terms: [women of color OR women of colour OR african american OR minority* OR black OR native american OR Asian OR Latin* OR Hispanic OR Indian AND sexually transmitted disease OR sexually transmitted infection OR STI OR STD AND prevention AND evidence based NOT male NOT men NOT drug NOT parents NOT pregnancy NOT cancer (full text[sb] AND "last 10 years"[PDat] AND Humans[Mesh] AND English[lang] AND Female[MeSH Terms]). The search terms were used in the hope of identifying all studies related to women of color, which was inclusive of all non-White and was not specific to the United States.

Data Collection Process

Abstracts were screened for eligibility by the authors using the aforementioned criteria. Full text articles fitting the criteria and/or that were questionable were read by each author. If there was disagreement about eligibility, a third reviewer read the article and we discussed until consensus was reached.

When the final set of 17 articles was selected, the two authors independently extracted data from the articles. Information related to STI/HIV behavioral interventions among women of color were extracted and included (a) author(s)/year, (b) sample characteristics, (c) design/intervention, (d) theoretical framework guiding the intervention, (e) how they were culturally grounded, (f) outcomes and measures, and (g) results of the study. The authors met and compared data extractions, combining them into a single evidence table. Disagreements were discussed until consensus was reached.

Synthesis of Results

We first identified study characteristics such as design, sample size, and demographics. We then examined how the intervention was developed including the use of theoretical frameworks. We also focused on how the intervention was delivered (e.g., individual vs. group, electronic vs. face-to-face, single session vs. multiple sessions) as well as outcomes measured in the study. Finally, we investigated how authors described incorporating culture into their interventions.

Results

Search Outcome

The search resulted in 128 potential articles. Forty-four abstracts appeared to meet criteria or required further review. After review of the 44 full text articles, 17 met criteria and were included in the review. Figure 1 depicts both the PRISMA process and results of the search. The results section focuses on the characteristics of studies reviewed and approaches to integrating culture.

Figure 1. PRISMA Diagram (Liberati et al., 2009).

Characteristics of Included Studies

We found 17 studies that incorporated culture into their interventions. All studies used a quantitative design. Most (n = 11) of the studies were experimental randomized control trials and assessed the efficacy of their interventions (Champion & Collins, 2012; DiClemente et al., 2009; Dolcini, Harper, Boyer, & Pollack, 2010; Helion, Reddy, Kies, Morris, & Wilson, 2008; Ito, Kalyanaraman, Ford, Brown, & Miller, 2008; Klein & Card, 2011; Klein, Kuhn, Altamirano, & Lomonaco, 2017; Marion, Finnegan, Campbell, & Szalacha, 2009; Thurman, Holden, Shain, Perdue, & Piper, 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Four studies used a quasi-experimental design (Bertens, Eiling, van den Borne, & Schaalma, 2009; Danielson et al., 2013; Harper, Bangi, Sanchez, Doll, & Pedraza, 2009; Hawk, 2013), and two used a descriptive design (Broaddus, Marsch, & Fisher, 2015; Davidson et al., 2014).

Sample sizes ranged from 7 to 715. Many studies had 100-200 or more females in their samples (Bertens et al., 2009; Broaddus et al., 2015; Champion & Collins, 2012; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Hawk, 2013; Helion et al., 2008; Klein & Card, 2011; Klein et al., 2017; Marion et al., 2009; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Study participants ranged from 13 to 65 years of age. A summary of study characteristics is presented in Table 1. Although the purpose of our review was to focus on populations of women of color, two studies included White women in their samples (Helion et al., 2008; Ito et al., 2008). Only four studies had samples of Latina women (Davidson et al., 2014; Harper et al., 2009; Klein et al., 2017; Wingood, DiClemente, Villamizar, et al., 2011), and eight studies focused solely on African American women (Broaddus et al., 2015; Danielson et al., 2013; DiClemente et al., 2009; Dolcini et al., 2010; Hawk, 2013; Klein & Card, 2011; Marion et al., 2009; Wingood, Card, Er, et al., 2011). Two studies included both African American and Latina women (Champion & Collins, 2012; Thurman et al., 2008), one study focused on Latina, White, and African American women (Ito et al., 2008), and the Bertens et al. (2009) sample consisted of Afro-Caribbean women, which included Afro-Surinamese and Dutch Antillean women.

Table 1-a. Study Characteristics (

Interactive Delivery

Seven studies used interactive media formats including computer web-based, compact disc read-only memory, digital optical disc storage format, and text message (Broaddus et al., 2015; Danielson et al., 2013; Helion et al., 2008; Ito et al., 2008; Klein & Card, 2011; Klein et al., 2017; Wingood, Card, Er, et al., 2011); only two of those interventions showed significant changes (Danielson et al., 2013; Wingood, Card, Er, et al., 2011). Broaddus et al. (2015) conducted an acceptability study of text message versus small group intervention delivery modalities and found that both interventions were acceptable to African American females.

Twelve studies included elements of a face-to-face intervention including individual or group-based counseling (Bertens et al., 2009; Broaddus et al., 2015; Champion & Collins, 2012; Davidson et al., 2014; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Hawk, 2013; Marion et al., 2009; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Seven of those studies were effective in reducing STI risk behaviors (Bertens et al., 2009; Champion & Collins, 2012; DiClemente et al., 2009; Harper et al., 2009; Marion et al., 2009; Thurman et al., 2008; Wingood, Card, Er, et al., 2011), and three studies looked at and reduced STI/HIV incidence (Champion & Collins, 2012; DiClemente et al., 2009; Thurman et al., 2008). Two studies found improvements in attitudes, beliefs, and STI/HIV knowledge (Bertens et al., 2009; Harper et al., 2009).

Ten studies had group-based interventions (Bertens et al., 2009; Champion & Collins, 2012; Davidson et al., 2014; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Hawk, 2013; Marion et al., 2009; Thurman et al., 2008; Wingood, DiClemente, Villamizar, et al., 2011). Seven were effective in reducing STI/HIV risk behavior (Bertens et al., 2009; Champion & Collins, 2012; DiClemente et al., 2009; Harper et al., 2009; Marion et al., 2009; Thurman et al., 2008; Wingood, DiClemente, Villamizar, et al., 2011). In contrast, six studies had interventions targeted only to the individual; no statistically significant differences were found between the control and intervention groups in those studies (Broaddus et al., 2015; Danielson et al., 2013; Helion et al., 2008; Ito et al., 2008; Klein & Card, 2011; Klein et al., 2017). Wingood, DiClemente, Villamizar, et al. (2011) created an individualized computer-based intervention in conjunction with a 15-min small group session and found that their intervention increased STI/HIV protective behaviors.

Twelve studies used multiple session interventions, and nine of them reported a reduction in STI/HIV risk behaviors (Bertens et al., 2009; Champion & Collins, 2012; Danielson et al., 2013; Davidson et al., 2014; DiClemente et al., 2009; Harper et al., 2009; Klein & Card, 2011; Klein et al., 2017; Marion et al., 2009; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Four studies used single session interventions and none of those studies were effective in reducing STI/HIV risk behaviors (Dolcini et al., 2010; Hawk, 2013; Helion et al., 2008; Ito et al., 2008).

Theories Used

All but one study (Helion et al., 2008) used theory in the development of the intervention. Because most interventions were adapted from Sisters Informing Sisters about Topics on AIDS (SISTA) and Sisters Informing Healing Living and Empowering (SIHLE), the most commonly used theories were Social Cognitive Theory and the Theory of Gender and Power (Broaddus et al., 2015; Danielson et al., 2013; Davidson et al., 2014; DiClemente et al., 2009; Klein & Card, 2011; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Five studies used the AIDS Risk Reduction Model (Champion & Collins, 2012; Dolcini et al., 2010; Harper et al., 2009; Klein et al., 2017; Thurman et al., 2008). Although a majority of studies used theory to develop their interventions, only three studies explicitly stated that the theories they used were culturally adapted (DiClemente et al., 2009; Marion et al., 2009; Thurman et al., 2008). DiClemente et al. (2009) adapted the Theory of Gender and Power for young African American women. Thurman et al. (2008) used AIDS Risk Reduction Model and stated it was culturally adapted for minority women. Marion et al. (2009) adapted the Interaction Model of Client Health Behavior (IMCHB). All three interventions using culturally adapted theories were effective in reducing STI/HIV risk behaviors (DiClemente et al., 2009; Marion et al., 2009; Thurman et al., 2008).

Outcomes

Outcomes varied greatly across studies from psychosocial, behavioral, and biologic markers. A majority of studies used self-reported measures, which included both psychosocial and behavioral outcomes. Psychosocial outcomes included STI/HIV knowledge; perceptions; attitudes toward condom use; intentions to use condoms; social support/friendships, communication, and self-efficacy; gender; and cultural norms (Bertens et al., 2009; Danielson et al., 2013; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Klein & Card, 2011; Klein et al., 2017; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Behavioral outcomes included condom protected sex, number of sexual partners, consistent condom use, and self-reported STI history or diagnosis (Bertens et al., 2009; Danielson et al., 2013; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Klein & Card, 2011; Klein et al., 2017; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). In four studies, the primary outcome was biologically confirmed STI incidence (Champion & Collins, 2012; DiClemente et al., 2009; Marion et al., 2009; Thurman et al., 2008).

Integrating Culture Into Interventions

There are currently no standards for evaluating the incorporation of culture into an intervention. Although all studies included in our review infused elements of culture into the intervention, their methods varied. We described the variation in these methods, as many authors did not specifically state how they evaluated the cultural components of their interventions, only that they adapted an EBI. Some researchers took a more comprehensive approach such as using qualitative methods (i.e., focus groups, ethnography) to capture the cultural experiences of women and integrated them into the interventions (Bertens et al., 2009; Davidson et al., 2014; DiClemente et al., 2009; Dolcini et al., 2010; Harper et al., 2009; Hawk, 2013; Klein et al., 2017; Wingood, DiClemente, Villamizar, et al., 2011).

Three studies conducted focus groups to make their interventions culturally grounded in the knowledge, beliefs, values, and languages of the population of interest (Davidson et al., 2014; Klein et al., 2017; Wingood, DiClemente, Villamizar, et al., 2011). Harper et al. (2009) and Wingood, DiClemente, Villamizar, et al. (2011) used community-based participatory research and narrative ethnographic methods to reveal community and cultural narratives. DiClemente et al. (2009) used qualitative methods for their intervention, which was field-tested in the community to assess gender and cultural appropriateness. Hawk (2013) described using community input to increase cultural relevance. Klein et al. (2017) provided extensive descriptions on how their intervention incorporated cultural specificities of Latinas such as the use of telenovela-style videos. Dolcini et al. (2010) used a friendship intervention for African American women that incorporated social and cultural factors (i.e., cultural exercises) into the structure and content of programs as described by Wilson and Miller (2003). Overall the studies that provided more comprehensive approaches using both community-based and qualitative research to integrate culture into their interventions were shown to be more successful in reducing STI/HIV risk behaviors (Bertens et al., 2009; DiClemente et al., 2009; Harper et al., 2009; Wingood, DiClemente, Villamizar, et al., 2011).

Other studies used minimal efforts to integrate culture through the use of only intervention facilitators. Some studies stated that the facilitators of the interventions matched the race of participants (Dolcini et al., 2010) or gave participants the opportunity to choose a culturally appropriate host to facilitate the intervention (Helion et al., 2008; Ito et al., 2008). None of those studies showed any differences between the control and intervention groups.

Only one study by Davidson et al. (2014), which adapted an EBI, explicitly stated that they used the Assessment Decision Administration Production Topical Experts-Integration Training Testing framework to assess cultural modifications, which included speaking with key stakeholders and gathering information from the desired population by conducting focus groups (Wingood & DiClemente, 2008). Nine studies were adapted from culturally tailored EBIs, but did not specifically describe how they retained cultural elements of these interventions (Broaddus et al., 2015; Champion & Collins, 2012; Danielson et al., 2013; Davidson et al., 2014; Klein & Card, 2011; Klein et al., 2017; Thurman et al., 2008; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011). Therefore, we conducted an ancestral search to determine how the authors incorporated cultural elements from EBIs into their interventions. Three studies (Broaddus et al., 2015; Wingood, Card, Er, et al., 2011; Wingood, DiClemente, Villamizar, et al., 2011) adapted SISTA created by DiClemente and Wingood (1995), which emphasizes ethnic and gender pride. Of these studies Wingood, DiClemente, Villamizar, et al. (2011) was the only one to show differences between the control and intervention group. Three studies (Danielson et al., 2013; Davidson et al., 2014; Klein & Card, 2011) were adapted from SIHLE developed by DiClemente et al. (2004). Researchers from SIHLE collaborated with African American adolescent girls in the community to create a culturally tailored intervention (DiClemente et al., 2004). All studies found participant satisfaction with the interventions, but no statistically significant differences were found in the control group. Three studies (Champion & Collins, 2012; Klein et al., 2017; Thurman et al., 2008) adapted Project Sexual Awareness For Everyone created by Shain et al. (1999). Project Sexual Awareness For Everyone incorporated ethnographic data about women's lifestyles, values, beliefs, and other measures, and also had a multiethnic team to help design the intervention (Shain et al., 1999). Champion and Collins (2012) and Thurman et al. (2008) both reported lowered incidence of STIs. Overall only three adapted EBIs significantly reduced STI/HIV risk behaviors or incidence (Champion & Collins, 2012; Thurman et al., 2008; Wingood, DiClemente, Villamizar, et al., 2011).

Additionally, an in-depth search of Hawk et al. (2013) revealed that the researchers used a bottom up approach to increase cultural relevance to develop an intervention created by African American women for African American women (Hawk, 2015). Bertens et al. (2009) developed Uma Tori and conducted an in-depth cultural exploration to create their culturally appropriate intervention, which incorporated core cultural, social, historical, environmental, and psychological factors that grounded the content of their intervention into the context, experiences, values, beliefs, and norms of the priority group (Bertens, 2008). Marion et al. (2009) designed the Well Woman Program, which adapted the interaction model of client health behavior, to create a culturally specific STI prevention framework for African American women but provided no information about how it was culturally grounded.

Discussion

The purposes of this review were to (a) examine the current state of STI/HIV behavioral interventions for women of color, (b) evaluate how culture was incorporated into these interventions, and (c) identify gaps in the literature. Our findings revealed face-to-face, group-based, multiple session, and theory-based interventions were more effective than single session, interactive, media interventions for reducing STI/HIV risk behaviors. We also found that studies that provided more comprehensive approaches using both community-based and qualitative research to integrate culture into their interventions were shown to be more successful at reducing STI/HIV risk behaviors (Bertens, 2008; DiClemente et al., 2009; Harper et al., 2009; Wingood, DiClemente, Villamizar, et al., 2011). Additionally, we found inconsistencies in how to infuse cultural components into behavioral interventions, how to adapt EBIs to retain cultural components, and how to culturally adapt theory-guided interventions.

Although interactive media interventions may be more cost effective and increasing in use, only one study (Wingood, Card, Er, et al., 2011), using a computer-based intervention, was found to be effective, but they also included a small group face-to-face component. Danielson et al. (2013) found pre to post changes in STI/HIV behaviors in their web-based delivery. Additionally, the effects of these interventions were not tested past 6 months. No significant differences in intervention effects were attributed to amount of time (hours included in the session) used to deliver the interventions. Although the majority of interactive media interventions were adapted from culturally tailored EBIs, the authors were not explicit about how cultural tailoring was retained in the interactive media format. Only the studies by Klein & Card (2011) and Klein et al. (2017) described how cultural components from the EBIs were translated into multimedia formats for their interventions.

Our review found that single session interventions were not as effective as multiple session interventions. A systematic review by Lin, Whitlock, O'Connor, and Bauer (2008) supported this finding. Multiple session formats may reinforce prevention strategies and create community among participants (Broaddus et al., 2015; DiClemente et al., 2009). In contrast, a meta-analysis by Crepaz et al. (2009) found that interventions with fewer sessions (e.g., 1 session) were as effective at reducing STI/HIV risk behaviors as interventions with more sessions. Another study by Jemmott, Jemmott, and O'Leary (2007) found that the effects of their single session, culturally sensitive intervention could be sustained up to 12 months after intervention implementation.

Overall group-based, multi-session, theory-based interventions that were delivered face-to-face were most effective at reducing STI/HIV risk behaviors (Bertens, 2008; Champion & Collins, 2012; DiClemente et al., 2009; Harper et al., 2009; Marion et al., 2009; Thurman et al., 2008; Wingood, DiClemente, Villamizar, et al., 2011). Researchers have found group-based interventions to be effective in women of color because of the benefit of shared experiences, social support, and reducing social stigma (Bertens et al., 2009; Broaddus et al., 2015). Other researchers have found group-based interventions that included elements of skill building, role playing, and interactive exercises to be effective for women of color (DiClemente et al., 2004; Jemmott et al., 2007; Shain et al., 1999). Findings from the meta-analysis by Crepaz et al. (2009) revealed that interventions for African American women should be culturally specific and focus on empowerment, which DiClemente and Wingood (1995) addressed in the cultural adaption of the Gender and Power Theory. Although many studies were effective at reducing STI/HIV risk behavior, we found three major gaps: (a) sustainability of interventions needs to be considered, (b) theoretical frameworks are not adapted for women of color, and (c) there is a lack of clarity about how to infuse culture into an intervention.

Sustainability

Given the fact that STI/HIV rates are high and consequences of these diseases are devastating for women of color, sustainability of interventions needs to be considered. Only four studies (Champion & Collins, 2012; DiClemente et al., 2009; Marion et al., 2009; Thurman et al., 2008) tested post-intervention follow-up to 12 months; the majority of the studies included in our review had not assessed follow-up beyond 2-3 months, so there were minimal data to assess whether the interventions were effective in the long term. A recent systematic analysis by Wetmore et al. (2010) shared similar findings that behavioral interventions were effective in the short term, but they could not demonstrate long-term effectiveness of their interventions because of data limitations. Given that young Black women are vulnerable to STI reinfection, researchers should focus on longer-term outcomes (Craft-Blacksheare, Jackson & Graham, 2014). Despite efforts to make the information about STI/HIV transmission and prevention available to women of color, STI rates, including repeat infections among African American and Hispanic women, remain the highest of any other racial groups (Centers for Disease Control and Prevention, 2016). Sustainability in behavioral interventions needs to examine how long-term behavior change can be sustained and to identify barriers to sustaining safe sexual behaviors (Wetmore et al., 2010). Additionally, researchers need to explore why women return to former behavior patterns. One way researchers can understand women's barriers to long-term sexual risk-related behavior change is through qualitative research. For example, Sales, DiClemente, Davis, and Sullivan (2012) explored why condom use did not increase post intervention in a subset of young African American women. Using grounded theory, the authors found barriers (e.g., nonstable relationships, substance abuse) to changing sexual behaviors for this population. Qualitative work such as that of Sales et al. (2012) can serve to inform future interventions by documenting what is relevant to women and the barriers to behavior change.

Limitations of Theoretical Frameworks

Although a majority of researchers used theory in the development and implementation of their interventions, only three studies used culturally adapted theories (DiClemente et al., 2009; Marion et al., 2009; Thurman et al., 2008). The majority of the studies we reviewed used theories that had not been adapted for women of color. For example, Social Cognitive theory, which was used by three researchers (Broaddus et al., 2015; Danielson et al., 2013; DiClemente et al., 2009), was developed and tested in White populations (Bandura, 1986). Many of the theories used in the development of interventions did not explicitly consider culture and cultural values, which have rarely been explored or explained in theoretical terms (Bertens, 2008). Culture and cultural values are important to explore in theory development because they can influence perceptions of sexuality and sexual behavior, communication, and negotiation of safer sex practices with partners (Gipson & Frasier, 2003). A meta-analysis by Logan et al. (2002) shared similar findings in critiquing the use of theory in interventions that did not consider the dyadic nature of sexual behavior or the range of important sociocultural factors (i.e., environment, culture) that have influenced women's sexual behavior. The intersection of race and gender/sex makes women of color especially vulnerable to historical trauma and discrimination as a result of being a part of marginalized groups (Crenshaw, 1991; Short & Williams, 2014). Women of color have been affected by various additional sociocultural conditions that may increase STI risk (Gipson & Frasier, 2003; Logan et al., 2002). Theorists should explore how these sociocultural conditions could be integrated into theory and then into interventions (Logan et al., 2002), and new theories should be developed specifically for women of color or extant theories should be culturally adapted to support the needs of under-represented populations (Bertens, 2008; Mize, Robinson, Bockting, & Scheltema, 2002).

How to Infuse Culture Into Interventions

The terms used by studies to describe the incorporation of culture into interventions were "culturally appropriate," "culturally based," "culturally relevant," "culturally tailored," "culturally specific," and "culturally congruent." We demonstrated that there was no consistency on how researchers infused culture into their interventions and there were no standard guidelines to do so. Although many of the interventions were adapted from EBIs, these EBIs were developed from 10 to nearly 20 years ago (DiClemente & Wingood, 1995; DiClemente et al., 2004; Shain et al., 1999). Such methods are likely outdated, as sexual behaviors and cultural norms change over time. Researchers and other practitioners should be more specific and explicit in how they adapt culturally tailored EBIs to address the sociocultural conditions of their populations (Logan et al., 2002). Research should, therefore, explore a clear, consistent definition of culturally adapting EBIs and then examine the effect of the revised or new culturally tailored interventions in the context of current social norms.

There are also discrepancies in how to adapt an EBI and what adaption means. Adaptation has been defined as "the degree to which an innovation is changed or modified by a user in the process of its adoption and implementation" (Rogers, 2003, p. 180) or "deliberate or accidental modification of a program" (Center for Substance Abuse Prevention, 2001, p. 7). Adaptation can include deletions or additions, modifications of existing components, changes in the manner or intensity of components, or cultural modifications required by local circumstances (Center for Substance Abuse Prevention, 2001).

There are very few recommended processes or best practices for adapting EBIs to conditions different from those presented in the original research (McKleroy et al., 2006; Wingood & DiClemente, 2008). The CDC has developed a map of the adaption process to adapt EBIs, which includes five action steps: assess, select, prepare, pilot, and implement (McKleroy et al., 2006). The adapt-ITT (assessment decision administration production topical experts-integration training testing) model created by Wingood and DiClemente (2008) described how to adapt evidence-based HIV interventions in eight phases: assessment, decision, administration, production, topical experts, integration, training, and testing. Although this is a valuable contribution to the literature, the process is quite extensive, which may limit its use in research. There is no concrete definition of how to culturally adapt interventions (McKleroy et al., 2006; Wingood & DiClemente, 2008). This is important to distinguish, as researchers who adapt EBIs aren't distorting the core cultural components/elements of an intervention. As interactive media increases in popularity, new research should focus on how the adaption of cultural components of EBIs translate into interactive media formats. It is also important for researchers to consider populations in which EBIs work before dissemination into communities.

Limitations

A first limitation was that although this review was inclusive of women of color, many of the studies included in our review focused primarily on African American women and some on Hispanic women. This limitation may be due to our inclusion criterion for English-language-only studies. This highlighted a gap in STI/HIV prevention research conducted with Asian, Native, and Indian populations. A second limitation was that the majority of the research focused on heterosexual relationships. Our ancestral searches revealed that a culturally adapted theory does not necessarily make an intervention culturally tailored. As discussed previously, infusing elements of culture and doing a more in-depth exploration of culture from the perspective of the population of interest would be necessary to create a culturally tailored intervention.

Recommendations

Although we found multiple session interventions to be more effective at reducing STI/HIV risk for women of color, additional studies are needed to further assess the efficacy of single-session to multiple-session interventions. This is necessary to determine which intervention format is better suited for populations of women of color. In addition to the session format, it is important to consider sustainability of the interventions. Given that studies in our review did not consider sustainability past 2-3 months and that most re-infections occurred 3-6 months following initial diagnosis/treatment, sustainability should be documented to determine long-term effectiveness of interventions. Researchers should consider the sustainability of their interventions given that STI/HIV prevention requires long-term behavioral change. Additionally, researchers should determine which interventions work for each age group as certain prevention strategies may be more effective in older versus younger populations of women of color.

Culturally adapting frameworks, theories, and interventions are necessary to support the needs of women of color. There is a lack of interventions that explicitly address sociocultural conditions related to culture, despite evidence that has demonstrated their effectiveness. In order to develop culturally sensitive interventions, health promoters should first aim to understand the cultural structures that influence their populations (Bertens, 2008; Resnicow et al., 2000; Wilson & Miller, 2003). The objectives of the intervention should address the needs of the population, and theoretical methods and practical strategies should match structures of the population (Bertens, 2008; Resnicow et al., 2000; Wilson & Miller, 2003). The lack of culturally adapted interventions may create barriers between health care providers, researchers, and participants and patients (Marsiglia & Booth, 2015). Using a culturally grounded approach in a clinical setting could lead to more equitable and productive relationships between patients and providers by grounding interventions in patient-lived experiences (Marsiglia & Booth, 2015). Culturally grounding interventions can be achieved in research by using or developing theories or frameworks that speak to the lived experiences of women of color, such as intersectionality. Intersectionality is a theoretical framework that considers how different identities, such as race and gender, are not separate, but interact to affect the lives of women of color, including their health (Crenshaw, 1991). Grounded theory could be used to create a framework that is developed and grounded in the perspectives and experiences of women of color. Studies that consider these recommendations could greatly inform prevention efforts in high-risk groups such as women of color.

Health care providers play a crucial role in STI/HIV prevention and treatment because they regularly encounter patients and can support the linkage to care in STI/HIV programs. Although women of color are at higher risk for STIs, it is important that providers do not overscreen minority women for STIs or target minority women without understanding the social and cultural context of higher STI rates, as the effects can result in labeling, shaming, and anxiety, having a negative effect on the patient-provider relationship (Zakher, Cantor, Pappas, Daeges, & Nelson, 2014). A recent review on overscreening for STIs in urban women suggested that although personal history of STI and younger age were important clinical variables, they did not necessarily result in a positive test (Jackson, McNair, & Coleman, 2015). Based on our findings, it is important for clinicians to discuss and explore possible risks for STI/HIV exposures that are culturally and contextually specific to the individual despite age, race, and environment.

Conclusion

Our review can facilitate discussion and research about integrating culture into STI/HIV interventions. This is important, as women of color are at particular risk for STI/HIV and culturally tailored behavioral interventions have been found to reduce risk. We found that group-based, multiple-session, theory-based interventions with infused culture were more effective than single-session, interactive, media interventions for reducing STI/HIV risk behaviors. The best approaches to integrate culture into interventions are the use of community-based research and qualitative research methods. Researchers should compare single- to multiple-session interventions and examine sustainability of the interventions to reduce STI recurrence. There is a lack of research on standards for integrating culture into extant interventions, the creation of new theories, or adapting theories specifically for women of color. Researchers should consider the sustainability of their interventions given that STI/HIV prevention requires long-term behavioral change.

Key Considerations

* There is a need to consider how to infuse culture into interventions designed for women of color who are at risk for STI/HIV.

* The combination of group-based, multiple-session, theory-based interventions with infused culture was more effective than single-session interactive media interventions at reducing STI/HIV risk behaviors.

* There is a lack of research on standards for integrating culture into interventions and adapting evidence-based interventions to retain the cultural components necessary to address the sociocultural conditions that impact STI/HIV risk in women of color.

* Clinicians should discuss and explore possible risks for STI/HIV exposures that are cultural and context specific to the individual despite age, race, and environment.

Disclosures

The authors report no real or perceived vested interests related to this article that could be construed as a conflict of interest.

Acknowledgments

This research was funded by the National Institute of Nursing Research (PI, N. Crooks, Grant #1F31NR016624-01A1) and by the University of Wisconsin-Madison, School of Nursing Robert and Carroll Heideman Research Award. The authors would like to give special thanks to Mary Hitchcock for her guidance in data sources and to Lisa Bratzke for her contributions in reviewing selected articles for the literature review and comments on the manuscript.

References