The Health, Enlightenment, Awareness, and Living (HEAL) Intervention: Outcome of an HIV and Hepatitis B and C Risk Reduction Intervention

African American women have among the highest HIV/AIDS and hepatitis B and C incidence rates in the United States, especially among those homeless or incarcerated. The objective of this study was to evaluate the Health Enlightenment, Awareness and Living Intervention, designed to decrease HIV/AIDS, hepatitis and related risky behaviors. The thirteen-session intervention was implemented among homeless and formerly incarcerated low-income African American women, ages 18 to 55, in Atlanta, Georgia from 2006 to 2010. A single group repeated measures study design was employed and consisted of a pre-test (n = 355) group, an immediate post-test (n = 228) group with a response rate of 64%, and a six-month follow up (n = 110) group with response rate of 48%, completing a 135-item survey. Paired-sample t-tests, McNemar tests, and repeated measures ANOVA were applied to compare survey results. Participants demonstrated statistically significant increases in hepatitis B and C knowledge over time (p < 0.001). Statistically significant decreases were also reported for unprotected sex in exchange for money, drugs or shelter (p = 0.008), and sex under the influence of drugs or alcohol (p < 0.001). Reported substance use decreased with statistical significance for alcohol (p = 0.011), marijuana (p = 0.011), illegal drugs (p = 0.002), and crack/cocaine (p = 0.003). Findings broaden the evidence base related to the effectiveness of HIV/AIDS and hepatitis risk reduction interventions designed for homeless and previously incarcerated African American women.


Introduction
African American women are at an increased risk for contracting sexually transmitted infections (STIs), including HIV/AIDS and hepatitis. At the end of 2013, of the total estimated number of women living with diagnosed HIV/AIDS, 61% were African American, 17% were white, and 17% were Hispanics/Latinas [1]. In addition, the rate of new HIV infections among African American women is 20 times higher than white women and almost five times higher than that of Hispanic/Latina

The Intervention
In response to the aforementioned health disparities, Wholistic Stress Control Institute, Incorporated (WSCI) developed and implemented the Health, Enlightenment, Awareness, and Living (HEAL) Intervention for African American formerly incarcerated and homeless women ages 18-55. WSCI, a non-profit community-based organization located in Southwest Atlanta, Georgia, initiated a community-academic partnership with the Morehouse School of Medicine Prevention Research Center (MSM PRC), evaluators of the HEAL intervention [27]. MSM PRC evaluations have been designed to address needs mutually identified by the funder, grantees and target populations to assure that program activities (1) are audience-driven, (2) foster sustained ownership of data collection processes, (3) implement participatory processes, and (4) are perceived as central to program planning, implementation and sustainability. This approach is applied to community-based, regional and national evaluations addressing cancer screening, HIV /AIDS, maternal and child health, mental health, substance abuse, and violence disparities [28][29][30].
The HEAL intervention was designed to address HIV/AIDS and hepatitis risk reduction, incorporating culturally relevant concepts through a life skills enhancement model. Funded by the Substance Abuse and Mental Health Administration (SAMHSA), the overarching goal of the intervention was to build the community's service capacity to prevent and reduce the onset of substance abuse and transmission of HIV/AIDS and hepatitis among minority homeless and formerly incarcerated populations.
The HEAL Intervention consisted of a 13-session curriculum components addressing substance abuse, hepatitis, and HIV/AIDS, as well as life skills related to parenting, stress management, and vocational assessment towards job readiness. The curriculum was developed based on key components of well-established curricula including: "Living in Balance: Moving from a Life of Addiction to a Life of Recovery" [31], "Beyond Trauma: A Healing Journey for Women" [32], "the HIV/AIDS-SISTA Curriculum" [22,23], and "the Hepatitis Integration Training Manual" (HIT'M) ( Table 1) [33,34].
Substance Abuse Prevention: (1) Living in Balance is a research-based intervention that provides a comprehensive, group-oriented framework for people who have substance abuse challenges involving alcohol and other drugs [31]; (2) Beyond Trauma: A Healing Journey for Women promotes a strength-based approach that seeks to empower women and increase their sense of self and uses cognitive-behavioral techniques, expressive arts, and is based on the principles of relational therapy [32].
HIV Prevention: SISTA is a CDC Diffusion of Effective Behavioral Intervention designed to prevent HIV infection among African American women through promoting and reinforcing safe behaviors and strong interpersonal skills in negotiating and sustaining behavior change [22,23].
Hepatitis B and C Prevention: Hepatitis Integration Training Manual (HIT'M) by the American Liver Foundation (ALF) and the New York State Department of Health in collaboration with the CDC Viral Hepatitis Training of Trainers program, is a comprehensive guide for youth and adults that provides Hepatitis B and C education in the areas of HIV/AIDS, sexually transmitted disease, harm reduction and/or drug treatment [33,34].
The HEAL intervention also provided other social and health supports including referrals for substance abuse treatment, transitional housing, adult education, career support, or recovery programs and services. WSCI health educators and community coordinators referred participants for HIV/AIDS and hepatitis B and C screening or hepatitis B immunization to Grady Hospital/Grady Infection Disease Program, the Department of Fulton County Health and Wellness program, AIDS Survival Project, and Mercy Mobile Health Care. WSCI health educators also offered individual assessments for substance abuse and mental health referrals. Participants also received vocational referrals, job placement assistance, and were provided with personal resources upon request (i.e., Bible/prayer partners, hygiene products, dentist referrals). Stress management, coping and relaxation techniques were intentionally and strategically integrated into curriculum to address potential triggers associated with trauma and past experiences. Independent counselors were also available at intervention delivery sites for those needed more extensive supports. Curriculum details are included in Table 1.

Design
A single group repeated measures design was utilized for this study, which consisted of a self-administered HEAL intervention survey comprised of 135 questions assessing participant changes in awareness, knowledge, behavior, and perceptions relative to HIV/AIDS, hepatitis B and C, and substance abuse. At pre-test, participants were assigned a unique code number that allowed for both anonymity in the collection of data and longitudinal participant tracking through to follow-up data collection. Three hundred fifty five (355) pre-test surveys and two hundred twenty eight (228) immediate post-test surveys were administered, representing a 64% response rate. Additionally, participants completed the survey six months after intervention completion; one hundred-ten (110) follow-up surveys were administered, representing a 48% response rate. Those who completed follow-up survey all completed the immediate post-test survey. To reduce attrition, WSCI health educators have kept in contact with participants during and after sessions by calling cell phones, sending emails, providing follow up cards, etc. This was an evaluation study conducted in partnership with those implementing the intervention (WSCI) as detailed earlier. Study was designed for program improvement and was not to be used towards generalizable knowledge but with implications for public health practice as detailed in the sections that follow. Therefore, a determination was made that these activities did not constitute human subject research as outlined by the Office of Human Research Protections guidelines (45CFR 46.102.(d)). Consent forms used to invite participation detailed risks and benefits of the study, as well as contacts for potential triggers that might require further mitigation.

Study Population and Participant Recruitment
Participant eligibility criteria included residence in Fulton and DeKalb counties of Atlanta, GA and self-identification as African American women, age from 18 to 55, who had been previously incarcerated or homeless. The program implementation was conducted from October 2006 to September 2010. The sessions were held twice a week at each site for a total of seven weeks. WSCI health educators utilized a purposive non-probability sampling plan from correctional facilities and community-based sites designed to serve homeless women. WSCI collaborated with correctional facilities, residential treatment facilities, and local community agencies accepting formerly incarcerated and homeless women. Health educators provided the shelters, agencies, and penal facilities with flyers, brochures and information packets that included the interventions curriculum outline and incentives. The site counselor or member of administrative team referred participants to the HEAL intervention continuously on an ongoing basis. If individuals met the eligibility criteria, WSCI health educators offered the opportunity to be enrolled and went through the informed consent process face-to-face to ensure that all questions and concerns associated with participation were addressed. Upon enrollment, participants were grouped in several cohorts. Up to twenty participants, per cohort, met twice a week for 2 hours at residential treatment facilities or at a local agency for formerly incarcerated women. WSCI started a new cohort at least two weeks following intervention completion by a previous cohort. Participants were classified as formerly incarcerated or homeless based on information collected from the host facility by HEAL community coordinators prior to intervention initiation. The intervention had averaged 5 cohorts per year. From 379 participants who were initially invited (114 previously incarcerated and 265 homeless), 355 were enrolled in the study and took the first survey.
Formerly incarcerated populations included adult women being released from prison for project services. The HEAL intervention offered services to formerly incarcerated women at three sites. Two HEAL health educators and one community coordinator worked closely with employees at correctional facilities prior to participants release and with probation officers after participants' release, to recruit and identify participants for the HEAL intervention. Health educators followed up with formerly incarcerated participants by personal visit and/or phone contact to recruit them once they are in the community. For homeless women, the intervention was advertised in community-based agencies, residential treatment facilities, and shelters designed to serve homeless women. Information cards, brochures, and flyers were distributed to all sites during orientations. WSCI health educators met with participants at community sites to explain the services and to register interested participants. Incentives used to increase intervention retention included a $20 gift card following completion of post-test and a $10 gift card at 6-month follow-up. WSCI and MSM PRC worked collaboratively to obtain written informed consent.
Intervention sessions were delivered by a team of two female African American educators to groups from 10 to 20 participants. Two WSCI health educators had 24 hours of training on HIV/AIDS prevention interventions and relationship development curriculum that comprised of SISTA intervention, SisterLove, Inc., AID Atlanta, and Red Cross. Additionally, WSCI health educators had 24 hours training on stress management utilizing WSCI Stress Management curriculum, and 8 hours training on substance abuse prevention during WSCI in-service trainings and supervision conducted by a Licensed Professional Counselor who has addiction certification. To account for participants' attendance and intervention session completion, a daily sign-in form was utilized by an internal data manager to record attendance. A maximum of 13 dosages or intervention sessions were offered during the eight-week period. A minimum of eleven sessions were required for program completion determining eligibility for pre-test, immediate post-test and follow surveys. An internal data manager compiled an attendance tracking log to display total attendance for each session for all participants throughout the program duration. This log allowed participants, staff, and evaluators to track missed sessions and overall completion rates for participants.

Measures
The instruments administered were approved and validated by the Government Reporting and Results Act, Performance Assessment Rating Tool, Office of Management and Budget, Substance Abuse and Mental Health Administration (SAMHSA), National Outcome Measures (NOMS), and Minority AIDS Initiative. Intervention outcomes were measured by the NOMS HIV Cohort 6 survey instruments and assessed HIV/AIDS and hepatitis knowledge, perceptions of risky behaviors, and intentions to use condoms [35][36][37][38][39][40].
The outcome domains analyzed for intervention efficacy were:

HIV/AIDS and Hepatitis Knowledge
A series of 18 questions were asked assessing the participants' knowledge of HIV/AIDS and hepatitis B and C [33][34][35]. Responses included "Yes", "No", or "Don't Know" choices. Correct responses received a score of 1 and incorrect answers a score of 0. These responses were summed to form a two knowledge scores that could range from 0 to 6 for HIV/AIDS and 0 to 12 for hepatitis B and C.

Condom Use
Condom use was measured by series of 6 questions. Participants were asked question about protected vaginal sex "Have you had vaginal sex in the past 30 days?" Responses included "Yes", "No", or "Don't Know" choices. If participant reported sexual activity in last 30 days, the next question was asked: "The last time you had vaginal sex, was it protected or unprotected?" A similar set of questions were asked about oral and anal sex [40].

Sexual Behavior
Participants were asked various questions about their sexual practices in the past 30 days and in the past 3 months. Sexual behaviors included oral, vaginal, anal, or unprotected sex ("Have you had oral sex in the past 30 days? The last time you had oral sex, was it protected or unprotected?"); sex with men, women, or injection drug users; individuals who possibly had an STD or HIV/AIDS (In the past 3 months, have you had unprotected sex (vaginal, anal, or oral) with a partner you know had, or suspected of having a sexually transmitted disease (STD)?); sex in exchange for money, drugs, or shelter (In the past 3 months, have you had unprotected sex (vaginal, anal, or oral) with someone in exchange for money, drugs, or shelter?) or sex under the influence of drugs or alcohol (In the past 3 months, have you had unprotected sex (vaginal, anal, or oral) with someone under the influence of drugs or alcohol?). Participants were also asked about their use of protection the last time they had vaginal, oral, anal sex and about their intentions to use condoms in the next 6 months with responses ranging from 1 ("Not at all likely") to 4 ("Very Likely") [40].

Risk Perception
Participants were asked about the risks of unprotected oral, vaginal, and anal sex. The responses were on a scale of 1 ("No Risk") to 4 ("Great Risk") [39,40].

Substance Use
Substance abuse was measured by a series of 10 questions. Participants were asked about the number of days in the last 30 days that they used substances, including cigarettes, tobacco, alcohol, marijuana, illegal drugs, crack/cocaine, prescription drugs, injected drugs, and methamphetamines [40].

Statistical Analysis
Matched pre-test, post-test, and follow-up surveys were utilized for analysis.A few participants did not complete all questions on the 135 item survey; hence, missing data were excluded from the analysis. Pearson's chi-square test, paired-sample t-test, McNemar, and ANOVA tests were used to assess change among intervention objectives. Data analysis was conducted using PASW Statistics for Windows, Version 18.0 (SPSS, Inc., 2009, Chicago, IL, USA).
We performed a posthoc retrospective power calculation for sample of 228 participants matched of pre-post and immediate post-test, mean difference 0.5, within group standard deviation 3, and alpha levels 0.05 for paired T test. Our calculation resulted in 70.8% power of this study. Additionally, we calculated a post-hoc retrospective power calculation for matched pre-test, post-test and follow up sample of 110 participants, mean difference 0.8, within group standard deviation 3 and alpha levels 0.05 for paired T test. Data indicate that the difference in the response of matched pairs is normally distributed with standard deviation 3. Our calculation resulted in 79.2% power. We used PS software to calculate power of the study [41].

Participant Characteristics
Our analytic sample comprised 355 pre-test surveys, 228 immediate post-test surveys, and 110 follow-up surveys. On pre-test, 105 adults were formerly incarcerated and 250 homeless. On pre-test (n = 355), participants ranged in age from 19 to 55, with a mean age of 37.7. Most (80.7%) participants described their ethnicity as African-American, a little over half of participants reported completing grade 12 or higher (62.6%), and most (81.1%) reported a household income of $10,000 or less. About one-third (34.8%) reported being homeless at the time of survey administration. Table 2 provides details on the characteristics of participants by formerly incarcerated and homeless status.

HIV/AIDS and Hepatitis B and C Knowledge
Prior to the intervention, among paired pre-test and post-test surveys, the mean HIV/AIDS score was 4.79 and after the intervention the mean score was 4.94. Though this resulted in a 3.1% increase in knowledge it was not statistically significant. This may be explained by a relatively high level of baseline knowledge. The mean number of correct responses to the twelve Hepatitis items at pre-test was 5.97 and increased to 7.04 at immediate post-test survey. This increase of 17.9% was also statistically significant (t (116) = −4.62, p < 0.001).
When examining differences in knowledge over time, there was a significant change in hepatitis B and C knowledge from pre-test, to immediate post-test, and follow-up knowledge scores (F (1.88) = 17.631, p < 0.001).Post hoc tests indicate that hepatitis B and C scores at post-test were significantly greater than pre-test scores (p < 0.001) and six-month follow-up scores were significantly greater than pre-test scores (p < 0.001). HIV/AIDS scores also significantly differed over time (F (1.94) = 3.177, p < 0.05) using Greenhouse-Geisser (chi square = 2.37, p > 0.05). However, post hoc tests revealed that only differences from pre-test to post-test were statistically significant (p < 0.05) ( Table 3).

Condom Use
Participants were asked about their use of protection the last time they had vaginal, oral, or anal sex. Of those who reported having vaginal sex in the past 30 days, reported condom use increased by 12.0% from pre-test to post-test survey, however the increase was not significant (McNemar test p = 0.478). Among those who reported having oral sex in the past 30 days, condom use increased by 2.6% from pre-test to immediate post-test survey (McNemar test p = 0.655). Raw numbers reflecting self-reported anal sex were extremely low and account for the large percentage decrease. Condom use continued to show an increase over time with protected anal (35.9%), vaginal (17.2%) (McNemar test p = 0.655), and oral (13.9%) (McNemar test p = 0.317) sex demonstrating increases from pre-test to follow-up survey (Table 4).

Sexual Behavior
All forms of sexual behaviors decreased from pre-test to immediate post-test survey. Significant decreases were indicated for the number of participants who: had unprotected sex in exchange for money, drugs or shelter (6.1%) (McNemar test p = 0.008), had unprotected sex with a partner having an STD (3.6%) (McNemar test p = 0.039), and had sex while under the influence of drugs or alcohol (14%) (McNemar test p < 0.001) from pre-test to immediate post-test survey.
When examining risky sexual behaviors over time, HEAL Intervention participants demonstrated a decrease in unprotected sex with partners having an STD (McNemar test n = 196, p = 0.039). For unprotected sex in exchange for money, drugs or shelter and sex under the influence of drugs or alcohol, decreases were statistically significant from pre-test to immediate post-test (McNemar test n = 196, p = 0.008 and n = 196, p < 0.001) and from pre-test to follow-up survey (McNemar test n = 76, p = 0.039 and n = 73, p = 0.007) ( Table 4).

Sexual Risk Perceptions
At pre-test, all forms of unprotected sex were seen as "Moderately Risky", with pre-test means falling between 3 ("Moderate Risk") and 4 ("Great Risk") with 3.  (Table 3).
Mean number of days of substance use was also examined over time, comparing pre-test, immediate post-test, and six-month follow-up data for adult participants. The results show that the mean number of days for the use of tobacco products (F (1.847) = 6.70, p = 0.002) among pre-test, immediate post-test, and follow-up time points differed significantly. Post hoc tests revealed that for tobacco products, the pre-test mean was significantly higher than the immediate post-test mean (p = 0.022) as well as the follow-up mean (p = 0.010). Overall, comparing aggregate pre-test and follow-up means, mean number of days of substance use continued to show a decrease from pre-test to follow-up for all substances.
In addition to the outcomes demonstrated by survey results, HIV testing behavior was also monitored for HEAL Intervention participants. Over the span of intervention implementation, 138 women were tested for HIV. Of this number, two women tested positive for HIV and were subsequently referred for medical treatment. Furthermore, among 212 women that were screened for hepatitis B or C, 5 women admitted to have positive results for hepatitis C and were subsequently referred for medical treatment.

Discussion
The HEAL intervention demonstrated the efficacy of the intervention for HIV/AIDS prevention and decrease associated risk factors, including substance abuse and unprotected sex in African American women, ages 18-55, residing in Atlanta, GA. Participants demonstrated statistically significant increases in the percentage of hepatitis B and C knowledge over time. The percentage of protected oral and vaginal sex had increased after the intervention though changes did not reach statistical significance. Results also indicated statistically significant decreases in unprotected sex in exchange for money, drugs or shelter and sex under the influence of drugs or alcohol. Substance use decreased for all categories, with statistically significant decreases in the mean number in the past 30 days for the use of cigarettes, tobacco products, alcohol, marijuana, illegal drugs, and crack/cocaine.
The HEAL Intervention also facilitated HIV and hepatitis screening and referral for treatment in community outreach settings, representing a multilevel approach that targeted both the individual participants and the context within which critical supports were needed. Recent research shows that HIV testing is important in identifying undiagnosed infections and is a critical complement to interventions that increase knowledge, awareness and capacity building toward individual risk reduction [42]. Delays in diagnosis result in lost opportunities for prevention and treatment, resulting in poorer health outcomes, further exacerbating racial and ethnic HIV/AIDS disparities [43].
Beyond the positive implications of results described above, additional strengths of this study are noteworthy. First, a satisfactory response rate (48%) to a 135-item survey with very sensitive and personal questions was observed. This is likely attributable to WSCI health educators who were socially and contextually congruent to participants. More specifically, health educators had a combined 45 years of planning and implementation in gender specific programming for incarcerated, homeless and high-risk female populations. Their experience represented prior work in the capacity of a health educator or counselor for numerous social service programs designed for indigent populations including the homeless and formerly incarcerated female adults and those in substance abuse recovery. They also offered on-going case management and relationship building to assuage participants' concerns regarding disclosure of responses. Second, confidential survey administration allowed for pre-test and post-test matching by non-personal identifiers with results that were likely stronger than if anonymous survey administration had been implemented. The intervention's 6-month follow-up supported longer-term assessment of intervention effectiveness [44]. Finally, the addition of HIV/AIDS and hepatitis screening and referral represented a multilevel approach, including both individual and systems linkages toward secondary and tertiary HIV/AIDS prevention and risk reduction.
Limitations of this intervention are also acknowledged. All surveys administered represented participants' self-report which is subject to several bias including: selective memory, telescoping (i.e., recalling events that occurred at one time as if they occurred at another time), attribution (i.e., the act of attributing positive events and outcomes to one's own agency while attributing negative events and outcomes to external forces), and exaggeration [45]. Another limitation of the intervention was a post-hoc power calculation. While ideally power should be conducted prospectively, our retrospective calculation indicated that the study had a sufficient power. In addition, while the strength of 6-month follow-up is aforementioned, lower response rates were observed. Differences in the intervention sites and populations served may also contribute to the findings. This issue could have been addressed through sub-group analysis, however due to insufficient sample size sub-group analysis was not conducted. Finally, the long survey instrument (135 items) may have served as a potential barrier to completion.

Conclusions
The HEAL Intervention reflects an evidence-based community intervention that was culturally designed for African American women. Findings point to strengths of integrating HIV/AIDS, substance abuse, and hepatitis knowledge with discussion of consequential outcomes associated with risk taking behaviors. These elements were complemented by other components including life skills development, stress management, group and facilitator relationship building and an open environment. Results hold implications for the design of community-based interventions and the community-campus partnerships that assess them.
Acknowledgments: The Health, Enlightenment, Awareness and Living Intervention was funded by grant award #5 SP 13435-04 from the Substance Abuse and Mental Health Administration (SAMHSA) and through research infrastructure grants including the Centers for Disease Control and Prevention (CDC) Health Promotion and Disease Prevention Research Center (Grant #1U58DP005945-01), the National Institute on Minority Health and Health Disparities (NIMHHD) (Grant #S21MD000101) and the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR000454. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIMHHD, NCATS, NIH or the CDC. Authors wish to acknowledge the communities, partners and women who helped to facilitate a broader understanding of HIV/AIDS hepatitis, and substance abuse prevention interventions.
Author Contributions: Tabia Henry-Akintobi was responsible for guiding evaluation and research implementation and conceptualizing the manuscript foci, including biostatistical supports central to data analysis, results and discussion presented. She co-authored or edited each section of this manuscript. Jennie Trotter developed and conceptualized the HEAL Intervention and developed sections related to the description of the intervention. Tarita Johnson was responsible for management of HEAL intervention activities, was the regular interface and liaison between the programmatic, evaluation and research teams and advised development of the methodology section specifically detailing the intervention implementation process. Nastassia Laster and DeBran Jacobs were responsible for recruitment and retention activities related to evaluation and research data collection and analysis. Nastassia Laster and DeBran Jacobs were responsible for evaluation and research data collection and helped write the methods section. Assia Miller was responsible for analysis of data and synthesizing the methods and results sections. Each author helped to gather literature on the subject matter for this manuscript per their programmatic and research expertise. Each author was central to revisions of the manuscript towards final submission.

Conflicts of Interest:
The authors declare no conflict of interest.