Family-based HIV prevention and intervention services for youth living in poverty-affected contexts: the CHAMP model of collaborative, evidence-informed programme development

Family-based interventions with children who are affected by HIV and AIDS are not well established. The Collaborative HIV Prevention and Adolescent Mental Health Program (CHAMP) represents one of the few evidence-based interventions tested in low-income contexts in the US, Caribbean and South Africa. This paper provides a description of the theoretical and empirical bases of the development and implementation of CHAMP in two of these countries, the US and South Africa. In addition, with the advent of increasing numbers of children infected with HIV surviving into adolescence and young adulthood, a CHAMP+ family-based intervention, using the founding principles of CHAMP, has been developed to mitigate the risk influences associated with being HIV positive.

Even in contexts where access to antiretroviral treatment (ART) and preventative interventions are more plentiful, such as the US or Europe, the HIV epidemic continues to take a toll on the health and wellbeing of children and adults. Th ose aff ected by this still lifethreatening and stigmatizing disease disproportionately reside in urban communities of colour, aff ected by high rates of poverty, substance abuse, and exposure to community and familial violence [12,13].
In the US, for example, the majority of HIV/AIDS cases are in large inner-city communities; African Americans comprise 51% of all newly reported HIV infections, with an additional 18% accounted for by Latinos [14]. Almost one-half of the more than 40,000 new HIV infections in the US each year are among people aged 25 years and under.
Conversely, the introduction of widespread HIV counsel ling, testing, and ART use during pregnancy and the birth process in countries with access has led to a dramatic drop in the rate of vertical transmission [15,16]. Access to ART has also meant that many HIV-infected children who were not expected to outlive their childhood are entering adolescence [17] and are presenting with: (1) serious mental health diffi culties [18,19]; (2) high-risk sexual behaviours and substance use [20][21][22]; and (c) non-adherence to ART [23][24][25]. Even brief episodes of ART non-adherence can permanently undermine treatment and lead to increased resistance to medica tions. Th us, perinatally infected adolescents may be living with a multidrug resistant virus and have poor health outcomes.
Th is grim reality becomes a serious public health issue as youth transition though adolescence, a time of increased experimentation with sexual risk behaviour and drug use. Unfortunately, few family-based programmes focused on the prevention of risk behaviour have been developed or tested with this population in high-or low-resource countries [26].

HIV prevention and intervention eff orts across the globe
Over the past three decades, there have been targeted eff orts to decrease the risk for HIV infection among uninfected youth [27,28]. Despite some of the early HIV prevention eff orts leading to improvements in youth knowledge regarding the signifi cance of HIV and modes of transmission, and short-term changes in sexual risk behaviour [28,[29][30][31], long-term behavioural change has been diffi cult to maintain [32]. Further, in a recent review of preventative interventions delivered in sub-Saharan Africa, no programme was associated with a signifi cant decrease in actual rates of HIV infection [28,33].
As the epidemic entered its second decade, there were increasing calls for more complex models of HIV prevention and intervention programming, particularly those capable of targeting both risky and protective relational and contextual infl uences on youth behaviour, such as multi-level HIV prevention and care models for youth that incorporated strong partnerships with families and communities [34]. Marshalling family, social network and community-level resources around vulnerable urban youth was thought to be a critical HIV prevention and health promotion strategy [28,31,34,35].

Case description
Although a number of family-based HIV prevention programmes have been developed and evaluated, few have actually been implemented and tested in lowresource settings where the burden of HIV exists and where the focus has been on school-based and community-based programmes targeting youth [36,3,2]. CHAMP [4,36,37,38] is one of the few HIV preventative eff orts that was initially focused on vulnerable youth and their families in the US, and then adapted for multiple international settings.
Th e fi rst family-based programme was developed in the mid-1990s based on critical streams of infl uence: (1) adolescent developmental models; (2) ecologically focused models that include multi-level factors (e.g., knowledge, skills and mental health characteristics of youth and their adult caregivers; interactional qualities with key protective resources, such as parents; social support systems; health-oriented institutions; and healthpromoting infl uences of families and communities); and (3) existing empirical fi ndings and intensive collaboration with youth, families and target community members.

Adolescent developmental models
Initially, CHAMP embraced the developmental model with two basic views: (1) for HIV prevention to be successful, programmes need to intervene with youth prior to the initiation of sexual and drug risk-taking behaviour, specifi cally in pre-and early adolescence; and (2) adolescent sexual decision making occurs within social relationships and refl ects a combination of social and psychological factors that need to be addressed [39].
More specifi cally, family and peer relationships significantly predict high-risk sexual and drug use behaviours in adolescents [40,41]. For example, family availability and monitoring are critical protective factors for reducing high-risk behaviours, while family confl ict and low levels of communication are associated with increased sexual and drug use behaviour [42,43,44,]. Also, research with youth has indicated that peers are a strong infl uence on sexual activity and the use of condoms, and friendships with peers who are not involved in problem behaviours are also protective factors for reduced sexual risk behaviour [12].

Ecological theories of youth risk
As prevention eff orts shifted from fi rst generation models, a number of more complex ecological theories were employed.
Th e Triadic Th eory of Infl uence (TTI) [45,46] is organized along two dimensions: levels of causation; and streams of infl uence. It thus represents both: (1) a theory of the problem in which the focus is on explanation and prediction of health behaviour change; and (2) a theory of action that emphasizes guiding the development of health-promoting interventions. Th ree relatively distinct streams of infl uence are proposed: intra-personal infl uences that contribute to: one's self-effi cacy regarding specifi c behaviours; interpersonal social infl uences, the social situations and/or contexts that contribute to social normative beliefs about specifi c behaviours; and culturalenvironmental infl uences, which constitute multiple socio-cultural macro-environmental factors that contribute towards attitudes about specifi c behaviours.
Th e theory proposes that some variables (such as intentions) have a direct eff ect on behaviour and are causally proximal, while others, like motivation to comply, have eff ects mediated through numerous other variables, such as social normative beliefs, and are considered to have a more distal infl uence.
Th e TTI has been translated into seven community fi eld principles to provide a conceptual framework for the adaptation of CHAMP for South African uninfected youth [47][48][49][50]. Th e seven fi eld principles included: (1) reestablishing the village (social networks); (2) providing access to health care (referral service); (3) improving bond ing, attachment and connectedness dynamics (parent ing styles and communication skills); (4) improving self-esteem (developing self-understanding and know ledge); (5) increasing social skills; (6) re-establishing the adult protective shield through monitoring (parental monitoring); and (7) minimizing residual eff ects of trauma (promoting supportive community networks).
Social Action Th eory (SAT) [51] is an alternative model of behaviour change that also emphasizes the context in which behaviour occurs, but also refers to the developmentally driven self-regulatory and social interaction processes, and the mechanisms by which these variables result in adaptive and risky health behaviours. It was developed for uninfected populations, but has been used in studies with populations infected and aff ected by HIV and multiple life stressors [52,53].
Most recently, an adapted SAT model has been used to posit that HIV prevention and care outcomes for perinatally infected youth are infl uenced by: (1) context (e.g., family and living situation, life events, service systems); (2) self-regulation processes that promote adaptive behaviours (e.g., child capabilities and motivation factors and self-effi cacy for treatment or prevention); and (3) social regulation factors (e.g., family and community support resources, caregiver supervision and involve ment, social stigma of illness) [54]. Th is model was used to inform the development of the CHAMP+ programme within both the US and South Africa.

Existing empirical evidence guiding youth-oriented HIV prevention
In addition to theoretical models, the CHAMP model of programme development also prioritizes basic research studies to inform interventions. More specifi cally, two studies -CHAMP I, a longitudinal study of 400 innercity pre-and early adolescents living in a high seroprevalence community, and Child and Adolescent Self-Awareness and Health (CASAH), a longitudinal study of 200 perinatally HIV-infected and 150 uninfected by perinatally HIV-exposed youth -were highly infl uential in informing CHAMP and CHAMP+, respectively.
CHAMP I data found that the following variables were associated with risk behaviour in uninfected youth: (1) family processes (e.g., communication, decision making, confl ict, supervision/monitoring, support); (2) outside family parental support network resources; (3) youth and family HIV/AIDS knowledge and comfort discussing sensitive issue; and (4) youth communication, social problem solving, and refusal skills. Th us, the fi ndings suggest that HIV prevention programmes targeting inner-city young adolescents need to focus on these variables in order to reduce opportunities for initiation of sexual experience and reduce risk for HIV [55].
Few HIV prevention programmes or determinant studies of behaviour exist for perinatally HIV-infected youth. CASAH was developed to identify the mental health and risk behaviour prevention needs of this population. In CASAH, high rates of psychiatric disorder were found among the predominantly African American and Latino youth living in inner-city communities, with higher rates (60%) in HIV-positive youth as compared to HIV-negative youth (47%, p=0.05). Among the HIVpositive youth, 10% had initiated sexual behaviour, with one-third of those youth reporting unprotected sex, and among those on ART, 50% reported recent non-adherence to ART. Family variables (e.g., communication, supervision, and caregiver mental health) predicted behavioural outcomes, suggesting a need to focus family-based interventions on this population of youth to improve mental health and reduce sexual risk behaviour [19,21,22,56].

Community collaborations
A critical component of CHAMP is the high level of intensive involvement of stakeholders in the design of the intervention for each community. Th us, within the CHAMP model of programme development, data from previous studies is placed in the hands of key stakeholders to inform the design of interventions that are culturally and contextually relevant and that can be suffi ciently fl exible to navigate the barriers within targeted communities. Th is process was used to develop the fi rst CHAMP intervention and for subsequent iterations, including CHAMP+ [34,57,41].
Collaborative design, delivery and testing of HIV prevention programmes has been emphasized as a means of overcoming the signifi cant obstacles to reaching vulnerable youth and their families [58]. In particular, HIV continues to be highly stigmatizing, and specifi c cultural concerns arise when health-related programmes are lead by "outsiders" that can signifi cantly impede HIV prevention eff orts [59]. As a result, community-based participatory research methodology has emerged as a critical research tool for developing and sustaining effi cacy-based interventions.
Th us, in each context, CHAMP has consistently sought out: (1) community representatives as advice and consent givers; (2) infl uential community representatives as endor sers of the research programme; (3) community members as advisors (e.g., hired as front-line staff ); and (4) community members as participants in the direction and focus of the research [4].

CHAMP and CHAMP+ results in the US
Th e CHAMP+ family-based intervention is currently delivered through multi-level group modalities, which include both multiple family sessions and parent/child group sessions. Sessions focus on: (1) parent-youth commu nication and decision making, particularly around sensitive topics and sexual possibility situations; (2) parental supervision and involvement; (3) family support; and (4) youth problem solving and negotiation skills. Th is is in addition to more traditional HIV prevention activities, including HIV knowledge.
Outcome fi ndings available to date are summarized in multiple articles, including 17 recently published [4]. In brief, signifi cant changes in parental reports of key family-level variables have consistently been associated with CHAMP participation relative to comparison families in the following domains: family decision making, with parents more likely to make decisions within the family for CHAMP participants; parental monitoring; family communication; and comfort related to family communication. Further, pre-adolescent youth have reported signifi cantly less exposure to situations of sexual possibility at post-test relative to comparison youth, and parents have reported signifi cant decreases in youth externalizing behavioural diffi culties in the programme condition relative to comparison youth.
Th e CHAMP+ intervention represents an adaptation of the CHAMP primary prevention programme to meet the needs of HIV-positive youth and their adult caregivers. Th e intervention protocol focuses on: (1) the impact of HIV on the family; (2) loss and stigma associated with HIV disease; (3) HIV, health, and antiretroviral medication protocols; (4) family communication about puberty, sexuality and HIV; (5) parental supervision and monitoring related to sexual possibility situations and sexual risk-taking behaviour; (6) helping youth manage their health and medication; and (7) social support and decision making related to disclosure.
In CHAMP+, there was a clear need communicated by the target community to address issues that are specifi c to HIV before discussion related to family processes, such as family communication and supervision and monitoring, can proceed. Th us, HIV-specifi c topics, such as coping, stigma, loss, disclosure, medication taking, health and risk behaviours, were created for use with infected populations.
Th e adaptation process resulted in: (1) signifi cant consumer involvement with regards to programme content; (2) strong sense of programme ownership from health care sites; and (3) high participation rates in CHAMP+. Post-intervention fi ndings for CHAMP+ participants relative to comparison youth and adult caregivers included: increases in child reports of caregiver supervision and monitoring of peer-based activities; decreases in selected youth depression symptoms; decreases in caregiver reports of diffi culties with youth; and improvements in HIV knowledge and communication about HIV with others. Manuscripts summarizing results are currently in preparation or under review and fi ndings have been presented at multiple national and international conferences (e.g., [60]).

CHAMP and CHAMP+ results in South Africa
South Africa adopted similar strategies to the original CHAMP and CHAMP+ in the US, namely to establish strong community and institutional partnerships so that prevention eff orts are supported by communities and institutions, and to use empirical evidence refl ecting relevant experiences of youth and families in the local setting to form the basis of the intervention. Key issues emerging from focused ethnographic studies for uninfected and infected South African youth [62,65] were used to inform the adaptation of the US-based programme for the South African context.
In particular, caregivers of uninfected youth in South Africa complained of disempowerment, which was a product of the erosion of traditional norms and social practices associated with protective parenting, as well as poor levels of HIV knowledge and information. A lack of trust and investment in community networks was also found to limit protective parenting in the target community [62]. For infected youth, similar psychosocial diffi culties to those found in US samples emerged, with loss of biological parents to AIDS being a key issue given the late roll out of ART in South Africa [65].
In keeping with other CHAMP interventions, CHAMPSA and CHAMP+SA are developed, manualized, family group interventions focusing on intrapersonal, family/interpersonal infl uences and community infl uences to strengthen family processes at each of these levels [36]. An innovation to the programmes in the South African context is the use of open-ended participatory cartoon narratives, given low literacy levels and to facilitate small group participatory experiential learning [64,66].
Th e CHAMPSA intervention results showed that, compared to controls, intervention families had signifi cantly better knowledge of AIDS transmission, had less stigmatizing attitudes towards people with HIV, and talked more and had greater comfort in talking about sensitive issues to their children, as well as increased monitoring of their children. In addition, they utilised their social networks more eff ectively in soliciting social support [38]. Community protective infl uences were also strengthened through facilitating greater informal social controls and promoting social actions to create a more health-enabling community for youth [63].
Preliminary fi ndings of the impact of CHAMP+SA suggest that families engaged with the programme reported positive experiences in helping families cope better with the diagnosis of HIV. Th ey also reported being able to better identify problems and possible solutions [66]. Analysis of follow-up data is currently underway ( Table 1).
In each context, CHAMP is implemented by three to four facilitators who co-lead the groups, allowing for separate adult and youth sub-groups for part of the sessions. Th e manualized intervention allows the use of lay facilitators, such as trained parents or lay counsellors, in most settings, with or without psychologists. In South Africa, given the shortage of mental health specialists, psychologists are utilized mainly in training and supervisory capacity in keeping with the concept of task shifting suggested for low-resourced settings [67].

Lessons learned
Th e development and implementation of CHAMP and CHAMP+ has suggested a number of important lessons for the fi eld of family-based HIV prevention and mental health treatment. Th ese include: 1. Intervention eff orts are likely to be more successful and sustainable if they are collaborative in nature and involve a community advisory board that participates in the design and delivery of the intervention. 2. Universal principles of intervention based on science can be applied across continents and diff erent contexts; yet these must be informed by local knowledge and empirical evidence to ensure cultural congruence. 3. An ecological framework within a developmental context is important in understanding complex family processes and cultural contexts, regardless of the micro-level theories used to inform specifi c behaviour change strategies within the ecological levels. 4. Family-based interventions should be group based to enhance social networking to enable the collective renegotiation of social norms regarding protective parenting practices. 5. Harnessing these social networks is important in fostering social support, which can enhance protective parenting, particularly in poor communities, as well as protective peer support networks for youth. 6. Social networks developed through group and community collaborative processes are important to build protective community environments, including re-building social controls to strengthen parental or adult supervision and care. 7. Lay facilitators can be successfully utilized to deliver the intervention with the support and supervision of mental health specialists in keeping with the move towards task shifting to increase access to mental health services in low-resourced settings.

Conclusions
Th ere is a substantive need for family-based HIV prevention and intervention programmes across the globe; yet few family-based programmes have been tested. CHAMP and CHAMP+ represent a model of family-based HIV prevention and mental health treatment that has been used across contexts (Chicago, New York, South Africa, Trinidad and Argentina) and with a range of target populations (youth in need of preventative services, HIVpositive youth, homeless youth). Further, the resulting programmes are informed by existing empirical fi ndings and data drawn directly from the target youth and/or families, as well as collaboration with key stakeholders. Th e model is based on the understanding that in order to impact youth HIV risk outcomes (attitudes, beliefs, knowledge, behaviour), interventions need to target both risk and protective factors at the level of the child, family and context.
Using this model of intervention development, the content of the intervention can be modifi ed to address the specifi c needs of youth and their families situated in unique contexts. Th e collaborative model of development enhances the chances that by co-designing, co-delivering and co-testing interventions with collaborative partners, including members of the target community, agency or medical setting, programmes and services can reach highly vulnerable youth and families that would otherwise be missed.
Further, the resulting effi cacy-based programmes can refl ect the cultural values and priorities that can be both universal and specifi c and ensure that programmes can be integrated into the settings they were developed for after the research phase.