Household-focused interventions to enhance the treatment and management of HIV in low- and middle-income countries: a scoping review

The review was designed to answer two research questions: (1) What is the impact of household-focused interventions on the management of HIV in the context of HIV competence in LIMCs? We adopted the Population, Intervention, Comparison and Outcome (PICO) framework (Table 1) to determine the eligibility of the review question; (2) What are the mechanisms involved in generating the outcomes of these household-focused interventions? Our goal with this question is to understand how and why household-focused interventions would (or would not) improve the HIV competency of the households of PLHIV.

FCM and CM systematically searched six databases – Web of Science, PubMed, Medline, Psych-ARTICLES, Academic Search Complete and Cumulative Index of Nursing and Allied Health (CINAHL) – to identify relevant studies that report on interventions targeting households/families with at least one HIV-positive member to enhance primarily the adherence to ART and retention in care behaviours of PLHIV in LIMCs. We developed a generic Boolean phrase “(Famil* OR household*) [AND] (intervention* [OR] program*) [AND] HIV” to search the identified databases. In April 2018, we searched in the title and abstract or abstract only – some of the search engines did not allow for a search in title and abstract concurrently, so we selected the abstract option assuming that words in a title are most likely to appear in the abstract. There were no language restrictions in our search as our team is multilingual. Google translate was used to translate the titles and abstracts of languages that none of the team members was familiar with.

While formulating the Boolean phrase, ‘family’ and ‘household’ were considered interchangeable. Nevertheless, we have to note that these two terms are not conceptually the same [30‐32]. The use of ‘family’ as applied in implementation research has been challenged by various authors citing the lack of clarity in terms of definition and conceptualisation. This is particularly so within the African context, where ‘family’ is fluid, complex and extends both geographically and by degrees of relationship than the household [33‐38]. Considering that it is those closest to the PLHIV who are the most likely to provide the support they require, it is likely to be the people within physical proximity who are also commonly close blood relations and household members. This understanding is in line with the definition by Rudie (2005) in Niehof [39]: who defines a household as a “co-residential unit, usually family-based in some way, which takes care of resource management and primary needs of its members” (p. 490). In this article, we use the term ‘household’ as a co-residential unit, most probably family-based, targeted by HIV-related interventions. Therefore, the mention of household-focused interventions should be considered as encompassing any existing family unit.

The references obtained from each database search were imported into the Zotero® reference manager. The software was then prompted to organise the imported references alphabetically. Each author was allocated a range of alphabets to screen the titles and abstracts. For instance, the first author screened from A-D. When unsure of whether a title was appropriate, Zotero® offers the option to view the abstract, which provides the reviewer more information to inform selection. Authors were asked to highlight each title/abstract they were uncertain about. FCM and AD rescreened the highlighted titles and decided on their inclusion/exclusion. Articles that qualified for inclusion based on the title and abstract screening were downloaded and screened by four authors (FCM, AD, NS and LTD). After the full text screening was completed, all four authors met to discuss and finalise the list of articles selected for inclusion. Disagreements were resolved by a majority vote amongst the four authors and if split, CM provided the final decision.

The National Heart, Lung, and Blood Institute’s (NIH-NHLBI) Quality Assessment of Systematic Reviews and Meta-Analyses (a collection of different assessment tools based on different study designs) was used to rate the quality of the articles included in the review. Studies that adopted different research designs were assessed using the appropriate tools (Table 1) [41]. Two of the authors (AD and NS) appraised each article independently and then a third author (FCM) reviewed their judgement and reconciled their differences. The Quality Assessment tool was used to rate the quality of studies as good, fair or poor (Additional file 1). Following the NIH-NHLBI guidelines, a grade of 75% or more was considered a good evidence; a grade of 60 to 75% was considered fair evidence. Any score below 60% was considered poor evidence.

Table 2 shows the 14,596 identified references from various databases using various Boolean combinations. The 14,596 search hits including articles in all languages were imported into the Zotero® referencing software Versions 8.0. Using this software, we ran an electronic deduplication operation of the references identifying 9,489 duplicates. After the electronic deduplication, we hand-searched the remaining references and identified a further 289 duplicates: 9778 references were therefore removed through the deduplication process. The remaining 4818 references were eligible for the title and abstract screening.

Eleven articles were included for analysis (Fig. 1). The selection of studies for final inclusion was informed by the nature of the intervention (prevention vs treatment and management focused) and the location of the intervention implementation (facility, community vs homebased). Herein, we focused on proposed, piloted and implemented ART treatment and management interventions targeting the households either in part or entirely. Three of the interventions were rated as providing ‘good’ evidence, five as providing ‘fair’ evidence and the other three were unassessed as they described how the interventions were developed based on the NIH-NHLBI guidelines.

Table 3 illustrates the characteristics of the 11 articles [46‐56] obtained after the comprehensive and systematic search of the literature. We categorised these articles in relation to the type of evidence, the research approach adopted and the study design. These papers describe and/or evaluate nine interventions designed to improve aspects of support for households affected by HIV.

Eight of the designed interventions [46‐49, 53‐56] had been piloted and evaluated at a small scale while the other three articles [50‐52] described the process of developing the intervention.

Eight of the 11 articles applied a dyadic principle to select the intervention participants. Five of the studies [48, 50, 51, 55, 56] particularly focused on PLHIV younger than 18 years and their caregivers as the dyads. Four of the studies [46, 47, 53, 54] did not have age limitations for the PLHIV and enrolled any other household member aware of the HIV status of the PLHIV to complete the dyad. Two studies [49, 51] were individual-based interventions, focusing on the PLHIV in the household (Table 4).

The 11 articles included in the review described nine different intervention packages targeting a household member living with HIV and another member who is either a caregiver of the person living with HIV or a dependent of the PLHIV.

Of the nine interventions identified, two were evaluations of existing programmes focused on providing information on HIV treatment adherence, counselling and home-based care [46, 48]. The other seven interventions [46, 47, 50, 51, 53‐55] were designed to improve aspects of support for families affected by HIV. Two papers, both reporting on the ‘Family Strengthening Intervention’ [53, 54], were focused on supporting parents by encouraging strong parenting skills through facilitated discussions. Another intervention, ‘The Family Matters!’ intervention, had similar goals to the ‘Family Strengthening Interventions’, and targeted 9–12 year-olds and their caregivers to promote positive parenting practices and effective parent-child communication about sexuality and sexual risk reduction [52]. Van Rooyen and colleagues [50] designed a study to assess the feasibility of expanding a home-based HIV counselling and testing model for adults to the whole family in a family-based counselling and testing intervention. Although the primary goal was to increase the uptake of HIV testing and linkage to care [50], the study also sought to improve family cohesion through addressing intergenerational communication challenges. Puffer et al. [56] in their study described a community-based intervention for family members from different households designed to strengthen family communication through modules on economic, relationship and HIV-related topics and improving the mental health of adolescents.

Other household-focused interventions were dedicated to promoting mental health among PLHIV and their affected household members. Li et al. [47, 49] described two multi-level family-focused interventions to improve the mental health (depressive symptoms) of both the PLHIV and their family members.

Five of the interventions used more than one approach to deliver different components of the intervention [48‐52]. For improving knowledge on various aspects of HIV and adherence, counselling through lay counsellors was the primary delivery approach [46, 48, 50]. To improve family functioning and cohesion, group discussions with the PLHIV and caregivers were predominantly used [47, 49, 51]. For caregivers of PLHIV, the researchers focused on providing knowledge that would empower the caregivers on how to support the PLHIV [52]. Three interventions were delivered in the form of a facilitated discussion among the PLHIV and their caregivers or had aspects of facilitated discussion as part of the delivery approach [47, 49, 51]. Table 5 further illustrates the other characteristics of the interventions and their frequencies in the identified studies.

Four of the interventions [46, 50, 53, 54] used trained Lay Counsellors or Community Healthcare Workers for delivery. In three studies [47, 52, 55], health educators with formal bachelor’s degree qualifications were used to deliver the interventions. One study indicated the use of a bachelor level counsellor [55], and another study used a certified facilitator (formal bachelor’s degree qualification) to deliver their interventions [52]. The Puffer et al. [56] study used a community advisory committee to deliver the intervention. Nevertheless, the community advisory committee had people from different professional backgrounds, not particularly trained in delivering health care and health promotion related services.

Regarding the point of delivery of each intervention, nine of the interventions described reported that either part or the whole intervention was delivered within the home of PLHIV [46‐48, 50, 51, 53‐55]. Another four studies reported delivering at least parts of the intervention at a healthcare facility [46, 47, 49, 52]. In most instances, the health care facility was the point of recruitment of the study participants and for obtaining baseline data from the study participants. Following the recruitment and baseline information, the designed intervention is delivered at the homes of the sampled participants. Four of the studies reported that parts of or all of their interventions were delivered out-of-clinic and out of PLHIV’s household [47, 49, 52, 56]. The study conducted by Winskell et al. [52] had components of the intervention delivered at the facility and the other parts in the community. Some of the intervention aspects of the study conducted by Fatti et al. [46] were delivered at the local health facility and the other aspects were delivered at the household of the study participants. Li et al. [47, 49] described the “Together for Empowerment Activities (TEA)” intervention that had different components delivered at the health care facility, community and household.

The outcomes reported below were obtained from the articles in which a formal analysis was conducted to investigate the impact of the intervention (eight interventions and articles). The other three articles only described the process of intervention development. The outcomes of the interventions were explored with regard to our primary outcomes (retention in care and adherence to medication) and secondary outcomes. The secondary outcomes were further dichotomised to individual and family-level outcomes (Table 6).

Our findings showed that only one study reported on our primary intended outcomes (Table 6) [46]. This study reported on the eight-year outcomes of adherence to medication, retention in care and the mortality rate of PLHIV receiving a home-based adherence and psychosocial support intervention. The study showed improved long-term ART outcomes among patients receiving an integrated community/home-based care intervention in South Africa. It also showed lower chances of being lost to follow-up (adjusted risk ratio; 0.74 [95%CI: 0.66–0.84; P < .0001]) compared to those not enrolled in the intervention. For those on ART, the risk of not achieving viral suppression was 11.4% for patients using the intervention, and 19.4% among patients in standard care (adjusted risk ratio = 0.47 [95% CI: 0.26–0.86; P = .015]) [46].

In the studies conducted by Betancourt et al. [53, 54] and Li et al. [49], mental health was the primary focus, as well as being the secondary focus of two other studies [47, 56]. Whether considered a primary or secondary outcome , improvements in subjective measures of mental health were observed in four of the five studies [47, 49, 53, 54]. The interventions reported in these articles showed statistical significance in reducing depressive symptoms and the occurrence of anxiety (Table 6). The Puffer et al. [56] study evaluating a family- and church-based intervention for adolescents living with HIV, did not find a significant impact of the intervention on mental health, likely due to the low endorsement of symptoms at baseline as they did not specifically target adolescents with mental health concerns [56].

Five articles [48, 49, 53, 54, 56] reported on individual-level outcomes such as improvements in the level of acceptance of one’s HIV status, building resilience in the face of challenges such as stigma, coping with HIV infection and self-management of HIV. Another four studies [47, 53‐55] reported significantly improved perceived social support by the PLHIV. Statistically significant results were also observed with improved quality of life [57] and reduced risky behaviour [56] – substance use, violence, sexual.

Regarding household outcomes, only one study included a component relevant to the disclosure of HIV status to household members. The relationships between the intervention and disclosure to household members was statistically significant, which means the intervention improved the rate at which PLHIV disclosed to other household members [48]. Chaudhury et al. [55] reported a statistically significant reduction in intimate partner violence among caregivers when they consumed less alcohol and their findings were supported by qualitative reports of improved family functioning. Family functioning – family relationship, system maintenance, and personal growth as a family/household outcome at the family/household-level – was discussed in two articles [47, 54]. Betancourt et al. [54] and Li et al. [47] found that the household-focused interventions they evaluated significantly improved family functioning. Li et al. [47] also found this relationship was significant for the PLHIV but they found no significant change in family function for the caregivers in a stratified analysis. Puffer et al. [56] did not assess the impact of the intervention on family functioning but did show improved communication within the household.

Other intervention outcomes for the family such as safe sex practices by other family members (caregiver) significantly improved in one of the studies [56]. Betancourt et al. [54] also reported a significant improvement in the ownership of and acceptance of the disease by the caregiver, this was indicated as improvements in caregiver-reported child perseverance/self-esteem.

Interventions are theory incarnate [58], meaning that the design of any intervention carries with it an assumption of how and why the intervention is expected to work. It is postulated that for interventions to work, the interventions’ participants must engage with the opportunities, resources and restraints that these interventions provide [58]. The reasoning, interpretation and actions that the actors adopt when exposed to the intervention modalities are assumed to cause the intervention outcomes [59]. Mechanisms of action, therefore, describe these causal forces, powers, processes or interactions that generate change within an intervention—including the choices, reasoning and decisions that people make as a result of the resources provided [59]. We focused on how the information and activities provided as part of the intervention influenced changes in the reasoning and actions of the participants [45] to explain how these interventions were expected to work.

While it was straightforward to identify the theory or theories that informed most of the interventions within the literature, some of the papers were adapted from previously designed parent studies, requiring review a of the original intervention study to identify the theory/theories. Interventions modified from the same parent interventions, therefore, had the same scaffolding theory such as the ‘Together for Empowerment Activities’ interventions [47, 49] and the “Family Strengthening Intervention” [39]. The different theories that underlie the development of the associated interventions and the mechanisms provided or activated by these interventions are indicated in Table 7.

Our analysis identified three primary mechanisms ‘activated’ by the interventions. A first mechanism identified speaks to empowering the PLHIV to disclose their HIV status and adopt health-enhancing behaviours. Empowerment as an essential generative mechanism refers to a sense of personal control, mastery, and power to effect change such as maintaining adherence to medication [60]. A second mechanism by which the intervention was proposed to work was improving the perceived social support of the PLHIV through improved inter-personal relationships. Perceived social support speaks to the feeling of being supported be it emotional, physical or practical support [60]. Improving perceived self-efficacy was the third identified mechanism by which these interventions were proposed to work. Self-efficacy refers to one’s judgement of their ability and capabilities to carry out critical tasks towards achieving a particular goal. As Bandura [61] puts it, self-efficacy is the belief in one’s ability to influence events that affect one’s life and control over the way these events are experienced.