Characteristics of psychosocial interventions to improve ART adherence in people living with HIV: A systematic review

The HIV/AIDS pandemic continues to be a significant global public health crisis. The main HIV/AIDS treatment is the antiretroviral therapy (ART), which is highly effective but depends on the patient’s adherence to be successful. However, the adherence to antiretroviral therapy remains unsatisfactory across different populations, which raises considerable difficulties at both individual and collective levels. Suboptimal adherence to ART can be overcome through multidisciplinary management that includes evidence-based psychosocial interventions. Existing reviews on these interventions have focused mainly on studies with experimental designs, overlooking valuable interventions whose evidence comes from different study designs. Here, we aimed to carry out a comprehensive review of the current research on psychosocial interventions for ART adherence and their characteristics including studies with different designs. We conducted a systematic review following PRISMA guidelines. We searched five databases (Pubmed, EBSCO, LILACS, WoS and SCIELO) for articles reporting a psychosocial intervention to improve treatment adherence for people living with HIV (adults). The quality of each study was analyzed with standardized tools, and data were summarized using a narrative synthesis method. Twenty-three articles were identified for inclusion, and they demonstrated good to fair quality. Individual counseling was the most frequent intervention, followed by SMS reminders, education, and group support. Most interventions combined different strategies and self-efficacy was the most common underlying theoretical framework. This review provides insight into the main characteristics of current psychosocial interventions designed to improve ART treatment adherence. PROSPERO number: CRD42021252449.


Introduction
Human Immunodeficiency Virus (HIV) and Acquired Immune Deficiency Syndrome (AIDS) continue to be a significant global public health crisis [1][2][3]. According  Organization (WHO) and the United Nations Joint Programme on HIV/AIDS (UNADIS), 37.7 million people were living with HIV/AIDS (PLWHA) worldwide in 2020; 1.5 million people were newly infected, and approximately 680.000 people were dying of HIV/AIDS [4]. The treatment cornerstone of PLWHA is Antiretroviral therapy (ART). The effectiveness of ART therapy has been translated into a significant decrease in morbidity and mortality [5][6][7][8]. For ART to be successful, adherence to its prescribed dose is critical [9][10][11][12]. Optimal adherence, which is often considered to correspond to 95% or more of the prescribed doses taken [13], increases the effectiveness of the treatment, allowing an undetectable viral load to be reached [14]. ART operates by inhibiting virus replication, measured by viral load (i.e., the quantity of virus in the blood). When viral load is undetectable (i.e., below the detection level of the technique), the treatment is considered successful. Reaching undetectability has two benefits: the individual, which allows the patient's immune system to recover, and the collective, which prevents the virus transmission and new infections. In contrast, suboptimal adherence to treatment favors the appearance of mutations in the virus, generating subtypes resistant to ART, which become predominant, enabling the transmission of resistant viruses in subsequent infections.
However, many patients have suboptimal adherence. A meta-analysis of 84 studies found that only 62% of patients took their prescribed doses at least 90% of the time [15]. It has also been found that adherence percentages vary between 27% and 80% across various cultural settings [16,17].
Suboptimal adherence to ART is a multi-factorial and dynamic process directly associated with more health complications, mortality, and a higher risk of HIV transmission [18].
Causes of suboptimal adherence to ART vary between and within individuals, over time, and according to how adherence is defined or measured [16]. Although aspects of regimen complexity (e.g., daily dosing frequency, pill burden, intake requirements or prohibitions associated with food, and side effects) influence adherence [19], adherence to ART is most strongly associated with psychosocial factors [19,20].
Several systematic reviews and meta-analyses have shown that the strongest predictors of treatment adherence are self-efficacy (i.e., a person's belief in their capacity to perform the behaviors needed to reach a given goal), concerns about ART effectivity, trust/satisfaction with the HIV care provider, depression symptoms, HIV stigma, and social support [19,[21][22][23][24].
For this reason, non-pharmacological interventions that focus on psychological/social factors (i.e., psychosocial interventions) have been widely promoted as one of the best options to improve treatment adherence in people with HIV [25,26].
However, previous reviews focused mainly on the intervention's outcomes and, thus, they only included studies with an experimental design, such as Randomized Controlled Trials (RCT). Nevertheless, as far as psychosocial interventions are concerned, multiple types of evidence (e.g., clinical observation, qualitative research, systematic case studies) are considered valid to nourish evidence-based practice, according to the American Psychological Association (APA).
For this reason, it might be fruitful to review the literature on interventions to improve ART adherence exhaustively. Thus, the general objective of the present study was to carry out a comprehensive review, including various research designs (i.e., randomized controlled trials, controlled intervention studies, case-control studies, before-after studies with no control group and qualitative studies) of the current research on psychosocial interventions for ART adherence and their characteristics. Specifically, in each of the selected papers, we aimed to identify the intervention's: a) type; b) frequency and duration; c) Target group; d) underlying theory; c) supporting evidence and d) setting of delivery.
Therefore, we conducted a systematic review of published literature to show the diversity of interventions that aim to improve adherence to ART in PLWHA and describe the main characteristics of these studies to enrich the current perspective on different designs.

Methods
We conducted a systematic review of the peer-reviewed academic literature. We registered the study protocol with the International Prospective Register of Systematic Reviews (PROSPERO; CRD42021252449). Additionally, the protocol and detailed methods are available in the Open Science Framework (osf.io/hctrf; 10.17605/OSF.IO/9MKRU). This review focused specifically on ART adherence. Other relevant concepts in HIV treatment were not included (e.g. retention in care, treatment adherence, self-efficacy, self-management) as this study is part of a larger project to improve ART adherence in Chile.

Eligibility criteria
We included studies published in peer-reviewed journals at any time. Searches were conducted on 1 October 2021; the review itself was completed from 1 October 2021 to 1 November 2021. We aimed to include studies that presented a psychosocial intervention aiming to improve ART adherence in people living with HIV or AIDS. We understood psychosocial interventions as interventions using a social, psychological or behavioral approach, or a combination of these two, as defined by Laurenzi et al. [36]. The primary outcome for the studies included in this review was improvement in ART adherence assessed with any type of measure of ART adherence. Studies eligible for inclusion could include male or female adults and young adults living with HIV/AIDS as participants. We also included studies whose participants' age ranges were below 18 years old if the average age was above 18 years old. As shown in Table 1, studies eligible for inclusion could include evaluations conducted as RCT, case studies, qualitative studies, pilot studies, protocols for RCT, prospective studies, pre-post studies and comparison trials. Studies with any comparison intervention (e.g., care as usual, other intervention, untreated) were included. There was no exclusion based on the primary comparator. We included all languages and geographical regions. More detailed information of inclusion exclusion criteria in Table 1.

Search method for identification of studies
Following PRISMA guidelines [37], we identified relevant studies by using a set of predetermined search terms (S1 Text) to systematically search numerous academic databases, Two reviewers (S.C.C and A.G.C) independently screened each article for eligibility by title and abstract. Cases of disagreement were discussed in a meeting between three authors (S.C.C, A.G.C and J.D) until reaching consensus.

Data extraction
Three review authors (S.C.C, A.G.C and J.D) performed data extraction independently using the same online forms to codify: (i) Information about the studies (i.e., design, location, and results), and (ii) Information about the interventions (i.e., type of intervention, description, duration and frequency, target population, underlying theory, supporting evidence and setting). After independent review, data extraction was compared and disagreement resolved by consensus.

Study quality and risk of bias assessment
Two researchers (S.C.C and A.G.C) independently conducted a quality assessment for each study using the NHBLI quality and risk of bias assessment Tool [40]. Each study was assessed with the design-pertinent scale (i.e., controlled intervention studies, case-control studies, before-after studies with no control group). In the case of RCT protocols, and pilot/feasibility studies we used the same scale as for full RCT, but we considered only the first five items, which were applicable before the implementation of the intervention. For qualitative studies, we used the quality assessment for the systematic review of qualitative evidence [41]. Disagreements between reviewers were resolved through discussion and consensus. (S1 Table).

Synthesis method
We used a narrative synthesis method [42] to summarize the extracted data. A narrative synthesis method is a textual approach to the synthesis process and is widely used in systematic reviews, especially those focusing on broad questions beyond the interventions' effectiveness [42]. Studies were grouped by intervention, population, outcomes and study design.

Study selection
Our initial search yielded 382 potentially eligible articles. After removing duplicates, the article titles were screened for inclusion/exclusion, and 237 articles were removed. The abstracts were reviewed, and eight more were excluded. The full texts of the remaining 93 were reviewed, and a further 70 were removed. As illustrated in Fig 1, a total of 23 studies met all of the inclusion criteria and were thus included in this review.

Study characteristics
The studies reflected a variety of research designs. The most common design (n = 8) was protocols of Randomized Controlled Trial (RCT) followed by RCT (n = 6), pilot and feasibility studies (n = 5), qualitative studies (n = 2), and quasi-experimental (n = 2). The majority of the studies were conducted in North America (N = 11) or sub-Saharan Africa (N = 8). Only one study was conducted in Asia and two in Europe.
Regarding the findings, the majority of the RCT and quasi-experimental studies found a significantly bigger improvement in ART adherence in the intervention group [43][44][45][46][47][48][49]. Only in one RCT did both control and intervention groups improve their ART adherence levels, but with no significant differences [50]. Most pilot/feasibility studies were acceptable and feasible.
The main points of the reviewed studies are summarized in Table 2.

Quality assessment and risk of bias
Overall, the studies demonstrated good to fair quality. S1 Table provides details regarding the methodological quality assessment and overall risk of bias within and across studies for each outcome.

RCT United States
Participants with moderated health literacy levels improved ART adherence but participants with low health literacy levels did not improve.

RCT United States
Counselling showed to be effective in enhancing ART adherence but there were no differences between groups with or without SMS.

Bogart et al. (2017) RCT United States
Larger effect on ART adherence over time in the experimental group. Importance of culturally tailored interventions.

Dulli et al. (2020)
RCT Nigeria ART adherence and retention in care improved significantly higher in the experimental group. Qualitative India Perceived informational support from peer facilitators has greater impact and credibility than informational support from medical providers.

Houston et al. (2015)
Qualitative United States Both SMS and phone calls were perceived as signs of "care" from the clinic. Two-way communication was preferred to one-way. Caution with risk of unintentional disclosure of HIV status.

Canada
ART adherence improved in both groups.
The most common features within the fair-quality studies were the lack of validated measures to assess the study's outcomes and the absence of blinding (although these criteria were not part of the assessment in the qualitative studies). The only research qualified as "poor quality" was an RCT protocol that did not report randomization techniques, blinding, and concealed treatment allocation.

Intervention characteristics
Intervention type. Of the 23 interventions, 10 were counseling; three were educational interventions; six were SMS based interventions, and two were group support interventions. Two interventions were composed of a combination of interventions (both included counseling but not as the central intervention).
The main components of the interventions are described below (see also Table 3). Intervention's main components. Counseling. Among the interventions using counseling, two integrated an educational component [44,63], one included SMS as reminders of ART taking [43], and one directly addressed alcohol consumption and stigma [52]. One counseling was exclusively focused on the pre-ART treatment period [66]. Two counseling interventions concentrated on the participant's partners; one was a couple's counseling [65], and another offered counseling for women's male partners [58]. Two counseling interventions   were peer-delivered [58,60], and two were part of a combination of other interventions [49,58]. Two counseling interventions were delivered by phone [43,52], and one was delivered by a virtual computer-based counselor [51]. SMS. Across SMS interventions, three consisted of health-related motivational messages and reminders [56, 59,61]. Of these, one had bi-directional messages (i.e., the participant had the option to ask and receive an answer), and one had, in addition, bi-directional interactive voice calls. Two included psychoeducational messages [62,64], and one included information regarding the participant's adherence levels (as measured with pillbox technology) and their peers' participants' adherence levels [53].
Educational. There were three educational interventions; two of them used a platform with a virtual nurse guiding the users through client-centered information material [48,50] and one used culturally adapted informational videos combined with counseling sessions [63].
Group support. Among group support interventions, there were two interventions delivered in a group format [46,47]. One of them [47] consisted of an 8-sessions manualized group therapy led by two trained counselors, and they were focused on risk reduction in drugs using PLWHA. The other was a social media private group composed mainly of young PLWHA and led by an adult LWHA [46].
The Information-Motivation-Behavioral (IMB) skills model was also a common underlying theory, and it was used in two counseling interventions [63,66] and one group support [47], which was also framed considering social support theories. Social support theory was also the theoretical base of a peer-delivered counseling intervention [47,60]. The Social-Ecological Model was used in two interventions: one counseling [45,58] and one facility-based intervention that included peer counseling sessions [45,58]. The health belief model was the theoretical framework of two SMS based interventions [56,64], while the other similar SMS interventions used the Nudge Theory [62]. Behavioral economics is the underlying theory in the intervention using SMS with information about the participant's levels of adherence and that of their peers [53]. Principles of Cognitive behavioral Therapy were the basis of the couples counseling intervention [65]. The Technology Acceptance Model was used in the intervention employing a computer-based counselor [51].
Finally, four of the reviewed articles did not report the underlying theory of their interventions [46,49,57,61].
Supporting evidence. We assessed the presence or absence of previous evidence for the intervention described in the reviewed studies. Previous evidence included former RCTs evaluating the same intervention in other contexts and developing and testing previous cultural adaptations. Of the total number of interventions assessed (n = 23), 18 presented supporting evidence [44-52, 54, 57, 59, 62-66], only one had provided partial evidence [43], and five did not present any evidence [53,56,58,60,61].

Discussion
According to the American Psychological Association (APA), evidence-based practice is composed of multiple types of research evidence. However, most systematic reviews of psychosocial interventions only include experimental designs in their analysis, as they are considered the highest quality evidence in medical practice. To address this gap, the present study carried out a comprehensive review of all current evidence on psychosocial interventions for ART adherence, regardless of study design. We analyzed the studies characteristics and assessed their quality to complement the information provided by systematic reviews that only include studies with experimental designs such as RCTs (e.g., [20,29,30,32]). The latter is essential since all available tools and resources are important when improving adherence to ART in PLWHA.
The first finding that stands out from our review is that despite RCTs being considered the strongest type of evidence (type 1 evidence), our quality assessment shows that diverse study designs are of good quality. Our results reveal that studies with RCT, RCT protocols, pilot/ feasibility, qualitative and quasi-experimental design scored high in quality. These findings show that a high-quality score is not something reserved for RCTs. For example, all qualitative studies scored for high-quality evidence. These studies provide relevant information for understanding in-depth adherence of ART for PLWHA. Qualitative methods applied to intervention's development allow researchers to access and reveal participants' experiences, providing essential information about what works well and how. This type of study needs to be developed to understand better what aspects of the interventions, the context, participants, and delivery are essential for adherence. Therefore, qualitative study designs should be included when analyzing the available evidence-based psychosocial interventions for a given problem. In this sense, reviewing intervention types and characteristics might be beneficial, especially in psychosocial interventions, to weigh potentially effective interventions.
Another interesting finding is that most interventions were centered on individual counseling and SMS reminders, while only two of the twenty-three studies focused on group intervention. Focusing on the individual rather than the group seems contradictory to evidence showing that support groups are the best way to help PLWHA deal with the stigma and discrimination [69]. Likewise, group treatment has been shown to provide good or excellent evidence for different mental health disorders such as depression and anxiety [70], which highly correlate with symptoms suffered by PLWHA. Similarly, although evidence shows that peer support is a substantial intervention with several advantages [71][72][73], only six studies included peer-related components (e.g., peer delivered counseling, peer group, among others). Peer support is crucial since it helps patients open up and feel understood by someone who has undergone similar experiences. In addition, implementing this kind of intervention is highly cost-effective [74][75][76].
Both group and peer-support components have been shown to boost the effectiveness of interventions [30,45,54,59,70,77]. Hence, combining these types of intervention might be something to investigate further, since participants often report that having a sense of intimacy and being in a tailor-made intervention is what makes counseling appealing [78][79][80].
Another point to consider is that studies lack a more deep consideration of the training of the facilitators. Only half of the studies mentioned that the facilitators received training; however, developing clear protocols and formal guidance might help standardize the delivery of the interventions. Likewise, evidence shows that receiving formal training encourages commitment in the intervention's facilitators [81][82][83]. Hence, this might be a feature to consider when designing interventions to improve adherence in PLWHA.
Our systematic review presents important limitations. First, we focused specifically on ART adherence. Future studies should consider adding other relevant aspects of HIV (e.g. retention in care, treatment adherence, self-efficacy, self-management). Second, we included only interventions customized for adults. Future reviews might include broader age ranges as interventions for populations such as young people living with HIV are extremely important. Third, our search strategy lacked the collaboration of a librarian, which would have helped develop a more precise and exhaustive inspection of scientific databases Overall, our study provides evidence that supports the idea that there is a diversity of interventions that aim to improve adherence to ART in PLWHA whose quality is not necessarily linked to one type of study design. This is relevant when thinking about different ways to assess effectiveness in adherence to treatment and to design studies that better represent the target population.

Conclusions
This systematic review aimed to provide a comprehensive review of the evidence on psychosocial interventions for ART adherence and their characteristics. We included diverse types of study designs, assessing the quality of the studies, identifying different kinds of interventions, target populations, the health facility in which the intervention was delivered, and the intervention's facilitator and their training.
This review emphasizes the diversity of research designs reporting and assessing interventions to improve ART adherence in PLWHA. This diversity of approaches may, in turn, enrich both clinical practice and future research programs.
This review highlighted several factors researchers should consider when planning further research into adherence to treatment in PLWHA. First, studies using qualitative methodologies should be considered since they have proven useful to deepen the understanding of participant behavior and, considering what this review shows, they scored highly in the quality assessment. Second, despite the evidence showing its effectiveness, the low number of group interventions indicates the need for additional work with this type of intervention. Third, further detail regarding the target population and the health facilities should be included when thinking about improving treatment adherence. Finally, training facilitators emerged as a common factor in these interventions; thus, training should be further investigated as a relevant factor.
While the body of work available in the literature emphasizes the problem of adherence to ART in PLWHA, it does not appear to have thoroughly examined the contribution of different types of research designs that report and assess interventions. Thus, a specific assessment of each type of study design is recommended.
Supporting information S1 Text. Search string used on PUBMED.