Barriers and Facilitators to Antiretroviral Therapy in Northern Tanzania in the Era of Test and Treat

Background Tanzania adopted WHO’s universal test and treat policy in 2016, recommending antiretroviral therapy (ART) for all people living with HIV (PLHIV). However, many individuals continue to confront challenges initiating or continuing ART. The study aims to understand how barriers and facilitators to ART have evolved since Test and Treat policy in Tanzania. Between March and June 2018, 25 semi-structured in-depth interviews were conducted among PLHIV who were out of care, including PLHIV who never initiated ART and PLHIV who were formerly in care and discontinued ART. Participants were 18-years-old or older, identied from clinic databases at three health facilities and through home-based care providers. Data were coded and interpreted using content analysis and the socio-ecological framework. Results Low HIV literacy, poor quality of health messaging and comprehension, stigma, structural factors, and poverty were identied as main barriers to care. HIV literacy appeared higher among those formerly in care and some misconceptions about ART differed whether or not participants had previously initiated treatment. Participants expressed desire for more information about their health and treatment, including more engagement with their healthcare providers. Conclusions Our results are consistent with the barriers and facilitators to care reported prior to the implementation of the Test and Treat policy. Despite universal access to care, limited progress has been made in addressing the underlying challenges that PLHIV confront in initiating and continuing ART. Efforts such as increase in HIV literacy, improve health messaging, and strengthen healthcare provider-to-patient communication will facilitate access to care. The policy implication of these ndings is that the landscape for barriers to ART has not changed drastically, efforts towards engagement to care need to be further rened and tailored to facilitate access to care.


Introduction
Sub-Saharan Africa (SSA) is home to ~2  . These barriers include demographic characteristics (male, youth and adolescents), treatment-related concerns (fear of negative side effects), social environment (stigma, fear of disclosure), health systems (poor treatment from healthcare providers and clinic experiences), and poverty (Connelly 2011;Govindasamy et al. 2014; Wachira et al. 2014). Interventions targeting these barriers have used a variety of approaches to facilitate access and engagement in care such as home-based care provider (HBCs) engagement, mobile technologies, peer and family support, point-of-care testing, and the use of incentives (Franke et

Study Population and Recruitment
Participants were identi ed through clinic databases of three health facilities (regional hospital, district hospital, and health center) and through HBC outreach activities in Northern Tanzania. Individuals 18 years or older, living with HIV infection, and designated as out of care between January 2016 and May 2018, were eligible to participate. Out of care was de ned as PLHIV who never initiated ART or were formerly in care (de ned as two weeks since a missed scheduled HIV clinic visit or 90 days since last scheduled clinic visit).

Study design
This qualitative assessment was nested within a larger quantitative study (n = 75) that investigated the effectiveness of HBCs in identifying out of care PLHIV and successful linkage to ART. Included as part of the initial quantitative study was an assessment of participant's HIV health literacy, knowledge of HIV transmission, and how to stay healthy. After the quantitative survey was completed, HBCs counseled participants to return to care and then invited a subset (n = 30) to participate in in-depth interviews. These interviews were scheduled one month after the quantitative survey to allow time for participants to seek care, if they so desired.

Data Collection
A semi-structured interview guide, designed in alignment with the research question and socioecological framework domains, contained open-ended prompts to probe for detailed feedback while allowing the interviewer exibility. Two local team members (AM, KH), trained in qualitative data collection, completed the interviews; it took ~ 1.5 hours. The audio-recorded interviews were transcribed and translated from Swahili into English. Five thousand Tanzanian shillings (~$2) was provided to participants as compensation for their time.

Data Analysis
The codebook was created a priori based on the socio-ecological domains. Content analysis was conducted for each transcript and a list of codes was nalized by two researchers (OM and PN). Transcripts were then coded using NVIVO 12®, qualitative analytical software. Two researchers analyzed the transcripts (OM and RH); one coded all the transcripts while a second coder randomly selected transcripts (50%) to ensure inter-rater reliability (Kappa coe cient − 98%). Similarities and discrepancies observed during the coding process were discussed and additional codes were added as needed. The US and Tanzania research teams discussed main themes and sub-themes identi ed during the coding process for additional context and validation of ndings. Where possible, we explored whether the barriers and facilitators differed by PLHIV subgroups such as those never in care versus those formerly in care (hereafter "PLHIV groups").
The study was approved by the National Health Research Ethics Sub-Committee, National Institute for Medical Research (NIMR), Tanzania and the Committee for Protection of Human Subjects (CPHS) at University of California, Berkeley.

Results
Of the 30 invited from the larger quantitative study, 83% consented and completed interviews (60% -never in care and 40% -formerly in care). Fifty-two percent of respondents were female, 48% were male, with a median age of 34 years. All study participants completed primary education. Only 25% had re-engaged in care by the time of the interview. See Table 1 below We identi ed several barriers and facilitators to care post-Test and Treat: 1) low HIV literacy, 2) limited comprehension and quality of health messaging, 3) stigma, 4) awareness of Test and Treat, and 5) structural barriers.

Individual factors: Low HIV Literacy
Participant's knowledge of HIV diagnosis and ART medication regimen When participants were asked to discuss their HIV diagnosis or past treatment, their responses re ected low levels of HIV literacy. Comments and discussion about their diagnosis was limited; for example, most were unaware of the effects of the virus on their bodies or the importance of initiating or remaining in care. Participants who were formerly in care discussed how clinic health seminars, typically comprised of a 3day mandatory counselling session conducted prior to ART initiation or re-initiation, were a useful source of information on their HIV health. However, knowledge was generally limited to medication pick-up and reasons for ART use (e.g., to slow down the virus). Others shared dated or misinformed messages quoted from health seminars.
To provide additional context on HIV health literacy among this group of participants, we triangulated the qualitative interviews to their responses to literacy questions conducted as part of the quantitative assessment -when asked about HIV progression, CD4 count and viral load; less than 30% of qualitative study participants reported being familiar with these terms. Furthermore, when asked about HIV transmission via public toilets, needle sharing, kissing, etc., the majority of participants (~ 60%) could not accurately identify how HIV could be transmitted, independent of PLHIV group.
Among those never in care, a prevailing misconception was -it is better not to initiate ART than to initiate and then discontinue use. This was a slight distinction in health literacy between those never in care and those formerly in care. Those never in care, some expressed lack of readiness to seek care based on health seminar messaging, where they learned the importance of remaining in care and not starting and subsequently stopping ART. Furthermore, some mentioned not initiating treatment because of their "good" current state of health, failing to understand medication is needed in spite of "good health." It remains unclear how some of this messaging is delivered at the health seminars or received the communities from other PLHIV. These types of misinformation may serve as a barrier to whether or not PLHIV initiate ART or how quickly treatment is sought. Notwithstanding limited knowledge, participants expressed wanting to learn more about HIV, particularly about their care, side effects, and treatment management.
Interpersonal factors: Limited comprehension and quality of health messaging

Communication between healthcare providers and patients and health messaging
When asked about interactions with healthcare providers, those formerly in care or had initiated care expressed that there was limited discourse with their providers. The majority of participants reported that they generally did not share HIV-related questions with their providers; nonetheless, participants asked several HIV-related questions to the interviewers, suggesting that they sought information not previously received or unsatis ed with the information provided during their health visits.
When communication with healthcare providers occurred, miscommunication or misinformation were reported. Some participants believed that food was required to take their medication. Therefore, in cases of food insecurity, participants opted not to take their medication as they believed the absence of food was contraindicated for medications. Responses suggested that these misconceptions were often learned through community members and were not discussed with their healthcare providers.

Perceived Stigma
When discussing ART initiation or re-initiation in care, those formerly in care reported feeling shame when they visited the clinic or when HBCs visited their homes. Thus, to avoid seeing people they knew, a few participants mentioned making concerted efforts to visit the clinic at odd hours or travelled to clinics farther away when they were in care.

Awareness of National Test & Treat policy
When asked about the national Test and Treat policy, most participants expressed not knowing the government initiative by name but some reported knowing about their eligibility to initiate ART. Participants also reported hearing about the change in treatment guidelines through radio messaging, healthcare providers, health seminars, family, friends, or through outreach activities by other NGOs, suggesting that the core messaging of the initiative was reaching PLHIV.
In some cases, participants who knew of Test and Treat before getting diagnosed were motivated to get tested and sometimes initiated treatment, suggesting that Test and Treat messaging may facilitate initiation of care for some, despite being out of care at the time of the interview.

Health facility concerns
Those formerly in care discussed how health facility-related barriers such as long distances to the clinic, long wait times, overcrowding at the clinic, and occasional ART stock-outs in uenced their decision to discontinue care. Most emphasized that clinic visits often required spending all day at the clinic and noted they prioritized work or other responsibilities over treatment.

Socioeconomic barriers to care
Most participants discussed limited nancial means and resources as a barrier to ART engagement. Furthermore, misinformation about HIV diagnosis and treatment was prevalent; mostly received through various sources including, friends, family, HBCs, and pre-ART health seminar classes. For example, those never in care expressed it was better not to initiate treatment than to initiate and then discontinue care -a message received from the mandatory health seminar. This may be due to misunderstanding health messages or an unintended consequence of strict adherence messaging. Although some described health seminars as a valuable health source, it remains unclear if participants misunderstood the information or if misinformation occurred due to health staff's low literacy. Similar ndings were reported in a study conducted in Iringa, southern Tanzania, which assessed community HIV/AIDS knowledge and health communication. Understanding HIV/AIDS messaging varied signi cantly among participants with different levels of education and marital status (Rumisha et al. 2005). Even among those with high comprehension, poor application of HIV/AIDS messages, which could be attributed to culture, illiteracy, and/or poverty, was observed.
We found that the quality of participant-to-healthcare provider interactions were limited and participants expressed a desire to this strengthen communication. In this study, we observed that despite government's efforts, PLHIV had limited awareness of the Test and Treat policy in name. There have been widespread efforts to provide ART to all PLHIV in Tanzania; however, gaps in disseminating this national program remains, particularly for individuals in the rural areas who often have limited access to information and experience high levels of poverty. Those with access to information reported that health information campaigns, HBCs, and radio announcements were helpful in delivering information on the current recommendations for treatment. These modes of communication are common across SSA (Rumisha et al. 2005) and these ndings suggest their potential to provide targeted health messaging and outreach to PLHIV who may not be in care or at risk for discontinuing care. This study highlights the need to adapt modes of communication to reach targeted populations that may not have access to health messages delivered through popular channels.
It is important to note that these barriers revealed in this study have long been reported among PLHIV in SSA and in spite of Test and Treat. We observed that socio-structural factors such as poverty -limited access to money for food, transportation, and other basic needs further contribute to individuals not initiating or continuing in care despite universal access to ART. Distance to the clinic and long wait times at the clinics were also reported as barriers to care. These socioeconomic and structural barriers to ART observed in this study suggest that access for the most vulnerable PLHIV hinges on addressing socioeconomic barriers to care. Tanzania as the current study, found that short term cash and food assistance improved ART possession, appointment attendance, and reduced loss to follow up.
This study is one of the few studies that explored barriers and facilitators to care in Northern Tanzania post-Test and Treat implementation. The ability to triangulate qualitative results on health literacy with the quantitative assessment allowed for a more robust assessment. Notably, this study elicited perspectives from two key groups of PLHIV -PLHIV never in care and those formerly in care. However, exploring some of these questions among individuals that were in care may have allowed for key comparisons and added to the understanding of how this group differs from those who are in care. Additionally, although our assessment covered all constructs within the socio-ecological framework, we focused on resounding themes within each constructs as emphasized by the participants instead of highlighting all the barriers within each categories, as previously done in other studies.

Conclusion
Our study suggests that barriers and facilitators to care have remained relatively unchanged since the implementation of Test and Treat in Tanzania. Our results underscore the importance of health literacy and messaging for PLHIV, speci cally the need to focus efforts on improving patient understanding of diagnosis and treatment to subsequently in uence ART initiation and retention. Our ndings also highlight the need to strengthen healthcare provider-patient communication to improve engagement in care. Availability of data and materials

Declarations
The datasets analyzed in this study are available from the corresponding author on reasonable request.
Competing interests: All authors declared that they have no competing interest