Determinants of testing for HIV among young people in Uganda. A nested, explanatory-sequential study

Awareness of HIV serostatus helps individuals calibrate behaviour or link to care. Globally, young people (15-24years) contribute over 30% of new HIV infections. Despite progress in enhancing access to HIV services, HIV testing among young people in Uganda is below target. We determined the prevalence and factors influencing HIV testing among young people in a peri-urban district with the highest proportion of young people. We conducted a nested explanatory sequential mixed-methods study from March to May 2019 in Wakiso district. We used stratified cluster random sampling to select 397 rural and 253 urban young people from eight parishes. We collected data using questionnaires and subsequently conducted in-depth interviews with 16 purposively selected survey participants. The prevalence of testing for HIV was 80.2%. Young people related their decisions about HIV testing to self-evaluation of their risk and perceived ability to manage the consequences of a positive result. Participants reported high levels of support for HIV testing from peers, partners, and family members. They perceived health facilities as confusing, distant, expensive, and staffed by judgmental, older health workers as barriers. They felt that mobile testing points solved some of these problems, but introduced less privacy and greater confidentiality concerns. The prevalence of HIV testing among young people in Wakiso district was low compared to the UNAIDS 2030 target but among the highest in sub-Saharan Africa. Community-based programs resolve many concerns about testing at health facilities. However, there is a need to make these programs more comfortable and private.


Introduction
In 2018, approximately 38 million people were living with HIV/AIDS, with nearly 21 million living in eastern and southern Africa [1]. In that same year, about 1.7 million people became a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 stratified random sampling to select three of six parishes in Nabweru [Kazo(N = 33,424), Maganjo(N = 45,847), and Nansana (N = 52,107)] and five of ten parishes in Namayumba [Kitayita (N = 3,402), Kyampisi (N = 1,963), Kyanuuna (N = 4,960), Luguzi (N = 6,144), and Lutisi (N = 2,077)] [13,14]. From each parish, we randomly selected six villages and then 14 households from each village with an eligible young person aged 15-24 years residing in the selected household. When we found a household with more than one eligible participant, we randomly selected one participant and obtained informed consent.
From the survey participants, we purposively selected 16 participants for in-depth interviews based on their HIV testing status, age, gender, and marital status.

Data collection
We selected participants from March to May 2019. We collected data during school holidays, when young people enrolled in school typically return home. We used interviewer-administered questionnaires to collect quantitative data and interview guides to collect qualitative data. The interview guides were designed using the Capability, Opportunity, Motivation-Behavior Model (COM-B) [15]. All tools were bilingual. Interviews took place in English or Luganda (the most widely spoken language), depending on participant preference. Authors DK and HM trained ten research assistants to collect quantitative data and introduced them to local village health team members (VHTs), who guided the team to local council leaders and in the enumeration areas.
In-depth interviews were conducted in English or Luganda by two bilingual research assistants, with training and experience in conducting qualitative interviews. Interviews were conducted at the nearest public health facility for 20 to 35 minutes and were audio-recorded. Participants were interviewed about their understanding, interpretation, and experiences regarding HIV testing.

Quality control
The questionnaire was pre-tested and the research assistants who administered it were trained and supervised during data collection. Authors DK and HM reviewed all questionnaires for completeness daily before storage. Double data entry was used to ensure quality.

Analysis
Data were entered into EPI-DATA 4.4 software and then exported to STATA-15 for cleaning and analysis. Descriptive characteristics were calculated as frequencies and percentages. The proportion of those who had ever tested was calculated with its 95% confidence intervals after adjusting for clustering at the sub-county and parish levels. We used modified Poisson regression to adjust for survey sampling and sample weights and report clustered robust standard errors because the outcome was not rare (>20%). Factors with a p-value <0.2 at bivariate analysis were considered for the multivariate analysis. Statistical significance was determined at a p-value <0.05.
Recorded interviews were transcribed and reviewed repeatedly alongside recordings to ensure that the content was transcribed verbatim. Luganda transcripts were subsequently translated into English. Transcriptions were imported into Atlas. ti 8 and applied open coding to inductively generate the initial set of codes. Codes were then iteratively reviewed and revised with BS, WM, AK, and MAH. The revised codes were applied to the rest of the data. The codes were grouped into categories and themes were identified. DK synthesized the emergent themes and selected illustrative quotations for each theme.

Ethics
We obtained approval from the School of Medicine Research and Ethics Committee (SOM-REC) of Makerere University (#REC REF 2019-052). We also obtained administrative clearance from the Wakiso District Health Officer, the Nansana Municipal Health Officer, local council leaders, and facility in-charges. Participants who were aged 18-24 years, and emancipated minors (individuals < 18 years who are pregnant, married, have a child, or cater for their livelihood) as categorized by national guidelines [16] individually gave informed written consent before participation. For participants < 18 years, we sought consent and assent from the guardian and the participants, respectively, before participation. For an illiterate participant, an impartial witness was invited to witness the consent. The consent was documented by providing a signature or thumbprint on the consent form after exchanging information between the researcher and research participants during the whole research process.
Participants were assigned unique identification numbers for confidentiality purposes and these numbers were maintained throughout the study.

Participant characteristics
A total of 650 young people participated in the study. Their average age was 19(±2.6) years and the majority were female (60.9%) ( Table 1).

Prevalence of HIV testing
The prevalence of ever testing was 80.2% (95%CI: 76.9-83.1%) while the prevalence of testing in the last year was 75.0% (95%CI: 71.1-78.6%). Among those who had ever tested, selfreported HIV positive status was at 5.2% (n = 26) but 16 did not mention their status. They listed their reasons for testing and not testing ( Table 2). Ever testing for HIV was significantly different between the female (83.6%) compared to the males (74.8%) (p-value = 0.006).

Interviews
"Why should I test?" Motivations for HIV testing. Young people reported making decisions about HIV testing based on their perceived risk of HIV infection. Individuals who abstained from sex or lacked a current sexual partner, those who did not feel sick or have symptoms, and those who believed their parents' HIV status to be negative, perceived themselves to be free of risk and therefore not in need of testing. One male participant described testing for HIV as a woman's responsibility, not his. Although sexually active, he believed he had no reason to test. Finally, some young people perceived themselves to be at low risk of infection because they trusted condoms. Trusting that their lack of symptoms, demands on sexual partners to test, and/or use of condoms resulted in low risk, these participants were unmotivated to test for HIV (Table 4).
Fears related to testing. Some young people explained that they feared retesting for HIV when they perceived themselves to be at elevated risk. They felt that inconsistent use of condoms, having multiple sexual partners, and engaging in compensated sex for money put them at a higher risk of HIV. Yet when they engaged in activities they perceived as risky, their motivation to test decreased because they feared being told that they were living with HIV. Others, who perceived themselves to be safe, also said they refrained from testing because they feared the results. Young people are also worried about their ability to manage the consequences of a positive result. They perceived the consequences of a positive result to be stress, growing thin, stigma, and swallowing big tablets. Finally, participants said they feared discomfort during the blood draw and loss of confidentiality at mobile testing points in the community and in both private and public facilities (Table 4B). These concerns also affected the willingness to test for HIV. Engagement with other health services facilitates testing. Young people mentioned being offered HIV testing when they visit health facilities for other medical services, such as

PLOS GLOBAL PUBLIC HEALTH
Determinants of testing for HIV among young people in Uganda. A mixed methods study   Id = 532, male, ever tested, 17).
(Continued ) when they suspect typhoid or malaria or during antenatal care (ANC). Some mentioned being first tested during ANC. First-time testing at ANC implies that testing was not routine for these individuals. Testing for HIV was also mentioned to be offered together with other medical services when someone is sick. However, young people also described feeling disengaged from health services and alienated by health facilities. Those who described receiving free HIV testing services at the public facility also said they experienced those health facilities as confusing, distant, expensive, and staffed by judgmental older health workers. Others who refrained from testing perceived health workers at public facilities to be unwelcoming, yet those attending private facilities were perceived to be expensive. Even the indirect financial costs of accessing "free" testing services at public facilities could be high due to long distances and transport costs (Table 4C). Mixed feelings regarding mobile testing outreaches. Young people reported that HIV testing outreaches enabled them to test for HIV because they resolved the problems of confusing, distant, and expensive health facilities. Community outreaches were reported to provide free testing services closer to the communities. However, young people also noted that these outreaches also introduced a lack of privacy. Young people said they feared loss of confidentiality because the entire community attended these outreaches. Young people also mentioned being tested at school, while others mentioned not testing because they were too busy with school. One participant suggested that HIV testing should be made compulsory at school (Table 4D).
Influence from peers, partners, and family members. Participants reported that receiving support from their peers, partners, or family members regarding HIV testing facilitated their testing for HIV. Even some who have never tested described encouragement to test from those close to them. Even in the context of a supportive social environment, though, some participants said they were too busy to test (Table 4E).

Discussion
In this study, we identified factors associated with testing for HIV among young people in a peri-urban district with the highest proportion of young people in Uganda, as well as the barriers and facilitators to HIV testing these young people perceive. Eight in ten of the young people in Wakiso had tested for HIV in their lifetime. This lifetime prevalence of HIV testing is among the highest in sub-Saharan Africa and is likely due to easy access and free testing services. The high prevalence can also be attributable to the difference in time and the fact that HIV testing varies considerably across different settings likely due to interventions and testing behaviour [17][18][19]. Similar studies have found the testing prevalence to be 59.3% in Kenya [20] and only 29% in Tanzania [21], though these studies included a wider age group of 13-24 years. A similar study among young people (18-24 years) in South Africa also reported a lower  " (Id = 704, female, never tested, 19). https://doi.org/10.1371/journal.pgph.0000870.t004

PLOS GLOBAL PUBLIC HEALTH
Determinants of testing for HIV among young people in Uganda. A mixed methods study prevalence of testing (52%) [22]. Nonetheless, Uganda's testing prevalence remains below the UNAIDS target of 95% [3].
In our study, young women were more likely to test than young men. This is consistent with a large body of literature finding that women have greater healthcare-seeking behaviour than men [23]. Similar findings were reported from South Africa [22] and four other sub-Saharan countries [24]. One reason for this may be that women are likely to be offered HIV testing during ANC; in in-depth interviews, young people said they are tested when they visit health facilities for other services, including ANC [25]. HIV testing programs should emphasize antenatal attendance while discouraging home births.
Young people aged 20-24 years were more likely to test for HIV than those aged 15-19 years. Older youth may be more knowledgeable about HIV/AIDS [26], more likely to have married and tested with their partners, and more likely to have become pregnant and tested during ANC [7,22,27].
From the interviews, decisions related to HIV testing are based on self-evaluations of the risk of HIV infection and personal capacity to manage the consequences of a positive result. Indeed, participants who had married or engaged in sexual intercourse were more likely to have been tested compared to their counterparts. This may be because they perceived themselves to be at risk of HIV or had more opportunities to do so. Previous studies found that many young people test for the first time upon marriage [27] and having at least two lifetime sexual partners is associated with HIV testing [19].
Young people who lived >10km from the nearest HIV testing facility were less likely to have tested than those who lived closer to it (<5km). Although Wakiso has many health facilities, distance remains a barrier to HIV testing for some young people. Previous studies have also found that distance restricts young people from accessing health care services including HIV testing services [21,28]. In our study, qualitative interviews indicated that participants felt community outreaches resolved many barriers to testing, such as cost and distance. However, they emphasized that outreach testing introduces new barriers, such as a lack of privacy. Youths do not want it known that they have gone for HIV testing.
Surprisingly, we did not detect a relationship between rural or urban status and testing. Opportunities for HIV testing may be equally available among both rural and urban youths in Wakiso. In contrast, others have reported that youth in urban areas are more willing to test for HIV compared to youth in rural areas [29].
Young people who were encouraged by their peers to test for HIV were more likely to have tested compared to those whose peers did not encourage testing. This implies young people are motivated to test for HIV by their peers. Peers may be particularly effective for encouraging people from stigmatized populations, who mistrust healthcare providers [30]. In interviews, young people also mentioned encouragement and support from partners and family members. Engaging parents, family, and peers may improve the uptake of HIV testing among young people. Interventions to provide parents and peers with more and correct information, such as through 'straight talk' programs, seminars, drama, and the provision of information, education and communication materials should be considered.
Young people who perceived HIV testing services as youth-friendly were more likely to have tested for HIV compared to their counterparts. In interviews, young people described health facilities as confusing, distant, expensive, and staffed by judgmental older health workers. Peer health workers can motivate, reduce mistrust in healthcare providers, and encourage young people to test for HIV [30]. The mere presence of health facilities is insufficient; there is a need to ensure facilities are also youth-friendly.
This study has some limitations. First, HIV testing was measured by self-report. This may result in over-reporting of testing. Second, social desirability bias may have influenced responses. Third, the survey component of this study did not study all potential factors associated with testing for HIV, such as involvement in commercial sex and drug use. However, we were able to probe respondents during the in-depth interviews about the contributors to the decision to test or not to test. Lastly, though we collected data during the holiday season, weekends, and evening hours, we were unable to capture relatively equal proportions of young people in school and those out of school.
Our study also has several strengths. First, we studied both urban and rural settings with a sufficient sample size of 650 and 84% power to detect a meaningful difference between those aged 15-19 years and 20-24 years. Second, we sequentially employed both quantitative and qualitative methods to gain a better understanding of these factors. Lastly, both quantitative and qualitative data were collected by fellow young people as research assistants trained on the protocol and how to collect data, which may have motivated participants to freely and frankly share their experiences.

Conclusions
The prevalence of HIV testing among young people in the Wakiso district is close to the UNAIDS 2030 target. Testing more frequently is needed to meet 95-95-95 targets since many of those ever tested are not aware of their status. Married women living near a testing site, those who had peer support, and those who had ever had sexual intercourse were more likely to test for HIV. Community testing programs were preferred for health facilities. However, there is a need to make these services more comfortable and private. This could include targeted community interventions to reach more young men living far from HIV testing sites and organizing outreaches at times young people are likely to be available and at appropriate venues that make young people feel safe enough to test. Finally, many young people who had previously tested for HIV were nonetheless uncertain about their HIV status and feared retesting. Further studies should investigate HIV status awareness among young people who have previously tested for HIV.