Children and adolescents account for approximately 10% of the 36.9 million people living with HIV worldwide, and about 80% come from sub-Saharan Africa.¹ Global paediatric ART coverage has increased from 44% in 2014 to 52% in 2017, but still falls short of the Start Free, Stay Free, AIDS Free framework of reaching 95% of children and adolescents living with HIV (ALHIV) with lifesaving antiretroviral therapy (ART).² Namibia is one of the 21 priority countries under the UNAIDS Start Free, Stay Free, AIDS Free framework, with targets to end AIDS among children, adolescents and young women and reach 95% treatment coverage by 2020.³ According to the study carried out by the Ministry of Health and Social Services (MoHSS)⁴ in Namibia, there are approximately 185 000 people living with HIV, of whom 9000 are children between 0 and 14 years and 4955 are adolescents between 15 and 19 years.³ The MoHSS⁵ further reported that improved access to HIV treatment and care for children has resulted in greater numbers of HIV-positive infants surviving into adolescence.³ The viral load suppression among young people aged 15–24 living with HIV is lower at 65.4% compared to the adult viral load suppression of 92.5%.³ This age group includes individuals who acquired HIV both perinatally and through other means, such as sexual transmission. Adolescents infected perinatally often face unique challenges, such as complex treatment histories, developmental issues and psychosocial factors associated with managing a lifelong illness.^6,7 On the other hand, those who acquired HIV during adolescence may encounter different barriers, including stigma, a lack of disclosure and unfamiliarity with healthcare systems.⁸ Tailoring adherence support to address these distinct challenges can help improve viral load suppression outcomes among young people living with HIV.

The MoHSS ART guideline indicates that viral suppression is higher among the 10–14-year-olds at 73% compared to 63% among the 15–19-year-olds.⁵ Data from the MoHSS⁴ show that viral load suppression among young people aged 15–24 years living with HIV is lower at 65.4% for females compared to 86% for males.² The National ART Guideline further suggests that older adolescents (15–19 years) adhere less to the medication than younger children. Evidence suggests that among those aged 15–24, many are individuals who have progressed from childhood HIV rather than newly diagnosed cases. These perinatally infected adolescents face unique adherence challenges, which can impact their understanding of the importance of adherence.^9,10 Newly diagnosed youth struggle with stigma and mental health issues, impacting their adherence to ART.¹¹ Recognising these distinctions is essential for tailoring adherence support to the specific needs of these subgroups.

Adherence to ART is even more complicated because it is very time sensitive.² Adherence to ART is critical for maintaining long-term virological suppression, and ultimately the quality of life. Maintaining high levels of adherence is a daily challenge and a dynamic process in the life of an adolescent living with HIV.² Adolescents face many challenges related to their age and the different biological, physiological, social and behavioural changes that they go through. This study has significant scientific and social value, as it addresses the pressing issue of ART adherence among adolescents, a group with notably lower viral suppression rates compared to adults. From a scientific perspective, understanding the specific adherence challenges faced by adolescents can inform targeted interventions that improve ART adherence and viral load suppression. This study also contributes to the knowledge base on HIV care in a sub-Saharan African context, where HIV prevalence is high, and where unique cultural and systemic factors impact adherence.

Socially, improving adherence among adolescents is vital for reducing HIV transmission within communities, enhancing the quality of life for young people living with HIV, and helping Namibia progress towards the UNAIDS 95-95-95 goals. The study unpacks the broader impact of non-adherence, highlighting the risk of drug resistance, continued transmission and psychosocial stress on affected adolescents. Ensuring that young people remain virally suppressed helps reduce stigma associated with HIV.¹² This study seeks to address the existing gap in knowledge concerning the specific adherence challenges faced by adolescents who have lived with HIV since childhood compared to those who are newly diagnosed during adolescence. Specifically, it aims to measure the level of adherence to ART among adolescents enrolled at the Omuthiya ART Clinic in the Oshikoto Region of Namibia, and to propose recommendations based on the findings.

This study targeted adolescents LWHIV. Initially, 125 adolescents aged 15–19 years were identified as those who met the inclusion criteria for the study. However, upon further review, only 100 adolescents actually participated. This exclusion was primarily because of two factors: some adolescents did not meet the age criteria, which included being aged 15–19 years, while others chose not to participate in the study. The study selection process was not intended to randomly exclude individuals; rather, it reflected the specific requirements set for the study, as well as the realities of participation. Ultimately, these 100 adolescents were selected as the study sample, to provide experiences of HIV-positive youth within this community.

Descriptive statistics were conducted to profile the study participants, revealing several important and statistically significant findings (see Table 1). The majority of participants (60%) were female, with a substantial proportion (73%) having completed secondary education. Notably, 81% of participants exhibited suppressed or low-level viral loads, indicating effective disease management, although 19% with high viral loads highlight a need for targeted intervention. Medication adherence varied, with 30% of those on daily doses demonstrating good adherence and 15% showing potential issues, while 9% of those on twice-daily doses exhibited significant non-adherence.

Patient-related barriers were notable; 87% managed their own medications, yet only 24% reported rarely forgetting to take their ARVs (see Table 1). Economic constraints were evident, as 46% had only two meals per day, potentially impacting overall health and adherence. Social support factors indicated that 64% lived with their parents, providing some potential assistance, but 52% felt that ARVs rarely affected their school performance. Despite potential stigma, 94% reported not using alcohol to an extent that interfered with their adherence. Support from health centres was adequate, with 54% receiving sufficient adherence information, and 93% never sent home without their medication. Most participants (73%) took only one pill per day, with 52% occasionally facing adverse reactions to their medication. In terms of adherence reminders, 45% utilised electronic reminders, and 48% relied on caregivers. Disclosure of HIV status to peers was low, with only 17% sharing this at school. Importantly, 96% did not mix ARVs with traditional herbs, indicating a commitment to adhering to their prescribed treatments.

The results in Table 2 present the p-values from a chi-square test examining the association between various explanatory variables and adherence to ART among adolescents at the Omuthiya ART Clinic. With the alpha level at 0.05, there is a significant association (p = 0.008) between the education level of adolescents and their adherence to ART, indicating that higher education levels positively impact adherence. Alcohol use has a significant impact on ART adherence (p = 0.022), suggesting that adolescents who use alcohol to the extent that it affects their ability to take ARVs are less likely to adhere to their treatment. The use of memory aids such as timers, alarm clocks or cell phones to remind adolescents to take their medication shows a significant association with adherence (p < 0.001). This indicates that memory aids are effective in improving ART adherence among adolescents. No significant associations were found between other personal characteristics, medical history, personal history, health centre assistance, stigma experiences or cultural beliefs and ART adherence in this study.

Data analysis to test which factors affect adherence was conducted using the binary logistic regression method because there were two outcomes (adherence and non-adherence). Table 3 shows the results of the analysis. The results from the binary logistic regression analysis reveal several key points about the model’s performance and the significance of variables. The model accurately predicted adherence 89.1% of the time and non-adherence 33.3% of the time, leading to an overall correct classification rate of 64%. This indicates that while the model performs reasonably well in identifying adherent individuals, it struggles to predict non-adherence effectively.

The constant term was not statistically significant (p = 0.318), meaning that it does not significantly differ from zero in predicting the outcome. Examining the specific variables, the coefficient for education level is positive (0.550), suggesting a trend towards increased odds of adherence with higher education levels. This effect is statistically significant (p = 0.001), indicating that education level has a reliable impact on adherence in this model. Similarly, the variable regarding the use of memory aids has a negative coefficient (-0.186), implying that using memory aids might be associated with lower adherence odds in those already non-adherents. This result is also statistically significant (p ≤ 0.001). Most other variables, including gender, religion, and various aspects of adherence behaviour (e.g. frequency of ARV taking, stigmatisation experiences), were not statistically significant in the binary logistic regression model at 95% level of significance.

This study aimed to identify factors associated with non-adherence to ART among adolescents at the Omuthiya ART Clinic in the Oshikoto Region of Namibia. The results highlight several key demographics, personal, and medical characteristics of the participants and their adherence behaviours. According to the national ART guideline, less than two daily dose pills count are indication of good adherence, 2–4 daily doses pills count portray fair adherence, and more than five was an indication of poor adherence. The study revealed that, most of the adolescents had notable suppressed or low viral loads, indicating effective management of their HIV. However, 19% had high viral loads and this did not include those that had defaulted from attending the clinic, necessitating targeted interventions. The majority of patients were on a first-line regimen, with fewer on second-line and third-line treatments, reflecting the typical progression of HIV treatment. A high percentage were on a DTG-based regimen, aligning with current treatment guidelines favouring DTG for its efficacy and tolerability. Viral load monitoring is an objective measurement of adherence in people on ART.

Our study found that most of the adolescents have two or more meals per day. Food availability and adherence to ART were found to be independent variables in this study. This was supported by the study performed by Nagata et al.¹⁵ Strong plausible evidence of the association between non-adherence to ART and food insecurity was found. Food insecure PLWHA were found to be twice as likely to display non-adherence to ART than their food secure counterparts. Studies in both developing and developed countries have indicated a strong association between non-adherence to ART and food security effects. For example, studies in Lake Victoria, Kenya¹⁵ and in Atlanta, USA showed that food insecure PLWHA miss doses of their daily medication.¹⁶ Similarly, studies in Jimma, southwest Ethiopia,¹⁷ and Congo¹⁸ established a strong association between non-adherence to ART and food insecurity. Because non-adherence to ART can risk developing viral resistance,^18,19 food insecurity poses a significant challenge to the HIV and/or AIDS response.^20,21,22

Chi-square tests revealed no significant association between most personal characteristics and adherence, except for education level (p = 0.008). Lack of sleep and stress because of HIV status were significantly associated with non-adherence (p = 0.022). Memory aids (timers, alarms, cell phones) were significantly associated with better adherence (p < 0.001). Stigma and cultural beliefs were not significantly associated with adherence. The findings align with several studies on adolescent ART adherence. For example, MacCarthy et al.²³ highlighted the role of psychological support in improving ART adherence among adolescents. Similarly, a study by Slogrove et al.²⁴ found that higher education levels correlate with better adherence, supporting the significant association found in this study.

However, some studies present conflicting findings. A study by Vreeman et al.²⁵ reported a stronger impact of stigma on ART adherence than observed in this study, suggesting that regional or cultural differences may influence the extent to which stigma affects adherence. In addition, while this study found no significant impact of cultural beliefs on adherence, other research, such as that by Hlophe et al.,²⁶ has shown that cultural beliefs can significantly hinder ART adherence in some communities.

The use of memory aids was significantly associated with better adherence in this study, which is consistent with findings by Chhim et al.,²⁷ who reported that electronic reminders significantly improve medication adherence among adolescents. This is similar to a study conducted by Gross et al.,²⁸ which showed that patients who use reminders to take their medications were more likely to have better adherence than those who did not use any reminders.²⁸ Patients who use memory aids were three times more likely to adhere to treatment than those who did not use any memory aids.²⁹ Among the adherent group in this study, those that use reminders to take treatment (91.7%) were more adherent than those that did not use any reminders (85.3%). This suggests that patients should be encouraged to use reminders regularly to ensure that they take medication as prescribed. However, among the non-adherent group, those that used the reminders to take treatment (8.3%), adhered less compared to those that did not use any reminder (14.7%). This suggests that other factors other than having reminders among the non-adherents are involved.²⁹

The model’s overall classification performance is moderate (64% accuracy), but it shows a strong ability to correctly identify adherent cases while failing to correctly classify non-adherent cases. The analysis suggests that education level and the use of memory aids are significant predictors of adherence. Specifically, those who use memory aids are more likely to adhere to their medication regimen. Many variables, including gender, religion, and various experiences related to stigma or treatment side effects, do not show significant relationships with adherence in this model. The results inform that there is a need to consider refining the model by including additional predictors or interactions between variables to improve the model’s ability to classify non-adherence. Explore significant predictors such as a focus on understanding how memory aids influence adherence to develop targeted interventions.

The results of the study may help the MOHSS to understand the challenges facing adolescents on ART and to tackle them in order to improve health of ALHIV. The inferences made from these data can be used to develop interventions that are relevant to the specific contexts and environments in which the patients live on a daily basis. When further developed and analysed from the patient’s perspective they can be used to inform policymakers on how quality of care can be improved and better understand the drivers of social determinants of health affecting young people.

This study was conducted at one local ART site. This limits the generalisation of the research findings to the rest of the ART clinics in Namibia. The reliance on convenience sampling may result in bias, as those participants choosing not to participate may well include patients with poor adherence. Furthermore, the study’s dependence on self-reported data via questionnaires could introduce social desirability bias, causing participants to respond in ways they believe are socially acceptable. The sample of the study was relatively small to represent all the ART adolescents in the country. Thus, in order to generalise the results from this study to other contexts, a bigger study needs to be conducted with the incorporation of other clinics in different regions of Namibia. Furthermore, the research instrument contained closed-ended questions, which meant it limited the response options for the participants. The fact that the model performs well for predicting adherence but poorly for non-adherence indicates that there might be additional factors influencing non-adherence that are not captured by the current model.

Another limitation was the lack of granular information on length of time on ART for new or perinatal exposure and the age of disclosure by care givers. These may have provided more information on factors affecting adherence. The research instrument also had a limitation in that the questionnaire contained only closed and open-ended questions, which meant limited response options, by implication. The research investigation was conducted in only one district, Oshikoto, the population and sample have limited statistic value and therefore results could not be generalised. To generalise the findings a bigger study should be conducted.

This study concludes that 24% of adolescent patients on ARV treated at the Omuthiya ART Clinic are non-adherent to treatment. It was concluded that there is no significant association between the personal characteristics of participants and adherence to ART. A significant association was found between education level and adherence; alcohol consumption and adherence; and the use of memory aids and adherence with p-values being less than 0.05. Further research is needed to explore other potential factors influencing adherence. Model limitations should be addressed by a further investigation on why the model fails to classify non-adherence accurately. It might be useful to gather more data or refine the existing predictors.