In these clinics from the Cameroonian context, the majority of adolescents receiving ART are monitored in urban health facilities (~ 90%). Interestingly, one urban clinic alone (The Mother and Child Centre [MCC] of the Chantal BIYA Foundation) enrols more than half of these adolescents. Of note, this is a specialised paediatric centre for children and adolescents whereas the other urban clinic (National Social Welfare Hospital—NSWH) is specialised in both paediatric and adult populations. Thus, this accounts for the small number of adolescents found in NSWH despite their huge active file. Furthermore, there are fewer number of adolescents managed in rural health facilities, this is partly due to referral of some cases to reference centres (often based in urban settings). The overall patient-staff ratio (131/1) indicates a heavy workload in health facilities managing patients on ART, especially those clinics located in rural settings [21]. Adolescents were older in urban settings (15 vs. 12 years), and this reflects the earlier launching of ART programmes in urban settings [18, 19].
Regarding on-time drug pick-up, only one site (from the urban setting) had an acceptable performance, and rural sites experienced a significantly lower performance for patients refilling their drugs on time. These findings are similar to those obtained in adult populations in Cameroon [18, 19, 22] and in other settings [17, 23,24,25,26,27]. In the rural areas, this delay could be due to the fact that patients live at far distances from the health facilities and have difficult transportation systems [18, 19]. Of note, within the Namibian ART programme, 82% of paediatric HIV clinics which attained a desired performance for timely drug pick-up were essentially those in the reference sites than those in the outreached facilities [20]. The poor outcome recorded in our study might be due to adherence challenges during adolescence in spite of the counselling sessions. This therefore calls for additional strategies to minimize missed appointments and to improve adherence. Similarly, 80.6% of paediatrics sites in Zimbabwe also had a poor performance [24]. In this frame, non-disclosure of HIV status during childhood plays a major role as counselling sessions are held mainly with parents/guardians. Thus, setting-up adolescent peer support groups would be relevant [17]. On-time drug pick-up was better in sites with “normal workload” as compared to those with “heavy workload”, suggesting longer waiting queue that might impair effective use of healthcare services [17, 18].
Regarding retention in care, facilities located in urban settings appear to have better performance as compared to those based in rural settings. Thus, to improve retention in care, there is need to distinguish and target adolescents who are lost to follow-up from those with voluntary treatment interruption. In the Namibian study, 82% of paediatric clinics attained the desired performance for retention in care with no significant difference between reference and outreached sites [20]. Poor performance on retention in care could be attributed to far distances to health facilities and high patient-to-staff ratio (i.e. higher workload) in rural settings. Of note, the problem of staff shortages could affect the recording system, which in turn may impair adherence monitoring [17, 18]. Of note, in rural sites, it was challenging to trace the files of some adolescents who were actively receiving therapy, especially with changes in the physical locations (a case study of the Mbalmayo district hospital). Limited access to patient data in these rural settings could also be explained by the lack of electronic storage systems, suggesting a high possibility of missing data or lost to follow-up due to the limitation of the systems in place [17, 18]. A much better approach to solve this issue will be the collection and storage of patient records in a well-protected electronic device, alongside staff training and considerations for the software device maintenance.
Regarding “drug stock outs”, poor performance was recorded in both urban and rural settings. Surprisingly, drug stock outs were significantly more frequent in urban as compared to rural hospital pharmacies. This observation is similar to that obtained by Billong et al. in Cameroon [21], Mutenda et al. in Namibia [20], and Dube et al. in South Africa [22]. Inversely, 96% of 929 ART clinics in Thailand had no event of drug stock outs [23]. This disparity shows that the drug supply machinery remains challenging in SSA, and this calls for SSA countries to learn from the example of non- African RLS for procurement and stock management. As previously observed in adult populations [17,18,19,20,21,22,23,24], ARV stock out is a key contributor to HIVDR emergence among adolescents living in both urban and rural communities of SSA.
Regarding dispensing practices, all the ART clinics attained the desirable performance of 100% good dispensing practices in both urban and rural settings, suggesting also good prescription practices from clinicians as per previous reports in the same country [12, 18, 19], in other SSA countries [23, 24], and in the WHO HIV drug resistance monitoring progress report of 2018 [17]. This is essentially due to the wide use of standard triple ART regimens and continuous training of healthcare providers [17,18,19,20].
Regarding viral load coverage, a desirable outcome was reported solely in facilities based in urban settings. This can be explained by the fact that viral load monitoring is readily available in urban sites whereas there is scarcity in remote, catchment or outreached facilities [18]. Of note, the WHO 2018 HIVDR progress report revealed that only 10% (3/31) of HIVDR focus countries attained the desired threshold for viral load coverage [17]. Thus, the advent of point-of-care viral load testing should improve coverage in rural settings by leveraging on integrated diagnostic assays [28, 29]. Systemic factors could also contribute to limiting access to viral load, such as poor prescription of viral load tests by clinicians, lack of reagents and/or lack of electronic records in several RLS [17].
Regarding viral suppression, an acceptable performance was reported only in 20% of the sites, an outcome similar to that reported by the WHO (29% desirable performance) [19]. Thus, a majority of countries still face issues with viral suppression, which is usually influenced by the poor viral load coverage as aforementioned [24,25,26]. When access to viral load testing is limited, there may be preferential selection of patients with advanced clinical disease conditions for conducting viral loads, thereby leading to selection bias and lowering the overall suppression rates.
Looking at the performance of “adequate switch to second line”, only one-third of the sites reached the desirable target for switching adolescents to second line therapy based on a confirmed virologic failure on first-line ART [17]. Workload does not seem to affect this performance. It therefore appears that poor access to viral load or limited awareness on the viral load algorithm prompt switch to second-line ART is based solely on clinical and/or immunological features.
This program evaluation was conducted within the frame of the project entitled Evaluation of Treatment Response, Drug Resistance and HIV-1 Variability among Adolescents on First- and Second-Line Antiretroviral Therapy in Cameroon: The READY-Study [27]. Advanced analysis from this study on 98 sequences out of 126 adolescents experiencing virologic failure (i.e. HIV viral load ≥ 1000 copies/ml) revealed an overall prevalence of HIVDR of 95.91%, with 95.91% NNRTI-resistance, 81.63% NRTI-resistance and 3.1% PI/r-resistance [28]. These findings underscore the need for systematic resistance testing in paediatric populations after an unsuppressed viral load as previously observed by Nasuuna et al. in a similar RLS [29]. Of note, in the latter study, only 23% of children who had a viral load above 1000 copies had a viral suppression after intensified adherence counselling (IAC) [29]. Furthermore, viral suppression rates appeared to be low among ART-treated children with virologic failure despite completion of the recommended three IAC sessions. Thus, meeting the third-90 among children/adolescents warrants innovative/adapted strategies for IAC, which take into account the psychosocial context, the educational level, and even gender disparities as they grow toward adulthood [29].
In order of priority, pharmacy stock outs recorded the poorest performance (very unsatisfactory), suggesting drug stock out as the leading cause of HIVDR among adolescents. This is followed by on-time drug pick-up, viral load suppression, adequate switch to second-line ART and viral load coverage whose performances were slightly higher but unsatisfactory. This constitutes the specific package of indicators that merit timely interventions to address HIVDR emergence among adolescents. In the frame of the aforementioned interventions, alleviating the workload would contribute in improving the performance for on-time drug pickup, retention in care, and viral load coverage.
Study limitation
Data from urban facilities was statistically higher as compared to those from the rural settings. This weakens data comparability by geographical locations. As an exhaustive sampling of eligible participants was conducted, abstracted data reflect the real-life estimates of adolescents monitored on ART in urban and rural settings. Furthermore, inappropriate recording systems and incomplete patient files also limited data abstraction. Stratified data of PQIs between early (10–14 years) and advanced age (15–19 years) adolescents should be considered in future studies.