Evaluating the contribution of operation triple zero to HIV viral load suppression and retention among the adolescents in TASO Uganda using RE-AIM framework: a before and after implementation science study

Abstract

Introduction

Adolescents living with HIV (ALHIV) aged 10–19 years endure sub-optimal viral load suppression (VLS) and retention in care in many settings. We implemented operation triple zero (OTZ) in The AIDS Support Organization (TASO) Soroti and Mbale Centers of Excellence (COEs) to improve VLS and retention. Thus, this study evaluated the contribution of OTZ to improving both treatment outcomes among the ALHIV in the two COEs at one year.

Methodology

This before and after study used Reach, Effectiveness, Adoption, Implementation and Maintenance (RE-AIM) framework to guide secondary data collection from the cohort of ALHIV active in care in the April-June 2022 quarter. Effectiveness was determined by computing the overall VLS rates basing on plasma RNA copies below 1000/ml while retention was based on being active in care at the end of June 2024. A self-report was used to gather fidelity data. Univariates were summarized as frequencies and proportions, Generalized Equation Estimate (GEE) to compute the effect of the model and associated factors at 95% confidence interval and P < 0.05 level of significance. Odds ratio was used to report levels of predictability.

Results

Out of the original 533 ALHIV, 510 were considered for the post-intervention analysis, 53.1% females, mean age of 15.27 (Standard deviation = 2.15). Overall, retention at 12 months improved from 95.9 to 97.3% while VLS from 84 to 92.7% [adjusted OR 1.26 95%CI (0.61–2.61) P = 0.036]. Importantly, there was zero death in the one year of implementation. After adjusting for confounders, adolescents in the facility-based group (FBG) were more likely to be retained in care [adjusted odds ratio (aOR)7.36 95% CI (2.35–23.10) P = 0.001]. Also, multi-month dispensing [aOR 11.65 95%CI (2.93–46.34) P < 0.001] and being in FBG [aOR 9.87 95%CI (4.08–23.88) P < 0.001] and community-based models [aOR 21.96 95%CI (2.68-179.84) P = 0.004] were predictive of good VLS while poor adherence [aOR 0.02 95%CI (0.0037-0.11) P < 0.001] and being male [aOR 0.5 95%CI (0.27–0.91) p = 0.024] were predictors of poor VLS. Fidelity was good, at 80%.

Conclusions

OTZ contributed to improved VLS in the setting possibly due to fidelity of enhanced implementation of adolescent friendly health services. We encourage OTZ adaptation in similar settings to strengthen improvements in VLS.

Introduction

The World Health Organization (WHO) recommends HIV RNA viral load testing as the gold standard for monitoring response to antiretroviral therapy (ART) [1, 2] among the people living with HIV (PLHIV). HIV RNA Viral load (VL) is classified as suppressed if below 1000 RNA copies/ml and non-suppressed if above that in most programs. Achieving sustained suppressed viral load is the ultimate goal of ART and a precursor for reaching the epidemic control by 2030. In 2023, 93% of all people living with HIV (PLHIV) and on ART had suppressed their viral load across the globe, still below the 95% target by 2025. Despite this progress among the general population, only 48% of children and adolescents aged below 15 years had attained VL suppression [3]. Indeed, adolescents living with HIV aged 10–19 years continue to lag and a recent systematic review reported a VL suppression rate of 55% in sub-Saharan Africa, including Uganda [4].

In Uganda, recent findings indicate viral load suppression and retention among children and adolescents living with HIV (CALHIV) of 80% and 87% respectively [5]. Another study reported a 12-month retention among the youth aged 15–24 years old of 65.7% and Izudi reported 90% following a quality improvement intervention in one facility [6, 7]. Further, Maena et al. reported viral load suppression among the adolescents of 69%, and a more recent study found an 81% rate in Uganda’s Kabale district [8, 9]. All these indicate sub-optimal retention and viral load suppression rates of less than the expected 95%. To attain sustained VL suppression, adolescents living with HIV (ALHIV), aged 10–19 years, need to remain in care and adhere well. This requires programs to implement adolescent health friendly services such as peer driven approaches that include the community adolescent treatment support (CATS) of Zimbabwe, the Youth and Adolescent Peer Supporters (YAPS) in Uganda and the operation triple zero in Kenya [10–12]. Uganda has implemented the YAPS model to improve retention and viral load suppression among the adolescents and young people living with HIV (AYPLHIV) since 2019. The YAPS are typically individuals aged 18–24 years, overcame internal stigma, able to read and write in English [13]. In addition, other modalities such as differentiated service delivery models including community drug delivery points, treatment optimization with Dolutegravir (DTG) and multi-month dispensing (MMD) have also been implemented to improve treatment outcomes [14]. Unfortunately, both retention and VL suppression remain sub-optimal in Uganda. In The AIDS Support Organization (TASO) Soroti and Mbale centers of excellence, viral load suppression was at 84%, far below the expected 95% target, and one year retention rate of 96% among the ALHIV [15].

This sub-optimal VL suppression rate necessitated a re-imagination of the current approach to adolescent peer support services. As per the success of the operation triple zero (OTZ) [12, 16] elsewhere, it was considered for implementation in TASO Soroti and Mbale COEs. The model aims to achieve zero missed appointment, zero missed pill and zero viral load among the adolescents and young people living with HIV [11, 16–18]. This asset-based model aims at improving the health of AYPLHIV by making them active participants of creating optimal health strategies through a multidisciplinary undertaking. The model entails empowering the adolescents and young people living with HIV, health workers and caregivers using their respective packages as provided for in the national guidelines. The package for caregivers offers basic knowledge about HIV and how to groom an adolescent and young persons while that for health workers includes training on the comprehensive ministry of health package for the same population. Finally, the AYPLHIV are provided with basic information about HIV, positive living concept and other coping mechanisms [11]. The typical implementation involves quarterly caregiver meetings and monthly meetings of the AYPLHIV led by the OTZ champions which can take place both in the facility and community [12, 19].

This study thus evaluated the one-year contribution of OTZ to improving VL suppression and retention among the sub-population in the TASO setting. The Reach, Effectiveness, Adoption, Implementation and Maintenance (RE-AIM) framework of implementation science was used to guide the evaluation [20]. The findings provide further insights toward scaling up of the model in similar settings.

Methodology

Study design

This was a before and after study design. We compared both retention and viral load suppression among the ALHIV before the implementation of OTZ and twelve months after the intervention. A formative approach was employed in this implementation science, from July 2023 to June 2024. Supported by the consolidated framework for implementation research (CFIR) [21], we conducted a mixed methods study to identify barriers and facilitators to retention and viral load suppression among the adolescents living with HIV (ALHIV) in the setting. These enabled the research team to adapt the operation triple zero (OTZ) to fit the context, using the ADAPT framework/process [22–24]. The adaptation plan was collectively developed with the facility teams, and documented using framework for reporting adaptations and modifications (FRAME) [25]. Facility teams were then oriented on the implementation of the adapted OTZ model and standard operating procedures developed to guide health workers.

Study setting

The study was conducted in TASO Mbale and Soroti COEs, home to more than 13,000 PLHIV as detailed in the protocol [17]. The two COEs provide comprehensive HIV services including prevention, care and treatment interventions to children, adolescents and adults. Both COEs are east of Kampala, capital city, approximately, 224 and 324 km away, respectively. Both COE serves people living with HIV who come from at least 23 districts, a radius of 75 km of each COE.

Study population

Adolescents living with HIV aged 10–17 years of age who were active in care in the April-June 2022 quarter. These were followed up longitudinally to establish their status for retention and viral load suppression at 12 months before and after the intervention.

Sample size estimation and sampling techniques

We used the census approach to enroll all the ALHIV who were active in care during the April-June 2022 quarter for the analysis.

The intervention description

General overview

A multi-disciplinary team composing of a clinician, counsellor and peer (leveraged the YAPS) was oriented on the adapted OTZ model for one day. The orientation included a refresher on Uganda 2022 national consolidated guidelines service package for ALHIV [14], communicating with adolescents, and how to nurture adolescents. The adapted components of OTZ included monthly engagement of adolescents with non-suppressed viral load, quarterly literacy sessions with caregivers of the ALHIV with high viral load, and other clinical services as required. These sessions were differentiated to both facility and community. During the engagements, the teams offered respective packages to the two groups, empowering them to take more responsibilities in improving the health outcomes of the ALHIV. To support the health workers, standard operating procedures were developed and users mentored to use them appropriately.

Engagement with ALHIV

During the monthly engagements, each adolescent was empowered to develop a written personal health improvement plan, and this was filed for future reference. The monthly meetings were championed by the peers who led various activities including treatment literacy sessions, sharing experiences among the adolescents to build each other’s self-esteem and empowering the clients to have a positive outlook into life. The peers also conducted pre-appointment reminders, follow-up of those who missed their appointment and community-based adherence support. Further, adolescents were also provided with financial literacy sessions, and in collaboration with Mbale Network of Young People Living with HIV/AIDS (MNYPA), trained 21 out of school adolescents with basic skills such as liquid soap making. To strengthen the implementation further, bi-weekly check-in by the research team and monthly by the technical team from International Pediatric HIV/AIDS Symposium in Africa (IPHASA) to offer additional guidance were conducted. Lastly, continuous quality improvement was also integrated into the intervention as well to ensure adequacy of the implementation.

The caregiver engagement

There were two caregiver meetings. The caregivers of ALHIV with unsuppressed viral load were gathered in one place (both facility and community) and provided with basic information about HIV including its transmission and management; adherence to ART, disclosing HIV status to an adolescent, stigma and discrimination, sexual and reproductive health issues and how to empower an adolescent to lead an independent life.

Data management

Data collection

Evaluation framework

We used the RE-AIM framework to determine the variables to consider. RE-AIM is a popular implementation science framework that helps standardize program evaluation data collection and it states that interventions are impactful if effective evidence-based intervention reaches a wide proportion of the public, through adoption by willing entities, implemented as intended and maintained over time [20, 26–28]. The model has been used elsewhere to evaluate programs such as integration of methadone into ART in Tanzania, economic empowerment interventions as HIV prevention programs among adolescent girls and young women in sub-Saharan Africa, and the continuum of care management in Brazil [27, 29, 30]. In this study, we used Reach, Effectiveness and Implementation dimensions to guide the evaluation of OTZ program. The three dimensions were best suited to guide data collection required to answer the research questions and the others (adoption and maintenance) were irrelevant to the current study, hence not considered. Table 1 summarizes the dimensions of RE-AIM used to guide measurement of the primary outcomes.

Table 1.

Dimensions of RE-AIM used to measure primary outcomes

Dimension	Description
Reach	Proportion of adolescents living with HIV who were enrolled into the model
Effectiveness	Proportion of individuals with HIV viral RNA copies of less than 1000/ml.
	Retention: Proportion of adolescents who were active in care by the end of June 2024. This was categorized into active (if ALHIV had their most recent clinic encounter is at least 28 days within the next scheduled appointment), transferred out for those who moved to other facilities, dead for those who died and lost for those who had no clinic contact for at least 28 days their last scheduled appointment.
Implementation	Fidelity (extent of the implementation of the adapted model). Fidelity is the extent to which an intervention is implemented as planned [25]. We used self report and observation to collect data on fidelity, basing on the adaptation plan developed at the start of the implementation of OTZ.

We used the same questionnaire as at baseline to collect post-implementation data as well. A virtual re-orientation of the data collection team was done prior. Data collection was done using Kobo collect, version 2.024.19. The variables included: Age, viral load status, ART status, HIV status disclosure status, school going status, differentiated service delivery model, presence of caregiver and Tuberculosis (TB) status.

For fidelity, self-report was used to collect data on some of the key implementation components basing on the adaptation plan. Three dimensions of adherence, dosage and quality of the implementation were considered. Each core component of OTZ was assigned one point except monthly OTZ meetings which received two marks. Arbitrarily, a score of below 50% was considered poor; 50–79%, fair and 80%+, good fidelity. Given that the two sites are generally similar and thus variation in implementation was negligible, overall fidelity for the two sites was considered.

Data analysis

Univariate analysis was reported as frequencies and proportions. The Generalized Estimating Equations (GEE) approach was used to build the multivariate model, a method suitable for analysing clustered data where observations are not independent, such as in longitudinal or group-based studies. The multivariate model was constructed in stages, starting with the bivariate analysis to identify factors that were statistically significant at the 0.2 level. A higher level was set in this initial phase to limit chances of excluding potentially significant factors early on. In this initial stage, variables that were found to be strongly associated with the outcomes of viral load suppression and retention in care were selected for inclusion in the multivariate model.

The inclusion criteria for the multivariable model were based on the significance of the variables in the bivariate analysis, as well as their theoretical relevance based on previous research. Variables with non-significant associations were excluded from the multivariate model to reduce model complexity and prevent overfitting. The final multivariate model was thus designed to examine the independent associations of adherence, gender, DSDM approach, caregiver presence, school-going status, and the after-intervention variable on viral load suppression and retention in care. Level of significance was determined using p < 0.05 at 95% confidence interval (CI). Odds ratio was then preferred for reporting magnitude and direction of predictability of a variable to influence retention and viral load suppression. We used Standard for reporting Implementation Studies (StaRI) guidelines [31] for reporting the work.

Results

Overall (Tables 2A and 2B), from the original 533, at post intervention, there were 510 ALHIV considered for the analysis, including 53.1% females, mean age of 16.27 years, interquartile range of 12–19 years. Overall, retention rate improved from 95.9 to 97.3%. importantly, there was zero death during the intervention period compared to four in the twelve months prior. The good retention could be attributed to enhanced multi-month dispensing of ART with 64.2% of the clients, receiving drugs for at least six months. Further, viral load suppression rate improved from 84 to 92.7%. The improvement could be attributed to a web of factors such as improved disclosure of HIV status (99.8%), improved adherence rate (99%) and multi-month dispensing of drugs. Overall, 134 (26%) ALHIV directly benefitted from OTZ intervention.

Table 2A.

Basic demographic characteristics of the ALHIV included in the study in the before and after study periods

Variables	BEFORE		AFTER
	Frequency (N = 533)	Proportion (percentage)	Frequency (N = 510)	Proportion (percentage)
Study centres
TASO Mbale CoE	304	57.0	287	56.3
TASO Soroti CoE	229	43.0	223	43.7
Current age
12–14 Years	199	37.3	122	23.9
15–19 Years	334	62.7	388	76.1
Sex
Female	289	54.2	271	53.1
Male	244	45.8	239	46.9
School going status
Not at school	165	31.0	77	15.1
At school	368	69.0	433	84.9
Caregiver present
No	8	1.5	2	0.4
Yes	525	98.5	508	99.6

Table 2B.

Clinical characteristics of the ALHIV included in the study in the before and after intervention periods

Variables	BEFORE		AFTER
	Frequency (N = 533)	Proportion (percentage)	Frequency (N = 510)	Proportion (percentage)
Current retention at 12 months
Active	511	95.9	496	97.3
Died	4	0.8	0	0.0
Transferred Out	8	1.5	13	2.5
Lost > 28days	10	1.9	1	0.2
Current viral load suppression (12 Months most recent)
Non-suppressed	87	16.3	35	7.1
Suppressed	446	83.7	459	92.7
No valid VL	0	0	1	0.2
Adherence scores
Good	494	92.7	505	99.0
Fair	27	5.1	0	0.0
Poor	12	2.3	5	1.0
Current Differentiated service delivery model (DSDM) approach
Community Client-led ART Delivery (CCLAD)	25	4.7	82	16.1
Community Drug Delivery Point (CDDP)	106	19.9	13	2.5
Facility-based groups (FBG)	354	66.4	370	72.5
Facility-based individual model (FBIM)	40	7.5	35	6.9
Fast-track drug refill (FTDR)	8	1.5	10	2.0
Tuberculosis (TB) status
No signs and symptoms	498	93.4	501	98.2
Presumptive	28	5.3	4	0.8
TB diagnosed	7	1.3	5	1.0
Orphaned and vulnerable children (OVC) status
Ever enrolled	323	60.6	389	76.3
Never enrolled	210	39.4	121	23.7
Benefited from OVC services
No	62	19.2	8	2.1
Yes	261	80.8	381	97.9
Multi-month dispensing (MMD)
< 3 months	51	10	3	0.6
3 to 5 months	261	51	174	35.2
6 + months	199	39	318	64.2
Disclosure status
Disclosed	522	97.9	509	99.8
Not disclosed	11	2.1	01	0.2

Viral load suppression

Participants at TASO Soroti CoE had significantly lower odds of achieving viral load suppression compared to those at TASO Mbale CoE (OR = 0.637, P = 0.027, 95%CI: 0.427–0.951) as indicated in Table 3. Overall, ALHIV at post-intervention were 2.4 times more likely to have suppressed viral loads (OR = 2.421, P < 0.001, 95%CI: 1.588–3.693), indicating the intervention’s positive contribution. Adherence was strongly linked to viral load suppression. Participants with fair adherence (OR = 0.014, P< 0.001, 95%CI: 0.004–0.049) and poor adherence (OR = 0.027, P < 0.001, 95%CI: 0.007–0.102) had much lower odds of suppression compared to those with good adherence. After intervention (OR = 1.519, P = 0.076, 95%CI: 0.958–2.411) there was a moderate increase in odds of suppression, though not statistically significant at the P=0.05 level.

Table 3.

Bivariate analysis

Variables	Viral load suppression			Retention
	Crude OR	P-value	Confidence interval	Crude OR	P-value	Confidence interval
Study centres
TASO Mbale CoE	1
TASO Soroti CoE	0.637	0.027*	0.427–0.951	1.448	0.311	0.707–2.964
After intervention	2.421	< 0.001*	1.588–3.693	1.453	0.292	0.725–2.915
Adherence scores
Good	1			1
Fair	0.014	< 0.001*	0.004–0.049	1	-	-
Poor	0.027	< 0.001*	0.007–0.102	0.276	0.106	0.058–1.313
After intervention	1.519	0.076	0.958–2.411	1.483	0.275	0.732–3.003
Current age
12–14 Years	1			1
15–19 Years	0.628	0.052	0.393–1.004	1.764	0.112	0.876–3.551
After intervention	2.57	< 0.001*	1.673–3.950	1.34	0.415	0.663–2.710
Sex
Female	1			1
Male	0.525	0.002*	0.350–0.788	0.613	0.167	0.306–1.227
After intervention	2.455	< 0.001*	1.607–3.748	1.45	0.295	0.724–2.907
School going status
Not at school	1			1
At school	1.269	0.285	0.820–1.965	0.27	0.04	0.077–0.944
After intervention	2.318	< 0.001*	1.513–3.552	1.696382	0.154	0.820 - 3.509
Caregiver present
No				1
Yes	1	Omitted		3.37	0.256	0.415–27.390
After intervention	2.45	< 0.001*	1.606–3.737	1.422	0.32	0.710–2.848
Current Differentiated Service Delivery Model (DSDM) approach
Facility-based individual management (FBIM)	1			1
Facility-based groups (FBG)	10.977	< 0.001*	6.248–19.284	3.426	0.011	1.319–8.897
Fast track drug refills (FTDR)	6.092	0.029*	1.200–30.917	0.771	0.762	0.143–4.165
Community client-led ART Delivery (CCLAD)	9.738	< 0.001*	3.699–25.640	7.359	0.066	0.879–61.602
Community drug delivery point (CDDP)	4.658	< 0.001*	2.314–9.377	1.408	0.574	0.427–4.636
After intervention	2.24	0.001*	1.417–3.540	1.18	0.67	0.552–2.524
Tuberculosis (TB) status
No signs and symptoms	1			1
Presumptive	1.140709	0.835	0.330–3.94	1.003	0.998	0.132–7.634
TB diagnosed	1.300565	0.802	0.167–10.120	0.344	0.327	0.041–2.910
After intervention	2.526997	< 0.001*	1.667–3.831	1.447	0.302	0.717–2.922
Orphaned and vulnerable children (OVC) status
Ever enrolled	1			1
Never enrolled	0.982	0.933	0.646–1.493	0.772	0.47	0.383–1.556
After intervention	2.418	< 0.001*	1.587–3.683	1.394	0.347	0.699–2.780
Benefited from OVC services
No	1			1
Yes	0.982	0.933	0.646–1.493	0.391	0.369	0.050–3.030
After intervention	2.418	< 0.001*	1.587–3.683	1.291	0.57	0.535–3.115
Multi-month dispensing of drugs (MMD)
< 3 months
3 to 5 months	4.159	< 0.001*	2.215–7.811
More than 5 months	9.537	< 0.001*	4.701–19.347
After intervention	1.598	0.046*	1.008–2.532
Disclosure status
Not disclosed	1			1
Disclosed	1	Omitted		1.498	0.539	0.740–3.034
After intervention	2.349	< 0.001*	1.539–3.585	2.121	0.324	0.477–9.430

Males had lower odds of attaining viral load suppression than females (OR = 0.525, P = 0.002, 95%CI: 0.350–0.788), suggesting gender-based disparities in suppression rates. Following the intervention, males showed a substantial improvement, with the odds of suppression increasing by 2.5 times (OR = 2.455, P < 0.001, 95%CI: 1.607–3.748). School-going status was not significantly associated with viral load suppression (OR = 1.269, P = 0.285, 95%CI: 0.820–1.965). However, after the intervention, participants who were at school had significantly higher odds of suppression (OR = 2.318, P < 0.001, 95%CI: 1.513–3.552).

The presence of a caregiver was significantly associated with increased odds of suppression (OR = 2.45, P < 0.001, 95%CI: 1.606–3.737). This highlights the critical role of caregiver support in facilitating viral load suppression. The DSDM approach was strongly associated with suppression, particularly in models such as FBG (OR = 10.977, P < 0.001, 95%CI: 6.248–19.284), FTDR (OR = 6.092, P = 0.029, 95%CI: 1.200–30.917), and CCLAD (OR = 9.738, P < 0.001, 95%CI: 3.699–25.640). The after-intervention variable showed an overall positive effect on suppression (OR = 2.24, P = 0.001, 95%CI: 1.417–3.540). Enrolling in OVC services and benefiting from OVC services were strongly associated with increased viral load suppression after the intervention (OR = 2.418, P < 0.001, 95%CI: 1.587–3.683). MMD was significantly associated with viral load suppression, with participants on 3–5 months and more than 5 months MMD having higher odds of suppression (OR = 4.159, P < 0.001, 95%CI: 2.215–7.811 and OR = 9.537, P < 0.001, 95%CI: 4.701–19.347, respectively). After the intervention, participants on MMD still had significantly improved odds of suppression (OR = 1.598, P = 0.046, 95%CI: 1.008–2.532).

The after-intervention variable had a significant impact on suppression for participants who had not disclosed their HIV status (OR = 2.349, P < 0.001, 95%CI: 1.539–3.585), suggesting the intervention’s positive effect on those who had not disclosed their status.

Retention

Results show that adolescents who are not in school [a0R 0.27 95%CI (0.077–0.944) P = 0.04] were more likely to remain in care, as school attendance appears to reduce retention odds. This suggests that for adolescents not attending school, there may be fewer competing priorities or barriers to engagement in care. Another significant factor was the DSDM approach, particularly the facility-based Group approach (FBG), which greatly improves retention outcomes. Adolescents receiving care through FBG had more than three times the odds of remaining in care compared to those in the FBIM approach. This approach seems to be a strong model for encouraging retention.

Multivariate modeling

Viral load suppression

The multivariable analysis in Table 4 indicates that Adherence remained one of the most significant predictors of viral load suppression. Participants with poor adherence [adjusted odds ratios [(aOR)0.03 95% CI (0.0038–0.23) P = 0.001] and fair adherence had [aOR = 0.02 95%CI (0.0037–0.11) P < 0.001)] were less likely to attain VLS. These findings underscore the critical role of adherence in ensuring effective HIV treatment. Regarding gender, males [aOR 0.5 95%CI (0.27–0.91) P = 0.024] were less likely than females to suppress their VL. Targeted interventions are required to support the males. For DSDM, participants in FBG [aOR = 9.87, 95%CI (4.08–23.88) P < 0.001], CCLAD [aOR 21.96 95%CI (2.68–179.84) and CDDPs [aOR 95%CI (3.80–70.14) P < 0.001] were more likely to achieve VLS compared to other models.

Table 4.

Multivariable analysis

Variables	Viral load suppression			Retention
	Adjusted OR	P-value	Confidence interval	Adjusted OR	P-value	Confidence interval
Study centres
TASO Mbale CoE	1	-	-	1	-	-
TASO Soroti CoE	1.46	0.411	0.59–3.61	1.26	0.703	0.38–4.15
Adherence scores
Good	1	-	-	1	-	-
Fair	0.02	< 0.001	0.0037–0.11	-	-	-
Poor	0.03	0.001*	0.0038–0.23	0.09	0.010*	0.02–0.57
Sex
Female	1	-	-	1	-	-
Male	0.5	0.024*	0.27–0.91	1.06	0.907	0.43–2.59
Current DSDM approach
Facility-based individual management (FBIM)	1	-	-	1	-	-
Facility-based groups (FBG)	9.87	< 0.001	4.08–23.88	7.36	0.001*	2.35–23.10
Fast-track drug refills (FTDR)	1	(omitted)	(omitted)	0.78	0.834	0.08–7.70
Community client-led ART Delivery (CCLAD)	21.96	0.004*	2.68–179.84	1	(omitted)	(omitted)
Community drug delivery point (CDDP)	16.34	< 0.001*	3.80–70.14	3.25	0.182	0.58–18.30
Benefited from orphaned and vulnerable children (OVC) services
No	1	-	-	1	-	-
Yes	0.29	0.11	0.06–1.33	0.61	0.644	0.08–4.85
Multi-month drug dispensing (MMD)
< 3 months	1
3 to 5 months	5.46	0.005*	1.66–17.93
More than 5 months	11.65	< 0.001*	2.93–46.34
Period
Before intervention	1
After intervention	1.26	0.036*	0.61–2.61	1.033836	0.949	0.375 2.853

*(P < 0.05)

As for MMD, individuals who received drugs for 3–5 months [aOR 5.46 (1.66–17.93) P = 0.005] and 6 + months [aOR 11.65 (2.93–46.34) P < 0.001] during their last clinic visit had better odds of achieving VLS. MMD likely reduces chances of treatment interruption, encourages drug availability leading to good adherence. Lastly, the multivariable model found that participants in the post-intervention group [aOR 1.26 (0.61–2.61) P = 0.036] had better chances of achieving viral suppression, compared to the pre-intervention period. However, given that CI appears to overlap 1, the intervention may not have strongly impacted the outcome in practical terms.

Retention in care

Adolescents receiving care through FBG [aOR 7.36 95%CI (2.35–23.10) P = 0.001] had higher likelihood for remaining in care, compared to other models, likely due to improved adolescent friendly services.

Regarding fidelity, Table 5 provides the overall score. For the components measured, the total score was 80%, considered good fidelity. This indicates that the core components of the model were well implemented.

Table 5.

Check list for fidelity scores

Dimension	Strategy	Status (Y/N)	Rating (/10)
Adherence	Kenyan team engaged	N	0
	Caregiver workshops conducted	Y	1
	Peers and health workers oriented/trained on OTZ	Y	1
	Monthly OTZ meetings conducted	Y	2
Dosage	Caregiver meeting held for at least 2 times	Y	1
	Monthly OTZ meetings held for at least six times	N	0
Quality of intervention delivery	Continuous quality improvement projects implemented	Y	1
	Project Monitoring and evaluation done	Y	1
	Support supervision done	Y	1
Total score			8

Discussion

Our study adapted and evaluated the contribution of operation triple zero (OTZ) to improving both retention in care and HIV viral load suppression rate among the adolescents aged 10–19 years old. After one year post implementation, viral load suppression improved from 84 to 92.7% while retention also slightly improved from 95.9 to 97.3%. The improvement in VLS rhymes well with findings from Nigeria (92%) and Ethiopia (92%) [16, 32]. Overall, these achievements were better than previous reports, perhaps because of the good fidelity of OTZ implementation. For VLS, the 92.7% was better than the 69% and 80% respectively reported in Mbale, eastern Uganda, and 81% in Kabale, southwestern Uganda [5, 8, 9]. On the other hand, the retention rate was much higher than the 65.7%, 87% and 90% reported previously in Uganda [5, 6, 33]. VLS typically requires provision of optimal ART and adherence [34] both of which appear to have improved during the OTZ implementation, similar to what has been reported elsewhere [35]. Uganda implements a test and treat all policy with DTG-based regimens as the preferred anchor drug, a highly effective ART, widely available across all HIV clinics in the country [14].

The improvement in VLS is attributable to the intervention which essentially led to good adherence. Poor adherence had been identified earlier by health workers, caregivers and adolescents themselves as a key barrier to viral load suppression [15]. Fortunately, achieving optimal adherence is one of the principle objectives of the OTZ, through empowering of both caregivers and the ALHIV [19] in order to attain zero missed pills. The monthly and quarterly engagements of ALHIV and their caregivers respectively, were aimed at improving adherence through counselling, sharing experiences, designing personal improvement plans and addressing the question of social support systems. Also, there were additional deliberate efforts to improve adherence such as regular phone calls to offer virtual adherence support, greater involvement of the peers, supporting disclosure of HIV status to the adolescents, celebrating those who re-suppressed and home-based counselling in collaboration with the OVC team. These interventions were guided by the identified barriers and facilitators to VLS [15]. Improved adherence through enhanced counselling has been found to improve viral load suppression rates among PLHIV by up to 82% [36]. Therefore, OTZ likely enabled an interplay of various adolescent friendly health services provided, which greatly improved the positive outlook to life and treatment literacy among both caregivers and the adolescents.

Relatedly, ALHIV who received their care through facility-based groups were more likely to suppress their viral load and remain in care. Differentiated service delivery is a person-centered approach that enables PLHIV to receive care from their preferred models [1, 37, 38]. It has been promoted as a key intervention for improving patient experience and onward treatment outcomes including viral load suppression [39–41]. Facility-based groups (FBG) include generally stable individuals who receive services as cohorts including children, adolescents, pregnant and breastfeeding women [14]. During the intervention period, ALHIV in the FBG received continued adherence support through peer-driven counselling, experience sharing with individuals who had suppressed their viral loads declaring their initiatives to attain sustained VLS status and also, caregiver literacy through health education. Indeed, social support groups have been reported to contribute to improved treatment outcomes [42, 43].

Further, ALHIV who received multi-month dispensing (MMD) of drugs had significantly better odds of attaining VLS during the intervention, similar to findings by Ainembabazi et al. [5]. Receiving drugs for at least three months enables PLHIV to adhere well for long because of sustained availability of the pills. However, ongoing counselling is important to strengthen adherence while clients are away from the clinics for long in order to sustain viral load suppression [5]. In this study, only 0.6% had received drug refills for less than three months during their last clinic visit. With the odds of up to 11.5 to achieve viral load suppression, this finding should therefore encourage health workers who are currently hesitant to implement MMD among ALHIV, to confidently do so. As reported elsewhere, MMD reduces work/study interruptions, saves people’s time, reduces workload on the health workers, improves adherence and a good antidote to stigma and discrimination [44].

Whereas association of gender and viral load suppression has been found to be mixed, our study found males were less likely to attain a suppressed VL compared to the females, consistent with findings from Uganda, South Africa, Zimbabwe and Kenya [9, 34, 45, 46]. This could be due to negative peer influence, reluctance of males to seek and/or utilize healthcare services [9, 34], gross internal stigma, inadequate social support and absence of male-friendly services in the facilities [45]. There is need to identify and address male specific issues in order to improve treatment outcomes by providing targeted interventions.

Finally, we acknowledge important study strengths such as: using routine program data likely reflects realities on the ground, use of RE-AIM offered a standardized way of conducting the evaluation, and a good study design that enabled us to compare the pre-and post-intervention periods. Conversely, study limitations included: implementation in COEs only, limiting potential applicability in public health sites. Whereas the setting maybe different, the same ministry of health package is implemented across all facilities in Uganda. Thus, the findings may yet be applicable across board despite the few variations. Also, only few ALHIV directly benefited from the model, limiting its potential wider effect on treatment outcomes. Despite this, all the adolescents benefitted from the intervention secondarily through improved skills from the health workers which led to enhanced implementation of adolescent friendly services including HIV status disclosure, differentiated service delivery and multi-month dispensing of drugs. Suffice to mention that while there were no policy changes during the implementation period, other contributing enhancers such as quality improvement and OVC models strengthened the intervention. Important to note that the intervention improved the implementation of the OVC in the settings.

Conclusions

Overall, OTZ improved retention and VLS in the setting due to implementation of adolescent friendly health services such as MMD, various DSDMs which improved the adherence considerably. We envisage that OTZ has the potential to improve person-centered care for the adolescents by individualizing care, empowering the service providers and caregivers as well.

Recommendations

We recommend a wider implementation of the model in Uganda to close the current VLS gaps among the adolescents living with HIV. Through this model, adolescent health friendly services such as multi-month dispensing, and community-based differentiated models can be scaled up to address poor treatment outcomes.

Abbreviations

ALHIV

Adolescent living with HIV

ART

Antiretroviral therapy

AYPLHIV

Adolescents and young people living with HIV

CATS

Community adolescent treatment supporter

CCLAD

Community client-led ART Delivery

CDDP

Community Drug Delivery Points

COE

Center of Excellence

DSDM

Differentiated Service Delivery Model

DTG

Dolutegravir

FBG

Facility-based group

FBIM

Facility-based Individual Management

FTR

Fast-track refill

GEE

Generalized Estimating Equations

MMD

multi-month dispensing

OR

Odds Ratio

OTZ

Operation Triple Zero

OVC

Orphaned and Vulnerable Children

PLHIV

People living with HIV

RE

AIM-Reach, Effectiveness, Adoption, Implementation and Maintenance

RNA

Ribonucleic acid

TASO

The AIDS Support Organization

VLS

Viral load suppression

WHO

World health organization

YAPS

Youth and adolescent treatment supporter

Author contributions

BO: Scientific direction, protocol development, resource mobilization, methodology design, interpretation of results, manuscript drafting, and overall coordination. AK, CA, BB, TM, DK and SMS contributed to developing the protocol, methodology design and writing the draft manuscript. ENM contributed to protocol development, reviewed the manuscript and mobilized resources. EBM, KM and YM reviewed the final draft and contributed significant input to the work. AIO: methodology design and data analysis. All authors read and approved the final manuscript.

Funding

This study was funded by the International Pediatric HIV Symposium in Africa (IPHASA) of the International AIDS Society (IAS), grant number 0184.

Data availability

Attached as supplementary material.

Declarations

Ethical approval

The study received ethical approval from TASO institutional review board (TASO‑ REC‑2022–176) and the Uganda National Council of Science and Technology (UNCST), registration number SS1610ES. A waiver for informed consent was attained given that only secondary data was used without directly contacting individual patients.

Competing interests

The authors declare no competing interests.