Abstract
Objectives:
To determine clinic-level and individual-level correlates of viral suppression among HIV-positive adolescents and young adult (AYA) aged 10–24 years receiving antiretroviral treatment (ART).
Design:
Multilevel cross-sectional analysis using viral load data and facility surveys from HIV treatment programs throughout Kenya.
Methods:
We abstracted medical records of AYA in HIV care, analyzed the subset on ART for more than 6 months between January 2016 and December 2017, and collected information on services at each clinic. Multilevel logistic regression models were used to determine correlates of viral suppression at most recent assessment.
Results:
In 99 HIV clinics, among 10 096 AYA on ART more than 6 months, 2683 (27%) had unsuppressed viral load at last test. Among 16% of clinics, more than 80% of AYA were virally suppressed. Clinic-level correlates of individual viral suppression included designated adolescent spaces [aOR: 1.32, 95% CI (1.07–1.63)] and faster viral load turnaround time [aOR: 1.06 (95% CI 1.03–1.09)]. Adjusting for clinic-level factors, AYA aged 10–14 and 15–19 years had lower odds of viral suppression compared with AYA aged 20–24 years [aOR: 0.61 (0.54–0.69) and 0.59 (0.52–0.67], respectively. Compared with female patients, male patients had lower odds of viral suppression [aOR: 0.69 (0.62–0.77)]. Compared with ART duration of 6–12 months, ART for 2–5, above 5–10 or more than 10 years was associated with poor viral suppression (P < 0.001).
Conclusion:
Dedicated adolescent space, rapid viral load turnaround time, and tailored approaches for male individuals and perinatally infected AYA may improve viral suppression. Routine summarization of viral load suppression in clinics could provide benchmarking to motivate innovations in clinic-AYA and individual-AYA care strategies.
Keywords: adolescents and young adults, clinic factors, individual factors, viral suppression
Introduction
Adolescents and young adults (AYA) living with HIV are at high risk of virologic failure and mortality [1,2]. AYA are twice as likely to have unsuppressed viral load, have higher risk of virologic rebound, and poorer adherence compared with adults [2]. Global estimates of rates of viral suppression among AYA vary widely (27–90%) because of differences in age disaggregation and age–period cohort effects [2–7]. Viral suppression in AYA in sub-Saharan Africa (SSA) remains poorly characterized as viral load testing has only recently been introduced for routine monitoring [4]. In evaluations of program data from Kenya and Uganda; which are among the top five countries globally with the highest burden of adolescent HIV, over a quarter of adolescents aged 10–19 years on antiretroviral therapy (ART) for more than 6 months had viral load at least 1000 copies/ml, indicating possible virologic failure [8,9]. To achieve UNAIDS 95–95–95 goals by 2030, there is a need to understand correlates of viral suppression to direct clinic practice and intervention development.
In previous studies among AYA in SSA, correlates of poor viral suppression have included male sex [9,10], younger age [9], poor adherence [9], tuberculosis (TB) diagnosis [9], older age at full disclosure of HIV status [11], and mental health factors [9] including depression, low self-efficacy, and poor social support. AYA in SSA are a mixed population with different modes of infection, with an estimated 56% of 10–14-year-olds and 14% of 15–19-year-olds estimated to be perinatally infected, though uncertainty exists because of lack of data on maternal HIV status [12]. Viral suppression differs by mode of infection [2,5]. Sub-optimal ART doses [13], perinatal ART exposure and acquired drug resistance [14,15], varied pediatric regimens or formulations, and caregiver and disclosure challenges [13,16,17] during childhood all may influence viral suppression. Transition to adult care in SSA remains poorly defined and its influence on viral suppression is unexplored [18,19].
Clinic-level factors may also influence AYA outcomes. In one study, AYA in clinics in high HIV prevalence areas had better viral suppression than clinics in settings with lower prevalence [20], likely a result of more experience and resources. It is also possible that high numbers of AYA in HIV care and high AYA to staff ratios reduce contact time, which decreases opportunities to identify and address ARTadherence. Clinic approaches to AYA care in SSA have evolved over time, and include different models of care that have been developed to address the needs of AYA [19,21]. These models include AYA-designated clinic days [19,21,22], including on weekends [23] and activities, such as peer clubs or treatment literacy counseling [19]. AYA-focused training, tools to track milestones, and topics to explore at clinic visits have also been developed [24]. For most of these innovations, impact on clinical outcomes remains unknown. Understanding clinic-level factors associated with viral suppression can help clinics to adopt approaches that offer more benefit to AYA.
We identified individual-level and clinic-level correlates of viral suppression among AYA enrolled in HIV care in program settings in Kenya during 2016–2017.
Methods
Study design
This study was conducted in the context of a larger study aimed at understanding transition practices in Kenya (Adolescent Transition to Adult Care for HIV infected adolescents [ATTACH] [NCT03574129]). Participating clinics were randomly selected from all clinics with greater than 300 total patients in care, and that were using electronic medical records in 2015. We estimated that these clinics would have at least 30 AYA aged 10–19 years enrolled in care. Prior to obtaining electronic medical records data, the actual number of AYA attending each clinic was unknown as clinic-level data were not disaggregated in AYA-specific age groups. Overall, 590 clinics met inclusion criteria. These were divided into tertiles based on the total clinic population and 34 clinics were randomly selected from each of the small, medium, and large tertile clinics. This approach ensured equal representation of different clinic sizes in the final sample. A total of 102 clinics throughout Kenya were selected.
Ethical considerations
The study was approved by the University of Washington Institutional Review Board (STUDY00001756) and the Kenyatta National Hospital/University of Nairobi Ethics and Research Review Committee (P248/05/2017). Approval from the Kenya Ministry of Health National AIDS and STI Control Program, County Departments of Health, and managers of participating clinics was also obtained.
Viral load data
Viral load data was obtained from clinic records in the national viral load program [8]. The study team did not test any samples for viral load. The study team accessed viral load records of AYA aged 10–24 during the period 1 January 2016 to 31 December 2017. Typically, patient information on current age, sex, ART regimen, ART initiation date, reason for viral load request, date of sample collection, processing and dispatch, and clinic identifiers (name of clinic, unique number, and county) was recorded at sample collection. Date of birth was also collected in some clinics. Laboratories use various methods of detection including Abbott m2000 systems and Roche Cobas Ampliprep/Cobas TaqMan [25]. Viral suppression was defined as viral load of less than 1000 copies/ml. The lower limit of detection varied, at 40, 250 or 550 copies/ml, and was recorded either as LDL (lower than detection limit) or as actual value. Clinics could access results through a secure link to the national program. The data was then transferred to individual patient records. To obtain the most recent and accurate records available, the study utilized data obtained directly from the national viral load program. Viral load records of AYA who were on ART for less than 6 months or were missing ART duration were excluded.
Clinic data
In participating facilities, data on clinic size, staffing, models of AYA care, and AYA services were collected. After written informed consent, healthcare workers responsible for AYA HIV care and who had been in the clinic for at least 6 months were interviewed. Surveys were conducted in-person or by telephone and data was recorded on paper records and then entered into a REDCap [26] database.
Data analysis
Continuous variables were summarized using medians and interquartile ranges (IQR) and categorical data using counts and proportions. Variables of interest from the viral load database included: age (classified as 10–14, 15–19, and 20–24 years), sex (male or female), ART duration in years (categorized as 6–12 months, 1–2, 2–5, 5–10, and over 10 years) and ART regimen [nonnucleoside reverse transcriptase inhibitor (NNRTI)-based, protease inhibitor-based, other)]. Proportion of perinatally infected was estimated using an age cut-off of age 12 years or less at ART initiation [5]. Clinic-level factors considered included number of adolescents in care, clinic type (county referral, subcounty, health center/dispensary, mission/foundation), number of AYA in follow-up, AYA to healthcare worker (HCW) ratio, HIV clinic staffing (availability of a nutritionist, social worker or community health worker), AYA model of care (AYA clinic days, weekend clinics, incentives), separate adolescent clinic space, staff training on AYA, use of the adolescent checklist [24], all day versus half day clinics, ability to identify adolescent records, availability of mental health screening tools and participation in community or school activities. County HIV prevalence information was obtained from county estimates, and classified as hyper-endemic, high, medium, or low endemicity per national classification (prevalence >11.1, 5.0–11.0, 2.1–4.9, and 0.1–0.2%, respectively) [27]. Viral suppression was defined using the last recorded viral load. Per national guidelines, levels of ≥1000 and <1000 copies/ml were defined as unsuppressed and suppressed, respectively [28].
Multilevel models using mixed effects logistic regression with random clinic-level intercepts were used to estimate odds ratios (ORs) for hypothesized correlates of viral suppression with level 1 factors (individual-level factors) and level 2 factors (clinic-level factors). A null model was first constructed to assess the contribution of clinics to variability in viral suppression. The second model added individual-level factors, and the third model included both individual-level and clinic-level factors after exclusion of collinear variables. To assess collinearity, separate models comparing pairwise all possible combinations of the variables of interest were built. Variables were defined as collinear if, after addition of the second variable, the standard error of the coefficient for the first variable changed by more than 10%. ORs and 95% confidence intervals were reported. Exploratory stratified univariate analysis by age-group and sex were also performed. We used chi-squared tests and student t tests to compare clinic characteristics by overall clinic viral suppression (<80 versus ≥80%). All data analysis was conducted using Stata version 14.
Results
Among 102 clinics sampled, this analysis included viral load records from 99 clinics; 3 clinics were excluded as their viral load data lacked information on ART duration. Overall, 12 676 individual adolescent routine viral load records were abstracted. Of these, 2580 AYA on ART for less than 6 months or with records missing ART duration were excluded. Of 10 096 individual AYA records among AYA who had received ART for over 6 months, 2683 (27%) had unsuppressed viral load at their last recorded viral load test.
Adolescents and young adult characteristics
A total of 3399 (34%), 2316 (23%), and 4381 (43%) AYA were in the 10–14, 15–19, and 20–24-year age groups, respectively. The majority of AYA (70%) were female individuals and 4341 (43%) were categorized as perinatally infected. Almost all [9033 (90%)] were on first-line NNRTI-based regimens. AYA had been on ART for 6–12 months [1603 (16%)], 1–2 years [1977 (20%)], 2–5 years [3275 (32%)] and 5–10 years [2698 (27%)] (Table 1). Compared with female patients, male patients were more likely to be perinatally infected (68 versus 32% P < 0.001).
Table 1.
Adolescent demographic characteristics at last viral load.
| Nb | All N = 10 096 | Suppressed N = 7413 (73%) | Unsuppressed N = 2683 (27%) | |
|---|---|---|---|---|
| Adolescent characteristic | n (%) or median (IQR)c | n (%) or median (IQR)d | n (%) or median (IQR)d | |
| Age-group | 10 096 | |||
| 10–14 | 3399 (34%) | 2284 (67%) | 1115 (33%) | |
| 15–19 | 2316 (23%) | 1572 (68%) | 744 (32%) | |
| 20–24 | 4381 (43%) | 3557 (81%) | 824 (19%) | |
| Sex | 10 060 | |||
| Male | 3050 (30%) | 1993 (65%) | 1057 (35%) | |
| Female | 7010 (70%) | 5391 (77%) | 1619 (23%) | |
| Mode of infection | 10 096 | |||
| Perinatal | 4341 (43%) | 2886 (66%) | 1455 (34%) | |
| Horizontal | 5755 (57%) | 4527 (79%) | 1228 (21%) | |
| ART regimen | 10 046 | |||
| NNRTI | 9033 (90%) | 6644 (74%) | 2389 (26%) | |
| PI | 898 (9%) | 642 (71%) | 256 (29%) | |
| Othera | 115 (1%) | 86 (75%) | 29 (25%) | |
| ART duration | 10 096 | |||
| 6 months to 1 year | 1603 (16%) | 1301 (81%) | 302 (19%) | |
| 1–2 years | 1,977 (20%) | 1567 (79%) | 410 (21%) | |
| 2–5 years | 3275 (32%) | 2388 (73%) | 887 (27%) | |
| 5–10 years | 2698 (27%) | 1781 (66%) | 917 (34%) | |
| Over 10 years | 543 (5%) | 376 (69%) | 167 (31%) | |
| Turn-around time | 10 067 | 12 (7–20) | 12 (7–19) | 12 (7–21) |
IQR, interquartile range; NNRTI, nonnucleoside reverse transcriptase inhibitor.
Integrase inhibitor, triple nucleotide, unspecified.
Data on sex, antiretroviral therapy (ART) regimen and turnaround time not available for all adolescents and young adults.
Column (%).
Row (%).
Clinic characteristics
Of included clinics, 27 (27%), 14 (14%), 55 (56%), and 3 (3%) were in counties classified as HIV hyper-endemic, high, medium, or low prevalence, respectively. The median number of AYA per clinic was 61 (IQR: 36–130) with a median AYA : HCW ratio of 6 (IQR: 3–10). Over half of the clinics had nutritionists (61%), laboratory staff (75%), and community health workers (66%); fewer had social workers (37%). The majority of the clinics had adolescent days, often on weekend days (57%). Separate adolescent spaces were uncommon (18%). Most clinics were open all day (75%), and almost half participated in school or community HIV activities (42 and 47%, respectively) (Table 3).
Table 3.
Clinic characteristics stratified by overall clinic viral suppression status.
| Percent suppressed in clinic | ||||
|---|---|---|---|---|
| Clinic characteristic | All n = 99 n (%)/median (IQR) | ≥80% n = 16 n (%)/median (IQR) | <80% n = 83 n (%)/median (IQR) | P value |
| County HIV prevalence | 0.04 | |||
| Hyper-endemic | 27 (26%) | 9 (56%) | 18 (22%) | |
| High | 14 (14%) | 2 (13%) | 12 (14%) | |
| Medium | 55 (56%) | 5 (31%) | 50 (60%) | |
| Low | 3 (3%) | 0 | 3 (4%) | |
| Clinic size and resources | 0.04 | |||
| County | 11 (11%) | 0 | 11 (13%) | |
| Sub-county | 36 (36%) | 3 (19%) | 33 (40%) | |
| Health center/dispensary | 50 (51%) | 12 (75%) | 38 (46%) | |
| Mission/foundation | 2 (2%) | 1 (6%) | 1 (1%) | |
| AYA in follow-up | 61 (36–130) | 58 (35–128) | 61 (36–130) | 0.77a |
| AYA: HCW ratio | 6 (3–10) | 5 (2–13) | 6 (3–10) | 0.85 |
| Separate adolescent space | 18 (18%) | 6 (38%) | 12 (14%) | 0.03 |
| Staffing in HIV clinic | ||||
| Nutritionist | 60 (61%) | 8 (50%) | 52 (63%) | 0.34 |
| Laboratory staff | 74 (75%) | 11 (69%) | 63 (76%) | 0.55 |
| Social worker | 37 (37%) | 1 (6%) | 36 (43%) | 0.004 |
| Community health worker | 65 (66%) | 11 (69%) | 54 (65%) | 0.78 |
| Adolescent services | ||||
| Model of care | ||||
| Adolescent days | 90 (91%) | 14 (88%) | 76 (92%) | 0.64 |
| Weekend | 51 (57%) | 7 (50%) | 44 (58%) | 0.58 |
| Snacks | 12 (13%) | 4 (29%) | 8 (11%) | 0.07 |
| More staff | 15 (17%) | 4 (26%) | 11 (15%) | 0.24 |
| Staff trained in adolescent care | 42 (43%) | 7 (44%) | 35 (43%) | 0.94 |
| Using adolescent checklist | 62 (63%) | 11 (69%) | 51 (62%) | 0.62 |
| All day clinics | 74 (75%) | 13 (81%) | 61 (73%) | 0.75 |
| Can identify AYA records | 74 (75%) | 12 (75%) | 62 (75%) | <1.0 |
| Can screen mental health | 72 (78%) | 13 (87%) | 59 (77%) | 0.51 |
| Community activities | ||||
| School activities | 42 (42%) | 9 (56%) | 33 (40%) | 0.22 |
| Community activities | 46 (46%) | 9 (56%) | 37 (45%) | 0.39 |
| Group summary variables | ||||
| Viral load turnaround time (days) | 12 (9–14) | 9 (6–12) | 12 (10–15) | 0.0008a |
AYA, adolescent and young adult; HCW, healthcare worker.
P values from a Student t test.
Bolded values are statistically significant.
Univariate analysis
In univariate analysis, AYA aged 10–14 years [OR: 0.49 (95% CI 0.44–0.54)] and 15–19 years [OR: 0.50 (0.44–0.56)] had lower odds of viral suppression compared with AYA aged 20–24 years (P < 0.001). Compared with female patients, male patients [OR: 0.57 (0.52–0.63)] had lower odds of suppression (P < 0.001). Perinatally infected AYA [OR: 0.56 (0.51–0.61)] had lower odds of viral suppression compared with horizontally infected AYA (P < 0.001). Adolescents on protease inhibitor regimens were less likely to have viral suppression (OR: 0.81 [95% CI 0.72–0.99], P < 0.001) compared with those on NNRTIs. Viral suppression significantly decreased with increasing duration on ART (chi-squared test for trend P < 0.001), with significantly lower odds of viral suppression in the 2–5, 5–10, and over 10-year duration, compared with 6–12 months of ART, respectively (Table 2).
Table 2.
Individual-level and clinic-level correlates of viral suppression.
| Univariate | Model 1: individual-level | Model 2: individual-level and clinic-level | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | |
| Adolescent characteristics | ||||||
| Age group | ||||||
| 10–14 | 0.49 | 0.44–0.54 | 0.61 | 0.54–0.69 | 0.61 | 0.54–0.69 |
| 15–19 | 0.50 | 0.44–0.56 | 0.58 | 0.51–0.66 | 0.59 | 0.52–0.67 |
| 20–24 | Ref | Ref | - | - | - | |
| Sex | ||||||
| Female | Ref | - | Ref | - | - | - |
| Male | 0.57 | 0.52–0.63 | 0.72 | 0.65–0.79 | 0.69 | 0.62–0.77 |
| Mode of infectiona | ||||||
| Horizontal | Ref | - | ||||
| Perinatal | 0.56 | 0.51–0.61 | ||||
| ART regimen | ||||||
| NNRTI | Ref | - | Ref | - | - | - |
| PI | 0.85 | 0.72–0.99 | 1.03 | 0.88–1.21 | 1.00 | 0.85–1.18 |
| ART duration | ||||||
| 6 months to 1 year | Ref | - | Ref | - | - | - |
| 1–2 years | 0.89 | 0.75–1.05 | 0.91 | 0.77–1.08 | 0.91 | 0.76–1.09 |
| 2–5 years | 0.63 | 0.54–0.73 | 0.69 | 0.59–0.80 | 0.69 | 0.59–0.81 |
| 5–10 years | 0.47 | 0.40–0.54 | 0.60 | 0.51–0.71 | 0.60 | 0.51–0.70 |
| Over 10 years | 0.55 | 0.44–0.69 | 0.68 | 0.54–0.86 | 0.70 | 0.55–0.89 |
| Clinic characteristics | ||||||
| County HIV prevalence* | ||||||
| Hyper-endemic (>11.1%) | Ref | - | Ref | - | ||
| High (5–11.1%) | 0.84 | 0.65–1.08 | 0.93 | 0.73–1.20 | ||
| Medium/low (<4.9%) | 0.69 | 0.58–0.82 | 0.84 | 0.69–1.02 | ||
| Clinic size and resources | ||||||
| County | Ref | - | Ref | - | ||
| Subcounty | 1.06 | 0.84–1.34 | 1.05 | 0.83–1.33 | ||
| Health center/dispensary | 1.40 | 1.11–1.76 | 1.22 | 0.94–1.59 | ||
| Mission/foundation | 1.90 | 1.14–3.18 | 2.07 | 1.32–3.25 | ||
| AYA in follow-up (per 100 AYA)b,c | 1.02 | 0.97–1.07 | ||||
| AYA: HCW ratio (10 unit increase)c | 1.05 | 0.96–1.16 | 1.06 | 0.97–1.16 | ||
| Separate adolescent spaced | 1.19 | 0.95–1.50 | 1.32 | 1.07–1.63 | ||
| Nutritionistd | 0.89 | 0.75–1.07 | 1.00 | 0.85–1.18 | ||
| Laboratory staffd | 0.95 | 0.78–1.17 | 0.98 | 0.82–1.19 | ||
| Social workerd | 0.77 | 0.65–0.91 | 0.76 | 0.66–0.88 | ||
| Community health workerd | 1.01 | 0.84–1.21 | 0.95 | 0.80–1.12 | ||
| Adolescent services | ||||||
| Adolescent daysb,d | 1.18 | 0.82–1.69 | ||||
| Weekendd | 1.02 | 0.86–1.21 | 0.97 | 0.84–1.13 | ||
| Snackd | 1.06 | 0.83–1.36 | 0.88 | 0.71–1.09 | ||
| More staffd | 1.01 | 0.80–1.28 | 0.95 | 0.78–1.15 | ||
| Staff trained in adolescent cared | 0.99 | 0.83–1.17 | 1.02 | 0.87–1.19 | ||
| Using adolescent checklistd | 1.06 | 0.89–1.27 | 0.97 | 0.82–1.15 | ||
| All day clinicsd | 1.01 | 0.83–1.23 | 0.95 | 0.79–1.14 | ||
| Can identify AYA recordsd | 1.01 | 0.82–1.23 | 0.94 | 0.77–1.16 | ||
| Has mental health toolb,d | 1.07 | 0.85–1.35 | ||||
| Other activities | ||||||
| School activitiesd | 1.10 | 0.92–1.31 | 1.08 | 0.92–1.26 | ||
| Community activitiesd | 1.16 | 0.98–1.38 | 0.98 | 0.83–1.16 | ||
| Turn-around time (10 day longer) | 0.95 | 0.93–0.97 | 0.94 | 0.92–0.97 | ||
Categories defined by Kenya Ministry of Health. Bolded: statistically significant, P values less than 0.05.
ART, antiretroviral regimen; AYA, adolescent and young adult; HCW, healthcare worker; NNRTI, nonnucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
Mode of transmission not included in models 1 and 2 because of collinearity.
AYA in follow-up, model of care, has mental health not included in model 2 because of collinearity.
Compared with a 1 unit higher in covariate.
Compared with the group without the covariate of interest.
Low or medium county HIV prevalence [OR: 0.69 (95% CI 0.58–0.82), P < 0.001], and having a social worker [OR: 0.77 (95% CI 0.65–0.91), P = 0.002], were significantly associated with poorer viral suppression, while being in a health center/dispensary [OR: 1.40 (95% CI 1.11–1.76) P = 0.004], or mission/foundation clinic [OR: 1.90 (95% CI 1.14–3.18) P = 0.012] was associated with better suppression compared with county referral clinics. Longer viral load result turnaround time was associated with lower odds of viral suppression (OR: 0.95 [95% CI 0.93–0.97], P < 0.001, for every 10-day increase) (Table 2).
Multivariate analysis
In multivariable analyses, age 10–14 years (adjusted OR [aOR]: 0.61 [95% CI 0.54–0.69], P < 0.001), 15–19 years (aOR: 0.59 [95% CI 0.52–0.67], P < 0.001), male sex (aOR: 0.69 [95% CI 0.62–0.67], P < 0.001), and longer ART duration (2–5 years: aOR: 0.69 [95% CI 0.59–0.81], P < 0.001; 5–10 years: aOR: 0.60 [95% CI 0.51–0.70], P < 0.001; >10 years: aOR: 0.70 [95% CI 0.55–0.89], P = 0.004) remained significantly associated with lower odds of viral suppression. AYA attending clinics with a social worker [aOR: 0.76 (95% CI 0.66–0.88), P < 0.001) and with longer viral load turnaround time [aOR: 0.94 (95% CI 0.92–0.97), P < 0.001] remained significantly less likely to be suppressed, and those attending mission/foundation clinics [aOR: 2.07 (95% CI 1.32–3.25), P = 0.001) or clinics with separate adolescent space [aOR: 1.32 [95% CI 1.07–1.63], P = 0.009) were more likely to be suppressed (Table 2).
Stratified analysis
In stratified univariate analysis, male individuals of all age groups had significantly lower odds of viral suppression compared with females. The magnitude of the reduced odds of suppression with longer ART duration was more pronounced in the 20–24-year olds than in other age groups (Fig. 1, panel a). Stratified by sex, similar patterns of viral suppression were observed in both male and female individuals, with lower suppression among those with perinatal infection and longer ART duration in female individuas (Fig. 1, panel b). The association between attending a clinic with a social worker and reduced odds of suppression remained only among perinatally infected AYA when analysis was stratified by mode of infection (data not presented).
Fig. 1. Individual-level correlates of viral suppression stratified by age and sex.
Panel a: stratified analysis age-group comparing sex, mode of HIV transmission, antiretroviral (ART) regimen and ART duration between adolescents and young adults with and without viral suppression. Estimates are presented as odd ratios (ORs). Total populations and percentages are presented. Panel b: sex-stratified analysis comparing age, mode of HIV transmission, ART regimen and ART duration between adolescents and young adults with and without viral suppression. Estimates are presented as ORs. Total populations and percentages are presented.
Overall clinic viral suppression
For overall clinic viral suppression, described as the proportion of AYA with viral load results that were suppressed, 16 (16%) clinics had at least 80% viral suppression among AYA attending that clinic, and 83 (84%) had less than 80% (Table 3). One clinic that contributed only four viral load records had 100% viral suppression. Clinics with at least 80% viral suppression were more likely to be in higher endemic counties [56 versus 22% and 31 versus 60% for hyper-endemic and medium endemic counties, respectively, (P = 0.04)]. Twelve (75%) clinics with at least 80% viral suppression were in health centers or dispensaries, whereas majority of those with less than 80% were distributed in subcounty (40%) or health centers and dispensaries (46%) P = 0.04. Clinics with at least 80% viral suppression were more likely to have separate adolescent space and a shorter viral load turnaround time (39 versus 15% P = 0.03 and 9 versus 12 days, P = 0.0008, respectively). Compared with clinics without social workers, clinics with social workers were less likely to have viral suppression of at least 80% (6 versus 43%, P = 0.004) (Table 3).
Discussion
In this study of over 10 000 AYA from 99 clinics throughout Kenya who had been on ART for over 6 months, 27% were not virally suppressed. Individual-level factors, including younger age, male sex, perinatally acquired HIV, and longer ART duration were associated with lower odds of suppression. At the clinic level, AYA attending clinics in higher HIV prevalence counties, lower level facilities or mission/foundation hospitals, or those with separate AYA spaces had higher odds of viral suppression. Conversely AYA in clinics with social workers had lower odds of suppression. Clinics with more than 80% of AYA virally suppressed were more likely to be in higher prevalence regions, to have shorter viral load turnaround time, and be mission/foundation or lower level facilities.
Viral suppression status in our study was well below the 2030 UNAIDS 95–95–95 target, and below the 2018 global estimates for adults 15 years and older (86%) [29]. In contrast to the UNAIDS target of having only 5% of those receiving treatment not virally suppressed, 27% of AYA in our study were not suppressed. Only 16% of the clinics had more than 80% of attending AYA virally suppressed. Similar to previous studies, we found that younger AYA were less likely to be virally suppressed than older AYA [8,9]. This could be because of the larger proportion of perinatally infected children in the younger age-groups. In analysis stratified by mode of infection, this association remained only for horizontally infected AYA, even after adjusting for ART duration. Pediatric ART challenges include unpalatable medication, inaccurate dosing because of rapid weight changes or difficulties administering treatment, which contribute to inadequate drug levels and drug resistance [13,30]. Parental loss or illness, challenges with disclosure, adherence challenges and stigma and an overall higher risk of transmitted drug resistance, contribute to virologic failure in children [13,16,31]. In a recent study, intensive adherence interventions resulted in suppression among 23% of children and adolescents [32], compared with 70% suppression in adults, which may reflect the higher likelihood of resistant virus in children and adolescents [33]. With the growing population of perinatally infected children surviving to adolescence [12], there is need to adopt a more aggressive approach to manage treatment failure in this population. With limited treatment options available for children and AYA, adherence support remains critical, but must be accompanied by appropriate and timely regimen switch particularly for perinatally infected AYA.
We found that male sex was independently associated with poor viral suppression. In the overall analysis, this could be partially because of a higher proportion of male AYA having perinatal HIV infection than female AYA. However, in age-stratified analyses, male AYA had poorer viral load suppression than female AYA in all age groups. Among adults in SSA, there is good evidence of poorer viral suppression among male individuals [9,34,35]. This was initially thought to be a result of late presentation, poor adherence, and higher substance abuse. However, there is emerging evidence of possible biological differences in drug metabolism [36]. Boulle et al. [34] found that male individuals in Cameroon were more than twice as likely to have poor viral suppression even after adjusting for plasma drug concentrations. Male sex has been associated with nonsuppression in several, but not all pediatric studies in SSA [9,10]. In a study from South Africa, sex differences were noted in regimen-related CD4+ recovery and cholesterol levels but not viral suppression [37]. Among perinatally and behaviorally infected AYA in Canada and the United States, no difference [38] or an opposite effect was noted, with male individuals being more likely to suppress than female individuals [39,40]. Further studies that include standardized age-cutoffs and accurate data on mode of infection will be useful for better understanding the potential role of sex differences in AYA viral suppression.
We hypothesized that clinic AYA care practices could result in better adherence and viral suppression. We found that designated adolescent clinic space was associated with better viral suppression. However, none of the other elements in adolescent specific models of care were associated with viral suppression. Separate safe clinic spaces may promote helpful adherence discussions. Although offered in a minority of clinics, space is an important component of youth friendly services advocated for in national guidelines [41]. AYA services were available in almost all clinics, and AYA-specific days were common; however, utilization of services provided in these clinics remains low [22] and clinic practices vary [19]. Counterintuitively, we found that AYA in clinics with social workers were less likely to have viral suppression. This may be because social workers trace out-of-care AYA who are less likely to be suppressed and bring them back into care where they contribute to the denominator of viral suppression. We found that clinics in high HIV prevalence regions had higher levels of viral suppression in AYA, perhaps because of more care experience and resources to support care. Efforts to support regions with low HIV prevalence are needed despite challenges in resource distribution and prioritization.
We found that AYA in mission/foundation hospitals had better viral suppression, even after adjusting for AYA : HCW ratio. However, this data is from only two mission/foundation clinics. These facilities may have better resources in terms of leadership, space and staff motivation, and AYA activities may differ. More evaluation of the differences in care in these facilities could inform interventions. We found that faster viral load turnaround time was associated with better suppression. Viral load testing is conducted in centralized facilities that are located in high burden areas. Transportation logistics from clinic to the laboratory may contribute to longer turnaround time, particularly with increased distance from clinic to laboratory [42], which may in turn result to delays in implementing appropriate care. Without prompt viral load results, healthcare workers may not implement adherence strategies or regimen change, contributing to lack of viral suppression. Point-of-care technologies for viral load testing may close the gap in turnaround time and could be explored, particularly in low HIV-burden counties.
Overall, only 16% of clinics had more than 80% of AYA on ART suppressed at last viral load. The clinic level viral suppression is an overestimate of suppression among AYA as it only includes AYA attending clinic and with viral load measured. As countries aspire to the third 95% of the 95–95–95 goals, it may be useful for clinics to routinely assess the rates of viral suppression among AYA on ART and to consider clinic strategies that may support optimal adherence counseling and prompt detection and management of viral failure in AYA.
Our study has limitations. The national viral load database had few variables available on AYA characteristics. Our population is composed of AYA who were retained in care, who had samples collected for viral load testing, and who were enrolled in clinics using electronic medical records, and may not generalize to other clinics. Age cut-offs have been used to estimate timing of infection, for example, specifying that ART initiation at age 10 years or less or 12 years or less signifies perinatal infection; however, this approach may misclassify perinatal infections [43,44]. Data on CD4+ counts and height (because of stunting in perinatally infected) could improve estimation of timing of infection; however, these data are often unavailable in program databases [45]. Strengths of our study include that we accessed national viral load data from 99 clinics throughout Kenya, enabling us to evaluate critical individual-level and clinic-level differences among AYA. We used two different approaches to estimate clinic-level effects, a multilevel model that included individual and clinic-level factors and comparisons of high suppression to lower suppression clinics, which identified consistent clinic-level factors contributing to viral suppression.
In conclusion, we found that age, sex, ART duration, separate AYA spaces, clinic type, and viral load turnaround time were important correlates of viral suppression. As children survive pediatric HIV and age into adolescence, there is need to comprehensively address viral suppression, and better understand age and sex differences that will help build targeted strategies that address barriers to viral suppression. Clinics in low/medium prevalence counties may require additional support for better outcomes, care approaches in mission/foundation clinics may be useful to emulate, and enhancing viral load turnaround time may enhance viral suppression in AYA.
Acknowledgements
D.W. and G.J.S. obtained grant funding. I.N., J.N., K.B.S., C.Mb., C.Mu., J.I., and A.O. developed study materials. A.O., J.I., C.Mb., CMu were involved in data collection. D.W., G.J.S., I.N., J.N., D.B., J.I., A.W., Y.E., B.G., B.A., and L.O. conceptualized the analysis and supported data analysis. I.N. wrote the first draft of the manuscript. All authors read the manuscript draft, provided feedback, and approved the final submitted manuscript. We acknowledge the study participants, caregivers, and the research administrative, clinical, and data teams for their dedication and support.
Funding: The National Institute of Child Health and Development (NICHD) 1R01HD089850-01, 5K24HD054314-09 to G.J.S. and F32HD088204 to A.D.W., and the Fogarty International Center (FIC) D43TW009783 to I.N. Additional support was provided by the UW Global Center for Integrated Health of Women, Adolescents and Children (Global WACh), the University of Washington CFAR (P30 AI027757). Partial support for this research came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant, P2C HD042828, to the Center for Studies in Demography & Ecology at the University of Washington, Seattle, United States of America. The contents are solely the responsibility of the authors and do not represent the official views of the funders.
Footnotes
Conflicts of interest
There are no conflicts of interest.
References
- 1.Slogrove AL, Mahy M, Armstrong A, Davies MA. Living and dying to be counted: what we know about the epidemiology of the global adolescent HIV epidemic. J Int AIDS Soc 2017; 20 (Suppl 3):21520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nachega JB, Hislop M, Nguyen H, Dowdy DW, Chaisson RE, Regensberg L, et al. Antiretroviral therapy adherence, virologic and immunologic outcomesin adolescents compared withadults in Southern Africa. J Acquir Immune Defic Syndr 2009; 51:65–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zanoni BC, Archary M, Buchan S, Katz IT, Haberer JE. Systematic review and meta-analysis of the adolescent HIV continuum of care in South Africa: the Cresting Wave. BMJ Glob Health 2016; 1:e000004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ferrand RA, Briggs D, Ferguson J, Penazzato M, Armstrong A, MacPherson P, et al. Viral suppression in adolescents on antiretroviral treatment: review of the literature and critical appraisal of methodological challenges. Trop Med Int Health 2016; 21:325–333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Slogrove AL, Sohn AH. The global epidemiology of adolescents livingwith HIV:timefor moregranular datatoimproveadolescent health outcomes. Curr Opin HIV AIDS 2018; 13:170–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kiseleva G, Volokha A, Malyuta R, Thorne C, Judd A, Nizova N, et al. Virological suppression among young people living with perinatally and behaviorally-acquired HIV infection in Ukraine In: IAS 2018, Amsterdam, Netherlands, 2018. [Google Scholar]
- 7.Haghighat R, Cluver R, Bungane N, Toska E. 90–90–48: the reality of viral suppression among ART initiated adolescents in South Africa In: 10th International Workshop on HIV Pediatrics. Amsterdam, The Netherlands, 2018. [Google Scholar]
- 8.Mwau M, Syeunda CA, Adhiambo M, Bwana P, Kithinji L, Mwende J, et al. Scale-up of Kenya’s national HIV viral load program: findings and lessons learned. PLoS One 2018; 13:e0190659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bulage L, Ssewanyana I, Nankabirwa V, Nsubuga F, Kihembo C, Pande G, et al. Factors associated with virological nonsuppression among HIV-positive patients on antiretroviral therapy in Uganda, August 2014–July 2015. BMC Infect Dis 2017; 17:326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Umar E, Levy JA, Bailey RC, Donenberg G, Hershow RC, Mackesy-Amiti ME. Virological nonsuppression and its correlates among adolescents and young people living with HIV in Southern Malawi. AIDS Behav 2019; 23:513–522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Beima-Sofie KM, Brandt L, Hamunime N, Shepard M, Uusiku J, John-Stewart GC, O?Malley G, et al. Pediatric HIV disclosure intervention improves knowledge and clinical outcomes in HIV-infected children in Namibia. J Acquir Immune Defic Syndr 2017; 75:18–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ng’eno BN, Kellogg TA, Kim AA, Mwangi A, Mwangi M, Wamicwe J, Rutherford GW. Modes of HIV transmission among adolescents and young adults aged 10–24 years in Kenya. Int J STD AIDS 2018; 29:800–805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Haberer J, Mellins C. Pediatric adherence to HIV antiretroviral therapy. Curr HIV/AIDS Rep 2009; 6:194–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Boerma RS, Sigaloff KC, Akanmu AS, Inzaule S, Boele van Hensbroek M, Rinke de Wit TF, Calis JC. Alarming increase in pretreatment HIV drug resistance in children living in sub-Saharan Africa: a systematic review and meta-analysis. J Antimicrob Chemother 2017; 72:365–371. [DOI] [PubMed] [Google Scholar]
- 15.Huibers MHW, Moons P, Cornelissen M, Zorgdrager F, Maseko N, Gushu MB, et al. High prevalence of virological failure and HIV drug mutations in a first-line cohort of Malawian children. J Antimicrob Chemother 2018; 73:3471–3475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fetzer BC, Mupenda B, Lusiama J, Kitetele F, Golin C, Behets F. Barriers to and facilitators of adherence to pediatric antiretroviral therapy in a sub-Saharan setting: insights from a qualitative study. AIDS Patient Care STDS 2011; 25:611–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Vreeman RC, Gramelspacher AM, Gisore PO, Scanlon ML, Nyandiko WM. Disclosure of HIV status to children in resource-limited settings: a systematic review. J Int AIDS Soc 2013; 16:18466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Asad H, Sabin C, Collins I, Goodall R, Judd A. Severe immunosuppression and viral failure in adult care among antiretroviral therapy experienced young people with perinatal HIV in the UK In: 10th International Workshop on HIV Pediatrics, Amsterdam, The Netherlands: 2018. [Google Scholar]
- 19.Dahourou DL, Gautier-Lafaye C, Teasdale CA, Renner L, Yotebieng M, Desmonde S, et al. Transition from paediatric to adult care of adolescents living with HIV in sub-Saharan Africa: challenges, youth-friendly models, and outcomes. J Int AIDS Soc 2017; 20 (Suppl 3):21528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kahana SY, Jenkins RA, Bruce D, Fernandez MI, Hightow-Weidman LB, Bauermeister JA, Adolescent Medicine Trials Network for HIV/AIDS Interventions. Structural determinants of antiretroviral therapy use, HIV care attendance, and viral suppression among adolescents and young adults living with HIV. PLoS One 2016; 11:e0151106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Badejo OA, Menson WNA, Sam-Agudu NA, Pharr J, Erekaha S, Bruno T, et al. Pediatric to adult healthcare transitioning for adolescents living with HIV in Nigeria: a national survey. PLoS One 2018; 13:e0198802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nyabigambo A, Muliira JK, Atuyambe L, Babikako HM, Kambugu A, Ndoleriire C. Determinants of utilization of a no-cost HIV transition clinic: a cross-sectional study of young adults living with HIV/AIDS. Adolesc Health Med Ther 2014; 5:89–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Njuguna IN, Beima-Sofie K, Mburu C, Guthrie B, Mugo C, Neary J, et al. Transitioning adolescents to adult HIV care: Health facility models of care and transition practices in Kenya In: 10th Workshop on HIV Pediatrics, Amsterdam, The Netherlands: 2018. [Google Scholar]
- 24.National AIDS and STI Control Program (NASCOP). Adolescent Package of Care in Kenya: a healthcare provider guide to adolescent care. Nairobi, Kenya: 2014. Available at: https://faces.ucsf.edu/sites/faces.ucsf.edu/files/AdolescentPackage.pdf. (Accessed 7 April 2019) [Google Scholar]
- 25.Mwau M, Syeunda CA, Adhiambo M, Bwana P, Kithinji L, Mwende J, et al. Scale-up of Kenya’s national HIV viral load program: findings and lessons learned. PLoS One 2018; 13:e0190659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009; 42:377–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kenya HIV Estimates: report 2018. National AIDS Control Council. Available at: https://nacc.or.ke/wp-content/uploads/2018/11/HIV-estimates-report-Kenya-20182.pdf. (Accessed 14 May 2019)
- 28.Ministry of Health, National AIDS & STI Control Program. Guidelines on use of antiretroviral drugs for treating and preventing HIV infection in Kenya 2018 edition. Nairobi, Kenya: NASCOP, August 2018. [Google Scholar]
- 29.Marsh K, Eaton JW, Mahy M, Sabin K, Autenrieth C, Wanyeki I, et al. Global, regional and country-level 90-90-90 estimates for 2018: assessing progress towards the 2020 target. AIDS 2019; 33 (Suppl 3):S213–S226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Clay PG, Nag S, Graham CM, Narayanan S. Meta-Analysis of studies comparing single and multi-tablet fixed dose combination HIV treatment regimens. Medicine (Baltimore) 2015; 94:e1677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Muller AD, Bode S, Myer L, Stahl J, von Steinbuchel N. Predictors of adherence to antiretroviral treatment and therapeutic success among children in South Africa. AIDS Care 2011; 23:129–138. [DOI] [PubMed] [Google Scholar]
- 32.Nasuuna E, Kigozi J, Babirye L, Muganzi A, Sewankambo NK, Nakanjako D. Low HIV viral suppression rates following the intensive adherence counseling (IAC) program for children and adolescents with viral failure in public health facilities in Uganda. BMC Public Health 2018; 18:1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Bonner K, Siemieniuk RA, Boozary A, Roberts T, Fajardo E, Cohn J. Expanding access to HIV viral load testing: a systematic review of RNA stability in EDTA tubes and PPT beyond current time and temperature thresholds. PLoS One 2014; 9:e113813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Boulle C, Kouanfack C, Laborde-Balen G, Boyer S, Aghokeng AF, Carrieri MP, et al. , Stratall ANRS 12110ESTHER Study Group. Gender differences in adherence and response to antiretroviral treatment in the Stratall Trial in Rural District Hospitals in Cameroon. J Acquir Immune Defic Syndr 2015; 69:355–364. [DOI] [PubMed] [Google Scholar]
- 35.Kipp W, Alibhai A, Saunders LD, Senthilselvan A, Kaler A, Konde-Lule J, et al. Gender differences in antiretroviral treatment outcomes of HIV patients in rural Uganda. AIDS Care 2010; 22:271–278. [DOI] [PubMed] [Google Scholar]
- 36.Ofotokun I, Chuck SK, Hitti JE. Antiretroviral pharmacokinetic profile: a review of sex differences. Gend Med 2007; 4:106–119. [DOI] [PubMed] [Google Scholar]
- 37.Shiau S, Kuhn L, Strehlau R, Martens L, McIlleron H, Meredith S, et al. Sex differences in responses to antiretroviral treatment in South African HIV-infected children on ritonavir-boosted lopinavir- and nevirapine-based treatment. BMC Pediatrics 2014; 14:39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kahana SY, Fernandez MI, Wilson PA, Bauermeister JA, Lee S, Wilson CM, et al. Rates and correlates of antiretroviral therapy use and virologic suppression among perinatally and behaviorally HIV-infected youth linked to care in the United States. J Acquir Immune Defic Syndr 2015; 68:169–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Palmer A, Gabler K, Rachlis B, Ding E, Chia J, Bacani N, et al. , Canadian Observational Cohort (CANOC) Collaboration. Viral suppression and viral rebound among young adults living with HIV in Canada. Medicine (Baltimore) 2018; 97:e10562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Mave V, Gahunia M, Frontini M, Clark R, Mushatt D. Gender differences in HIV drug resistance mutations and virological outcome. J Womens Health (Larchmt) 2011; 20:117–122. [DOI] [PubMed] [Google Scholar]
- 41.National guidelines for provision of adolescent and youth friendly services in Kenya. 2016. Available at: https://faces.ucsf.edu/sites/faces.ucsf.edu/files/YouthGuidelines2016.pdf. (Accessed 17 May 2019)
- 42.Thiha S, Shewade HD, Philip S, Aung TK, Kyaw NTT, Oo MM, et al. Factors associated with long turnaround time for early infant diagnosis of HIV in Myanmar. Glob Health Action 2017; 10:1395657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Collaborative Initiative for Paediatric HIV Education and Research (CIPHER) Global Cohort Collaboration. Slogrove AL, Schomaker M, Davies M-A, Williams P, Balkan S, et al. The epidemiology of adolescents living with perinatally acquired HIV: a cross-region global cohort analysis. PLoS Med 2018; 15:e1002514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Lowenthal ED, Bakeera-Kitaka S, Marukutira T, Chapman J, Goldrath K, Ferrand RA. Perinatally acquired HIV infection in adolescents from sub-Saharan Africa: a review of emerging challenges. Lancet Infect Dis 2014; 14:627–639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Tsondai P, Davies MA, Songtaweesin W, Technau K, Kariminia A, Chimbetete C, et al. An algorithmn to determine likely mode of infection in adolescents living with HV enrolling in care at 10–15 years In: International Workshop on HIV Pediatrics Amsterdam, Netherlands: 2018. [Google Scholar]