Summary:
Background:
The number of South African adolescents receiving HIV care and treatment in South Africa is growing. We used routinely collected laboratory data from South Africa’s National HIV Programme to: 1) quantify the numbers of adolescents accessing HIV care and treatment over time; 2) characterize the role of perinatal infection in these trends; and 3) estimate proportions of adolescents seeking HIV care and antiretroviral treatment (ART) in South Africa’s public sector.
Methods:
National Health Laboratory Service (NHLS) conducts all laboratory monitoring for South Africa’s National HIV Programme. We conducted a descriptive cohort study of children and adolescents (aged 1-19 years) accessing care in South Africa’s public sector HIV treatment program from 2005-2016 with a CD4 count or viral load recorded in the NHLS database. We estimated the total number entering HIV care (number with CD4/viral load test result) by calendar period, as well as proportion in care and on ART (at least one VL test result). We stratified analyses by gender and by whether the patient entered care <15 years (likely perinatally infected) or at 15-19 years (likely infected in adolescence).
Findings:
The cohort included 730,882 patients aged 1-19 years at entry to care. Fifty-four percent of patients (n=209,205) entering care <15 years were female while 88% (n=301,242) of those entering care aged 15-19 were female. During the study period, the number of virologically monitored ART patients aged 15-19 years increased 10-fold, from 7,949 in 2005-2008 to 80,918 in 2013-2016. Still, just two-thirds (n= 92,783/140,028) of 15-19-year olds seeking care started ART by 2016, well below UNAID’s target of ART for 90% of those diagnosed. We project the number of adolescents on ART will continue to rise.
Interpretation:
Large increases in numbers of adolescents (aged 15-19 years) on ART reflect aging of children entering care at ages 1-14 years and increases in care-seeking care among 15-19 year-olds, presumably horizontally infected. However, many adolescents seeking care do not start ART, suggesting an urgent need for interventions to increase uptake of ART and improve services for this growing population.
Introduction
There is a growing awareness that epidemic control in South Africa will require large expansion of treatment for adolescents, in order to break HIV transmission cycles and reach the promise of an AIDS-free generation.1 In 2016, there were an estimated 10.2 million adolescents aged 10-19 years in South Africa.2 Despite widespread access to HIV treatment and prevention services, South African youths aged 15–24 face marked challenges in achieving the national targets of 90% tested, 90% on ART, and 90% virally suppressed by 2020. Adolescent girls have the highest HIV incidence of any demographic group.3 In addition, the HIV disease burden previously noted in vertically-infected children has been predicted to shift to adolescent age groups as widespread access to paediatric antiretroviral treatment (ART) reduces mortality and slow-progressors survive untreated. 4,5 A recent global meta-analysis estimated median age at ART initiation of 7.9 years among perinatally-infected children. 6 There is also evidence suggesting adolescents in Southern Africa are a vulnerable group with unique challenges in accessing testing, 7,8 timely initiation of antiretroviral therapy and successful adherence to treatment and retention in HIV care programs. 7–12
The ability to monitor trends in adolescent care-seeking has been limited by the absence of national longitudinal data. In partnership with South Africa’s National Health Laboratory Service (NHLS), we created a National HIV Cohort through novel record linkage of the complete laboratory records of the national HIV program13. This longitudinal database includes all CD4 counts and HIV viral loads (VL) of all children and adolescents accessing public sector HIV care and treatment. Analyzing these data, we: (1) quantify increases in number of children and adolescents ages 1-19 years accessing HIV care and treatment, 2005-2016, (2) characterize trends in the age distribution of this population over time with respect to two distinct sub-groups: patients who entered HIV care between the ages of 1-14 years (likely perinatal mother-to-child transmission) and patients who entered HIV care aged between 15-19 years (likely adolescent sexual transmission); and (3) estimate proportions of patients in care who are on ART for each of these sub-populations. Finally, we project numbers of adolescents on ART through 2021 based on current trends.
Methods
Data sources
NHLS conducts all laboratory monitoring for South Africa’s national public sector HIV program. The NHLS Corporate Data Warehouse (CDW) database includes all CD4 counts and viral loads dating back prior to large scale-rollout of HIV in 2004 (with the exception of the province of KwaZulu-Natal which contributed data from 2010 onwards). Data are captured from each individual laboratory requisition form at the time a blood sample is taken and include demographic information such as name, surname, gender and date of birth. As the CDW database lacks a routinely-collected unique patient identifier, we developed and validated a linkage algorithm to identify unique patients (and their associated laboratory results) through the demographic data available14–15 using traditional deterministic and probabilistic as well as network-based linkage methodology.16, 17 The algorithm was developed and validated for application to the approximately 54 million CD4 counts and viral load tests in the NHLS during the period 2004-2016 and attained sensitivity of 94%, specificity of close to 100% and positive predictive values of 99% relative to manually matched data. This linkage enables analysis of the NHLS database as a National HIV Cohort, with longitudinal follow-up of all lab-monitored patients (including adolescents) accessing care in South Africa’s public sector HIV program.
Study design and population
The study population for this analysis included all children and adolescents (aged 1–19 years) who entered HIV care in South Africa’s public sector with a CD4 count or viral load between 1 January 2005 and 31 December 2016 (and from 2011 in KwaZulu Natal (KZN) province). Lab records for children <1 year of age were excluded as infants are often identified in the NHLS database using information about the mother, reducing linkage accuracy. We excluded patients whose first CD4 or VL date was prior to the study period (2005) or prior to 2011 for KZN. We additionally excluded patients whose first recorded test was a suppressed VL, as they most likely started ART in the private sector before transferring into the national program or had a prior CD4 count that was not correctly matched to them.
This study involved secondary analysis of de-identified data collected as part of routine care. Analysis was approved by the Human Research Ethics Committee of the University of the Witwatersrand (M150429) and Boston University Medical Campus Institutional Research Board (H-31968).
Variable definitions
Key steps of the care cascade can be imputed from laboratory data based on national treatment guidelines. During the study period, CD4 counts were assessed at clinical presentation in order to determine eligibility for ART initiation. VL monitoring was indicated for all patients on ART with the first VL at initiation (pre-2008) or at six months after initiation (2008 and later).18 During the study period, patients in care but not yet on ART should have had at least one CD4 count per year and patients on ART should have had at least one VL per year.
We defined “entry into HIV care” as the date of the first CD4 count or viral load test result in the NHLS database (as this would signify, at a minimum, accessing HIV testing and blood tests to determine eligibility for ART initiation). We defined “currently in HIV care” as having at least one CD4 count or viral load test result in the NHLS database in a given year and “currently accessing ART” was defined as the presence of a VL in a given year. We additionally distinguished between whether a person “newly initiated ART” in a given year or whether they were “continuing on ART”, having been on ART the prior year.
We defined “probable mode of transmission” based on age at entry into HIV care, with patients entering care as children (1-14 years) most likely “vertically infected” via perinatal mother-to-child transmission and patients who entered care as older adolescents (15-19 years) most likely “horizontally infected” via sexual transmission. To evaluate the accuracy of using age of entry to define these categories, we computed the gender ratio among patients entering care at each age. Infant girls are more at risk of vertical infection19, 20 and adolescent girls are more likely to become newly infected with HIV than boys.21, 22
Statistical analysis
Patient characteristics were summarized using simple proportions and medians with interquartile ranges (IQR), stratified by probable mode of infection. We then assessed changes in the composition of the population in care and on ART over time. We computed numbers of patients newly initiating and continuing on ART each year, 2005-2016, stratifying by probable mode of infection. We also assessed the age distribution of children and adolescents on ART at three specific calendar years: 2008, 2012 and 2016, chosen to represent an early, middle, and late cross-section in the scale-up of treatment in South Africa. Finally, we projected numbers of adolescents (15-19 years) on ART through 2021 based on past trends, under three scenarios: assuming no improvement in ART uptake or retention; assuming improvement in retention only; and assuming improvement in both ART uptake and retention (see Appendix A for details). Analyses were performed using SAS statistical software (version 9.4) and Microsoft Excel (version 16.16.5).
Role of the funding source
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results
After exclusions (Figure 1), the NHLS National HIV Cohort included 730,882 patients aged 1-19 years who entered South Africa’s public-sector HIV program with a first CD4 count or HIV viral load during 2005-2016 (Table 1). These patients had 2.6 million CD4 counts and 2.2 million viral load results during the study period. Adolescents aged 15-19 years comprised nearly half (47%, n=342,443) of the study population.
Figure 1:
Flow chart of patients included in analysis
Table 1:
Characteristics of 730,882 persons 1-19 year of age at entry to HIV care in South Africa from 2005-2016
| Probable Vertical Infections | Probable Horizontal Infections | ||||
|---|---|---|---|---|---|
| Age of entry into care | 1-4 years n=166,211 | 5-9 years n=125,012 | 10-14 years n=97,216 | 15-19 years n=342,443 | |
| Gender | Female, N (%) | 86,256 (52%) | 66,739 (53%) | 56,210 (58%) | 301,242 (88%) |
| Male | 79,955 (48%) | 58,273 (47%) | 41,006 (42%) | 41,201 (12%) | |
| CD4 count at entry to care (cells/mm3) | Median (IQR) | 791 (434-1319) | 478 (255-785) | 354 (168-598) | 409 (265-585) |
| 0-100, N (%) | 7,640 (5%) | 13,179 (11%) | 15,826 (16%) | 22,395 (7%) | |
| 101-200 | 8,363 (5%) | 10,934 (9%) | 12,358 (13%) | 31,629 (9%) | |
| 201-350 | 16,006 (10%) | 20,411 (16%) | 19,656 (20%) | 81,818 (24%) | |
| >350 | 134,202 (81%) | 80,488 (64%) | 49,376 (51%) | 206,601 (60%) | |
Numbers entering into care increased steadily each calendar year until 2010, when they began to decline (Table 2). The share of adolescents (15-19 years) in the total population (1-19 years) presenting for care increased over time, from 32% in 2005 to 55% in 2016, reflecting increases in adolescent care-seeking as well as large declines in new patients entering care at younger ages, coinciding with the successful scale-up of PMTCT.
Table 2:
Numbers of 1-19 year-olds entering HIV care in South Africa stratified by calendar year
| Probable Vertical Infections | Probable Horizontal Infections | ||||
|---|---|---|---|---|---|
| Age of entry into care | 1-4 years n=166,211 | 5-9 years n=125,012 | 10-14 years n=97,216 | 15-19 years n=342,443 | |
| Year at entry to care | 2005, N (%) | 8,618 (34%) | 6,254 (24%) | 2,384 (9%) | 8,278 (32%) |
| 2006 | 12,213 (33%) | 8,036 (22%) | 3,569 (10%) | 13,164 (36%) | |
| 2007 | 14,321 (33%) | 8,767 (20%) | 4,337 (10%) | 16,459 (38%) | |
| 2008 | 15,741 (29%) | 9,743 (18%) | 5,770 (11%) | 22,305 (42%) | |
| 2009 | 15,642 (28%) | 9,751 (17%) | 6,262 (11%) | 24,429 (44%) | |
| 2010 | 15,771 (25%) | 10,531 (17%) | 7,622 (12%) | 28,803 (46%) | |
| 2011* | 19,842 (20%) | 18,574 (19%) | 15,095 (15%) | 44,793 (46%) | |
| 2012 | 15,710 (19%) | 13,871 (17%) | 12,232 (15%) | 40,408 (49%) | |
| 2013 | 13,367 (19%) | 11,774 (16%) | 10,431 (14%) | 36,525 (51%) | |
| 2014 | 12,521 (18%) | 10,390 (15%) | 10,085 (15%) | 35,942 (52%) | |
| 2015 | 11,864 (17%) | 9,409 (14%) | 10,204 (15%) | 37,325 (54%) | |
| 2016 | 10,601 (17%) | 7,912 (13%) | 9,225 (15%) | 34,012 (55%) | |
(Note data from KwaZulu Natal included only from 2011 onwards); data include all children and adolescents with at least one CD4 count or viral load test.
Gender composition differed by age at entry into care, likely reflecting different modes of infection (Figure 2). Females comprised just over half of patients entering care before age 10, 58% of those entering care at 11-14 years, and 88% of those entering care at 15-19 years. These patterns support our choice of age 15 at entry into care as a cut-off for probable vertical versus probable horizontal infection.
Figure 2:
Gender distribution by age in years at entry to HIV care program
Next, we stratified the age distribution of patients in care in 2016 in South Africa’s public sector HIV program by probable mode of infection (Figure 3). The peak of the age distribution at 10-14 years among children seeking care prior to age 15 is consistent with a large cohort of children vertically infected in 2002-2006, a time when PMTCT had not fully scaled up but when survival was increasing due to the rollout of pediatric ART. We note that in 2016, this vertically-infected population was just entering adolescence. The peak of the distribution is expected to shift to the 15-19-year age group over the next five years and then decrease.
Figure 3:
National distribution of current age among children and adolescents in care in 2016 in South Africa’s public sector HIV program stratified by probable mode* of infection.
*Note: age <15 years at entry to care classified as probable vertical/perinatal transmission; age 15+ years at entry to care classified as probable incident horizontal infection. The figure depicts the current age distribution of all patients in care in 2016 who entered care aged 1-19 years at any point during the period 2005-2016.
Among patients ages 1-19 years who entered HIV care 2005-2016, a total of 415,882 patients (57%) initiated ART as indicated by record of at least one VL in the NHLS database (Table 3). When stratified by age category at entry to care, the proportions initiating ART were 69% of boys and 68% of girls entering at ages 1-4 years, 72% of boys and 68% of girls entering at ages 5-9 years, and 68% of boys and 62% of girls entering at ages 10-14 years. Among patients entering care as older adolescents (15-19 years), the proportion of those initiating ART dropped for both girls (45%) and boys (42%).
Table 3:
Numbers of patients on antiretroviral therapy aged 1-19 years in South Africa 2005-2016 stratified by calendar period and age category at entry to care
| Probable vertical infections | Probable horizontal infections | ||||
|---|---|---|---|---|---|
| AGE AT ENTRY INTO CARE | 1-4 years n=166,211 | 5-9 years n=125,012 | 10-14 years n=97,216 | 15-19 years n=342,443 | |
| Initiated*on ART 2005-2016 (all available data) | Total ART initiators | 114,187/166,211 (69%) | 87,632/125,012 (70%) | 62,526/97,216 (64%) | 151,537/342,443 (44%) |
| Total Male ART initiators | 55,555/79,955 (69%) | 41,943/58,273 (72%) | 27,849/41,006 (68%) | 17,265/41,201 (42%) | |
| Female ART initiators | 58,632/86,256 (68%) | 45,689/66,739 (68%) | 34,677/56,210 (62%) | 134,272/301,242 (45%) | |
| Total initiated on ART 2005-2008 | 25,803/50,893 (51%) | 17,171/32,800 (52%) | 7,130/16,060 (44%) | 7,949/60,206 (13%) | |
| Total initiated on ART 2009-2012 | 43,415/66,965 (65%) | 33,475/52,727 (63%) | 21,912/42,211 (52%) | 28,688/138,433 (21%) | |
| Total initiated on ART 2013-2016 | 44,969/48,353 (93%) | 36,986/39,485 (94%) | 33,484/39,945 (84%) | 114,900/143,804 (79%) | |
| AGE AT ENTRY TO CARE | 1-4 years n=141,105 | 5-9 years n=99,196 | 10-14 years n=72,580 | 15-19 years n=264,399 | |
| Initiated*on ART 2005 – 2016 (excluding KZN data) | Total initiated on ART 2005-2016 | 94,483/141,105 (67%) | 68,373/99,196 (69%) | 45,028/72,580 (62%) | 112,385/264,399 (43%) |
| Total initiated on ART 2005-2008 | 25,803/50,893 (51%) | 17,171/32,800 (52%) | 7,130/16,060 (44%) | 7,949/60,206 (13%) | |
| Total initiated on ART 2009-2012 | 36,224/55,496 (65%) | 25,308/39,147 (65%) | 15,291/29,521 (52%) | 23,518/108,525 (21%) | |
| Total initiated on ART 2013-2016 | 32,456/34,716 (93%) | 25,894/27,249 (95%) | 22,607/26,999 (84%) | 80,918/95,668 (72%) | |
ART initiator defined as having at least one viral load test observed during the period of entry to care as specified
The age distribution of children and adolescents on ART has changed over time in two important ways (Figure 4). First, the absolute number of children and adolescents on ART increased substantially from 2008 to 2012 to 2016 with expansion of the national treatment program. Second, there was a rightward shift in the age distribution as vertically-infected patients aged and as the numbers of horizontally-infected patients seeking care and starting ART increased.
Figure 4:
National distribution of age among children and adolescents on ART in South Africa’s public sector HIV program, stratified by calendar year
These two trends have resulted in very large increases in the number of adolescents aged 15-19 years on ART (Table 3). Contrasting the period 2005-2008 with the period 2013-2016 (and excluding KZN data to ensure comparability), the number of ART patients ages 1-4 increased by 25% from 25,803 to 32,456. Over the same period, the number of ART patients ages 15-19 grew from 7,949 patients on ART in 2005-2008 to 80,918 on ART during 2013-2016 – a 10-fold increase in the number of adolescents on ART. Focusing just on new ART initiates, the number of patients aged <15 years starting ART rose steadily from 6,262 to 16,875 new initiates in 2010, and then declined (Table 4). Comparatively, numbers of new ART initiates among patients entering care at ages 15-19 years (presumably horizontally infected) increased from 519 in 2005 to 7,213 in 2010 – and then continued to rise, to 39,053 in 2016. By 2016, adolescents aged 15-19 (presumably horizontally infected) alone accounted for >40% of all new ART initiators aged 1-19 years (Table 4).
Table 4:
Prevalent numbers on ART stratified by calendar year, probable mode of infection and age category
| Probable vertical infection: Age at entry to care 1-14 years | Probable horizontal infection: Age at entry to care 15+ years | |||||||
|---|---|---|---|---|---|---|---|---|
| Age at test: 1-14 years | Age at test: 15-19 years | Age at test: 15-19 years | ||||||
| Proportion on ART of those currently in care# | New ART Initiation* | Continuing on ART** | New ART Initiation* | Continuing on ART** | Proportion on ART of those currently in care# | New ART Initiation* | Continuing on ART** | |
| 2005 | 6,262/17,256 (36%) | 6262 | 0 | 0 | 0 | 519/8,278 (6%) | 519 | 0 |
| 2006 | 16,144/32,228 (50%) | 12227 | 3869 | 23 | 25 | 1,761/15,288 (12%) | 1562 | 199 |
| 2007 | 25,270/45,780 (55%) | 14432 | 10604 | 60 | 174 | 3,033/21,753 (14%) | 2291 | 742 |
| 2008 | 35,655/60,280 (59%) | 16986 | 18027 | 114 | 528 | 8,180/32,126 (25%) | 3577 | 1570 |
| 2009 | 42,111/71,756 (59%) | 14782 | 25965 | 131 | 1233 | 6,210/39,519 (16%) | 3525 | 2685 |
| 2010 | 50,057/84,680 (59%) | 16875 | 30894 | 209 | 2079 | 7,213/52,269 (14%) | 3784 | 3429 |
| 2011 | 74,419/115,745 (64%) | 33376 | 36991 | 495 | 3557 | 13,310/78,177 (17%) | 8422 | 4888 |
| 2012 | 90,106/124,688 (72%) | 32057 | 51169 | 877 | 6003 | 21,722/87,840 (25%) | 12957 | 8765 |
| 2013 | 103,520/132,544 (78%) | 30446 | 62144 | 1211 | 9719 | 53,337/96,377 (55%) | 16919 | 14696 |
| 2014 | 115,060/140,126 (82%) | 28672 | 70124 | 1438 | 14826 | 45,942/109,222 (42%) | 23263 | 22679 |
| 2015 | 126,625/150,048 (84%) | 26052 | 76863 | 1921 | 21789 | 73,317/126,540 (58%) | 37752 | 35565 |
| 2016 | 135,058/155,229 (87%) | 23612 | 80794 | 2087 | 28565 | 92,783/140,028 (66%) | 36966 | 55817 |
Currently in care population defined as number of patients with at least one CD4 count or viral load observed during the specified calendar year.
New ART initiator defined as first VL observed in database during that calendar year
Continuing on ART defined as any viral load subsequent to the first viral load observed during year after the initiating year
We estimated projected numbers of adolescents aged 15-19 years on ART in the public sector through 2021, based on the further aging of vertically-infected children into the 15-19-year age group and assumptions about care seeking, ART initiation, and retention in different age groups (Figure 5). Based on existing trends, the number of adolescents on ART is expected to increase substantially. And if both uptake of ART and retention on ART increased to their respective 90% targets, the number of adolescents on ART aged 15-19 years could nearly double in the 5 years between 2016 and 2021.
Figure 5:
Projected estimated numbers on ART* among adolescents aged 15-19 years
- The solid black line above the columns for 2005-2016 reflects observed trends in numbers on ART
- All data beyond 2016 are projections of numbers on ART based on assumptions in the text:
- Columns from 2016-2021 reflects projected numbers on ART stratified by mode of infection assuming no improvement in ART uptake or retention
- Dashed line from 2016-2021 reflects projected numbers on ART assuming 90% annual rate of retention
- Dashed/dotted line from 2016-2021 reflects projected numbers on ART assuming 90% uptake of ART and annual rate of retention at 90%
Discussion
Our results suggest three key findings. First, the number of adolescents (15-19 years) on ART in South Africa increased 10-fold between 2005-2008 and 2013-2016 and is expected to continue to grow. Second, this adolescent treatment bulge is due to the confluence of two factors: 1) the aging of perinatally-infected infants who entered HIV care in childhood; and 2) rising numbers of older adolescents (15-19 years) seeking care for the first time and who were likely infected through sexual transmission. Finally, our analysis also found low rates of ART initiation (<50%) among 15-19 year olds seeking care, suggesting substantial scope to improve HIV care and treatment services for adolescents.
The success of past HIV programs is reflected in the bulge of adolescents accessing HIV care in South Africa. First, the increase in the absolute number of adolescents on ART confirms that, at least in part, the changes in eligibility criteria to include less-severely immune-compromised children and adolescents living with HIV have resulted in the expansion of the HIV treatment program over time. Second, the success of paediatric screening 23 and subsequent ART initiation among perinatally-infected children has dramatically increased survival of HIV-infected children. Confirming a recent meta-analysis, 6 we found that many children enter care only in late childhood, having survived for many years without ART. Still, few would be expected to survive into adolescence without paediatric ART. Third, successful implementation of PMTCT has led to a large overall decrease in numbers of perinatally-infected children over time. Third, increases in numbers of adolescents entering care and accessing ART in the age group 15-19 are driven by two distinct groups: one group entering care <15 years of years who were most likely perinatally-infected children who have survived into adolescence; as well as a second group entering care at 15 years or older. The increases in this group are likely influenced by 1) new horizontal infections, reflected in the sharply contrasting gender distribution in the age group as well as the recent surveys confirming this age group to have the highest annual incidence rates of HIV in South Africa; 3 or 2) successful implementation of the 2010 testing campaigns in this age group (though low rates of testing in this age group estimated in recent surveys would suggest otherwise) or 3) expansion of the coverage of ART in this age group as eligibility thresholds for initiation of ART has increased over time.
However, these successes should not overshadow the persistence of high incidence of sexual transmission of HIV among adolescent girls, which has also contributed to the adolescent treatment bulge. Females were over-represented in all age groups and represented 88% of patients entering care at age 15-19 years in our data. These results are in keeping with increased risk of vertical transmission of HIV for female infants19, 20 as well as patterns of horizontal transmission already described in the 2017 HSRC survey and elsewhere in sub-Saharan Africa,21, 24 where young women have higher incidence than young men. Though we are not able to differentiate fully how much this is a change in incidence in these groups or differences in access to testing or care-seeking behavior for girls either independently or through pregnancy and PMTCT-related services, it is likely a composite of these. There remains an urgent need to address the inequalities in education, sexual and reproductive health, poverty and food security that impact on a young woman’s risk of acquiring HIV, particularly during this key adolescent period, and to ensure that young men who are HIV-infected seek care.
In addition, low rates of ART initiation after seeking care signal that current models of HIV care are not meeting the unique needs of adolescents. We observed only two thirds (66%) of 15-19-year-old adolescents (presumably horizontally infected) who entered an HIV care program as having initiated ART by 2016. In contrast, 87% of 1-14 year olds (presumably vertically infected) initiated ART after seeking care by 2016; rates of ART initiation comparable to those described previously in sub-Saharan Africa25 among those seeking care. The proportion of presumably horizontally infected adolescents aged 15-19 observed initiating ART did increase sharply during the 2013-2016 period; possibly related to corresponding changes in guidelines increasing eligibility thresholds and also fast-tracking of certain groups onto treatment. Despite this positive upswing in numbers initiating ART, several factors persist that can negatively contribute to low rates of uptake of ART observed in the 15-19-year-old adolescent group. These include low testing rates (estimated at 50% and 46% 26 for female and male 15-19 year-olds, respectively – the lowest testing rates for any age group in South Africa), frequent clinic visits to establish and maintain ART, issues of disclosure of HIV status and increasing responsibilities at home particularly among those orphaned by HIV.27
Our findings highlight the importance of preparing South Africa’s health systems to deliver effective HIV care to rising numbers of adolescents. Several factors will impact on the numbers of adolescents accessing HIV care in the coming years. On the one hand, the effective implementation of PMTCT since 2003/4 will result in fewer vertically-infected adolescents in the future. Further, the growth in the number of 15-19 year olds in care will begin to slow if increases in HIV testing and ART uptake among adolescents begin to slow, either due to saturation or program failures. If, on the other hand, South Africa reaches 90-90-90 targets, our projections indicate that, the adolescent population on ART could nearly double in the next five years. This expansion could have a powerful impact in breaking HIV transmission chains among adolescents, leading ultimately to lower ART burdens in the long run. If, however, adolescents continue to be the fastest growing group of people living with HIV, while also the least likely to start ART timeously and therefore fall behind on progress towards the second 90 (providing ART to 90% of those diagnosed with HIV), achieving the full 90-90-90 targets nationally in South Africa remains unlikely. Health systems should be evaluated to determine how the unmet needs of adolescents, particularly horizontally-infected, could be addressed through interventions in service delivery among this growing population. Differentiated models of care have shown promise in improving outcomes for adolescents living with HIV 28, 29 and interventions to increase uptake of testing and treatment encompassing both facility-based and self-testing, as well as provision of adolescent-friendly services such as youth clubs and convenient clinic times should be considered.
While our analysis is the largest to date among South African adolescents in HIV care, there are some limitations that should inform the interpretation of these results. First, though extensively validated record linkage techniques were used, the process may still lead to errors in matching at the patient level so that the possibility exists that incorrect links have been made (i.e. overmatching) or that actual links have been missed (i.e. under-matching). In attempts to quantify the extent of over- or under-matching, we validated the linkage algorithm relative to a representative sample of manually-coded data as well as relative to national ID numbers available for a small, non-representative proportion of laboratory specimens. The algorithm had high sensitivity (94%) and positive predictive value (99%) relative to the manually-matched quasi-gold standard and 99% sensitivity relative to national ID numbers, where available.13 Second, as described in the methods section, our definitions of events in the care cascade are not based on actual clinical visit dates, but rather are inferred from routine laboratory testing protocols. As such, a patient whose blood samples were not taken or not processed, would not be observed in our dataset. The likelihood of such underestimation of those in care and on ART may vary by facility type and age group (lower tier facilities with fewer skilled pediatric staff may experience lower coverage of routine blood sampling). Third, we do not directly observe mode of transmission. Although our data supported the age of 15 years as a reasonable cut-off to distinguish between sexual and perinatal modes of transmission, horizontal infections prior to age 15 would be misclassified as “presumably vertically infected” in this analysis. Finally, there are some groups not observed directly in the data presented here. In particular, we do not report on lab results of infants <1 year of age, patients accessing care in the private sector, laboratory monitoring done outside of the NHLS among patients who migrate into or out of the South African public-sector HIV care program and, of course, those who do not test and do not enter the treatment program will also not be observed. In addition, KwaZulu-Natal is not observed prior to 2011. However, the proportions observed initiating ART by age category and mode of transmission do not change substantively with the inclusion of the data from KZN (Table 3).
Despite these limitations, the analysis has several strengths. First, the NHLS database is national in scope, covering all children and adolescents seeking care in the public sector HIV program with a CD4 count or VL. Second, the laboratory data are very high quality. The data come directly from NHLS and do not depend on facility-level staff copying results into patient charts, nor on research staff extracting data from charts – both of which are opportunities for data loss and error. Third, record linkage process used to construct the NHLS National HIV Cohort occurs at a national level (not just district or clinic level), leading to far greater deduplication than other available datasets. Because all laboratory tests are observed for individual patients regardless of where they seek care, the dataset is robust to so-called “silent transfers” where a patient’s undocumented transfer to another facility resulting in potential double-counting and misclassification of the patient’s status – recorded as lost to follow up at the original facility and a new initiate at the receiving facility. Fourth, the Cohort enables identification of key steps of the cascade, including entry to care among non-ART initiators, who are often excluded from existing datasets. As we are able to observe the first CD4 count or viral load test recorded in the database regardless of whether the patient returned for care, we can quantify the population entering care independent of retention in care thereafter and probabilities of being on ART conditional on being in care. Fifth, the longitudinal dimension of the Cohort enables identification of age at which patients sought care - regardless of where they sought care within the public sector system – which allows us to infer probable route of transmission, something prior studies have not been able to do at scale. Sixth, the ability to observe exact age at each laboratory sample allows for finer stratification by age group – in particular we are able to observe those aged 15-19 years – a group highly pertinent to the adolescent population in SA but not reported on directly in the DHIS datasets.
Large-scale HIV treatment for adolescents is an important recent phenomenon in South Africa. Adolescent treatment programs are forecast to grow in the coming years, highlighting the need to adequately prepare for and allocate resources to this high priority population. Programs and interventions directly addressing the unique social and developmental challenges faced by adolescents should be prioritized. Increasing testing and ART uptake among adolescents will be critical to meeting 90-90-90 targets, breaking transmission cycles, and reaching epidemic control.
Supplementary Material
Research in context.
Evidence before this study
We conducted two searches on the PubMed database for articles published between January 1st 2004 and April 30th 2019 using the search terms (1) “HIV”, “adolescents” “treatment outcomes” and “South Africa” and (2) “HIV”, “adolescents” “transmission” “South Africa”. There is growing awareness of the importance of adolescent HIV infected populations in South Africa and their key role in national epidemic control, particularly among young females. The 2017 HSRC South African National HIV Prevalence, Incidence, Behavior and Communication survey estimated that while 5.8% of the 2.5 million girls aged 15-19 in South Africa were HIV infected, the prevalence increased to 15.6% among 20-24 year-old females. However, less is known about adolescent care-seeking patterns as few large-scale cohorts are available. Recently, data from a newly-linked HIV cohort developed from South Africa’s National Health Laboratory Service (NHLS) database has been used to describe care-seeking and retention in HIV care programs among adult HIV-infected patients accessing care in South Africa’s national treatment program.
Added value of this study
We describe and quantify the adolescent population accessing care in South Africa, using the complete lab records from South Africa’s public sector treatment program. Our results show a ten-fold increase in numbers of adolescents aged 15-19 years on HIV treatment in the past decade. This increase is due to the survival of perinatally-infected children on ART and an increase in care-seeking among adolescent girls and young women. Finally, we identify low proportions progressing to treatment initiation after entering HIV care among 15-19 year olds.
Implications of all the available evidence
Increased HIV care-seeking among adolescents is encouraging, provided the heath system is sufficiently prepared to meet the needs of this growing population. However, low rates of successful ART initiation among 15-19 year olds entering care suggest an urgent need to improve services for adolescents.
Acknowledgements
Role of the Funding Source
This work was supported by grants 1R01AI115979-01 and 1K01MH105320-01A1 from the National Institutes of Health. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This study was also supported by the generous support of the American People and the President’s Emergency Plan for AIDS Relief (PEPFAR) through USAID under the terms of Cooperative Agreements AID-674-A-12-00029 and 72067419CA00004 to HE2RO. WBM and SC were supported by USAID through Cooperative Agreement AID- 674-A-12-0020 from the United States Agency for International Development (USAID). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The contents are the responsibility of the authors and do not necessarily reflect the views of PEPFAR, USAID or the United States Government.
Funding
National Institutes of Health, the American People and the President’s Emergency Plan for AIDS Relief (PEPFAR) through USAID.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data Sharing statement
Access to primary data is subject to restrictions owing to privacy and ethics policies set by the South African Government. Requests for access to the data can be made to the National Health Laboratory Services directly (http://www.nhls.ac.za/) and require a full protocol submission. Inquiries can be made via the Office of Academic Affairs and Research at NHLS.
Competing interests
The authors have declared that no competing interests exist.
Contributor Information
Mhairi Maskew, Health Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, University of the Witwatersrand, Johannesburg, South Africa.
Jacob Bor, Health Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, University of the Witwatersrand, Johannesburg, South Africa; Department of Global Health, Boston University School of Public Health, Boston University, Boston, United States; Department of Epidemiology, Boston University School of Public Health, Boston University, Boston, United States.
William MacLeod, Health Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, University of the Witwatersrand, Johannesburg, South Africa; Department of Global Health, Boston University School of Public Health, Boston University, Boston, United States.
Sergio Carmona, National Health Laboratory Service, Johannesburg, South Africa.
Gayle G. Sherman, Department of Paediatrics and Child Health, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; National Institute for Communicable Disease, a division of the National Health Laboratory Service
Matthew P. Fox, Health Economics and Epidemiology Research Office, Department of Internal Medicine, School of Clinical Medicine, University of the Witwatersrand, Johannesburg, South Africa; Department of Global Health, Boston University School of Public Health, Boston University, Boston, United States; Department of Epidemiology, Boston University School of Public Health, Boston University, Boston, United States
REFERENCES
- 1.Oliveira T de, Kharsany ABM, Gräf T, Cawood C, Khanyile D, Grobler A, et al. Transmission networks and risk of HIV infection in KwaZulu-Natal, South Africa: a community-wide phylogenetic study. Lancet HIV 2017; 4(1): e41–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Statistics South Africa (STATS SA). Demographic Profile of Adolescents in South Africa. 2018. Available from: http://www.statssa.gov.za/publications/Report03-00-10/Report03-00-102016.pdf Accessed 1st March 2019. [Google Scholar]
- 3.Human Sciences Research Council. The Fifth South African National HIV Prevalence, Incidence, Behaviour and Communication Survey, 2017. 2018. Available from: http://www.hsrc.ac.za/uploads/pageContent/9234/SABSSMV_Impact_Assessment_Summary_ZA_ADS_cleared_PDFA4.pdf Accessed 1st March 2019 [Google Scholar]
- 4.Johnson LF, Davies M-A, Moultrie H, Sherman GG, Bland RM, Rehle TM, et al. The Effect of Early Initiation of Antiretroviral Treatment in Infants on Pediatric AIDS Mortality in South Africa. Pediatr Infect Dis J 2012; 31(5): 474–80. [DOI] [PubMed] [Google Scholar]
- 5.Ferrand RA, Corbett EL, Wood R, Hargrove J, Ndhlovu CE, Cowan FM, et al. AIDS among older children and adolescents in Southern Africa: projecting the time course and magnitude of the epidemic. AIDS 2009; 23(15): 2039–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Slogrove AL, Schomaker M, Davies M-A, Williams P, Balkan S, Ben-Farhat J, et al. The epidemiology of adolescents living with perinatally acquired HIV: A cross-region global cohort analysis The Collaborative Initiative for Paediatric HIV Education and Research (CIPHER) Global Cohort Collaboration. PLoS Med 2018; 15: 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Govindasamy D, Ferrand RA, Wilmore SMS, Ford N, Ahmed S, Afnan-Holmes H, et al. Uptake and yield of HIV testing and counselling among children and adolescents in sub-Saharan Africa: A systematic review. J Int AIDS Soc 2015;18(1): 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brown K, Williams DB, Kinchen S, Saito S, Radin E, Patel H, et al. Status of HIV Epidemic Control Among Adolescent Girls and Young Women Aged 15–24 Years — Seven African Countries, 2015–2017. MMWR Morb Mortal Wkly Rep 2018; 67(1): 29–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nglazi MD, Kranzer K, Holele P, Kaplan R, Mark D, Jaspan H, et al. Treatment outcomes in HIV-infected adolescents attending a community-based antiretroviral therapy clinic in South Africa. BMC Infect Dis 2012; 12(1): 21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Nachega JB, Hislop M, Nguyen H, Dowdy DW, Chaisson RE, Regensberg L, et al. Antiretroviral therapy adherence, virologic and immunologic outcomes in adolescents compared with adults in southern Africa. J Acquir Immune Defic Syndr 2009; 51(1): 65–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bygrave H, Mtangirwa J, Ncube K, Ford N, Kranzer K, Munyaradzi D. Antiretroviral therapy outcomes among adolescents and youth in rural Zimbabwe. PLoS One 2012; 7(12): e52856. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bock P, Boulle A, White C, Osler M, Eley B. Provision of antiretroviral therapy to children within the public sector of South Africa. Trans R Soc Trop Med Hyg 2008; 102(9): 905–11. [DOI] [PubMed] [Google Scholar]
- 13.Bor J, MacLeod W, Oleinik K, Potter J, Brennan A, Candy S, et al. Building a National HIV Cohort from Routine Laboratory Data: Probabilistic Record-Linkage with Graphs. bioRxiv 2018; doi:(10.1101):450304. Available from: https://www.biorxiv.org/content/early/2018/11/02/450304 Accessed 4th December 2018. [Google Scholar]
- 14.Carmona S, Bor J, Nattey C, Maughan-Brown B, Maskew M, Fox MP, et al. Persistent High Burden of Advanced HIV Disease among Patients Seeking Care in South Africa’s National HIV Program: Data from a Nationwide Laboratory Cohort. Clin Infect Dis. 2018; 66(supplement 2). S111–S117. doi: 10.1093/cid/ciy045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fox MP, Bor J, Brennan AT, MacLeod WB, Maskew M, Stevens WS, et al. Estimating retention in HIV care accounting for patient transfers: A national laboratory cohort study in South Africa. PLoS Med. 2018; 15(6). e1002589. doi. 10.1371/journal.pmed.1002589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fellegi IP, Sunter AB. A Theory for Record Linkage. J Am Stat Assoc 1969; 64(328): 1183–210. [Google Scholar]
- 17.Jaro MA. Probabilistic linkage of large public health data files. Stat Med 1995; 14(5–7): 491–8. [DOI] [PubMed] [Google Scholar]
- 18.South Africa National Department of Health. South Africa Antiretroviral Treatment Guidelines. Africa: 2004. Available at http://apps.who.int/medicinedocs/documents/s17758en/s17758en.pdf Accessed 4th December 2018. [Google Scholar]
- 19.Mwau M, Bwana P, Kithinji L, Ogollah F, Ochieng S, Akinyi C, et al. Mother-to-child transmission of HIV in Kenya: A cross-sectional analysis of the national database over nine years. PLoS ONE 2017; 12(8): 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Taha TE, Nour S, Kumwenda NI, Broadhead RL, Fiscus SA, Kafulafula G. Gender Differences in Perinatal HIV Acquisition Among African Infants. Paediatrics 2005; 115(2): e167–172. [DOI] [PubMed] [Google Scholar]
- 21.Gregson S, Nyamukapa CA, Garnett GP, Mason PR, Zhuwau T, Caraël M, et al. Sexual mixing patterns and sex-differentials in teenage exposure to HIV infection in rural Zimbabwe. Lancet 2002; 359(9321): 1896–903. [DOI] [PubMed] [Google Scholar]
- 22.UNAIDS. Global AIDS Update. 2016. https://www.unaids.org/sites/default/files/media_asset/global-AIDS-update-2016_en.pdf Accessed 4th December 2018.
- 23.Massyn N, Pillay Y, Padarath A. District Health Barometer 2017/2018. Health Systems Trust. 2019. Available at: https://www.hst.org.za/publications/District%20Health%20Barometers/DHB+2017-18+Web+8+Apr+2019.pdf Accessed 1st March 2019. [Google Scholar]
- 24.UNAIDS. Global AIDS epidemic update. 2016. [Google Scholar]
- 25.Rosen S, Fox MP. Retention in HIV Care between Testing and Treatment in Sub-Saharan Africa: A Systematic Review. PLoS Med. 2011; 8(7). e1001056. doi. 10.1371/journal.pmed.1001056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.National Department of Health, Statistics South Africa (STATS SA), South African Medical Research Council, ICF. South Africa Demographic and Health Survey 2016. 2016. Available at: https://www.statssa.gov.za/publications/Report%2003-00-09/Report%2003-00-092016.pdf Accessed 1st March 2019.
- 27.Lowenthal ED, Bakeera-Kitaka S, Marukutira T, Chapman J, Goldrath K, Ferrand R a. Perinatally acquired HIV infection in adolescents from sub-Saharan Africa: a review of emerging challenges. Lancet Infect Dis 2014; 14(7): 627–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Grimsrud A, Walker D, Ameyan W, Brusamento S for the Child Survival Working Group: IAS, UNICEF, WHO Providing differentiated delivery to children and adolescents. Approaching 2020: Scaling up key interventions for children and adolescents living with HIV 2018; 1–4.Available from: https://www.who.int/hiv/pub/paediatric/diff-delivery-children-hiv/en/ Accessed 10 June 2019.
- 29.Mackenzie RK, Lettow M Van, Gondwe C, Nyirongo J, Singano V, Banda V, et al. Greater retention in care among adolescents on antiretroviral treatment accessing “Teen Club” an adolescent-centred differentiated care model compared with standard of care : a nested case – control study at a tertiary referral hospital in Malawi. JIAS 2017; 20: e25028. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.