Lower self-reported ART adherence among adolescents in boarding schools compared to day schools

Abstract

Introduction

Adolescents living with HIV (ALH) have poorer adherence to antiretroviral therapy (ART) than adults. Many ALH in sub-Saharan Africa (SSA) are enrolled in boarding schools where stigma is pervasive and may impact adherence.

Methods

We collected sociodemographic data, school information, medical history, and viral load (VL) data from ALH age 14–19 in 25 HIV clinics in 3 counties in Kenya. Using generalized estimating equations, we compared ART adherence in ALH attending day and boarding schools.

Results

Of 880 ALH, 798 (91%) were enrolled in school, of whom 189 (24%) were in boarding schools. Of those in school, median age was 16 (IQR: 15, 18), 55% were female, 78% had a parent as a primary caregiver, and 74% were on DTG-based ART. Median age at ART initiation was 6 years (IQR 3, 10).

Overall, 227 (29%) ALH self-reported missing ART when school was in session (40% in boarding and 25% in day school). After adjusting for sociodemographic and HIV care characteristics, ALH in boarding schools were significantly more likely to self-report missing ART than those in day schools (adjusted Prevalence Ratio (aPR): 1.47, 95% CI 1.18, 1.83, p=0.001). Among 194 ALH, only 60% had undetectable (<20 copies/ml) HIV viral load (62% day schools and 51% boarding schools) (p=0.097).

Conclusion

ALH had high self-reported non-adherence overall, with worse adherence among those in boarding schools. Schools remain a critical untapped resource for improving ALH outcomes.

Introduction

Antiretroviral therapy (ART) has increased life expectancy for adolescents living with HIV (ALH), transforming HIV from a life-threatening to a chronic condition. ALH on current ART regimens have better life expectancy than in the pre-ART era. However, compared to adults, ALH have poor viral suppression, and lower retention in care^1–4 with self-reported non-adherence among ALH as high as 45% in sub-Saharan Africa (SSA)⁵.

A high proportion of ALH (60–98%) in SSA are enrolled in school^6,7. In many SSA countries, children typically attend local day primary schools and transition around age 14 to more distant secondary schools. In Kenya, roughly half of the 9000 public secondary schools are boarding schools where youth spend ~9 calendar months in residence^8,9.Stigma in school has been identified as an important cause of poor adherence ^10–12.Stigma from teachers and peers^6,12,13 may result in ALH concealing their HIV status limiting support for adherence^6,10,11. Several specific school-related practices, particularly in boarding schools, such as requirements to keep medication with the school nurse and regular pill ‘raids’ to confiscate unauthorized drugs pose a barrier to adherence^7,10–16. A lack of or incorrect HIV knowledge among teachers^13,14, rigid school policies regarding lateness, absenteeism, and excused absence^10,12,13, and an insufficient level of privacy and confidentiality^6,10,15 have been identified qualitatively as barriers to adherence in the both the boarding and day school environment.

Many community or clinic-based interventions have been developed to support ALH services^17–19 but few are school based. The objective of this study was to quantitatively determine if ALH in day school had poorer adherence than those in boarding school. Data from this study would help build targeted programs to address adherence for ALH in day and boarding schools as well as provide supportive data for counseling ALH as they transition to different school types.

Methods

This study is part of a larger mixed methods study (Timiza study [NIH 1K43TW011422–01A]) aimed at understanding the school environment and its contribution to HIV outcomes. The Timiza study incorporated school surveys in select regions, an enrollment and 1 year follow-up ALH survey to understand adherence and retention in care, qualitative data collection with school staff, ALH and their caregivers and a user centered intervention development workshop to develop school-based interventions to support youth living with HIV. The primary study is ongoing. The study was approved by the Kenyatta National Hospital/ University of Nairobi, Ethics and Research Review Committee.

We report results from the enrollment youth survey whose eligibility criteria were: ALH (age 14–19 years) attending care in selected clinics, ever enrolled in school, knows their HIV status and on ART while in school. Study sites included 25 HIV clinics in Homabay, Kajiado, and Nairobi counties in Kenya. Participants were recruited using a standard recruitment script during routine clinic visits. Written informed consent was collected from ALH ≥18years, while assent was collected from caregivers of adolescents aged 14–17 who came to clinic with caregiver or unaccompanied. A structured questionnaire was administered by trained research assistants to obtain data on sociodemographic characteristics, school history, medication use while in school and barriers to medication use and clinic attendance, and family support. ART regimen and viral load data was obtained from medical record abstraction. In Kenya, secondary school enrolment is based on academic merit. Top performers are selected into national schools that serve students from all regions and selection continues to other school levels until each level is filled up. No data was collected from schools. The primary outcome, adherence was defined as ALH report of ever missing ART while school is in session while enrolled in their current school. Sociodemographic variables collected included age, gender (male/female), caregiver type (parent/non-parent), orphaned and vulnerable child (OVC) defined as ALH with one or both parents deceased or not involved with the ALH care. Social support was measured using the Multidimensional Scale of Perceived Social Support (MSPSS)²⁰ where responses were categorized as medium/low perceived support (score <61) and high support (score >60). Clinical factors included service point (pediatric/youth clinic), ART regimen, age at ART initiation, clinic visit schedule, attending clinic alone and HIV viral load. The PHQ-9 (Patient Health Questionnaire-9)^7,21 was used to screen for depression (classified as mild [1–9], moderate [10–14] and severe depression [15–27]) and the Youth Brief Stigma scale was used to assess stigma (personal stigma [maximum score 15], disclosure concerns [maximum score 10], negative self-image [maximum score 20) and public attitudes [maximum score 5)]; with higher scores indicating more stigma²². School characteristics included school type (day or boarding), HIV status disclosure in school, ever missing ART or clinic appointments when school was in session, perceived school support for ART use, awareness of a school counselor in their school, peer support and whether the school tracks ART adherence or clinic appointments.

The primary outcome, adherence, was defined as ever missing ART while school is in session while enrolled in their current school.

Baseline characteristics were summarized using medians and interquartile range (IQR) for continuous variables and proportions for categorical variables. Characteristics of ALH in boarding and day schools were compared using chi2 test or Student’s t-test. Generalized estimating equations (GEE) clustered by county (where ALHs were enrolled since HIV prevalence and availability of resources varies by county) were used to estimate prevalence ratios and 95% confidence intervals (CIs) comparing adherence among ALH in day and boarding school, adjusting for baseline difference in boarding and day school participants. All p values ≤0.05 were considered statistically significant. All analysis was conducted using Stata version 17 (StataCorp, College Station, Texas, USA)

Results

Of 880 ALH, 798 (91%) were enrolled in school. Eighty-two ALH not enrolled in school had completed schooling or were waiting to join secondary school or college 45% or did not have school fees 30%. Other reasons included not wanting to stay 7%, pregnancy 6%, poor health 4% and others (marriage, need to work, school too far, all <3%).

Of 798 ALH enrolled in school, the median age was 16 (IQR: 15, 18), 55%, were female 54% OVC, and 78% had a parent as a primary caregiver. The majority (74%) were on DTG based ART regimens and had a median age at ART initiation of 6 years. Overall, 189 (24%) were enrolled in boarding school. Overall, 95% were presumed perinatally infected defined as starting ART at an age of less than 15 years (Table 1).

Table 1:

Correlates of boarding or day school among ALH enrolled in school (univariate analysis)

	All N=798 n(%)	Day school n=609 (76%) n(%)	Boarding school n=189 (24%) n(%)	p-value(t-test or chi2 test)
Socio-demographic
Age in years	16 (15, 18)	16 (15, 17)	17 (16, 19)	<0.001
Female	436 (55%)	330 (54%)	106 (56%)	0.647
Orphaned and vulnerable child	434 (54%)	315 (52%)	119 (63%)	0.007
Primary caregiver parent	610 (76%)	472 (78%)	138 (73%)	0.204
MSPSS score
Medium/low support	654 (84%)	515 (86%)	139 (76%)
High support	129 (16%)	85 (14%)	44 (24%)	0.002
HIV care
Service point
Pediatric clinic	156 (20%)	135 (22%)	21 (11%)
Youth clinic	635 (80%)	470 (77%)	165 (87%)	0.002
ART regimen				0.960
DTG based	584 (74%)	446 (74%)	138 (73%)
PI based	11 (1%)	8 (1%)	3 (2%)	-
NNRTI based	197 (25%)	150 (25%)	47 (25%)	-
Age at ART initiation	6 (3, 10)	6 (3, 9)	7 (3, 11)	0.142
On ART before age 15	761 (95%)	583 (96%)	178 (94%)	0.323
Attends clinic alone	734 (92%)	557 (91%)	177 (94%)	0.333
3 monthly or longer visit schedules	509 (64%)	365 (61%)	144 (77%)	<0.001
Missed visit in the last 6 months	84 (11%)	66 (11%)	18 (10%)	0.632
Missed ART in the last 30 days	169 (21%)	134 (22%)	35 (19%)	0.306
Viral suppression (undetectable) (n=194)	117 (60%)	98 (62%)	19 (51%)	0.216
Attends peer support group	563 (71%)	434 (72%)	129 (69%)	0.448
Depression screen				0.130
Minimal depression	743 (94%)	567 (94%)	176 (94%)
Mild	46 (6%)	37 (6%)	9 (5%)	-
Moderate to severe	5 (<1%)	2 (<1%)	3 (2%)	-
Stigma scale
Total score (n=331)	17.8 (7.8)	17.4 (7.7)	19.1 (8.0)	0.082
Personal stigma (n=331)	4.9 (2.8)	4.7 (2.7)	5.3 (3.1)	0.09
Disclosure concerns (n=781)	4.2 (2.2)	4.1 (2.2)	4.4 (2.2)	0.108
Negative self-image (n=797)	7.5 (3.6)	7.4 (3.5)	7.9 (4.0)	0.055
Public attitudes (n=795)	2.6 (1.6)	2.4 (1.6)	3.0 (1.6)	<0.001
School factors
Disclosed HIV status to school	385 (49%)	230 (38%)	155 (82%)	0.008
Need to attend clinic	472 (61%)	372 (64%)	100 (53%)	0.007
Ever missed clinic school time	112 (24%)	96 (26%)	16 (16%)	0.036
Ever missed ART when in school	227 (29%)	151 (25%)	76 (40%)	<0.001
Aware school has counselor	387 (51%)	257 (45%)	130 (71%)	<0.001
School supports ART use	323 (43%)	196 (34%)	127 (69%)	<0.001
Has peer support	89 (12%)	66 (12%)	23 (13%)	0.752
Tracks clinic appointments	55 (7%)	28 (5%)	27 (15%)	<0.001
Tracks adherence	37 (5%)	19 (3%)	18 (9%)	0.001

ALH in boarding school were older, more likely to be OVC and had higher social support. For HIV care, ALH in boarding school were more likely to be attending youth clinic (vs pediatric clinic) and have 3-monthly or longer intervals between clinic visits. ALH in boarding schools had higher stigma scores in the public attitudes’ domain but not in the overall stigma score. Only 5 ALH (<1%) had moderate to severe depression symptoms, 3 in boarding and 2 in day school. There were no differences in ART regimen comparing those in day and boarding school. Compared to those in days school, ALH in boarding school were more likely to have disclosed their HIV status to schools, more likely to report HIV care support in school (Table 1).

Overall, 227 (29%) ALH reported missing ART when school was in session at their current school. ALH in boarding schools were more likely to report missing ART doses when school was in session in univariate analysis (PR: 1.49, 95% CI 1.22, 1.80, p<0.001) and in analysis adjusted for age, OVC status, social support, stigma score and clinic schedule, disclosure to school and school support for ART (adjusted Prevalence Ratio (aPR): 1.37, 95% CI 1.08, 1.75, p=0.010).

Among 194 ALH with viral load (VL) data 3 months before and 6 months after study enrollment date, only 60% had undetectable (<20 copies/ML) viral loads (98/157 (62%) day schools and 19/37 (51%) boarding schools); with a trend to poorer viral suppression among those in boarding school. (p=0.097, Table 1). There was no difference in viral non-suppression by reported adherence (35% who self-reported nonadherence had unsuppressed VL versus 42% unsuppressed who did not report nonadherence, p=0.423).

Among 151 day school ALH who reported missing ART when in school, common nonadherence reasons were related to school/medication schedules (59%), forgetting (23%), deciding not to take (8%) and stigma or lack of food at home (3% each). Among 76 boarding school ALH, common reasons for non-adherence included school/medication schedule challenges (45%), stigma (30%), lack of access to medicine (18%) and forgetting (12%) (Figure 1).

Figure 1:

Reasons for missed ART and missed clinic visit by school type.

Discussion

We observed overall high self-reported nonadherence among ALH with almost one-third of ALH reporting non-adherence, and significantly higher nonadherence (40% versus 25%) among ALH enrolled in boarding schools.

Nonadherence is common among ALH, with some studies reporting 45% self-reported nonadherence among ALH⁵. Common reasons for nonadherence include individual, family, and health care factors²³. As ALH survival has improved, school related factors have emerged as reasons for nonadherence²⁴. Most (90%) of the ALH we enrolled were in school with a quarter in boarding schools. Previous qualitative work has highlighted school-related challenges experienced by ALH^10,13,15including keeping medication with the school and regular pill ‘raids’ to confiscate unauthorized drugs. On the other hand, day school ALH report challenges with rigid school policies regarding lateness, absenteeism, and excused absence,¹⁰ and an insufficient level of privacy and confidentiality¹⁵.

Overall viral suppression was 60% in our cohort, mirroring findings from other studies^25,26; While boarding school ALH had a lower proportion suppressed, interpretation of this finding is limited by missing viral load information related to supply chain challenges during the COVID-19 pandemic^27,28. The majority of ALH were on DTG-based regimens at enrollment; however, the data collection period coincided with national guideline change to switch to DTG, and we did not collect data on switch dates. As a result, the results do not reflect the viral suppression associated with DTG-based regimens, which has been observed to be higher ^29,30. Surprisingly, ALH who self-reported poorer adherence did not have poorer viral suppression, possibly due to sample limitations with viral load results.

Modifiable factors to prevent nonadherence among ALH in school could include supporting ALH manage school/medication schedules. We observed that ALH were more likely to need to attend clinic during school time, likely as clinics do not consider aligning day school ALH clinic visits with school schedules. With once daily DTG based regimens ^31,32 adjusting medication timing to fit the school schedules is feasible. We recently identified that schools in Kenya lack medication use policies making it challenging to ensure access when needed³³. Identification of strategies to ensure confidentiality and access to ART are urgently needed in school. Stigma reduction interventions such as teacher education consistently demonstrate effectiveness and need to move to practice. Other out of school interventions include intrapersonal interventions like resilience building, community mass education campaigns and advocacy to promote national anti-stigma policies^{23,24,34,35,36}. Our study had some limitations. We relied on ALH self-report which is subject to social desirability bias. Although self-report overestimates adherence, almost one third of ALH stated they were nonadherent and provided reasons for their nonadherence, suggesting validity. Since ALH were enrolled in clinics, it was difficult to link challenges to school policies or structures. We did not have school data on student numbers, teacher-student ratio, educational outcomes, and health care services. The study was conducted during the COVID-19 pandemic period during which there were disruptions in school schedules, supply chain for ART commodities and household level changes that could have impacted adherence. Viral load results were difficult to interpret because of data missingness. Strengths of our study include the large sample size with detailed school information. Our study adds to the growing literature on school related barriers to adherences, by adding quantitative data to support existing qualitative data.

Conclusion

In summary, adherence remains a major challenge among ALH, addressing adherence among ALH needs to include the school environment where ALH spend majority of their daytime or year. ALH in boarding and day school need specific targeted approaches to maintain adherence when school is in session. Schools remain a critical untapped potential resource to improve outcomes for ALH.