Our results indicate that, after adjustment on sociodemographic factors, ALPHIV were 5 times more likely to experience school trajectory disruption than controls and half of academic delay ≥ 1 year or early dropout was attributed to HIV-infection. Furthermore, all other factors being equal, ALPHIV living in institutions had a 10-fold higher risk of having a disrupted school trajectory than controls. School trajectory disruption was also significantly associated with HIV-comorbidities such as mental or growth delay and stigmatization experiences.
In the TEEWA survey, school attendance was high, above 85%, for both ALPHIV and controls, in the upper range compared to other studies in various age groups (8–12, 14). However, among adolescents living in family settings, school attendance was lower among ALPHIV than among general population controls (83% vs. 94%, respectively). Also, as observed (8, 12, 28), the percentage of academic delay ≥ 1 year among ALPHIV was significantly higher than among controls (17% vs. 4%), similar to the proportion of 20% found in another study among HIV-infected children aged 6–12 years in Thailand (28).
We investigated the school trajectories of ALPHIV living in either family or institutional settings. For those living in institutions, school attendance was almost universal because of the strict institutional administrative rules. However, the risk of school trajectory disruption was 11 times higher when living in an institution than in family settings. Children living in institutions often experience chaotic life trajectories such as parental loss, caregiver turnover, neglect by relatives, or poverty, which makes them particularly vulnerable.
As usually observed, boys were more prone to disrupted school trajectory than girls (8, 14, 21, 23). Furthermore, ALPHIV from ethnic minorities were more likely to experience disrupted school trajectories than their Thai peers, probably because of Thai-language deficiencies or administrative issues causing delays in school enlistment (29).
Some studies have highlighted the negative effect of adverse living conditions on school enrolment and academic performance (6, 8, 9, 21–23). In our study, low caregiver educational level was significantly associated with disrupted school trajectories, but no significant difference was found according to the household’s financial situation or type of living area (rural or urban). This could be a consequence of the recent policies to improve access to education in Thailand, with free, compulsory enrolment in primary and lower-secondary school (19, 30).
Several studies carried out in North America have focused on academic achievement among HIV-infected or affected children (13, 16, 17). Garvie et al. have shown that academic performances of HIV-infected and exposed HIV-uninfected children aged 7–16 years were significantly lower than the general population standards but not significantly different from each other (17). In our study, a small number of HIV-uninfected adolescents living in institutions were born from HIV-infected mothers (n = 16 HIV-exposed uninfected controls). These adolescents were less likely to experience academic delay ≥ 1 year than their HIV-infected counterparts (31% vs. 74%, p < 0.001, data not shown), confirming a contributing role of HIV-infection per se to educational disadvantage.
Studies carried out in low-income countries also found a significant association between HIV-infection and disrupted schooling after adjustment on socio-economic factors (8, 9, 14). In a recent study in South Africa, Fotso et al. found that HIV-infection significantly reduced adolescent school attainment and that contextual factors could only explain 18% of the attainment gap (8). While neurocognitive impairments connected to HIV-infection can directly impact academic achievements, school absenteeism for frequent illness or HIV-related stigmatization could mediate the negative effect of HIV-infection on school life (6, 11, 12, 14). In the TEEWA survey, about 20% of ALPHIV mentioned episodes or extended periods of school absenteeism for medical reasons. In a recent cross-sectional study, Rukuni et al. found that HIV-infected children were not only more likely to repeat a grade but that their risk of social and educational exclusion increased in case of disabilities (12). In the TEEWA survey, after adjustment, delayed development was significantly associated with disrupted school trajectories.
In a literature review, Smith et al. emphasized the importance of early effective ART to reduce the negative effect of HIV-infection on neurodevelopment (2). In the TEEWA study, after adjustment, delayed ART initiation was not associated with school trajectory disruption. However, most ALPHIV had delayed ART initiation – median age 9 years.
Finally, as in other studies (12), we found that stigmatization was significantly associated with school trajectory disruption. Stigmatization, in the form of bullying, violent or humiliating behaviors and exclusion, was most frequently experienced at school. These traumatic events, instigated by teachers, other students or their parents, can lead to academic failure or school dropout. The experience of stigmatization was probably underreported by caregivers who may not have been informed by the children of all stigmatization episodes.
Our study presents some limitations. First, participating hospitals were not randomly selected based on regional HIV prevalence, limiting our ability to generalize the findings to the country level. Nevertheless, the survey coverage, around 10% of ALPHIV aged 12–19 years receiving ARTs in Thailand at the time, and the geographical distribution of the hospitals across the country provide a good representation of this population. Considering most perinatally HIV-infected children in this generation had no access to ART in early childhood, our survey comprised a selected population of survivors.
A strength of this study is its having targeted all ALPHIV living in family or institutional environments. Using a unique case-control study design, we also compared ALPHIV with appropriate controls of the same sex, age and place of residence.
Our survey covered a wide age range of adolescents facing different challenges. Younger participants were enrolled in primary school, while older participants had reached the upper limit of compulsory education. Our large sample, however, and the results of our sensitivity analyses strengthen our findings.
Finally, we created a composite outcome, including academic delay ≥ 1 year and early school dropout, to investigate factors associated with a disrupted school trajectory. We assumed these events derived from similar mechanisms of academic failure, and this hypothesis is supported by the literature (26). However, academic delay may result from other processes, such as delayed school entry, school interruption or repeat grades, all of which could be compounded.