EFFECTS OF HIV/AIDS ON CHILDREN'S EDUCATIONAL ATTAINMENT: A SYSTEMATIC LITERATURE REVIEW

Abstract

Over the last three decades, 35 million people have died of AIDS. As a result, HIV/AIDS has brought about a significant reduction in human capital, especially in sub-Saharan Africa. Several studies have examined the effects of HIV/AIDS on human capital, in particular educational attainment. These studies have examined different countries, datasets, and educational outcomes. This systematic literature review provides a comprehensive up-to-date overview of peer-reviewed papers published in English by focusing on the main mechanisms that influence the effects of HIV/AIDS on educational outcomes. These are sickness of the child, orphanhood, and sickness of parents. The results show that educational outcomes of HIV-infected children, AIDS orphans, and children with HIV-infected parents are affected differently. HIV-infected children mainly miss school days due to illness and medical appointments, and orphans mainly face financial problems and lack motivation in their education, while children with HIV-infected parents may have to take care of their sick parents or face financial problems that affect their education. Distinguishing these groups of children could help to formulate policies that adequately improve schooling outcomes of these vulnerable children.

1 Introduction

About 17 million children have lost one or both parents to HIV/AIDS since the eruption of the epidemic (USAID, 2016). Most of these children (about 90%) reside in sub-Saharan Africa (SSA). An estimated 3.4 million children under 15 years are currently living with HIV (USAID, 2016). HIV-infected children, AIDS-orphaned children, and children with HIV-infected parents may be deprived of opportunities that lead them to become economically productive adults (UNAIDS, 2016). In particular, HIV/AIDS may impede children's schooling, through childhood illness, orphanhood, and parental illness. The effects may further differ by gender, thereby inducing a gender gap in schooling among children affected by HIV. Given the various ways in which HIV disrupts children's schooling, it is important to know the impact of HIV/AIDS on life chances of boys and girls; specifically, on their educational attainment, enrollment, and attendance.

The literature on the educational attainment of orphans and children living with HIV-infected adults has been reviewed before by Guo et al. (2012). They include 23 quantitative studies that analyzed the impact of HIV/AIDS on educational attainment of children affected by HIV. Most of the literature in this review analyzed educational attainment of orphans only. The review showed that educational attainment of children differed by type of orphan (i.e., double orphan, maternal orphan, or paternal orphan). However, results on gender gaps in educational attainment were mixed and also varied by type of orphan. This review study excluded qualitative papers that capture issues that may not be identified by quantitative studies. These include pathways through which HIV affects children's education.

While the extant health literature has largely focused on inter alia, general disease burden (Lopez and Murray, 1998; Kassebaum et al., 2017; Rajasingham et al., 2017), health care reform (Simms et al., 2001; Jakovljevic et al., 2016), health policy (Yamada et al., 2015), universal health coverage (Largomarsno et al., 2012; Ranabhat et al., 2018), and HIV mortality (Wang et al., 2016), this study aims to provide a systematic review of the effects of HIV on children's educational attainment. Goldeberg and Short (2016) conducted a systematic review that included 45 articles. The study examined physical and emotional health as well as schooling of children living with HIV-infected or AIDS-ill adults. The studies highlighted factors that influence these outcomes, including poverty, transmission of opportunistic diseases, caring for sick adults, stigma, and lack of support. Only 10 out of 45 studies (nine quantitative and one qualitative) reported results on educational outcomes. This left out many important and relevant studies on educational attainment of HIV/AIDS-affected children. This review showed that children who live with HIV-infected parents and adults attended school less frequently and had deficits in grade progression. Other important forms of educational attainment such as dropout, enrollment, and years of schooling were not discussed in this review.

From the literature, three main underlying mechanisms by which HIV affects educational attainment of children can be identified. These are: sickness of the child, orphanhood, and HIV infection of parents (UNICEF, 2006). These mechanisms may have different effects on educational outcomes of children. By systematically reviewing peer-reviewed English language publications, we examine the extent to which HIV affects children's schooling differs by these three mechanisms. We focus on three main effects:

1. Effects on children: Due to illness, HIV-infected children may miss school days and perform academically less than non-HIV children. Correspondingly, orphans – even when they are not sick themselves – generally obtain less schooling compared to nonorphans (Case et al., 2004; Evans and Miguel, 2007; Ardington and Leibbrandt, 2010). Hence, it is important to further analyze the effects of HIV on various educational outcomes so as to identify the most vulnerable groups.

2. Effects of HIV on gender gaps in educational outcomes: Because of HIV, women and girls are more likely to lose jobs, lose income, miss school, and primarily take care of sick people due to patriarchal norms that subordinate women (Madiba and Ngwenya, 2017). In addition, women and girls are the majority of the HIV-affected population (UNWOMEN, 2016); therefore, they are more likely to suffer from the effects of this disease.

3. Intergenerational (parent-to-child) transmission of education in case of HIV/AIDS: Children with HIV/AIDS-ill parents may observe that their parents physically deteriorate from the disease and may experience both parents’ deaths, sometimes in quick succession. This may lead to posttraumatic stress syndrome, depression, poverty, and stigma (Cluver et al., 2012; Anabwani et al., 2016; UNAIDS, 2016). To the extent that children's educational attainment is highly correlated with parental education (Becker and Tomes, 1986; Björklund and Salvanes, 2011), HIV may interfere with intergenerational (parent-to-child) transmission of human capital. Without parental human capital investment, chances of upward social mobility may decrease (Spiegler, 2018).

Our systematic review adds to the aforementioned reviews by: (i) including papers that examine sick children, AIDS-orphaned children, children living with HIV-infected parents or caregivers, or children living with an HIV-infected family member; (ii) distinguishing three pathways through which HIV/AIDS affects children's educational attainment; (iii) focusing on direct effects of HIV on educational attainment by only including studies that have information on HIV/AIDS infection or confirm AIDS deaths of the parent/guardian; (iv) analyzing results based on the type of educational outcome; (v) including quantitative, mixed-methods, and qualitative studies that discuss educational attainment of children affected by HIV; and (vi) updating and expanding the literature in the aforementioned reviews. The review increases our knowledge and insight on the underlying mechanisms and is relevant for future studies and policy makers.

2 Methodology

This literature review follows the guidelines of the Preferred Reporting Items for Systematic Reviews (PRISMA). PRISMA guidelines were developed using an evidence-based approach. They consist of a 27-item checklist and a four-phase flow diagram of items that are deemed essential for a thorough and transparent systematic review (Moher et al., 2009).

The search of relevant articles was conducted in six databases, namely, EconLit, ERIC, PubMed, SocINDEX, Web of Science (WoS), and Google Scholar. The search and review of the articles was conducted between December 2017 and July 2018, and was updated in June 2019. Table 1 shows the exact search terms that were used in each database. In each column, the table shows: (i) terms for HIV, (ii) terms for schooling and socioeconomic outcomes, and (iii) terms for children and gender. In some cases, truncations were used to ensure that all relevant papers were included. To optimize the results, the searches in EcoLit and ERIC were performed in the title and abstract, and the searches in PubMed, WoS, and SocINDEX were performed in the title. Searches in Google Scholar used similar terms as in the other databases. A university librarian verified and approved the combinations used.

Table 1. Search Terms Used in Each Database

Database	HIV terms	Education and socioeconomic terms	Child and gender terms
EconLit, ERIC, SocINDEX, and Web of Science	HIV OR HIV/AIDS OR AIDS OR HIV-affected OR HIV-infected OR AIDS-ill OR HIV-positive OR AIDS-affected OR Parental-AIDS OR Parental-HIV OR Maternal-AIDS OR Maternal-HIV OR Paternal-HIV OR Paternal-AIDS	Education* OR School* OR Learn* OR “Human Capital” OR “Drop-out” OR Dropout OR Truancy OR Enro$l* OR Absenteeism OR Absence or “School leaving” OR Intergenerational OR Parental-education OR Maternal-education OR Paternal-education or Socio-economic OR Socioeconomic OR Economic	Child* OR Adolescent* OR Infant* OR Youth* OR Orphan* OR “Vulnerable Children” OR Boy* OR Girl* OR Gender-gap* OR Gender-difference*
PubMed	HIV [Title] OR HIV/AIDS [Title] OR AIDS [Title] OR HIV-affected [Title] OR HIV-infected [Title] OR AIDS-ill [Title] OR HIV-positive [Title] OR AIDS-affected [Title] OR Parental-AIDS [Title] OR Parental-HIV [Title] OR Maternal-AIDS [Title] OR Maternal-HIV [Title] OR Paternal-HIV [Title] OR Paternal-AIDS [Title]	School [Title] OR Schooling [Title] OR Education [Title] OR Dropout [Title] OR “Drop-out” [Title] OR Learning [Title] OR “Human Capital” [Title] OR Truancy [Title] OR Absenteeism [Title] OR Absence [Title] OR “School leaving” [Title] OR Enrollment [Title] OR Enrolment [Title] OR Educational [Title] OR “Educational Attainment” [Title] OR Intergenerational [Title] OR Parental-education [Title] OR maternal-education [Title] OR Paternal-education [Title] OR Socio-economic [Title] OR Socioeconomic [Title] OR Economic [Title]	Child [Title] OR Children [Title] OR Adolescent* [Title] OR Infant* [Title] OR Youth* [Title] OR Orphan* [Title] OR “Non-orphan*” [Title] OR “Vulnerable Children” [Title] OR Boys [Title] OR Girls [Title] OR Gender-gap* [Title] OR Gender-difference* [Title]

2.1 Inclusion and Exclusion Criteria

Papers were included in the review if they analyzed the direct relationship between HIV and schooling outcomes. There were no time or language restrictions imposed in the search. Articles were excluded if they were nonempirical, discussed the relationship between HIV and psychological or cognitive issues, only focused on perceptions of HIV risk, if there was no HIV-testing done on either parent or child, or if there was no confirmation of AIDS-related death of parent or guardian. Books, working papers, conference papers/abstracts, and meeting papers/abstracts were also excluded. We only included peer-reviewed papers because the peer review process adds to the quality of the studies reviewed. As in the other databases, papers in Google Scholar were included if the title and abstract discussed the relationship between HIV and educational attainment and if they were peer-reviewed. Papers were excluded if there was no information on HIV infection or no confirmed AIDS deaths (see Figure 1).

Figure 1

Flow Chart of the Selection Process. [Color figure can be viewed at wileyonlinelibrary.com]

2.2 Selection Process

The selection process consisted of several phases. Once all papers were extracted from the databases, duplicates were removed in EndNote. The first selection was based on the title and the abstract. Two reviewers independently screened the papers to be included in the review. The independently screened lists were compared and discussed, leading to a final selection of papers to be included in the review. The remaining papers were fully reviewed by one researcher. The results were discussed with all researchers to assure the consistency in the selection process. The reference lists of the selected papers and previous reviews were then screened for relevant publications applying the same inclusion and exclusion criteria. Again, the results were discussed with all researchers to make sure that no relevant papers were left out. The final list of publications included in the review can be found in Appendix A.

2.3 Analysis

The selected papers were categorized into quantitative, mixed-methods, and qualitative studies. We applied the method of directed qualitative content analysis (Hsieh and Shannon, 2005) for the analysis of the papers selected for the review. Specifically, we extracted information related to the key themes identified in the introduction (I) HIV-affected versus HIV-unaffected children, (II) gender gaps in educational attainment, and (III) intergenerational transmission of education. During the data extraction phase, all interim results were reviewed by all researchers and discussed to assure the quality of data extraction.

From the quantitative papers, the following information was extracted: country of analysis, children's level of schooling, education variables used, HIV status, gender and age of the child, income and education of the parents, income and education information of caregivers, type of orphan, information on intergenerational transmission of education, type of data, number of observations, method of analysis, comparisons made in the analysis, and summary of the results. From the qualitative papers, the following information was extracted: country of analysis, HIV status method of data collection, type of individuals interviewed, education outcome analyzed, and summary of results. From the mixed-methods papers, a combination of information from quantitative and qualitative results was extracted. The results were synthesized and were presented in the form of tables and narrative descriptions (see Appendix B).

2.4 Assessment of Publication Quality

The quality of the papers selected for the review was assessed using the Mixed-Methods Appraisal Tool (MMAT) (Pluye et al., 2011). The MMAT is used for complex systematic literature reviews that include quantitative, mixed-methods, and qualitative studies. It accounts for five common methodologies, qualitative (section 1), quantitative randomized (section 2), quantitative nonrandomized (section 3), descriptive (section 4), and mixed (section 5). Each section is composed of three to four questions related to data sources, data collection, and outcome data. A quality score between 0% and 100% is assigned for each study component. An answer of “yes,” “no,” or “I can't tell” is assigned for each question in the corresponding study component (see Appendix C). Papers of good quality met all criteria (i.e., had a score of 100%) and papers of poor quality did not meet any criteria (i.e., had a score of 0%). We also checked the quality of our review by using the PRISMA 2009 checklist (see Appendix D).

3 Results

Figure 1 shows a flow chart of the selection process as per PRISMA guidelines. Using the search terms in Table 1, a total of 5961 papers were extracted from EconLit, ERIC, PubMed, SocINDEX, WoS, Google Scholar, and references of selected papers. After duplicates were removed, the exclusion criterion was applied and full texts were read. This left 62 papers (46 quantitative, 7, mixed-methods, and 9 qualitative). Of these 62 papers, there were seven mixed-methods papers that were included in both the quantitative and qualitative groups distinguished in the analysis.

3.1 General Characteristics of the Selected Papers

Table 2 shows the characteristics of 52 selected quantitative papers (45 were purely quantitative and 7 were mixed-methods papers). The papers used randomized, nonrandomized, and descriptive methods. These studies were published between 1994 and 2019. About 60% (31 out of 52) of these papers reported studies conducted in Southern and East Africa. Most of the papers used attendance, enrollment, correct grade for age, and years of schooling as measures of schooling. In some cases, studies used multiple educational outcomes. Regression analysis was the most common form of analysis applied in these papers (51 out 52).

Table 2. General Characteristics of the Quantitative and Mixed-Methods Papers (Quantitative Part) Included in the Review (n = 52)

	Number of publications	Publication reference in Appendix A
Year of publication		N = 52
2015–2019	16	2, 16, 20, 21, 23, 24, 25, 35, 36, 40, 42, 45, 46, 51, 55, 56
2010–2014	21	1, 4, 5, 7, 8, 10,15, 18, 26, 29, 31, 33, 34, 41, 44, 48, 49, 53, 58, 59, 60
2005–2009	9	6, 11, 14,17, 39, 54, 57, 61, 62
2000–2004	2	12, 38
1995–1999	3	3, 9, 37
1990–1994	1	50
Region/country of study		N = 52
Multiple African countries (more than 1 country)	3	1, 25, 51
Southern Africa	19	2, 4, 7, 8, 14, 16, 17, 29, 32, 33, 35, 36, 40, 42, 44, 48, 49, 50, 55
East Africa	12	3, 6, 10, 23, 24, 31, 34, 39, 45, 54, 55, 62
West Africa (Guinea, Côte d'Ivoire)	2	11,15
East Asia (China)	8	20, 21, 26, 57, 58, 59, 60, 61
South Asia (India, Myanmar)	4	5, 18, 41, 46,
Latin America (Brazil)	1	53
North America (USA)	4	9, 12, 37, 38
Sample size		N = 52
0–10	1	12
10–50	1	53
50–100	6	3, 9, 18, 37, 38, 50
100–200	1	60
200–500	17	5, 10, 11, 15, 26, 29, 32, 34, 40, 42, 45, 46, 54, 58, 59, 61, 62
500–1000	10	2, 7, 20, 21, 23, 25, 36, 41, 44, 51
1000+	16	1, 4, 6, 8, 14, 16, 17, 24, 31, 33, 35, 39, 48, 49, 55, 56, 57
Analytical model/statistical methods		N = 52
RCT	5	20, 21,45, 54, 55
Regression (GLM/random effects/fixed effects, logit, maximum likelihood, difference in differences)	31	1, 4, 6, 7, 8, 10, 11, 14, 15, 16, 17, 23, 24, 25, 26, 31, 33, 34, 35, 36, 39, 42, 44, 46, 48, 49, 51, 56, 57, 58, 62
Statistical tests (Chi-square, T-test/	11	2, 3, 5, 12, 29, 32, 37, 41, 50, 53, 59, 61
F-test, Pearson's correlation, ANOVA, Fisher's exact test, Mann-Whitney U test)
Descriptive statistics	5
		9, 18, 38, 40, 60
Education variables analyzed		N = 89 (Some studies analyzed multiple outcomes)
Enrollment, dropout, years of schooling	25	1, 3, 5, 7, 15, 16, 18, 29, 31, 33, 35, 36, 37, 41, 44, 46, 48, 49, 51,53, 54, 58, 59, 60, 61
Attendance, absenteeism, truancy	2	1, 2, 5, 7, 11, 12, 14, 15, 18, 24, 29, 39, 40, 41, 44, 45, 46, 48, 49, 51, 53,56, 58, 59, 60, 61, 62
Grades, correct grade for age, highest grade level, grade repetition, grade progression	25	1, 2, 4, 5, 10, 12, 14, 16, 20, 21, 23, 26, 35, 36, 37, 38, 42, 46, 48, 49, 51, 53, 54, 56, 57
Other	4	6, 22, 29, 30

Table 3 shows the characteristics of the 16 papers that report qualitative results (nine were purely qualitative and seven were mixed-methods). The papers were published between 2005 and 2017. All of the studies were conducted in Southern and East Africa. Almost all (14 out of 16) of the studies included various forms of interviews and focus group discussions. The main education variables were attendance and dropout.

Table 3. General Characteristics of the Qualitative and Mixed-Methods Papers (Qualitative Part) Included in the Review (n = 16)

	Number of publications	Publication reference in Appendix A
Year of publication		N = 16
2015–2017	3	2, 13, 24
2010–2014	6	4, 7,19, 28, 29, 32
2005–2009	7	17,22, 27, 30, 43, 47, 52
Region/country of study		N = 16
Multiple Countries (Eswatini and South Africa)	1	47
Southern Africa	10	2, 4, 7, 13, 17, 22, 24, 29, 32, 43
East Africa	5	19, 27, 28, 30, 52
Method of data collection		N = 17 (Ref. #32 used both methods)
Interviews (interviews, semi-structured interviews, focus group discussions, in-depth interviews, exit interviews, informal interviews)	14	2, 4, 7, 13, 17, 19, 4, 2, 27, 28, 29, 30, 32, 43, 47
Other (letter writing, case studies)	3	22, 32, 52
Education variables analyzed		N = 28 (Some papers analyzed multiple variables)
Attendance	12	2, 4, 7, 13, 17, 19, 22, 24, 27, 30, 32, 47
Absenteeism	3	19, 29, 32
Dropout	6	7, 13, 29, 30, 32, 52
Enrollment	1	17
Other	6	2, 27, 28, 29, 43, 52

The quality of the quantitative, qualitative, and mixed-methods papers is assessed and presented in Appendix C. In short, the assessment tool provides a score based on the data and methodological quality of quantitative, qualitative, and mixed-methods papers. Approximately 52% (24 out of 46) of the purely quantitative papers had a score of 100% (i.e., they met all four criteria), followed by about 20% (9 out of 46) that met three criteria, 13% (6 out of 46) met two criteria, 11% met one criterion, and about 3% met none. Only one out of seven papers (14%) of the mixed-methods papers met all criteria, 57% (4 out of 7) met two, and 29% (two out of seven) met one. For the purely qualitative papers, about 33% (three out of nine) met all four, 44% (four out of nine) met three, and 22% (two out of nine) met two criteria.

3.2 Relevant Findings Reported in the Papers Reviewed

Table 4 shows the main findings of the 52 quantitative studies. There were seven mixed-methods studies included in this group that were also included in the qualitative study group. The three main results categories are (i) HIV-affected and HIV-unaffected children, (ii) gender comparisons in schooling outcomes, and (iii) and intergenerational transmission of education. The table lists the effects of HIV on five main educational outcomes and indicates whether effects or no effects were found.

Table 4. Main Findings of Selected Quantitative Papers (Publication Reference in Appendix A)

	Category 1: HIV-affected and HIV-unaffected			Category 2: Gender comparisons in schooling outcomes			Category 3: Intergenerational transmission of education
	Positive effect	Negative effect	No effect	Positive effect	Negative effect	No effect	Positive effect	Negative effect	No effect
Attendance		2, 8, 9, 14, 15, 32, 37, 48, 56, 62	2, 49		11	24		1, 8, 9	24, 39,44, 48
Enrollment		3, 8, 31, 41, 46			33		35	3, 8	17, 33,44
Dropout		5, 15, 41, 46, 50, 53						7	50
Correct grade for age		4, 8, 16, 23, 31, 38, 46, 49	48	48				1	33,44
Years of schooling			49
Other	6	53	12, 14, 33, 55	10, 25	6, 20, 21, 44		57	42	14, 17, 42,44

There were 16 papers included in the qualitative study group. This included nine purely qualitative papers and seven mixed-methods studies that are also found in the quantitative study group. Of these studies, 14 studies conducted interviews. Individuals who participated in the qualitative studies included HIV-infected children, HIV-infected parents, non-HIV parents, adult caregivers, child carers, teachers, and pupils.

• Summary of educational outcomes of HIV-affected children

Twenty out of the 28 quantitative results on HIV-affected children (Category 1 in Table 4) found that HIV-infected children and AIDS orphans attained less education than HIV-unaffected children. This is in comparison to one paper that found a positive effect and four papers that found no effect. Bhargava (2005) found that AIDS orphans were more likely to participate in school than children with parents who died of non-AIDS-related illnesses. However, Pufall et al. (2014a, 2014b) found no relationship between being HIV-positive and educational outcomes. Some studies (e.g., Bandason et al., 2013) showed that being HIV-infected delayed schooling. However, studies listed in the table went further by comparing different groups of HIV-affected children. For example, Cohen et al. (1997) compared HIV-infected children with mild, moderate, and severe symptoms, and found that children with severe symptoms missed more school. Mayes et al. (1996) compared 66 American boys diagnosed with hemophilia, of which 18 were HIV-positive. The results showed that HIV-positive boys missed more school days than non-HIV boys. However, there were no differences in academic grades. On the other hand, Delva et al.’s 2009 comparison of AIDS orphans and non-AIDS orphans to nonorphans showed that AIDS orphans in Guinea missed more school than nonorphans and other orphans. Kidman et al. (2012) went further by examining different types of orphans and found that double orphans and maternal orphans in Malawi experienced more educational deprivation compared to nonorphans. However, Orkin et al. (2014) found that HIV/AIDS orphanhood was not associated with nonenrollment or nonattendance in South Africa. They found that HIV/AIDS affected educational outcomes indirectly via orphanhood and parental/caregiver illness through poverty and internalization of problems.

Among the qualitative studies, Anabwani et al. (2016) found that HIV-infected children in Botswana reported no major problems in school performance. Other studies showed that HIV-infected children were missing school days due to illness and parental illness (Poulsen, 2006; Skovdal and Ogutu, 2009; Harms et al., 2010; Cluver et al., 2012; Bandason at al., 2013; Anabwani et al., 2016).

• Summary of gender comparisons of all children affected by HIV

Table 4 shows that 6 out of 10 (60%) papers found negative effects of HIV on gender differences in educational attainment within the group of HIV-infected children and children with HIV-infected parents. The results were mixed. Delva et al.’s (2009) analysis of different types of orphans in Guinea showed that regardless of orphan status, boys were significantly more likely to attend school on a daily basis than girls. Similarly, Harrison et al. (2017) found that HIV-affected girls in China reported lower grades and had less interest in school. However, Henning et al. (2016) found that gender did not affect school attendance in Zambia. Orkin et al. (2014) showed that HIV-affected boys in South Africa reported difficulties with grade progression. Pufall et al. (2014a) also showed that girls were more likely to be in correct grade for age compared to boys in Zimbabwe. Hensels et al. (2016) found that being a girl was significantly associated with better educational functioning and being a boy was associated with more educational risks. Zivin et al. (2009) is the one of the few studies that analyzes effects of antiretrovirals (ARVs) on children's education. They found that ARV treatment effects were significant for Kenyan girls in early ARV treatment stages, and not significant for boys.

Only two qualitative studies reported on gender gaps in educational attainment. Jepkemboi and Aldridge (2009) found that HIV-affected boys performed better in math and science. However, Jepkemboi and Aldridge (2014) found that HIV-affected girls were more persistent and had a more positive attitude toward school than HIV-affected boys.

• Summary of intergenerational (parent-to-child) transmission of education

Category 3 in Table 4 shows that 8 out of 16 quantitative papers that examined how HIV-infected parents transmit education to their children found that HIV had negative effects. Akbulut-Yuksel and Turan's (2013) comprehensive analysis of 11 countries in SSA showed that children with HIV-positive mothers attained 30% less education than the general population. Additionally, Cluver et al. (2013) and Mishra et al. (2007) showed that children with AIDS-ill parents had low attendance rates. Orkin et al. (2014) found that caregiver HIV/AIDS illness was associated with concentration problems via poverty and internalization of problems among South African adolescents. Zivin et al. (2009) found that children in early-stage ARV households and children in later-stage ARV households had similar increase in school attendance. Comparisons between orphans and children with HIV-infected parents by Tu et al. (2009) revealed that Chinese orphans have lower grades compared to children with HIV-infected parents. Ryder et al. (1994) also found that Congolese maternal orphans withdrew from school more often than children with HIV-positive mothers. On the contrary, Floyd et al.’s (2007) study in Malawi found no evidence of low grades among male and female children of HIV-positive individuals. Similarly, Grant (2008) found no differences in school enrolment between children with HIV-positive mothers and HIV-negative mothers.

Grant's mixed-methods study of mothers who were tested for HIV found that these parents were dedicated to ensuring that their children obtained their schooling, while they were still in control of their children's matters. Additionally, qualitative studies showed that caregivers (such as relatives and grandparents) reported that children were frequently out of school due to financial problems (e.g., Kakooza and Kimuna, 2006; Nyasani et al., 2009; Kembo, 2010; Fauk et al., 2017).

4 Discussion

Guo et al. (2012) and Goldberg and Short (2016) produced systematic literature reviews similar to this study. Our study adds to these earlier reviews by distinguishing three mechanisms through which HIV affects children's educational attainment. These are sickness of the child, orphanhood, and parental illness. Our study also adds to the literature by distinguishing three main effects of HIV on educational attainment of children: effects on sick children and orphans, effects on gender-gaps, and effects on the intergenerational transmission of education. This study employed more databases than the previous reviews that resulted in the inclusion of additional studies, including additional quantitative studies some of them published recently, as well as qualitative and mixed-methods studies. Results from these latter studies complemented the results of quantitative studies by providing explanations of the three mechanisms through which HIV affects educational attainment. In addition, papers were included in this review if there was confirmation of HIV/AIDS infection of the child, HIV/AIDS infection of the parent, AIDS death of the parent, or AIDS-illness in the family. This helped establish the direct effects of HIV on educational attainment of children.

The results from this systematic literature review show that the three mechanisms have different effects on different types of children and educational outcomes. However, in some cases, certain groups of children face similar issues. For example, HIV orphans and children with HIV-positive parents may be living with grandparents or other relatives. These two groups of children are likely to face similar problems with their education. One of the mechanisms that affect children's schooling is HIV-related sickness. Studies included in this review showed that HIV-infected children attended fewer school days (Mayes et al., 1996; Cohen et al., 1997; Anabwani et al., 2016), dropped out of school more frequently (Bele et al., 2011; Parchure et al., 2016), were more likely not to be in the correct grade for their age or to have repeated a grade (Bandason et al., 2013; Henning et al., 2018), and had low grades while in school (Ellis, 2004). These results indicate that physical illness is the main barrier to HIV-infected children's schooling. Anabwani et al. (2016) found that HIV-infected children's absenteeism from school was mainly due to frequent medical appointments and illness. One solution to this issue may be to increase access to ARV treatment and extra lessons for HIV-infected children. Souza et al. (2010) found that about 90% of Brazilian adolescents who were on highly active antiretroviral therapy were attending school. Voluntary teaching programs in some African countries (e.g., Zambia, Uganda, and Malawi) could also be a viable solution to the issue of absenteeism and attendance among HIV-infected children.

The results also showed that orphans were more likely to dropout or not be enrolled in school (Aaspas, 1999; Bele et al., 2011), experience grade delay (Kasirye and Hisali, 2010; Cluver et al., 2013), and have low attendance (Delva et al., 2009). This is in contrast to the fact that HIV-infected children faced delays to their education mainly due to illness. These results were complemented by qualitative results that showed that AIDS orphans’ education is interrupted due to financial problems (Nyasani et al., 2009; Kembo 2010; Fauk et al., 2017), lack of motivation (Jepkemboi and Aldridge, 2014), and disciplinary issues (Nyasani et al., 2009). These qualitative studies were based on interviews with AIDS orphans, their teachers, and their caregivers. They provided an overview of the issues faced by AIDS orphans. The studies showed that issues faced by orphans are complex and that the mere provision of food and shelter is not necessarily sufficient. They may also need assistance with school fees and school supplies. Some international organizations such as SOS Children's Villages have initiated programs that meet the educational needs of orphans. Their programs provide comprehensive services to orphans by building family environments and providing a holistic approach to child-centered education. Collaboration with such organizations may ensure that orphans and children at risk have the comprehensive care needed to achieve their educational goals. Free universal education may also help. For example, the Universal Primary Education Program in Uganda provides tuition assistance to all eligible primary school children. However, with such programs, many children need further assistance with school supplies and uniforms (Kakooza and Kimuna, 2006).

School children with HIV/AIDS-ill parents may miss school or may not be in the correct grade due to the need to provide care to parents (Harms et al., 2010; Cluver et al., 2012; Pufall et al., 2014a, 2014b). ARV treatment has been shown to reverse HIV-related adult morbidity and mortality (Zivin et al., 2009; Wang et al., 2016). Zivin et al. (2009) found that providing Kenyan children with HIV-positive parents ARV treatment led to a significant increase in weekly hours of schooling. Scaling up ARV treatment for parents living with HIV could help them remain healthy and economically active, thereby avoiding delays to their children's education (Delva et al., 2009). As in the case of orphans, children of HIV positive parents are also likely to live with their grandparents (Floyd et al., 2007). This is supported by the qualitative study conducted by Harms et al. (2010) who found that HIV/AIDS orphans stated that their orphanhood status started with the illness of their parents as opposed to the death of their parents. Children living with sick adults, particularly girls face the burden of providing care and performing adult chores (Yamano and Jayne, 2005). In addition, they are also more likely to be living with their grandparents or other relatives. Grandparents are likely to have only limited resources and may be too frail to work. Therefore, children living with grandparents, even when their AIDS-ill parents are alive, may face similar issues as AIDS orphans living with their grandparents. Floyd et al. (2007) suggested that foster carers (including grandparents) should be supported regardless of age and relationship to the child. Projects such as the Young Carers South Africa that help governments identify children who live in AIDS-sick homes and provide them with social welfare grants, home visits, and free school meals may help reduce these problems.

Most of the papers in this study examined effects of HIV on sick children, orphans, and children with HIV/AIDS-ill parents. A few studies (5 out of 57) mainly discussed effects on gender gaps in educational attainment (Bhargava, 2005; Poulsen, 2006; Zivin et al., 2009; Kitara et al., 2013; Hensels et al., 2016). Most of the studies that discussed gender issues controlled for the gender variable or gender of the household head. Bhargava (2005) found that girls who were maternal orphans were less likely to participate in school. On the other hand, Hensels et al. (2016) found that girls had better educational outcomes than boys. Kitara et al. (2013) examined nonorphaned, non-HIV-orphaned, and HIV-orphaned girls. They found that nonorphaned and non-HIV-orphaned girls had a more positive attitude toward school compared to HIV orphaned girls. These few studies indicate that results on intergender and intragender issues among children affected by HIV are complex and remain underexplored. Examination of gender gaps and gender issues among children affected by HIV requires attention, given the fact that HIV-affected girls are also likely to experience the effects of patriarchal norms and stigma that could significantly affect their education (Cluver et al., 2013; Madiba and Ngwenya, 2017).

Despite the advantages of our review design, there are also a few limitations that need to be acknowledged. Specifically, we only included studies that confirmed HIV infection of the child or parent/guardian and HIV/AIDS death of a parent. This leaves out studies that examined effects of orphanhood on educational attainment. Additionally, our systematic literature review includes quantitative, mixed-methods, and qualitative papers, which makes difficult to standardize the comparisons among the papers. Despite these limitations, we were able to have a comprehensive set of studies that provided insight on issues faced by children affected by HIV.

5 Conclusion

The results of our systematic review showed the mechanisms that influence the relationship between HIV/AIDS and children's education. Differences were observed between HIV-infected and uninfected children, between HIV-affected boys and HIV-affected girls, and children with HIV-infected parents and other groups of children. HIV-infected children mainly miss school days due to illness, orphaned children mainly lack financial means and motivation, and children with HIV-infected parents may care for their parents and or face similar issues as orphans. It is important to distinguish these mechanisms and groups of children so as to adequately formulate policy prescriptions (Evans and Miguel, 2007). Orkin et al. (2014) is the only study that conducted path analyses between familial HIV/AIDS and educational outcomes. They found that HIV/AIDS affected educational outcomes indirectly via orphanhood and parental/caregiver illness through poverty and internalization of problems. Therefore, it is advisable to focus on interventions that reduce stigma rather than targeting individual families (Orkin et al., 2014). More studies on path analyses are needed so as to further inform policy.

Only a few studies examined gender gaps in educational attainment among children affected by HIV. Therefore, there is no conclusive evidence on whether HIV-infected girls, female AIDS-orphans, or girls with HIV-positive parents face more delays in schooling compared to their male counterparts. Additionally, because issues faced by children affected by HIV are complex, more mixed-methods and qualitative studies are needed to further understand the pathways that influence the relationship between HIV/AIDS and educational attainment of children. In particular, qualitative studies (through interviews and focus groups) could provide insight on these mechanisms by highlighting stories of different groups of children and caregivers.

Appendix A

List for quantitative, mixed-methods, and qualitative papers included in the review

1. Akbulut-Yuksel, Mevlude, & Turan, Belgi. (2013). Left Behind: Intergenerational Transmission of Human Capital in the Midst of HIV/AIDS. Journal of Population Economics, 26(4), 1523–1547.

2. Anabwani, Gabriel, Karugaba, Grace, & Gabaitiri, Lesego. (2016). Health, schooling, needs, perspectives and aspirations of HIV infected and affected children in Botswana: a cross-sectional survey. BMC Pediatrics, 16(1), 132–132.

3. Aspaas, Helen Ruth. (1999). AIDS and orphans in Uganda: Geographical and gender interpretations of household resources. The Social Science Journal, 36(2), 201–226.

4. Bandason, T., Langhaug, L. F., Makamba, M., Laver, S., Hatzold, K., Mahere, S., … & Ferrand, R. A. (2013). Burden of HIV among primary school children and feasibility of primary school-linked HIV testing in Harare, Zimbabwe: a mixed methods study. AIDS Care, 25(12), 1520–1526.

5. Bele, Samir D., Valsangkar, Sameer, & Bodhare, Trupti N. (2011). Impairment of nutritional, educational status and quality of life among children infected with and belonging to families affected by human immunodeficiency virus/acquired immune deficiency syndrome. Vulnerable Children & Youth Studies, 6(4), 284–292.

6. Bhargava, Alok. (2005). AIDS epidemic and the psychological well-being and school participation of Ethiopian orphans. Psychology, Health & Medicine, 10(3), 263–275.

7. Cluver, L., Operario, D., Lane, T., & Kganakga, M. (2012). “I Can't Go to School and Leave Her in so Much Pain”: Educational Shortfalls among Adolescent “Young Carers” in the South African AIDS Epidemic. Journal of Adolescent Research, 27(5), 581–605.

8. Cluver, Lucie, Orkin, Mark, Boyes, Mark E., Sherr, Lorraine, Makasi, Daphne, & Nikelo, Joy. (2013). Pathways from parental AIDS to child psychological, educational and sexual risk: developing an empirically-based interactive theoretical model. Social Science & Medicine (1982), 87, 185–193.

9. Cohen, J., Reddington, C., Jacobs, D., Meade, R., Picard, D., Singleton, K., Hsu, H. W. (1997). School-related issues among HIV-infected children. Pediatrics, 100(1), E8–E8.

10. Curley, Jami, Ssewamala, Fred, & Han, Chang-Keun. (2010). Assets and educational outcomes: Child Development Accounts (CDAs) for orphaned children in Uganda. Children and Youth Services Review, 32(11), 1585–1590.

11. Delva, Wim, Vercoutere, An, Loua, Catherine, Lamah, Jonas, Vansteelandt, Stijn, De Koker, Petra, Annemans, Lieven. (2009). Psychological well-being and socio-economic hardship among AIDS orphans and other vulnerable children in Guinea. AIDS Care, 21(12), 1490–1498.

12. Ellis, Walter L. (2004). Factors Associated with the Academic Achievement of Perinatally HIV-Infected Elementary and Middle School Children. Negro Educational Review, 55(1), 51–58.

13. Fauk, Nelsensius Klau, Mwakinyali, Silivano Edson, Putra, Sukma, & Mwanri, Lillian. (2017). The socio-economic impacts of AIDS on families caring for AIDS-orphaned children in Mbeya rural district, Tanzania. International Journal of Human Rights in Healthcare, 10(2), 132–145.

14. Floyd, S., Crampin, A. C., Glynn, J. R., Madise, N., Mwenebabu, M., Mnkhondia, S., Fine, P. E. M. (2007). The social and economic impact of parental HIV on children in northern Malawi: retrospective population-based cohort study. AIDS Care, 19(6), 781–790.

15. Fofana, N. B., van Ophem, J. A., Niehof, A., & Antonides, G. (2014). Effects of HIV/AIDS and microfinance of women on income, medical expenditures and schooling in Côte d'Ivoire. African Development Review, 26(2), 322–332.

16. Fotso, A. S., Banjo, O., & Akmyemi, J. O. (2018). HIV and adolescents' educational attainment in South Africa: Disentangling the effect of infection in children and household members. South African Journal of Child Health, 12(SPE), s4–s9.

17. Grant, Monica J. (2008). Children's school participation and HIV/AIDS in rural Malawi: The role of parental knowledge and perceptions. Demographic Research, 19, 1603–1634.

18. Gupta, Anil K., Rawat, Nidhi, Rai, Kuldeep, Rana, Surendra, & Chakraborty, Sabyasachi. (2013). Orphan and vulnerable children infected or affected by HIV/AIDS in Delhi – situational analysis and state government's initiative of household economic strengthening. Vulnerable Children & Youth Studies, 8(2), 161–170.

19. Harms, Sheila, Jack, Susan, Ssebunnya, Joshua, & Kizza, Ruth. (2010). The orphaning experience: Descriptions from Ugandan youth who have lost parents to HIV/AIDS. Child and Adolescent Psychiatry and Mental Health, 4(1), 6.

20. Harrison, Sayward E., Li, Xiaoming, Zhang, JiaJia, Chi, Peilian, Zhao, Junfeng, & Zhao, Guoxiang. (2017). Improving School Outcomes for Children Affected by Parental HIV/AIDS: Evaluation of the ChildCARE Intervention at 6-, 12-, and 18-Months. School Psychology International, 38(3), 264–286.

21. Harrison, Sayward E., Li, Xiaoming, Zhang, JiaJia, Zhao, Junfeng, & Zhao, Guoxiang. (2018). A Randomized Controlled Trial of a Resilience-Based Intervention for Children Affected by Parental HIV: Educational Outcomes at 24-, 30-, and 36-Months. School Psychology International, 39(2), 170–195.

22. Hartell, C. G., & Chabilall, J. A. (2005). HIV/AIDS in South Africa: A study of the socio-educational development of adolescents orphaned by AIDS in child-headed households. International Journal of Adolescence and Youth, 12(3), 213–229.

23. Henning, Margaret J., Betancourt, Theresa S., & Khanna, Sunil K. (2016). Protective factors that contribute to improved school attendance for children that are HIV/AIDS affected in Zambia. International Journal of Health Promotion and Education, 54(6), 318–334.

24. Henning, Margaret, Kirk, Catherine M., Franchett, Emily, Wilder, Rose, Sezibera, Vincent, Ukundineza, Christian, & Betancourt, Theresa. (2018). Over-age and underserved: a case control study of HIV-affected children and education in Rwanda. Vulnerable Children & Youth Studies, 13(1), 81–93.

25. Hensels, IS, Sherr, L, Skeen, S, Macedo, A, Roberts, KJ, & Tomlinson, M. (2016). Do not forget the boys–gender differences in children living in high HIV-affected communities in South Africa and Malawi in a longitudinal, community-based study. AIDS Care, 28(sup2), 100–109.

26. Hong, Yan. (2011). Care Arrangements of AIDS Orphans and Their Relationship with Children's Psychosocial Well-Being in Rural China. Health Policy and Planning, 26(2), 115–123.

27. Jepkemboi, Grace, & Aldridge, Jerry. (2014). Effect of HIV/AIDS on Children's Attitudes toward Learning: Voices of Teachers and Caregivers in Western Kenya. Childhood Education, 90(3), 219–224.

28. Jere, Catherine M. (2012). Improving Educational Access of Vulnerable Children in High HIV Prevalence Communities of Malawi: The Potential of Open and Flexible Learning Strategies. International Journal of Educational Development, 32(6), 756–763.

29. Kakooza, J., & Kimuna, S. R. (2006). HIV/AIDS orphans' education in Uganda: the changing role of older people. Journal of Intergenerational Relationships, 3(4), 63–81.

30. Kasirye, Ibrahim, & Hisali, Eria. (2010). The socioeconomic impact of HIV/AIDS on education outcomes in Uganda: School enrolment and the schooling gap in 2002/2003. International Journal of Educational Development, 30(1), 12–22.

31. Kembo, Joshua. (2010). Social and economic consequences of HIV and AIDS on children: case study of a high-density community in Harare, Zimbabwe. Sahara J-Journal of Social Aspects of Hiv-Aids, 7(4), 39–46.

32. Kidman, Rachel, Hanley, James A., Foster, Geoff, Subramanian, S. V., & Heymann, S. Jody. (2012). Educational Disparities in AIDS-Affected Communities: Does Orphanhood Confer Unique Vulnerability? Journal of Development Studies, 48(4), 531–548.

33. Kitara, David Lagoro, Amongin, Hellen Christine, Oonyu, Joseph C., & Baguma, Peter K. (2013). Assertiveness and Attitudes of HIV/AIDS Orphaned Girls Towards Education in Kampala (Uganda). African Journal of Infectious Diseases, 7(2), 36–43.

34. Lucas, A. M., Chidothe, M., & Wilson, N. L. (2019). Effects of adult health interventions at scale on children's schooling: Evidence from antiretroviral therapy in Zambia. Economics of Education Review, 72, 107–120.

35. Luseno, W., Zhang, L., Rusakaniko, S., Cho, H., & Hallfors, D. (2015). HIV infection and related risk behaviors: does school support level the playing field between orphans and nonorphans in Zimbabwe? AIDS Care, 27(9), 1191–1195.

36. Mayes, S. D., Handford, H. A., Schaefer, J. H., Scogno, C. A., Neagley, S. R., Michael-Good, L., & Pelco, L. E. (1996). The relationship of HIV status, type of coagulation disorder, and school absenteeism to cognition, educational performance, mood, and behavior of boys with hemophilia. The Journal of Genetic Psychology, 157(2), 137–151.

37. Mialky, E., Vagnoni, J., & Rutstein, R. (2001). School-age children with perinatally acquired HIV infection: medical and psychosocial issues in a Philadelphia cohort. AIDS patient care and STDs, 15(11), 575–579.

38. Mishra, V., Arnold, F., Otieno, F., Cross, A., & Hong, R. (2007). Education and nutritional status of orphans and children of HIV-infected parents in Kenya. AIDS Educ Prev, 19(5), 383–395.

39. Mokgatle, Mathildah M, & Madiba, Sphiwe. (2015). The burden of disease on HIV-infected orphaned and non-orphaned children accessing primary health facilities in a rural district with poor resources in South Africa: a cross-sectional survey of primary caregivers of HIV-infected children aged 5–18 years. Infectious Diseases of Poverty, 4(1), 18.

40. Mon, M. M., Saw, S., Nu-Oo, Y. T., San, K. O., Myint, W. W., & Nge, P. T. (2013). Threat of HIV/AIDS in children: social, education and health consequences among HIV orphans and vulnerable children in Myanmar. WHO South-East Asia journal of public health, 2(1), 41.

41. Nicholson, Laura, Chisenga, Molly, Siame, Joshua, Kasonka, Lackson, & Filteau, Suzanne. (2015). Growth and health outcomes at school age in HIV-exposed, uninfected Zambian children: follow-up of two cohorts studied in infancy. BMC Pediatrics, 15, 66-66.

42. Nyasani, Evalyne, Sterberg, Erna, & Smith, Helen. (2009). Fostering children affected by AIDS in Richards Bay, South Africa: A qualitative study of grandparents' experiences. African Journal of AIDS Research, 8(2), 181–192.

43. Orkin, Mark, Boyes, Mark E, Cluver, Lucie D, & Zhang, Yuning. (2014). Pathways to poor educational outcomes for HIV/AIDS-affected youth in South Africa. AIDS Care, 26(3), 343–350.

44. Osuji, H. L., Nabunya, P., Byansi, W., Parchment, T. M., Ssewamala, F., McKay, M. M., & Huang, K. Y. (2018). Social support and school outcomes of adolescents orphaned and made vulnerable by HIV/AIDS living in South Western Uganda. Vulnerable children and youth studies, 13(3), 228–238.

45. Parchure, Ritu, Jori, Vijaya, Kulkarni, Sanjeevani, & Kulkarni, Vinay. (2016). Educational outcomes of family-based HIV-infected and affected children from Maharashtra, India. Vulnerable Children & Youth Studies, 11(4), 332–338.

46. Poulsen, Helen. (2006). The gendered impact of HIV/AIDS on education in South Africa and Swaziland: Save the Children's experiences. Gender & Development, 14(1), 47–56.

47. Pufall, Erica L., Gregson, Simon, Eaton, Jeffrey W., Masoka, Tidings, Mpandaguta, Edith, Andersen, Louise, Campbell, Catherine. (2014). The contribution of schools to supporting the well being of children affected by HIV in eastern Zimbabwe. AIDS (London, England), 28 Suppl 3, S379–387.

48. Pufall, Erica L, Nyamukapa, Constance, Eaton, Jeffrey W, Campbell, Catherine, Skovdal, Morten, Munyati, Shungu, Gregson, Simon. (2014). The impact of HIV on children's education in eastern Zimbabwe. AIDS Care, 26(9), 1136–1143.

49. Ryder, R. W., Kamenga, M., Nkusu, M., Batter, V., & Heyward, W. L. (1994). AIDS orphans in Kinshasa, Zaire: incidence and socioeconomic consequences. AIDS (London, England), 8(5), 673–679.

50. Santelli, J. S., Mathur, S., Song, X., Huang, T. J., Wei, Y., Lutalo, T., … & Serwadda, D. (2015). Rising school enrollment and declining HIV and pregnancy risk among adolescents in Rakai District, Uganda, 1994–2013. Global Social Welfare, 2(2), 87–103.

51. Sherr, Lorraine, Tomlinson, Mark, Macedo, Ana, Skeen, Sarah, Hensels, Imca Sifra, & Cluver, Lucie Dale. (2017). Can cash break the cycle of educational risks for young children in high HIV-affected communities? A cross-sectional study in South Africa and Malawi. Journal of Global Health, 7(2), 020409–020409.

52. Skovdal, Morten, & Ogutu, Vincent O. (2009). “I washed and fed my mother before going to school”: understanding the psychosocial well-being of children providing chronic care for adults affected by HIV/AIDS in Western Kenya. Globalization and health, 5, 8. https://doi.org/10.1186/1744-8603-5-8

53. Souza, Edvaldo, Santos, Nicole, Valentini, Sophia, Silva, Gerlane, & Falbo, Ana. (2010). Long-term follow-up outcomes of perinatally HIV-infected adolescents: infection control but school failure. Journal of tropical Pediatrics, 56(6), 421–426.

54. Ssewamala, Fred M., & Ismayilova, Leyla. (2009). Integrating Children's Savings Accounts in the Care and Support of Orphaned Adolescents in Rural Uganda. Social Service Review, 83(3), 453–472.

55. Ssewamala, Fred M., Wang, Julia Shu-Huah, Neilands, Torsten B., Bermudez, Laura Gauer, Garfinkel, Irwin, Waldfogel, Jane, … Kirkbride, Gwyneth. (2018). Cost-Effectiveness of a Savings-Led Economic Empowerment Intervention for AIDS-Affected Adolescents in Uganda: Implications for Scale-up in Low-Resource Communities. The Journal of adolescent health: official publication of the Society for Adolescent Medicine, 62(1S), S29–S36.

56. Toska, E., Cluver, L., Orkin, M., Bains, A., Sherr, L., Berezin, M., & Gulaid, L. (2019). Screening and supporting through schools: educational experiences and needs of adolescents living with HIV in a South African cohort. BMC public health, 19(1), 272.

57. Tu, Xiaoming, Lv, Yunfei, Li, Xiaoming, Fang, Xiaoyi, Zhao, Guoxiang, Lin, Xiuyun, … Stanton, Bonita. (2009). School performance and school behaviour of children affected by acquired immune deficiency syndrome (AIDS) in China. Vulnerable Children and Youth Studies, 4(3), 199–209.

58. Xu, T., Wu, Z., Duan, S., Han, W., & Rou, K. (2010). The situation of children affected by HIV/AIDS in Southwest China: Schooling, physical health, and interpersonal relationships. J Acquir Immune Defic Syndr, 53 Suppl 1, S104-110.

59. Xu, Tao, Wu, Zunyou, Rou, Keming, Duan, Song, & Wang, Huishan. (2010). Quality of life of children living in HIV/AIDS-affected families in rural areas in Yunnan, China. AIDS Care, 22(3), 390–396.

60. Xu, Tao, Wu, Zunyou, Yan, Zhihua, Rou, Keming, & Duan, Song. (2010). Measuring health-related quality of life in children living in HIV/AIDS-affected families in rural areas in Yunnan, China: preliminary reliability and validity of the Chinese version of PedsQL™ 4.0 Generic Core Scales. Journal of Acquired Immune Deficiency Syndromes (1999), 53(Suppl 1), S111.

61. Yang, H., Wu, Z., Duan, S., Li, Z., Li, X., Shen, M., … Stanton, B. (2006). Living environment and schooling of children with HIV-infected parents in southwest China. AIDS Care, 18(7), 647–655.

62. Zivin, Joshua Graff, Thirumurthy, Harsha, & Goldstein, Markus. (2009). AIDS Treatment and Intrahousehold Resource Allocation: Children's Nutrition and Schooling in Kenya. Journal of Public Economics, 93(7–8), 1008–1015.

Appendix B: Summary of Quantitative Results

 

Ref no.	Author (year)	Country	Children's age (years)	Education variable	HIV status or testing	Comparisons	Methods	Mixed-methods study	Gender comparison of educational attainment	Effects of parental HIV/AIDS on educational attainment	Main results
1.	Akbulut-Yuksel and Turan (2013)	11 SSA countries	13–17	Years of schooling, progress in school, and attendance	Parents and children were tested during survey	HIV-positive mothers vs. HIV-negative mothers	Cross-sectional data, N = 8992, fixed effects regression	No	No	Yes	Mother's HIV status has significant effects on inheritability of education. The association between infected mothers' education and their children's education is 30% less than the general population. Children with HIV positive mothers had decreased school progress and attendance.
2	Anabwani et al. (2016)	Botswana	6–18	Attendance and grades	HIV results for children were documented	HIV-infected children vs. children living with an HIV-infected child in the household	Cross-sectional data, N = 984, chi-square test, and z-test	Yes	No	No	About 99% of HIV-infected and about 97% of HIV-affected children were attending school. Sixty percent of HIV-infected children missed at least one day of school due to illness or medical appointments. Seventy-eight percent of HIV infected children and 62% of HIV affected children reported facing problems at school.
3	Aspaas (1999)	Uganda	0–17	Enrollment	Households had AIDS orphans	HIV/AIDS orphans vs. nonorphans	Cross-sectional data, N = 60, t-test	No	No	Yes	AIDS orphans in male-headed households were enrolled at lower rates than orphans in female-headed households.
4	Bandason et al. (2013)	Zimbabwe	11–13	Correct grade for age	Teachers, pupils, and their families received HIV testing	N/A	Cross-sectional data, N = 4386, logistic regression	Yes	No	No	Being HIV-positive was significantly associated with being behind by at least 1 class grade for age.
5	Bele et al. (2011)	India	5–11	Attendance, dropout, school performance	Children were HIV-infected or had lost a parent to HIV/AIDS	HIV-infected children vs. HIV orphans	Cross-sectional, N = 387, z-scores	No	No	Yes	The main reason for dropout in the HIV-infected group was illness. The main reason for dropout in the HIV orphan group was financial constraints.
6	Bhargava (2005)	Ethiopia	10+	School participation before and after mother's death	Children were AIDS orphans	AIDS orphans vs. non-AIDS orphans	Cross-sectional data, N = 1053, bivariate logistic regression	No	Yes	Yes	A mother's premature death hindered children's school participation. AIDS orphans were more likely to participate in school than non-AIDS orphans. Following the death of their mothers, girls were less likely than boys to participate in school.
7	Cluver et al. (2012)	South Africa	10–20	Attendance, dropout, and inability to concentrate	Children lived in a household with an AIDS-ill individual	Adolescents in homes with AIDS-sickness vs. adolescents in other-sick homes vs. adolescents in healthy homes.	Cross-sectional data, N = 599, ANOVA, t-test, and multivariate regression	Yes	No	No	Living in a home with an AIDS-sick person was significantly related to missing or dropping out of school. Living in a home with other sickness was also related to dropping out of school but this effect was less than that of living in a home with an AIDS-sick person.
8	Cluver et al. (2013)	South Africa	10–17	Enrollment, attendance, and correct grade for age	AIDS-orphaned children vs. children who lived with AIDS-ill parents vs. non-HIV/AIDS affected children	Children were AIDS orphans or lived with an AIDS-ill parent	Cross-sectional data, N = 6002, maximum likelihood estimation, chi-square test, and ANOVA	No	No	No	AIDS-orphaned children showed that more grade delay and those with AIDS-ill parents had low attendance and enrollment.
9	Cohen et al. (1997)	USA (Massachusetts)	5–17	Attendance	Children were infected with HIV	HIV-infected children with mild symptoms vs. HIV-infected children with moderate symptoms vs. HIV-infected children with severe symptoms	Longitudinal data, N = 92, descriptive statistics	No	No	Yes	49% of children missed two or more weeks of school, 12% missed more than eight weeks. Only three children missed school because of mother's illness. 75% were absent for less than two weeks, compared to 51% of children moderate symptoms and 27% of children with severe symptoms.
10	Curley et al. (2010)	Uganda	13.7	Grades.	Self-identified AIDS orphans	Experimental group orphans vs. comparison group orphans	Longitudinal data, N = 274, quasi-experimental design/OLS	No	Yes	No	Children in the experimental group are more likely to have better grades, and have positive changes in their educational plans. Girls are more likely to positive changes in educational plans than boys.
11	Delva et al. (2009)	Guinea	10–18	Attendance	Self-identified AIDS orphans	AIDS-orphans vs. non-AIDS orphan vs. nonorphans	Cross-sectional data, N = 397, logistic regression	No	Yes	No	Compared to nonorphans, non-AIDS orphans and AIDS orphans were less likely to attend school. A trend of decreased attendance could be observed among AIDS orphans. Boys were more likely to attend school on a daily basis compared to girls.
12	Ellis (2004)	USA (North Carolina)	9 years old	Truancy, and low grades	Children were perinatally infected with HIV.	None	Cross-sectional data, N = 9, bivariate analysis, t-test	No	No	No	HIV-related medical factors were not associated with children making low grades. However, poor-study habits, disruptive behavior, and peer pressure were found to be associated with low grades.
14	Floyd et al. (2007)	Malawi	6–28	Grades and attendance	Parents were tested for HIV during survey	HIV-positive parents vs. HIV-negative parents	Longitudinal data, N = 2525, logistic regression and linear regression	No	No	Yes	There was no evidence that the mean grade was lower among children of an HIV-positive individual, for both girls and boys. Secondary school attendance was lower in maternal and paternal orphans.
15	Fofana et al. (2014)	Côte d'Ivoire	6–14	Attendance	Mothers were HIV positive	HIV-positive women vs. HIV-negative women	Cross-sectional data, N = 439, Heckman's two-step regression	No	No	Yes	HIV was found to be the main cause of children dropping out of school. In households with HIV-positive people, fewer children go to school.
16	Fotso et al. (2018)	South Africa	10–19	Enrollment and school progression (correct grade for age)	Children were tested for HIV	HIV-positive vs. HIV negative children	Longitudinal data, N = 8835, logistic regression, multivariate decomposition	No	No	No	HIV contracted in childhood and early adolescence resulted in a significant school progression gap between HIV-negative and HIV-positive children.
17	Grant (2008)	Malawi	6–16	Enrollment	Mothers were tested for HIV during survey	HIV-positive mothers vs. HIV-negative mothers	Longitudinal data, N = 2308, logistic regression	Yes	No	Yes	A mother's HIV status is not significantly associated with school enrollment of 6–10-year-old children. Women who reported a medium or high likelihood of future HIV infection had children with higher odds of being currently enrolled compared to children of women who reported no chance of future infection. There was no difference in school participation between children with HIV-positive mothers and HIV-negative mothers.
18	Gupta et al. (2013)	India	2–18	Dropout and attendance	Children had lost one or both parents to HIV	Orphans and vulnerable children in institutional care vs. those in home-based care	Longitudinal data, N = 65, descriptive statistics	No	No	No	About 72% of Orphans and vulnerable children (OVC) in home-based care were going to school, whereas 98% of OVC in institutional care were attending school.
20	Harrison et al. (2017)	China	6-17	Grades	Children had at least one HIV-positive biological parent or were AIDS orphans	Experimental group of HIV/AIDS affected children vs. comparison group of HIV/AIDS affected children	Longitudinal data, N = 790, RCT, random, and fixed effects	No	Yes	No	Child and caregiver interventions displayed improvements in academic performance. Girls reported lower grades and less interest in school.
21	Harrison et al. (2018)	China	6–17	Grades	Children had at least one HIV-positive biological parent or were AIDS orphans	Experimental group of HIV/AIDS affected children vs. comparison group of HIV/AIDS affected children	Longitudinal data, N = 790, RCT, random and fixed effects	No	Yes	No	The number of HIV infections in the family had a significant negative impact on school satisfaction and on children's school interests. Boys reported more academic interests than girls.
23	Henning et al. (2018)	Rwanda	10–17	Correct grade for age	HIV positive children and parents were identified through an Electrical Medical Record for registered patients living with HIV	HIV-positive vs. HIV-affected vs. HIV-negative	Cross-sectional data, N = 681, logistic regression	No	No	No	HIV-positive children experienced higher levels of stigma and were twice as likely to be at least year older than their appropriate grade for age.
24	Henning et al. (2016)	Zambia	5–17	Attendance	Children lost one or more parents to HIV/AIDS or had a parent or household member who has been sick for three months or more	Orphans and vulnerable children vs. nonorphans and vulnerable children	Cross-sectional data, N = 1651, logistic regression	Yes	Yes	No	The number of orphans and vulnerable children in a household decreased the likelihood of school attendance. Gender did not affect school attendance. Teachers' support was critical to the attendance of HIV/AIDS-affected children.
25	Hensels et al. (2016)	South Africa and Malawi	4–13	Nonspecific educational outcomes	HIV status was obtained using parental report	HIV (and non-HIV) girls vs. HIV (and non-HIV) boys	Longitudinal data, N = 989, t-test, chi-squared test, linear regression	No	Yes	No	Results showed that boys and girls differed significantly in educational outcomes. A carer's HIV status had negative effect on children's educational outcomes. However, after controlling for school attendance, cognitive abilities, and carer HIV status, being a girl was significantly associated with better educational outcomes.
26	Hong (2011)	China	6–18	Grades	Children had lost both parents to AIDS	AIDS orphans living in kinship care (or extended family care) vs. AIDS orphans in orphanages, vs. AIDS orphans in community-based group homes	Longitudinal data, N = 296, linear regression	No	No	No	Children who lived in group homes had the best school grades followed by those in orphanages and those in kinship care (or extended family care).
29	Jere (2012)	Malawi	Standard 6 pupils	Dropout, and absenteeism	Children are HIV-positive, have lost one parent to AIDS, children have an HIV/AIDS-ill parent or guardian, or live in a household affected by HIV/AIDS	Experimental group of HIV/AIDS-affected children vs. comparison group of HIV/AIDS-affected children	Cross-sectional data, N = 259, Mann–Whitney U test	Yes	No	No	The mean of class dropout of the intervention group was lower than the control group. Pupil promotion was based on performance in school-based end-of-year exams. There was no significance in the difference in mean promotion rates between intervention and nonintervention.
31	Kasirye and Hisali (2010)	Uganda	6–17	Enrollment and grade progression	Children were HIV/AIDS orphans	HIV orphans vs. non-HIV orphans	Cross-sectional data, N = 1244, probit regression	No	No	No	HIV/AIDS orphans were about three years behind their appropriate grade. Poor HIV/AIDS orphans are likely to fall behind their appropriate grade.
32	Kembo (2010)	Zimbabwe	10–18	Dropout, attendance, and absenteeism	Children were AIDS orphans or lived with a chronically-ill person	HIV-affected children vs. HIV-unaffected children	Cross-sectional data, N = 386, t-test	Yes	No	No	About 72% of HIV-affected children were not in school compared to about 29% of HIV-unaffected children. The main reason was lack of money. About 13% of HIV-affected children were not in school because they did not have birth certificates.
33	Kidman et al. (2012)	Malawi	6–14	Enrollment and highest grade level attained	Children living with a parent with an AIDS-related illness were identified	Orphans vs. nonorphans	Cross-sectional data, N = 13,090, logistic and linear multilevel regression	No	Yes	Yes	There is little evidence that living with a chronically ill parent or in a household experiencing a recent adult death negatively impacts children's enrollment or grade attainment. Double and maternal orphans experience more educational deprivation compared to nonorphans. Being a maternal orphan, a stronger impact on enrollment on boys compared to girls.
34	Kitara et al. (2013)	Uganda	12–15	Attitude toward education	Children were AIDS orphans	HIV-orphans vs. non-HIV-orphans vs. nonorphans	Cross-sectional data, N = 255, t-test, ANOVA, Pearson's correlation, logistic regression	No	Yes	No	Nonorphans had a more positive attitude toward education. Non-HIV orphaned girls had more positive attitude toward school compared to HIV orphaned girls. This is attributed to HIV/AIDS-related stigma.
35	Lucas et al. (2019)	Zambia	7–15	Enrollment and correct grade for age	Children lived in households with HIV-positive adults	Children in households with HIV-positive heads vs. children in households with non-HIV-positive heads	Cross-sectional data, N = 12,128, difference in difference regression	No	Yes	Yes	Children who live in households with HIV positive household heads were more likely to be attending school.
36	Luseno et al. (2015)	Zimbabwe	15–21	Dropout and years of schooling	Children's HIV status were provided	Orphans vs. nonorphans	Cross-sectional and RCT, N = 751, logistic regression	No	No	No	Participants of experimental program had lower chances of school dropout compared to orphans from a nationally representative household survey.
37	Mayes et al. (1996)	USA	5–17	Grades and absenteeism	18 boys were HIV positive	HIV-positive hemophilic boys vs. HIV-negative hemophilic boys	Cross-sectional data, N = 66, t-test, Pearson's correlation coefficients	No	No	No	All 66 boys who participated in the study were boys diagnosed with hemophilia. HIV-positive boys missed more school days (14.8%) compared to HIV-negative boys (7.9%). Academic grades did not suffer significantly between the two groups.
38	Mialky et al. (2001)	USA (Philadelphia)	5–18	Grade repetition	Children were found to be HIV-positive	None	Cross-sectional data, N = 85, descriptive statistics	No	No	No	About 24% of HIV-infected children repeated at least one grade. 13.2% of caregivers described school performance of children as below average. The mean number of absences reported by caregivers was 5.2 times.
39	Mishra et al. (2007)	Kenya	6–14	Attendance	Parents and children were tested during survey	Orphans vs. fostered children vs. children with HIV-positive parents	Cross-sectional data, N = 6928, logistic regression	No	No	Yes	Orphaned children and fostered children are less likely to be attending school than children of HIV-negative parents. Children of HIV-positive parents are less likely to be attending school than children of HIV-negative parents.
40	Mokgatle and Madiba (2015)	South Africa	5–17	Attendance	Children were HIV-infected	None	Cross-sectional data, N = 406, descriptive statistics	No	No	No	96% of nonorphans and 96.7% of orphans were attending school.
41	Mon et al. (2013)	Myanmar	6–17	Attendance, dropout, absenteeism, and enrollment	Children were AIDS orphans	HIV orphans vs. non-HIV-related children	Cross-sectional data, N = 600, descriptive statistics, t-tests	No	No	Yes	There were no differences is schooling between in HIV orphans and non-HIV-related children. However, there were differences in dropout, absenteeism, and enrollment.
42	Nicholson et al. (2015)	Zambia	6–12	Maths and English grades	Mothers were HIV-positive and children were exposed to HIV in utero	HIV-exposed children (in utero) vs. HIV-unexposed children	Longitudinal data, N = 390, linear regression model	No	No	Yes	HIV-exposed children had lower math grades than HIV-unexposed children; however, there were no differences in English grades.
44	Orkin et al. (2014)	South Africa	11–25	Attendance, enrollment, and grade progression	HIV/AIDS-related parental deaths required a conservative threshold of three or more HIV/AIDS defining symptoms	None	Longitudinal data, N = 723, maximum likelihood estimation, Bayesian estimation, standardized regression	No	Yes	No	Neither caregiver HIV/AIDS-sickness nor HIV/AIDS orphanhood was associated with nonenrollment, nonattendance, concentration problems, or grade progression. Caregiver sickness was indirectly associated with nonenrollment educational variables through internalizing problems and poverty. Boys reported more concentration problems and difficulties with grade progression.
45	Osuji et al. (2018)	Uganda	13.8 years	Attendance and academic attendance	Children were AIDS orphans	None	Longitudinal, N = 346, multivariate regression analyses	No	No	Yes	Family cohesion, support from classmates was associated with more days of attending school. Support from caregivers, friends. and classmates was associated with fewer days of missed school.
46	Parchure et al. (2016)	India	6–16	Enrollment, dropout, and correct grade for age	Children were infected with HIV or had one or both parents infected with HIV (living or dead)	HIV-infected vs. HIV-affected	Cross-sectional data, N = 472, logistic regression	No	No	Yes	HIV-infected children were seven times more likely to be out of school compared to HIV-affected children. HIV-infected and maternal orphans were 2.82 and nine times more likely to not be in the correct grade for age, respectively. Child illness was that most common reason for low educational attainment.
48	Pufall et al. (2014a)	Zimbabwe	6–17	Attendance and correct grade for age	HIV tests were conducted on children aged 2–17	Vulnerable children (i.e., HIV-positive, orphans, or had HIV-positive parents) vs. nonvulnerable children	Cross-sectional data, N = 4577, logistic regression	No	Yes	Yes	Vulnerable Children are likely to be in correct grade for age but not in regular attendance. Older children were less likely to be in correct grade for age. Females were more likely to be in the correct grade for age compared to males.
49	Pufall et al. (2014b)	Zimbabwe	6–24	Completion of primary school, attendance, and correct grade for age	HIV tests were conducted on children aged 2–17	HIV-positive children vs. children with HIV positive parents vs. orphans vs. young carers	Cross-sectional data, N = 5520, logistic regression	No	No	No	Being HIV-positive was not associated with any education measures in youth and children. Young carers were less likely to attend secondary school. Orphans were less likely to be in the correct grade for age.
50	Ryder et al. (1994)	D.R. Congo	0–15	Dropout	Mothers were HIV-positive	HIV maternal orphans vs. children with HIV-positive mothers vs. children with HIV-negative mother	Longitudinal data, N = 78, Yates’ corrected chi-square test, and Fisher's exact test	No	No	No	55.6% of maternal orphans were forced to withdraw from school compared 25.0% of children with HIV-positive mothers and 40% children of HIV-negative mothers.
51	Sherr et al. (2017)	Malawi and South Africa	5–15	Enrollment, attendance, and correct grade for age	Child HIV status was determined using caregiver reports	HIV-positive children vs. HIV-negative children with/without receipt of cash grant	Longitudinal data, N = 854, t-test, chi-square test, logistic regression, and multivariate regression	No	No	No	Among HIV-positive children, receiving a cash grant was associated with much more struggling in school than children who did not receive a grant. No effect was found on attendance, being in correct grade for age, or being a quick learner.
53	Souza et al. (2010)	Brazil	10–19	Attendance, dropout, and failure	Children were HIV-positive	Adolescents with high viral loads vs. adolescents with low viral loads	Cross-sectional data, N = 49, ANOVA, t-test, chi-square test, and Fisher's exact test	No	No	No	School failure and school dropout were reported by 51% and 28.6% of all participants. However, 30.8% of adolescents with a low viral load and 26.1% of those with high viral load reported school dropout. 46.2% of adolescents with high viral load and 56.5% of children with high viral load reported school failure.
54	Ssewamala and Ismayilova (2009)	Uganda	11–17	Attendance and grades	Adolescents were AIDS orphans	Experimental group orphans vs. comparison group orphans	RCT, N = 277, analysis of variance	No	No	No	There was a 27-percentage point increase in experimental group's certainty of accomplishing educational plans. The intervention did not have a significant effect on school attendance.
55	Ssewamala et al. (2018)	Uganda	12 years (average)	Dropout and attendance	Adolescents were AIDS orphans	Experimental group orphans vs. comparison group orphans	RCT, N = 1410, multilevel regression	No	No	No	On average, adolescents receiving both interventions showed lower dropout rates, higher likelihood to take national exam, and score higher on the exam. However, intervention effects were not statistically significant.
56	Toska et al. (2019)	South Africa	Adolescents <18 years	Absenteeism and correct grade for age	Adolescents were living with HIV	Adolescents living with HIV vs. uninfected adolescents	Cross-sectional data, N = 1519, multivariate regression	No	No	No	Having disease burden resulted in more frequent need to miss school to attend clinics.
57	Tu et al. (2009)	China	6–18	Grades	Children were living with HIV-infected parents or lost one or both parents to AIDS	Orphan/children with HIV positive parent vs. nonorphan	Cross-sectional data, N = 1625, linear regression model	No	No	No	Orphans had lower grades compared to children with HIV-infected parents and children with HIV-negative parents. There was no difference in educational expectation or student leadership.
58	Xu et al. (2010a)	China	8–17	Attendance dropout and	Children had at least 1 HIV-positive parent or had lost on or both parents to AIDS.	HIV-affected vs. HIV-unaffected	Cross-sectional data, N = 225, multivariate regression	No	No	No	Children living with grandparents reported higher scores in school functioning. Caregiver spending more time accompanying the child increased school functioning.
59	Xu et al. (2010b)	China	8–18	Attendance and dropout	Participants were recruited through health service providers who treat HIV-positive patients.	HIV-affected vs. HIV-unaffected children and caregivers	Cross-sectional data, N = 225, chi-square test	No	No	No	19% of HIV-unaffected children and 15% HIV affected children dropped out of school.
60	Xu et al. (2010c)	China	8–18	Attendance and dropout	Children had at least one HIV-positive parent or had lost on or both parents to AIDS.	Orphans vs. nonorphans	Cross-sectional data, N = 114, Descriptive statistics	No	No	No	75% of nonorphans and 73.4 of orphans were attending school, whereas 25% of nonorphans and 26.6% of orphans dropped out of school
61	Yang et al. (2006)	China	0–15	Attendance, truancy, and dropout	Rural household members contracted HIV through drug use	Orphans vs. nonorphans	Cross-sectional data, N = 266, chi-square test, Fisher's exact test	No	No	No	HIV-infected drug users between 16 and 50 years old were recruited. Orphans and older children between 6 and 15 years old were less likely to attend school and skip class compared to nonorphans.
62	Zivin et al. (2009)	Kenya	8–18	Attendance	Parents were on ARV treatment	Experimental group of HIV/AIDS affected children vs. comparison group of HIV/AIDS affected children	Longitudinal data, N = 480 quasi-experimental design, fixed effects	No	Yes	Yes	There is a similar increase over time in school attendance for children in early-stage ARV households, relative to children in later-stage ARV households. Children in later-stage ARV households do not experience any significant change in attendance. ARV treatment effects are large and significant in early stages of ARV treatment for girls and not significant for boys.

 

Summary of Qualitative Results

Ref no.	Author (year)	Country	Method	HIV-positive population/HIV-testing	Description	Education variable	Mixed-methods study	Results
2	Anabwani et al. (2016)	Botswana	Focus group discussions	Children had documented HIV results	19 HIV infected and 6 HIV affected children	Attendance	Yes	HIV-infected children reported no major problem in school. However, some reported missing school for medical reasons.
4	Bandason et al. (2013)	Zimbabwe	Focus group discussions, informal interviews, and exit interviews	Children were HIV-infected	Three teachers, six pupils (aged 11–13 years), five parents, two counselors	Attendance	Yes	HIV-infected children missed school.
7	Cluver et al. (2012)	South Africa	Interviews	Children lived in a household with an AIDS-ill individual	659 children and youths (aged 10–20)	Attendance and dropout	Yes	Children missed school or dropped out of school to care for sick adults.
13	Fauk et al. (2017)	Tanzania	In-depth interviews	Children were HIV/AIDS orphans	20 heads of household caring for AIDS-orphaned children, two government staff, two from a nongovernmental organization	Attendance, dropout	No	Participants did not send their children to school due to school fees and other school-related expenses.
17	Grant (2008)	Malawi	Interviews	Mothers were tested for HIV	60 adults (aged 25–50 years) who were a parent of at least one child aged 6–18 years	Enrollment and attendance	Yes	Parents were committed to ensure that their children were enrolled in school, while their children's matters were still in their control.
19	Harms et al. (2010)	Uganda	Interviews	Youth had lost or both parents to HIV/AIDS	13 youth (with mean age of 15) who had lost one or both parents to HIV/AIDS who were affiliated with a nongovernmental organization providing support to orphans	Attendance and absenteeism	No	Six of the youth were not attending school at the time of the study. The most poignant losses were actual death of parent and loss of educational opportunities. The time of parental sickness was marked by extended periods of absenteeism.
22	Hartell and Chabilall (2005)	South Africa	Case studies	Adolescents were AIDS orphans	Six case studies of adolescent AIDS orphans we selected, observed, and interviewed in their natural settings	Attendance and dropout	No	Two of the participants’ attendance was affected due to taking care of their critically ill parents, but attendance increased after death of parents. Two of the participants dropped out of school due to pregnancy
24	Henning et al. (2016)	Zambia	Focus group discussions and interviews	Children have lost one or more parents to HIV/AIDS or have a parent or household member who's been sick for three months or more	Six focus groups with children (aged 10–18) at six different schools (8–10 children per group). 12 key informant interviews	Attendance	Yes	Households with HIV/AIDS-affected children were more likely to have all HIV/AIDS-affected children attending school if all of the HIV/AIDS-affected children were relate to household head. Key informant interviews focused on two key themes connected to school attendance: (1) training and (2) stigma. Teachers were overwhelmed by the number of HIV/AIDS-affected children.
27	Jepkemboi and Aldridge (2009)	Kenya	Interviews	Children had HIV-positive parents or were AIDS orphans	12 teachers and eight caregivers from seven orphanages were interviewed	Attendance, math, science, and language grades	No	No gender differences were noted in preschool. However, for older children, boys performed better in math and science and girls were believed to be better in languages. Girls were likely to be taken out of school to take care of family needs when a parent became ill.
28	Jepkemboi and Aldridge (2014)	Kenya	Interviews	Children were HIV/AIDS orphans or had at least one HIV/AIDS-ill parent	12 teachers and eight caregivers from seven orphanages participated in study	Attitude toward school	No	A majority of children do not like to come to school. Teachers noticed that children's attitude toward school improves after two years being in orphanage. Girls were more persistent and had a more positive attitude.
29	Jere (2012)	Malawi	Focus group discussions and interviews	Children are HIV-positive, have lost one parent to AIDS, children have an HIV/AIDS-ill parent or guardian, or live in a household affected by HIV/AIDS	Key informant interviews were performed on teachers and school heads	Dropout, absenteeism, pupil promotion	Yes	Teachers attested that pupils targeted but the intervention became more capable and confident learners. Evidence from the interviews showed that this was a result of pupils’ perception of their improved competency in English and Mathematics.
30	Kakooza and Kimuna (2006)	Uganda	Focus group discussions	Children were HIV/AIDS orphans	12 focus group discussions were held from two subcounties. Participants were heads of household, grandparents, caring or HIV/AIDS orphans, and were aged 50 years or over	Attendance, dropout	No	Grandparents had difficulty providing for children's education
32	Kembo (2010)	Zimbabwe	Semistructured interviews and letter writing	Children were AIDS orphans or lived with a chronically ill person	Semistructured interviews were conducted with children (aged 10–18 years). 12 children were asked to write letters to their parents, regardless of their survival status and tell them how they feel about their lives.	Dropout, attendance, absenteeism	Yes	Qualitative data show that some children affected by HIV/AIDS lack money to attend school. They also face hunger that affects their performance in school.
43	Nyasani et al. (2009)	South Africa	Focus group discussions and interviews	Grandparents were registered as foster-carers to orphans affected by HIV/AIDS	Focus group discussions were conducted with a total of 45 participants. The groups were included elderly female foster-carers, community leaders, urban and rural elderly foster-carers. In-depth and key informant interviews were also performed.	Educational needs	No	Rural grandparents were concerned with meeting educational needs of children. The data revealed that the prospects for orphans’ tertiary education were disquieting for both rural and urban grandparents.
47	Poulsen (2006)	South Africa and Eswatini	Semistructured interviews	Children were HIV/AIDS orphans, had parents who were AIDS-ill, or were HIV-infected	Interviews were conducted with head teachers, teachers, parents and caregivers, school committee members, members of Orphans and Vulnerable Children Committees, members of Child Care Forums, students, and out-of-school children	Attendance	No	Children affected by HIV/AIDS were missing school or dropping out due to parental illness, abuse, disrupted family lives, lack of money, lack of support from home, living with grandparents, and household duties.
52	Skovdal and Ogutu (2009)	Kenya	Case studies	Children provide care for people chronically ill from AIDS	Data collection involved photography and three highlighted case studies	Grade repetition and dropout	No	Being a child carer compromised education, for example, through grade repetition and dropout. Young carers juggle household duties, caregiving, and education.

Appendix C

 

The MMAT used in this study can be found on this link:

https://mixedmethodsappraisaltoolpublic.pbworks.com/w/file/fetch/84371689/MMAT%202011%20criteria%20and%20tutorial%202011-06-29updated2014.08.21.pdf

There are four criteria (in question format) to be met for each qualitative or quantitative study component. A score of 25% is assigned for each criterion met. For mixed-methods study papers, there is an additional mixed-methods study component that contains three criteria. A score of 0 in the mixed-methods study component is equivalent to a score of 1 (or 25%) in the other study components; a score of 1 in the mixed-methods study component is equivalent to a score of 2 (or 50%); etc. (Pluye et al., 2011). The overall quality score for a mixed-methods paper is the lowest score of any of the three (qualitative, quantitative, or mixed) study components. For example, if the qualitative component of a mixed-methods study has a score of 2, the quantitative component has a score of 2, and the mixed-methods component has a score of 0, then overall score will be 25%.

Quality Assessment for Quantitative Studies Using the Mixed-Methods Appraisal Tool

			Methodological quality criteria for randomized/nonrandomized/descriptive studies
Ref. no.	Author (year)	Type of study	Is there good description of randomization/ minimum selection bias/ relevant sampling strategy?	Is there clear description of allocation concealment /appropriate measurements/ representative sample?	Are there complete outcome data/ comparable participants/ appropriate measurements?	Is there low dropout rate/ complete outcome data/ acceptable response rate?	Score
1	Akbulut-Yuksel and Turan (2013)	Nonrandomized	Yes	Yes	Yes	Yes	100%
2	Aspaas (1999)	Descriptive	Yes	Yes	Yes	I can't tell	75%
5	Bele et al. (2011)	Nonrandomized	I can't tell	Yes	Yes	No	50%
6	Bhargava (2005)	Nonrandomized	Yes	Yes	Yes	Yes	100%
8	Cluver et al. (2013)	Nonrandomized	Yes	Yes	Yes	Yes	100%
9	Cohen et al. (1997)	Nonrandomized	Yes	Yes	Yes	Yes	100%
10	Curley et al. (2010)	Nonrandomized	No	Yes	Yes	Yes	100%
11	Delva et al. (2009)	Nonrandomized	No	Yes	Yes	Yes	75%
12	Ellis (2004)	Descriptive	No	No	Yes	Yes	50%
14	Floyd et al. (2007)	Nonrandomized	Yes	Yes	Yes	Yes	100%
15	Fofana et al. (2014)	Nonrandomized	Yes	Yes	Yes	No	75%
16	Fotso et al. (2018)	Nonrandomized	Yes	Yes	Yes	Yes	100%
18	Gupta et al. (2013)	Nonrandomized	No	No	No	I can't tell	0%
20	Harrison et al. (2017)	Randomized	No	Yes	I can't tell	I can't tell	25%
21	Harrison et al. (2018)	Randomized	No	Yes	I can't tell	I can't tell	25%
23	Henning et al. (2018)	Nonrandomized	Yes	Yes	Yes	Yes	100%
25	Hensels et al. (2016)	Nonrandomized	Yes	No	I can't tell	Yes	100%
26	Hong (2011)	Nonrandomized	No	Yes	Yes	Yes	100%
31	Kasirye and Hisali (2010)	Nonrandomized	Yes	Yes	Yes	Yes	100%
33	Kidman et al. (2012)	Nonrandomized	Yes	Yes	Yes	Yes	100%
34	Kitara et al. (2013)	Nonrandomized	Yes	Yes	Yes	Yes	100%
35	Lucas et al. (2019)	Nonrandomized	Yes	Yes	Yes	Yes	100%
36	Luseno et al. (2015)	Randomized and nonrandomized	Yes	I can't tell	No	Yes	50%
37	Mayes et al. (1996)	Descriptive	Yes	Yes	Yes	Yes	100%
38	Mialky et al. (2001)	Descriptive	Yes	I can't tell	No	No	25%
39	Mishra et al. (2007)	Nonrandomized	Yes	Yes	Yes	Yes	100%
40	Mokgatle and Madiba (2015)	Descriptive	Yes	No	Yes	Yes	100%
41	Mon et al. (2013)	Nonrandomized	Yes	No	No	I can't tell	25%
42	Nicholson et al. (2015)	Nonrandomized	No	Yes	Yes	Yes	75%
44	Orkin et al. (2014)	Nonrandomized	No	Yes	Yes	Yes	75%
46	Parchure et al. (2016)	Nonrandomized	I can't tell	Yes	Yes	I can't tell	50%
45	Osuji et al. (2018)	Randomized	No	Yes	Yes	No	50%
48	Pufall et al. (2014a)	Nonrandomized	Yes	Yes	Yes	Yes	100%
49	Pufall et al. (2014b)	Nonrandomized	Yes	Yes	Yes	Yes	100%
50	Ryder et al. (1994)	Nonrandomized	Yes	Yes	Yes	I can't tell	75%
51	Sherr et al. (2017)	Nonrandomized	No	No	I can't tell	I can't tell	0%
53	Souza et al. (2010)	Nonrandomized	Yes	Yes	Yes	Yes	100%
54	Ssewamala et al. (2018)	Randomized	Yes	Yes	Yes	Yes	100%
55	Ssewamala et al. (2018)	Randomized	Yes	Yes	Yes	Yes	100%
56	Toska et al. (2019)	Nonrandomized	Yes	Yes	Yes	Yes	100%
57	Tu et al. (2009)	Nonrandomized	Yes	Yes	Yes	Yes	100%
58	Xu et al. (2010a)	Nonrandomized	No	Yes	Yes	I can't tell	75%
59	Xu et al. (2010b)	Descriptive	Yes	No	Yes	I can't tell	25%
60	Xu et al. (2010c)	Descriptive	Yes	No	Yes	Yes	75%
61	Yang et al. (2006)	Nonrandomized	No	Yes	Yes	Yes	75%
62	Zivin et al. (2009)	Nonrandomized	No	Yes	Yes	I can't tell	50%

Quality Assessment for Mixed-Methods Studies Using the Mixed-Methods Appraisal Tool

			Methodological quality criteria for randomized/nonrandomized/ descriptive studies				Methodological quality criteria for qualitative component of mixed-methods studies				Methodological quality criteria for mixed-methods studies
Ref. no.	Author (year)	Type	Is there good description of randomization/ minimum selection bias/ relevant sampling strategy?	Is there clear description of allocation concealment /appropriate measurements/representative sample?	Are there complete outcome data/comparable participants/appropriate measurements?	Is there low dropout rate/complete outcome data/acceptable response rate?	Are sources of data relevant to address research question?	Is the process of analyzing data relevant to address research question?	Is consideration given to how findings relate to context?	Is consideration given to how findings relate to researchers influence?	Is a mixed methods research design relevant?	Is it relevant to use qualitative and quantitative data?	Is consideration given to limitation of integrating qualitative and quantitative data?	Score
2	Anabwani et al. (2016)	Quantitative descriptive and qualitative	Yes	Yes	Yes	Yes	Yes	I can't tell	Yes	Yes	Yes	Yes	No	75%
4	Bandason et al. (2013)	Qualitative nonrandomized and qualitative	Yes	Yes	Yes	I can't tell	Yes	Yes	Yes	I can't tell	No	Yes	No	50%
7	Cluver et al. (2012)	Quantitative nonrandomized and qualitative	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	75%
17	Grant (2008)	Quantitative nonrandomized and qualitative	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	100%
24	Henning et al. (2016)	Quantitative nonrandomized and qualitative	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	75%
29	Jere (2012)	Quantitative nonrandomized and qualitative	Yes	Yes	No	Yes	I can't tell	Yes	Yes	Yes	Yes	Yes	No	75%
32	Kembo (2010)	Quantitative descriptive and qualitative	No	Yes	No	Yes	No	Yes	I can't tell	Yes	Yes	I can't tell	I can't tell	50%

Quality Assessment of Qualitative Studies Using the Mixed-Methods Appraisal Tool

Ref. no.	Author (year)	Are sources of data relevant?	Is the process of analyzing data relevant?	Is consideration given to how findings relate to context?	Is consideration given to how findings relate to researchers’ influence?	MMAT score
13	Fauk et al. (2017)	Yes	Yes	Yes	I can't tell	75%
19	Harms et al. (2010)	Yes	Yes	Yes	Yes	100%
22	Hartell and Chabilall (2005)	Yes	Yes	Yes	I can't tell	75%
27	Jepkemboi and Aldridge (2009)	Yes	Yes	Yes	No	75%
28	Jepkemboi and Aldridge (2014)	Yes	Yes	I can't tell	No	50%
30	Kakooza and Kimuna (2006)	Yes	Yes	Yes	Yes	100%
43	Nyasani et al. (2009)	Yes	Yes	Yes	Yes	100%
47	Poulsen (2006)	Yes	Yes	I can't tell	I can't tell	50%
52	Skovdal and Ogutu (2009)	Yes	Yes	I can't tell	Yes	75%

Appendix D

 

Prisma 2009 Checklist

Section/topic	#	Checklist item	Reported on page #
TITLE
Title	1	Identify the report as a systematic review, meta-analysis, or both.	1
ABSTRACT
Structured summary	2	Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number.	1
INTRODUCTION
Rationale	3	Describe the rationale for the review in the context of what is already known.	2
Objectives	4	Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS).	3
METHODS
Protocol and registration	5	Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number.
Eligibility criteria	6	Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.	4
Information sources	7	Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.	4
Search	8	Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.	5
Study selection	9	State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).	5
Data collection process	10	Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.
Data items	11	List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.
Risk of bias in individual studies	12	Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.
Summary measures	13	State the principal summary measures (e.g., risk ratio, difference in means).
Synthesis of results	14	Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I^2) for each meta-analysis.
Risk of bias across studies	15	Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).
Additional analyses	16	Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, metaregression), if done, indicating which were prespecified.
RESULTS
Study selection	17	Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.	7
Study characteristics	18	For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.	10
Risk of bias within studies	19	Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12).
Results of individual studies	20	For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.
Synthesis of results	21	Present results of each meta-analysis done, including confidence intervals and measures of consistency.
Risk of bias across studies	22	Present results of any assessment of risk of bias across studies (see Item 15).
Additional analysis	23	Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, metaregression [see Item 16]).
DISCUSSION
Summary of evidence	24	Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).	12
Limitations	25	Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).	13
Conclusions	26	Provide a general interpretation of the results in the context of other evidence, and implications for future research.	13
FUNDING
Funding	27	Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review.

From: Moher et al. (2009).

Appendix E

Review Protocol

Review title:

Effects of HIV on Children's Educational Attainment: A Systematic Literature Review

Review question(s)

What are the effects of HIV on children's schooling outcomes? Specifically,

1. What are the effects on HIV-affected children?

2. What are the effects of HIV on gender gaps in educational outcomes?

3. What are the effects of HIV on intergenerational (parent-to-child) transmission of education in case of HIV/AIDS?

Searches

Main Databases: PubMed, ERIC, EconLit, SocIndex, and WoS. Supplementary database: Google Scholar

Keywords

HIV/AIDS, children, education, intergenerational transmission, human capital investment

Types of study to be included

Peer-reviewed papers, papers, and reports from international organizations. There is no time restriction. Papers will be included if they examine the relationship between HIV (or being HIV-affected through parents being sick or parental death) and educational attainment. Papers will be excluded if there is no confirmation of HIV-infection or AIDS death.

Condition or domain being studied

The disease being examined is HIV infection and the outcome is schooling, that is, school enrollment, dropout, absenteeism, and correct grade for age.

Participants/population

The participants are school-going aged children (orphaned and nonorphans), HIV-infected parents, and caregivers/guardians for HIV-infected children.

Intervention(s), exposure(s)

Papers that confirm HIV-infection or AIDS-related death (of parents) and papers that have interventions that target or translate to educational outcomes of children affected by HIV.

Comparator(s)/control

HIV-negative children (orphans and nonorphans), HIV-negative parents/guardians, and HIV-positive or HIV-negative children of the opposite gender.

Context

Papers that discuss cognitive issues in HIV-infected children and those that discuss HIV risk will be excluded.

Outcome(s)

Primary outcomes

Educational outcomes: years of schooling, enrollment, school dropout, absenteeism, correct grade for age, school grades, and school progression.

Data extraction (selection and coding)

This review follows the guidelines of the PRISMA. Papers will be selected by title and then exported to EndNote. Two reviewers will filter papers by abstract to select papers for full text reading.

Risk of bias assessment

The quality of the papers selected for the review was assessed using the MMAT (Pluye et al., 2011). The MMAT is used for complex systematic literature reviews that include quantitative, mixed-methods, and qualitative studies. It accounts for five common methodologies, qualitative (Section 1), quantitative randomized (Section 2), quantitative nonrandomized (Section 3), descriptive (Section 4), and mixed (Section 5). Each section is composed of three to four questions related to data sources, data collection, and outcome data. A quality score between 0% and 100% is assigned for each study component. An answer of “yes,” “no,” or “I can't tell” is assigned for each question in the corresponding study component. Papers of good quality met all criteria (i.e., had a score of 100%) and papers of poor quality did not meet any criteria (i.e., had a score of 0%).

Analysis of subgroups or subsets

None

Dissemination plans

The review will be published in a peer-reviewed journal for academics, policy workers, and members of civil society.

Contact details for further information

[e-mail]

Organizational affiliation of the review

United Nations University-MERIT and Maastricht University School of Governance

Funding sources

N/A

Conflicts of interest

None

Collaborators/review team

Tatenda Zinyemba, Maastricht University and United Nations University-Merit

Professor Wim Groot, Maastricht University and United Nations University-Merit

Associate Professor Milena Pavlova, Maastricht University, Maastricht University

Language

English

Country

The Netherlands

Stage of review

Complete

Stage of review at time of this submission

	Started	Completed
Preliminary searches	Yes	Yes
Piloting of the study selection process	Yes	Yes
Formal screening of search results against eligibility criteria	Yes	Yes
Data extraction	Yes	Yes
Quality assessment	Yes	Yes
Data analysis	Yes	Yes