The impact of HIV/AIDS on children’s educational outcome: A critical review of global literature

Abstract

The number of children losing one or both parents to HIV/AIDS has continued to rise in the past decade, with most of them being school-aged children. This study reviews global literature on the effects of HIV/AIDS (e.g., parental HIV-related illness or death) on children’s schooling. Systematic review procedures generated 23 studies for examination. Existing studies show educational disadvantages among children affected by AIDS in various educational outcomes, including school enrollment and attendance, school behavior and performance, school completion, and educational attainment. A number of individual and contextual factors potentially moderate or mediate the effect of HIV/AIDS on children’s education. These factors include gender of child, pattern of parental loss (maternal vs. paternal vs. dual), living arrangement (relationship with caregivers, gender of the household head), and household poverty. Current literature indicates limitations in number and scope of existing studies and in educational outcome measurements. There is a lack of studies with longitudinal design and data collection from multiple sources (e.g., students, teachers, caregivers), and a lack of studies on the relationship between psychosocial well-being of children affected by AIDS and their educational outcomes. Future studies need to employ more rigorous methodology and incorporate both individual and contextual factors for children affected by AIDS in various regions. More efforts are needed to design and implement culturally appropriate and context-specific approaches to improve the educational outcomes of children affected by AIDS.

Introduction

While in many countries the prevalence of HIV has decreased in the past decade, the number of children orphaned by AIDS continues to rise due to the time lag between HIV infection and HIV-related death (Salaam, 2005). The most recent Joint United Nations Program on HIV/AIDS (UNAIDS) global report (UNAIDS, 2010) estimated that the number of children orphaned by AIDS increased from 10 million in 2001 to 16.6 million in 2009. In the worst affected countries, the number of children orphaned by AIDS increased by almost threefold during the years 2000–2006 (UNICEF, 2009). In sub-Saharan Africa, the hardest hit region by HIV in the world, the number of children under 18 years of age orphaned by AIDS was estimated to be 15.7 million by 2010, accounting for 30% of the children orphaned from all causes in the region and 80% of children orphaned by AIDS worldwide (UNICEF, 2006).

While most studies examine only orphaned children, defined as the loss of one or both parents to AIDS, a number of studies note that living with a HIV-positive parent can also dramatically affect children’s psychological adjustment, care, and outcome (Sherr et al., 2008). Literature further shows that the loss of a father may have differential impacts on children’s educational outcomes than the loss of a mother (e.g., Evans & Miguel, 2007). In addition, some studies have revealed the severe impact of parental HIV/AIDS on children prior to the death of a HIV-positive parent (Ainsworth et al., 2005). We, therefore, examine children who are affected by HIV/AIDS and define them as being either orphaned or made vulnerable (living with one or both HIV-positive parents) by HIV/AIDS under the age of 18 years. Orphans are further categorized as paternal, maternal, and double (loss of both parents to HIV/AIDS) orphans. The elevating number of children affected by AIDS and their psychosocial well-being remain a great concern in the region and global society. The majority of children affected by AIDS are school-aged children, living in countries where education is not compulsory or free of fees (Monasch & Boerma, 2004; UNICEF, 2004).

Thus, the education of children affected by AIDS deserves great attention worldwide. Education is universally recognized to be essential for child development and has far-reaching impact on societies (Bredie & Beeharry, 1998; Stiglitz, 2005; UNICEF, 2009). Based on this recognition, the UN Convention in 1989 advocated the rights of children to education in article 28 of the Rights of the Child (UNICEF, 2009). As of 2004, 194 countries had ratified the Convention on Educational Rights of the Child (UNICEF, 2009). Although 8 years have passed, educational rights are still at stake for millions of children, especially for children affected by AIDS.

Even though an increasing number of empirical studies have examined the impact of HIV/AIDS-related parental death on children, especially in African countries where the HIV epidemic is most severe, there were only one article published in 2005 and one report published in 2008 that reviewed the empirical studies with a focus on the impact of the HIV epidemic on the schooling of children affected by AIDS (Bennell, 2005; Sherr, 2008). Bennell (2005) summarized main findings of three studies in three African countries (Botswana, Malawi, and Uganda) (Bennell et al., 2001; Hyde, Ekatan, Kiage, & Barasa, 2002; Kadzamira, Swainson, Maluwa Banda, & Kamlongera, 2001). Sherr (2008), in a report to the Joint Learning Initiative on Children and HIV/AIDS, reviewed 15 studies, including both peer-reviewed studies and working papers, on the important issues in the development of children affected by AIDS, with schooling being one of these issues. The current study expands the previous reviews to include more recent literature and examine the impact of parental HIV/AIDS on various educational outcomes of children in both African countries and other regions (e.g., regions in China). Educational outcomes are measured using a number of variables, including school enrollment and attendance, appropriate grade by age, school performance, school behavior, and school attainment.

We first review the global literature on the impact of HIV/AIDS on education of children affected by AIDS along a spectrum of educational outcome measures. We then examine a number of individual and contextual factors that potentially moderate or mediate the effect of HIV/AIDS on children’s education. These individual and contextual factors include types of parental loss (maternal, paternal, or double), children’s living arrangements (e.g., relationship with the caregiver, gender of the household head), household poverty, and gender of the orphans. In discussion, we highlight the important findings and identify the gaps in the existing literature and make recommendations for future research.

Methods

We conducted a systematic literature search on March 2011 to identify studies that examined schooling issues among children affected by AIDS. The inclusion criteria included (1) peer-reviewed studies published in English, (2) empirical studies using either qualitative or quantitative methodologies, and (3) education of children affected by AIDS as one of the primary outcomes/variables. Publications were retrieved from the following electronic databases: Medline/PubMed, Web of Science, and PsycInfo. Each database was searched by using various combinations of following keywords: HIV, AIDS, orphan, vulnerable children, education, school, and schooling. The search retrieved 80 articles from the databases; 20 of them met the inclusion criteria. We also hand searched the citations of the previous review articles and found one additional study. In addition, we hand searched the references of the eligible articles and found two more studies, resulting in 23 articles that met our inclusion criteria for the review.

Results

Data collection and sampling methods

Studies that met our inclusion criteria were summarized in Table 1. As shown in Table 1, all studies except one (Aspaas, 1999) were published in 2003 or beyond, 10 studies in 2007 or beyond. All studies except two were conducted in African countries, including Tanzania, Kenya, Zimbabwe, Uganda, Burkina Faso, Nigeria, Ethiopia, Ghana, Lesotho, and Malawi. Two studies were conducted in regions of China with high HIV prevalence (Tu et al., 2009; Yang et al., 2006). Many of these studies employed secondary data such as Demographic and Health Surveys (DHS) conducted in developing countries, End-of-Decade Multiple Indicator Cluster Surveys (MICS) in Africa, Latin America (e.g., Bolivia, Brazil, Colombia, Guatemala, Guyana, Nicaragua, Peru, Venezuela), and Asia (e.g., Lao PDR, Mongolia, Philippines, Vietnam), Integrated Household Surveys (IHS, such as National Socioeconomic Surveys or Living Standards Surveys) in Asia (e.g., Cambodia, Philippines, Vietnam), National Household Survey in Uganda, Migration and Urban Integration Survey in Burkina Faso, Africa Center’s Household Socio-Economic (HSE) Survey, and Kagera Health and Development Survey (KHDS) in Tanzania. Some studies collected their own data, both qualitative (e.g., Oleke, Blystad, Fylkesnes, & Tumwine, 2007) and quantitative (e.g., Tu et al., 2009; Yang et al., 2006). For example, a study in Uganda utilized school records, ethnographic field observations, and in-depth interviews with multiple sources such as community leaders, heads of households, and orphans (Oleke et al., 2007). Another study in rural China collected data from 1625 children and their schoolteachers (Tu et al., 2009).

Table 1.

Empirical studies on the educational outcomes of children affected by AIDS.

Reference Number	Author, year	Years of data collection	Location and data sources	Study population and sample size	Sampling method	Outcome measurement	Major findings
1	Ainsworth & Filmer, 2006	1992–2003	Fifty-one countries in Sub-Saharan Africa (35), Latin America (8), the Caribbean (2), and Asia (6). Sixty-seven Demographic and Health Surveys (DHS), 26 End-of-Decade Multiple Indicator Cluster Surveys (MICS), and nine Integrated Household Surveys (HIS) such as National Socioeconomic Surveys (SES) or Living Standards Surveys (LSS)	Children aged 7–14 years, sample sizes for the countries ranged between 1100 and 32,500, but most are on the order of 5000–10,000	Cross-sectional	Enrollment rate: percentage of children aged 7–14 years who are currently “in school”	Important variations were observed across the countries. Orphans sometimes had lower enrollment than nonorphans, but in many countries, there was small or no difference. Enrollment gap could be explained by the household economic status better than orphanhood status. There was no significant gap in enrollment between girls and boys.
2	Ainsworth et al., 2005	1991–1994 annual interview rounds	Kagera, Tanzania	Children aged 7–14 years: N = 1339	Longitudinal	(1) School attendance and (2) hours spent at school for each of the 7 days preceding the interview	Attendance was delayed for maternal orphans and children in poor households with a recent adult death; no evidence was found that children aged 7–14 years dropped out of primary school due to orphan status or adult deaths. Among children who attended school, school hours were significantly lower in the months prior to an adult death in the household and recovered following the death. Girls greatly reduced school hours immediately after losing a parent.
3	Aspaas, 1999	1993	Jinja District in Eastern Uganda	Household heads interviewed: N = 103	Cross-sectional	Education index, comparing the number of school-age orphans to the number of orphans who were in school in each household	Indigenous children in men-headed households were more likely to be enrolled at school than school-age orphans in the same households, while no such difference was found between indigenous children and orphans living in the households headed by rural women who had the lowest levels of education themselves.
4	Beegle et al., 2009	First round: 1991–1993; second round: 2004	Kagera, Tanzania, Kagera Health and Development Survey (KHDS)	Children who became orphaned when < 15 years: N = 718	Longitudinal	Number of completed years of formal education	Children who had lost their mother before the age of 15 years suffered a deficit of 1 year of final attained schooling. There was no such causal link between paternal death and school years attained.
5	Bhargava, 2005	2001–2002	The National Survey of Prevalence and Characteristics of Orphans in Ethiopia	Maternal or double AIDS orphans: N = 479; other orphans: N = 574	Cross-sectional	School participation before and after the death of the mother	Income and good feeding and clothing conditions were significant predictors of school participation. School participation before maternal death was an important predictor of subsequent school participation probabilities. Girls were less likely to be participating in school after maternal deaths.
6	Bicego et al., 2003	1995–2000	DHS in 17 countries and in-depth study in five African countries (Zimbabwe, Kenya, Tanzania, Ghana, and Niger)	Orphans aged ≤ 15 years in five countries: 1992: N = 78,965; 1998: N = 63,198	Cross-sectional	Appropriate grade level	Losing one or both parents was significantly associated with diminished chances of being at the appropriate grade level for age.
7	Birdthistle et al., 2009	2004	Harare, Zimbabwe	Girls aged 15–19 years: N = 743	Cross-sectional	(1) Educational attainment and (2) current school attendance	Double orphans had significantly lower educational attendance and attainment than nonorphans. Maternal orphans had higher rates of school dropout than nonorphans, but the difference disappeared when adjusted for recent mobility. High overall levels of secondary school participation and school fee assistance provided to vulnerable families may have reduced the schooling disparities between orphans and nonorphans in Highfield, Harare of Zimbabwe.
8	Case et al., 2004	1992–2000	Ten African countries, 19 (DHS)	Children aged ≤ 14 years	Cross-sectional	Enrollment	Orphans were less likely to be enrolled than nonorphans with whom they lived, which could be largely explained by the fact that orphans were more likely to live with distant relatives or non-relative caregivers.
9	Case& Ardington, 2006	2001 first round, 2003–2004 second round	Umkhanyakude District of KwaZulu-Natal, Africa Center’s Household Socio-Economic (HSE) Survey	Children aged 6–16 years, currently enrolled: N = 18,670; completed education: N = 18,568	Longitudinal	(1) School enrollment and (2) educational attainment	The loss of a child’s mother was a strong predictor of poor schooling outcomes. Maternal orphans had significantly lower rates of school enrollment and less schooling years than children whose mothers were alive, after controlling for age. Maternal orphans appeared to be at an educational disadvantage when compared with nonorphaned children with whom they lived. The timing of mothers’ deaths had a causal effect on children’s educations. The loss of a child’s father was a significant correlate of poor household socioeconomic status.
10	Evans & Miguel, 2007	1998–2002	Kenya	Baseline nonorphaned children (sample) of 5–18 years enrolled at school: N = 24,111; longitudinal sample: N = 7815	Longitudinal (5-year panel data)	(1) School attendance and (2) enrollment	Parental death had a substantial negative impact on orphans’ primary school participation, with impacts being more than twice as large for maternal deaths than paternal deaths. Children with lower baseline academic test scores were more likely to drop out of school than children with better school performance after parental death.
11	Kasirye & Hisali., 2010	2002–2003	Uganda National Household Survey (UNHS)	Orphans: N = 1244; children aged ≤18 years: N = 28,792	Cross-sectional	(1) Current enrollment for children 6–17 years and (2) schooling gap (number of years a child is below an appropriate grade)	Children orphaned by AIDS were not significantly less likely to continue schooling but were by far more likely to fall below their appropriate grade, due to the interruption of parental illness and death (looking after ailing parents). Schooling gap decreased at higher level of household welfare status, which meant poor orphans were significantly less likely to continue schooling.
12	Kobiané et al., 2005	2000	Burkina Faso, Migration and Urban Integration Survey	Individuals aged 15–44 years: N = 6232; three cohorts, 1975–1985 birth cohort: N = 2336; 1965–1974 birth cohort: N = 2594; 1955–1964 birth cohort: N = 1870	Cross-sectional	School attendance	Double orphans had decreased chances of school enrollment, especially in rural areas. The negative effect of orphanhood had decreased over time by comparing three birth cohorts from a cross-sectional data. Orphans living in urban areas were more likely to be supported by nongovernmental organizations than those in rural areas. Male double orphans had higher rates of school enrollment than female double orphans.
13	Kúrzinger et al., 2008	2003	Tanzania, Burkina Faso	Children aged ≤ 18 years: Tanzania: N = 11,392; Burkina Faso: N = 11,856	Cross-sectional	(1) School enrollment and (2) proper grade level	Compared to nonorphans, orphans were less likely to attend school and more likely to be at lower grades. However, after controlling for the confounders such as age, religion, living arrangement, the relationship between the child and the household head and the household dependency ratio, the risk was no longer significant. Orphans’ relatively equivalent educational status to nonorphans might be explained by family-based or community-safety nets.
14	Monasch & Boerma, 2004	Thirty-seven countries during 1999–2002, three countries in 1997 or 1998	Twenty-three MICS and 17 DHS, 40 countries in sub-Saharan Africa	No. of children (aged 0–14 years) surveyed: 6200–47,516	Cross-sectional	School attendance	Orphans were approximately 13% less likely to attend school than nonorphans. Double orphans were most likely to be disadvantaged.
15	Nyamukapa & Gregson, 2005	1998–2000	Rural Zimbabwe, the “Manicaland Study”	Children aged < 16 years: N = 14,169; primary school completion by children aged 13–15 years: N = 2402	Cross-sectional	Primary school completion	Maternal orphans but not paternal or double orphans had lower primary school completion rates than nonorphans in rural Zimbabwe, which could be explained by the increased residence in female-headed households and greater access to external resources. Disadvantage in primary school completion among maternal orphans resulted from the lack of support from fathers and stepmothers and not being eligible for welfare assistance due to better socioeconomic household status. These negative effects were partially offset by increased assistance from maternal relatives.
16	Oladokun et al., 2009	2005–2006	Nigeria	HIV-positive children: N = 110	Cross-sectional	School enrollment	There was no significant difference in school enrollment between orphans and nonorphans among the HIV-positive children in the sample. Extended family system appeared to be coping with the orphan situation.
17	Oleke et al., 2007	2002–2003 and follow-up trip 2003–2004	Amach subcounty, Lira District, Northern Uganda	Households: N = 402; orphans: N = 1399	Longitudinal	(1) School enrolment; (2) school attendance; and (3) school performance	The widowed and single household heads were more likely to have all orphans in school than the married, and households that received external support offered better educational opportunities. Poverty, as indicated by lack of food at school and heavy involvement of orphans in domestic labor were identified as major constraints on orphan’s schooling.
18	Parker & Short, 2009	2004	Lesotho, DHS	Children aged 6–16 years: N = 10,641	Cross-sectional	School enrollment	Maternal orphans who lived with a grandmother were just as likely to be enrolled at school as children living with a mother. The protective effect of living with a grandmother existed for children whose mothers were alive but not affiliated with their households.
19	Sharma, 2006	2000 and 2004	Malawi	Children aged 5–15 years: N = 729	Longitudinal	School enrollment	There was no independent effect of orphanhood status on school attendance; orphans had higher rates of school dropout than nonorphans as grade-level increases. Orphans residing with nonparent or nongrandparent caregivers did not have a higher probability of dropping out. On the contrary, orphans under the care of a surviving parent or grandparents were less likely to attend school compared to those living with other relatives, which might be explained by the income loss and household poverty associated with parental death, and was also likely due to the fact that orphans were often adopted by relatively wealthier relatives.
20	Tu et al., 2009	2006–2007	Two rural counties in Henan Province, China	Children aged 6–18 years: N = 1625; single or double orphans: N = 755; children living with HIV-positive parents: N = 466; comparison children: N = 404	Cross-sectional	(1) Academic marks; (2) future educational expectation; (3) student leadership; and (4) school behavior	Orphans and vulnerable children (OVC) due to AIDS had disadvantages in school performances compared to those who did not experience AIDS-related death and illness in their family. Orphans had the lowest academic marks based on the reports of both children and teachers. Educational expectation was significantly lower among OVC than comparison children from teacher’s perspective. Children orphaned by AIDS were significantly more likely to demonstrate aggressive, impulsive and anxious behavior than nonorphans; they also had more learning difficulties. Vulnerable children were also at a disadvantage on most measures.
21	Yamano, 2007	2004	Kenya	All individuals: N = 6240; orphans: N = 786	Cross-sectional	The attained highest grade	Maternal orphans who were 16 years of age or older and lost their mothers before reaching the age of 15 years received 1 year less in educational attainment compared with nonorphans among those who started schooling before the Free Primary Education program introduced in 1974 but not among those who started schooling after 1974.
22	Yamano & Jayne, 2005	1997, 2000, and 2002	Kenya	Children aged 7–14 years: year 2000: N = 2565; year 2002: N = 2107	Longitudinal (3-year panel)	School attendance	Children experiencing working-age adult mortality since the first survey in 1997 to the last survey in 2002 were about 3% less likely to attend school than children who did not experience adult mortality during the same period. Adult mortality had a larger negative impact on girls than on boys. Both girls and boys in the relatively poor households were about 33% less likely to be in school in the 1- to 2-year period prior to adult mortality. Boys in relatively poor households were also 9% less likely to be in school some time after experiencing adult mortality, which could be because boys were compensating for the loss of family labor.
23	Yang et al., 2006	2001	Longchuan County, Yunnan Province, China	Children aged ≤ 15 years, living in families of HIV-positive drug users: N = 213	Cross-sectional	(1) Current school attendance; (2) school dropout; and (3) often skipping a class (yes/no)	Orphans were less likely to attend school and more likely to be truant if enrolled in school. One-half of the children experienced frequent parent discord and/or expressions of desire to divorce, family anxiety, and shame for having a parent who was an injection drug user and infected with HIV.

The sample sizes of the studies varied from 103 to 7815 for those with primary data collection and from 718 to 78,965 for those with secondary data collection. Most of the studies (16 of 23) used cross-sectional data whereas seven studies used longitudinal data. The most frequently used educational measurements were school enrollment and attendance, followed by schooling gap (whether a child was below an appropriate grade level, or number of years a child was below an appropriate grade) and the highest educational achievement (years of formal schooling). Other educational measurements included hours spent at school, schooling ratio (number of orphans who were in school to number of school-aged orphans in each household), school performance (academic scores, educational expectation, student leadership, and school behavior), school dropout, and primary school completion.

Educational outcomes of children affected by AIDS

School enrollment and attendance

Some studies reported that orphans were less likely to be enrolled in school or attend school (e.g., Case & Ardington, 2006; Case, Paxson, & Ableidinger, 2004; Monasch & Boerma, 2004; Yang et al., 2006). Using national representative data (23 MICS and 17 DHS) from 40 countries in sub-Saharan Africa, one study found that orphans were approximately 13% less likely to attend school than nonorphans (Monasch & Boerma, 2004). Another study in China found that children aged under 15 years whose parent(s) (one or both) used injecting drugs and died of AIDS were less likely to be enrolled in school or to attend school if enrolled (Yang et al., 2006). One study in Kenya utilizing five year longitudinal data revealed a substantial and significant impact of parental death on orphan’s primary school enrollment and attendance (Evans & Miguel, 2007). Compared with nonorphans and single orphans (i.e., children who lost one parent to AIDS), double orphans were found to have significantly lower educational enrollment and attendance, especially in rural Africa (Birdthistle et al., 2009; Case et al., 2004; Kobiané, Calvès, & Marcoux, 2005).

However, several other studies did not find an independent effect of orphanhood on school enrollment or attendance after controlling for other factors (Birdthistle et al., 2009; Kúrzinger et al., 2008; Oladokun et al., 2009; Sharma, 2006). These factors included age, gender, education (grade level), and religion of the child, living arrangement (i.e., whether the child was living with his/her own family), relationship between the child and the household head (e.g., parents, grandparents, relatives, nonrelatives), household child/adult ratio (i.e., number of children aged under 18 years to number of adults, an indication of childcare burden), and school fee assistance to families (Birdthistle et al., 2009; Kúrzinger et al., 2008; Sharma, 2006).

Difference in school enrollment and attendance between orphans and nonorphans varies across countries. Based on 102 national surveys (e.g., DHS, MICS2, HIS) from 51 countries in regions including sub-Saharan Africa, Latin America, the Caribbean and Asia, one study found that orphans in some countries had lower school enrollment and attendance rates than nonorphans, whereas in many other countries, the difference was small or nonexistent (Ainsworth & Filmer, 2006).

Schooling gap

A few studies examined the appropriate grade level for age. These studies showed that children who lost one or both parents to AIDS might continue schooling but had diminished chances of being at the appropriate grade level for age, due to the interruption of parental illness and death (Ainsworth, Beegle, & Koda, 2005; Bicego, Rutstein, & Johnson, 2003; Kasirye & Hisali, 2010). Younger children and double orphans were the most affected groups in terms of schooling gap (Ainsworth et al., 2005; Bicego et al., 2003). Households affected by AIDS were often delaying school enrollment of younger children, while trying to maintain the enrollment of older children if possible. Double orphans often experienced multiple interruptions of parental illness and death and therefore were more likely to fall below their appropriate grade (Ainsworth et al., 2005; Bicego et al., 2003).

School performance and behavior

Compared to studies on school enrollment and attendance, there was less research regarding the school performance and school behavior among children affected by AIDS. We found only one study that examined such outcomes among children in rural villages in Henan province, central China, a region with high HIV prevalence because of unhygienic commercial blood/plasma collection in the 1980s and 1990s (Tu et al., 2009). Using self-report data from 1625 children and evaluation data from their schoolteachers, the study indicated that children affected by AIDS had worse school performance compared to their peers (comparison children) from the same community who did not experience HIV-related parental illness or death (Tu et al., 2009). In particular, children orphaned by AIDS had the lowest academic marks among the three groups of children (orphans, vulnerable children who were living with one or both HIV-positive parents, and comparison children) based on reports from both children and teachers (Tu et al., 2009). From teachers’ perspective, children orphaned or made vulnerable by AIDS had significantly lower educational expectations than comparison children (Tu et al., 2009). In addition, children orphaned by AIDS were reported to demonstrate more aggressive, impulsive, and anxious behavior in school and have less confidence when interacting with other students than the comparison children. Children affected by AIDS were having more learning difficulties, such as difficulty in concentrating on their studies, due to the experience of stressful events and the resulting psychosocial problems, which might also explain their worse school performance (Tu et al., 2009).

School completion and educational attainment

Among the 23 studies, only one study has closely examined the issue of school completion among children affected by AIDS. The study was carried out in rural Zimbabwe that combined both quantitative data of 14,169 children aged under16 years and qualitative data from key informant interviews and focus group discussion (Nyamukapa & Gregson, 2005). This study showed that maternal orphans instead of paternal or double orphans had shown lower primary school completion rates than nonorphans.

Closely related to school completion was school dropout. Another study in Zimbabwe (urban area) found that among different types of orphans (paternal, maternal, and double orphans), only maternal orphans had higher rates of school dropout than nonorphans, but the difference disappeared when adjusted for recent mobility (Birdthistle et al., 2009). A longitudinal study in rural Malawi found that the likelihood of dropping out of school was higher for orphans than nonorphans as grade level increased (Sharma, 2006). Similar to maternal orphans, double orphans also had significantly lower educational attainment than nonorphans (Birdthistle et al., 2009).

Individual and contextual factors

Maternal versus paternal death

For children who have lost one of their parents to AIDS, existing literature has shown differential effects of maternal and paternal death on children’s educational outcomes, with maternal death often having greater negative impacts on children’s educational outcomes than paternal death (e.g., Evans & Miguel, 2007). While the loss of a father was often associated with a deteriorated household socioeconomic status and might consequently result in children dropping out of school (Case et al., 2004), the loss of a mother was a strong predictor of children’s poor schooling outcomes, including delayed school attendance, less likely to be enrolled at school, at a lower grade for age, worse school performance, and lower educational attainment (Ainsworth et al., 2005; Birdthistle et al., 2009; Case & Ardington, 2006; Evans & Miguel, 2007; Nyamukapa & Gregson, 2005).

Based on a five year panel data-set of over 20,000 children in rural Kenyan, Evans and Miguel (2007) found a substantial and highly significant negative impact of parental death on primary school participation, with impacts being more than twice as large for maternal deaths than for paternal deaths. In addition compared to paternal orphans, maternal orphans were less likely to complete their primary school education (Nyamukapa & Gregson, 2005).

There were several possible explanations for the stronger impact of maternal death than paternal death on children’s education. First, children in many African countries were less likely to live with their fathers and more likely to live with their mothers (Nyamukapa & Gregson, 2005). For example, a study in Malawi found that 70% of children lived with their biological fathers while their mothers were alive, and only 27% lived with biological fathers after losing their mothers (UNICEF, 2004). Second, mothers would make considerable sacrifices to safeguard their children’s education, which might be explained by the close day-to-day connections between mother and child (Nyamukapa & Gregson, 2005). By contrast, after mothers’ death, children’s immediate caregivers (fathers, stepmothers, other relatives) might be less committed to securing their education (Nyamukapa & Gregson, 2005). Third, more external assistance (e.g., from the government, NGOs, extended families) was sometimes available to paternal or double orphans residing in female-headed households (Birdthistle et al., 2009; Nyamukapa & Gregson, 2005). For example, the study in rural Zimbabwe found that primary school completion rates among paternal orphans and double orphans were similar to that of nonorphans because of the external assistance to female-headed households (Nyamukapa & Gregson, 2005). Such assistance was often denied to maternal orphans as they often lived in households with relatively better socioeconomic status (Nyamukapa & Gregson, 2005).

In addition, the loss of a mother has shown a long-term negative effect on children’s educational attainment (Beegle et al., 2009; Case & Ardington, 2006; Nyamukapa & Gregson, 2005). Based on longitudinal data from KwaZulu-Natal, South Africa, Case and Ardington (2006) found that maternal orphans had completed significantly fewer years of schooling than children whose mothers were alive and nonorphaned children who lived in the same households after controlling for age. Another longitudinal study in Tanzania found that children who had lost their mothers before the age of 15 years had one less year of formal schooling compared to nonorphans whereas such disadvantage was not observed among children who lost their fathers (Beegle et al., 2009). Among all types of orphans (i.e., maternal, paternal, and double orphans), double orphans seemed to be the most disadvantaged group. Based on a national survey, one study in Burkina Faso found that double orphans had the smallest chances of going to school compared to single orphans and nonorphans, especially in rural areas (Kobiané et al., 2005).

Living arrangements

Existing literature indicates that more closely related caregivers are more likely to invest in children’s education (Case et al., 2004; Parker & Short, 2009). Utilizing 19 DHS data-sets conducted in 10 sub-Saharan African countries between 1992 and 2000, Case et al. (2004) found that maternal orphans who lived with a grandmother were as likely to be enrolled at school as children who lived with a mother. The protective effect of living with a grandmother existed for children whose mothers were alive but not living in the same households (Parker & Short, 2009). Case et al. (2004) also found that orphans who lived with other relatives were less likely to be enrolled at school than those who lived with one or both parents or grandparents and that children who lived with non-relatives had the least chances of being enrolled at school among children in all living arrangements (Case et al., 2004). In addition, orphans who lived with distant relatives in the extended family or nonrelative caregivers were less likely to be enrolled at school compared with nonorphans living in the same households, and this enrollment gap increased with age and did not decline with increases in household wealth (Case et al., 2004). However, Case et al. (2004) did not adjust for household economic status when estimating the enrollment differentials associated with orphanhood status.

Gender of the household head appears to be also related to the educational differences between orphans and nonorphans in the same household. One study in eastern Uganda found that indigenous children in male-headed households had a higher rate of enrollment than school-aged orphans in the same households, whereas there was no such difference in female-headed households (Aspaas, 1999).

Household poverty

Existing research has indicated the importance of the interaction between household poverty, orphanhood status, and the educational outcomes of children affected by AIDS, as household economic status often moderate or mediate the impact of parental illness or death on the educational outcomes of children (Ainsworth & Filmer, 2006; Kasirye & Hisali, 2010). One study using three year panel data found that children in the relatively poor households were about 33% less likely to attend school one-two years prior to parental death (Yamano & Jayne, 2005). Another study with both qualitative and quantitative data in northern Uganda identified poverty as a major constraint on orphans’ schooling, with household poverty being measured as lack of food for orphans at school and heavy involvement of orphans in domestic labor (Oleke et al., 2007). By adding an interaction term of orphanhood status and household consumption (a measure of economic status), a recent study indicated that the impact of orphanhood on schooling gap differed by household socioeconomic status, as poorer orphans were more likely to fall below their appropriate grade than orphans from better-off families (Kasirye & Hisali, 2010).

Using large data-sets across countries and by comparing the enrollment difference between orphans and nonorphans and that between children from the poorest and richest household quintile, Ainsworth and Filmer (2006) found that the economic gap was statistically significantly larger than the orphan gap, and thus concluded that household economic status was a more important factor than orphanhood status in explaining the gap in school enrollment between orphans and nonorphans.

When adding a set of household economic status variables as explanatory variables, a study in rural Malawi found that orphans under the care of a single parent or grandparent/s were less likely to attend school compared to those living with other relatives (Sharma, 2006). It might be because of high HIV/AIDS prevalence in rural Malawi among income-earning young adults, the death of a parent and the resulting increase in household poverty would have a large negative impact on children’s school attendance. In the meanwhile, orphans were also likely to be adopted by relatively wealthier relatives in rural Malawi who were less constrained financially and were capable of taking in orphans.

Gender of the orphans and effect of parental death

Even though the research based on large cross-country data-sets did not find significant gap in enrollment rates between girls and boys (Ainsworth & Filmer, 2006), parental death appeared to be more detrimental for girls than for boys (Ainsworth et al., 2005; Kobiané et al., 2005). For children who attended school, both girls and boys had a substantial reduction in school hours three-six months prior to a parental death in the household. Following the parental death, boys resumed their normal levels of hours at school, but girls had a sharp decline in school hours within six months following their mothers’ deaths or when they became double orphans, as girls were expected to care for an ill parent prior to parental death and to take over their mothers’ responsibilities of housework and childcare immediately following their mothers’ death (Ainsworth et al., 2005). Similarly, a study in Ethiopia found that girls were less likely to attend school than boys following a maternal death (Bhargava, 2005). However, a study in Kenya found that boys in relatively poor households were 9% less likely to attend school some time after parental death (Yamano & Jayne, 2005). The reason could be that the boys were working in farms or participating in other income-generating activities to compensate for the loss of family labor (Yamano & Jayne, 2005).

Discussion

This review of global literature has revealed the educational disadvantages among children affected by AIDS compared with other children in various measures of educational outcomes, including school enrollment and attendance, school performance and behavior, school completion, and educational attainment. As indicated in the existing studies, the educational outcomes are the balanced results of various factors, including both risk and protective factors. These factors include resource constraints (e.g., household poverty), orphanhood status (e.g., vulnerable children living with HIV-positive parent, maternal, paternal, or double orphans), living arrangement (e.g., relationship with the caregiver, gender of the household heads), and external assistance (e.g., financial and policy supports from the government or NGOs). A better understanding of the underlying theoretical mechanisms of these factors is necessary to develop effective policy recommendations to improve the educational outcomes of children affected by AIDS.

While existing studies have provided valuable information on the educational outcomes of children affected by AIDS, there are several limitations in the existing literature. First, the number and scope of the existing studies on the educational outcomes of children affected by AIDS are limited. In this review, we were able to identify only 23 peer-reviewed studies published between the years 1999 and 2010 on the effect of HIV/AIDS on children’s education. Even though there has been a growing number of studies in recent years (19 studies since 2005), more research is needed to add to our understanding on the schooling of the ever elevating number of children affected by AIDS. In addition, studies in regions other than sub-Saharan Africa with high prevalence of HIV/AIDS, such as regions in Asia and Latin America are needed, as children in these regions may face different challenges of schooling. Research conducted in these regions will improve our understanding on the schooling of children affected by HIV/AIDS in various cultural contexts worldwide.

Second, a limited number of educational outcomes are examined in the existing literature. Most of the studies have only focused on children’s school enrollment and attendance. A few studies have examined the appropriate grade level and schooling gap of children affected by AIDS. Only one study has examined school performance and school behavior and one study on primary school completion. In the future, more studies are needed to examine various educational outcomes including school performance, school behavior, school completion (primary and secondary), and educational attainment of children affected by AIDS.

Third, there are limited studies that have examined the effect of psychosocial well-being of children affected by AIDS on their educational outcomes. There is evidence that children affected by AIDS experienced more psychosocial symptoms, which might cause them to have more learning difficulties at school and the consequent worse school performance (Tu et al., 2009). These children with poor school performance might be further disadvantaged by experiencing significant drop in their school participation, as households might decide to invest their limited resources on more promising children after a parental death (Evans & Miguel, 2007). Regarding the HIV status of the children, there is not a single study that has reported such data. As it has been well documented the impairing cognitive effects of HIV on child functioning (Sherr et al., 2008), children’s HIV status may have moderated or mediated their educational and behavioral outcomes. It is therefore a severe limitation in the existing literature for a lack of such data and discussion of possible educational needs particularly among HIV-positive orphans and vulnerable children. Future studies that examine the relationships between psychosocial factors (e.g., depression, isolation, perceived and internalized stigma, sense of future control, social interaction) and educational outcomes will contribute to the understanding of how to improve children’s educational outcomes through psychosocial interventions. Future studies that examine the effect of HIV infection are needed to point the way forward for more fine-tuned educational provision.

Fourth, there is a lack of longitudinal design to establish the causal relationships between educational outcomes and other individual and contextual factors among children affected by AIDS. For example, orphaned children may experience psychological distress of losing one or both parents to the extent that they cannot focus on study and subsequently lead to school dropout. However, it can also be possible that because of not being able to go to school like other children, orphaned children may experience additional psychological distress. By following a cohort of children over time, we could not only sort out the orders of such events and possible causal relationships but also describe changes in children’s educational outcomes and functioning before and after they become orphans and differences in their development paths (e.g., physical, psychosocial, emotional, intellectual, behavioral) in relation to those of nonorphans (Case & Ardington, 2006). There is no randomized controlled trial to improve the educational needs of children affected by AIDS or to address these shortcomings in the existing literature.

Fifth, most of the existing studies utilized nonorphaned children as comparison group and lacked a non-AIDS orphan comparison group (e.g., Tu et al., 2009), which limited our ability to differentiate the effect of parental HIV/AIDS infection from the effect of orphanhood status. Future research needs to include children orphaned by causes other than HIV/AIDS, so researchers can possibly identify the unique effect of parental HIV/AIDS infection on children’s schooling.

Finally, most of the existing studies rely on data collected from a single source (e.g., children or households). The incorporation of perspectives from teachers, caregivers, and community leaders may triangulate the data from children affected by AIDS and improve the quality of the studies. Future research on the schooling of children affected by AIDS needs to be carefully designed and executed to address these limitations and to help us reach a better understanding of the scope, extent, and mechanism of the impacts of HIV/AIDS on educational needs of these children.

Implications for future research and intervention

Several findings from the existing literature need to be highlighted and may shed a light on future intervention aiming to improve the educational outcomes of children affected by AIDS. First, as children orphaned by AIDS or living with HIV-positive parent(s) are likely to face economic constraint as a result of parental illness and death and the associated financial cost, and household poverty may be a more influential factor than orphanhood status on the educational outcomes of these children (Ainsworth & Filmer, 2006), strategies to alleviate or eradicate household poverty should be a priority to improve their educational outcomes. Since there is evidence for the independent effect of both household poverty and orphanhood status on the educational outcomes of children affected by AIDS (Ainsworth & Filmer, 2006; Case et al., 2004), limited resources should be prioritized to children affected by AIDS who live in poor families. Global literature has shown that such strategies (e.g., cash transfers, food allocation, microcredit program) are associated with improved school outcomes (Sherr, 2008; UNICEF, 2009).

Second, as losing a mother and losing a father might have different implications for children’s educational outcomes, effective interventions need to accommodate differential needs of maternal and paternal orphans accordingly. As the death of a father was often associated with poor household socio-economic status (Ainsworth et al., 2005; Birdthistle et al., 2009; Case & Ardington, 2006; Case et al., 2004; Evans & Miguel, 2007; Nyamukapa & Gregson, 2005), financial assistance to mothers or other caregivers for children who lost their fathers to AIDS would be much needed. For children who lost their mothers to AIDS, more adult attention and care in general and supervision on education are needed. Intervention efforts that encourage the involvement of fathers, stepmothers, other caregivers, or outside supervisors as well as teachers in orphans’ schooling are likely to be effective in improving maternal orphans’ educational outcomes.

Third, children’s living arrangements are important contextual factors that should be taken into account in future intervention. As many orphans live with their grandparents and global literature has shown a strong protective effect of grandmothers on children’s education (Parker & Short, 2009), financial assistance directed to children (e.g., school fee assistance) or other types of assistance (e.g., food, cash) to the poor households headed by grandmothers might be effective in helping children’s schooling. For orphans who live with other relatives or nonrelative caregivers, it may be more effective to direct educational assistance to orphans, for instance, through educational subsidies or nontransferable vouchers for schooling that are earmarked for orphans (Case et al., 2004).

Fourth, close attention should be paid to children who live with HIV-positive parent(s) at the final stage of the disease and who are experiencing a recent bereavement. As the burden of taking care of sick parents three-six months prior to parental death may consume much of children’s school time, more assistance should be provided for children who live with severely ill parent(s), including medical assistance to care for children’s sick parents, financial assistance to the family, and psychosocial counseling to help children go through these difficult times (Ainsworth et al., 2005).

Fifth, more attention should be paid to the importance of differentiating the gender of the affected children as well as of the deceased parent(s). It is a powerful and tragic finding that girls rather than boys have to sacrifice their schooling hours and take over their mothers’ extensive duties of housework and childcare following their mothers’ death (Ainsworth et al., 2005; Bhargava, 2005). Special intervention efforts are needed to arrange appropriate care (e.g., securing more help from female relatives or other helpers) for schoolgirls with a recent bereavement of their mothers and ensure their school attendance.

In summary, existing literature has clearly demonstrated the vulnerability and risk of children affected by AIDS in their schooling. Global research has indicated various educational disadvantages of children affected by HIV/AIDS compared to other children in terms of their school enrollment and attendance, school performance and behavior, school completion and educational attainment. Other factors also play important roles in influencing the educational outcomes of children affected by AIDS, including types of orphanhood (i.e., maternal, paternal, or double orphans), living arrangements (i.e., relationship with the caregiver, gender of the household head), household poverty, and gender of the children. Researchers, practitioners, policy-makers, and all stakeholders need to work together, informed by the existing knowledge base, to design and implement culturally appropriate and context-specific approaches to improve the educational outcomes of children affected by AIDS and to fulfill our collective promise to the educational rights of all children.