Risk and Resilience in Orphaned Adolescents Living in a Community Affected by AIDS

Abstract

The AIDS pandemic has resulted in a dramatic rise in the number of orphans in South Africa. This study was designed to investigate the associations between family, peer, and community factors and resilience in orphaned adolescents. Self-report questionnaires were administered verbally to 159 parentally bereaved adolescents (aged 10-19) in an economically deprived urban area. Questionnaires included measures of depression, anxiety, and self-esteem. The results of a hierarchical multiple regression analysis indicate that cumulative stress exposure, losing a parent to a cause other than HIV and AIDS, and being cared for by a nonrelative were associated with an increased risk of internalizing symptoms. Family regulation and respect for individuality, peer connection, and community connection and regulation were significantly associated with greater emotional resilience. The findings support a main-effects model of resilience in which risk factors and protective factors contribute additively to the prediction of the outcome, without interaction.

Keywords

orphans HIV and AIDS South Africa

One of the distressing consequences of the HIV and AIDS epidemic in South Africa is a dramatic rise in the number of orphans. Statistical models have predicted that, by 2015, approximately 2.5 million children under the age of 18 will have lost their mothers (Dorrington, Johnson, Bradshaw, & Daniel, 2006). According to Johnson and Dorrington (2001), the majority of these orphans will be adolescents. Most of them will come from poor socioeconomic backgrounds and have lost either or both parents to an AIDS-related illness.

The international literature on childhood bereavement suggests that the death of a parent places children at risk for internalizing problems such as depression, anxiety, and low self-esteem (Dowdney, 2000). Links between parental death and externalizing problems such as delinquency are much less consistent. Similarly, Southern African research provides growing evidence of internalizing problems among children orphaned as a result of AIDS but less evidence for externalizing difficulties (Cluver & Gardner, 2007a; Cluver, Gardner, & Operario, 2007). There is also evidence that children orphaned as a result of AIDS show higher rates of adjustment problems than do those orphaned as a result of other causes (Cluver et al., 2007).

A particular situation is typically considered a risk factor if it is statistically associated with an increased probability of negative outcomes or maladjustment (Luthar, Cicchetti, & Becker, 2000). Thus, the above research suggests that parental bereavement is a risk factor for internalizing difficulties. However, it is also clear that there are large individual differences in children’s responses to risk. The construct of resilience is widely used to refer to a relatively good outcome despite experiences of serious stress or adversity (Masten, 2001; Rutter, 2000). “Good” outcomes may be assessed using external adaptation criteria (such as academic achievement), internal adaptation criteria (such as an absence of psychopathology or low levels of symptoms), or some combination of the two (Masten, 2001).

One possible explanation for the variation in children’s responses to parental bereavement is that individual orphans may not have experienced equivalent levels of risk. It is common for risk factors to co-occur, and the accumulation of multiple risk factors poses far greater risks to development than any single negative experience (Masten, 2001; Rutter, 2000). Thus, some children in apparently high-risk contexts may have experienced less adversity than others (Luthar et al., 2000).

Even when youth are exposed to comparable conditions of risk, however, some individuals fare better than others (Rutter, 2000). Resilience is facilitated by protective factors—assets or resources of children, their families, and their wider social environments that distinguish high-functioning children at risk from those who are negatively affected (Luthar et al., 2000). Person-based approaches to studying resilience typically entail comparing two groups from the same high-risk-sample who have adaptive and maladaptive outcomes. Variable-based approaches, in contrast, use multivariate statistics to test for linkages between risk factors, protective factors, and an outcome-criterion variable. In main effects or additive models of resilience, risk factors and protective factors contribute additively to the prediction of the outcome, without interaction. The protective factor confers an advantage regardless of the level of risk or adversity and helps to “neutralize” or compensate for the effect of stresses. In interaction or moderator models, a protective factor interacts with (moderates) a risk factor and prevents a decline in competence under high levels of stress. A protective factor may have a direct effect on adjustment under conditions of low risk, but it has a stronger effect under high-risk conditions (Masten, 2001).

Preliminary evidence suggests that individual differences in children’s responses to the death of a parent may be influenced by personal characteristics such as gender, preexisting mental health, and cognitive appraisals (Dowdney, 2000; Lin, Sandler, Ayers, Wolchik, & Luecken, 2004). However, it is the functioning level of the child’s primary caretaker after the death, their relationship with the child, and the quality of care and support they are able to offer that appear to be of primary importance (Cerel, Fristad, Verducci, Weller, & Weller, 2006; Lin et al., 2004). Effective social support from friends and adults outside the family has also been found to be a protective factor for bereaved children (Hough, Brumitt, Templin, Saltz, & Mood, 2003). In contrast, low socioeconomic status and the presence of additional stressors have been associated with poorer child functioning (Cerel et al., 2006).

To date, however, many factors that may contribute to the variation in children’s responses to parental bereavement remain insufficiently explored (Dowdney, 2000). In addition, most of the existing research has been conducted with predominantly White American children and adolescents living with a surviving parent (usually their mother). The findings of such studies may not be generalizable to communities affected by AIDS in Africa, where parental bereavement often occurs in a context of multiple losses of caregivers and family members, widespread and deepening poverty, poor access to services, and stigma and social exclusion (Richter & Foster, 2005). The vast majority of orphans in South Africa live with extended family, particularly grandmothers and aunts (Madhavan, 2004; Richter & Desmond, 2008). However, there is increasing concern about the capacity of extended families to absorb orphans in the face of serious economic constraints. Many families in AIDS-affected communities rely heavily on social grants. In some areas, faith-based, community-based, and nongovernmental organizations also provide food and other essential services (Madhavan, 2004). However, increasing number of orphans and endemic poverty are placing these support systems under increasing strain, which means that little attention is paid to psychosocial issues such as bereavement and loss (Davids & Skinner, 2006). Two qualitative studies conducted in South Africa (Cluver & Gardner, 2007b; Rochat & Hough, 2007) have supported the importance of the presence of caring and supportive primary caregivers, extended families, friends, and community members in contributing toward well-being in orphaned children. However, they also identified additional risks related to poverty, crime and violence, lack of access to education, and perceived stigma and discrimination that may be unique to this group.

The small amount of quantitative evidence available from research in Africa suggests that female gender, an unsympathetic foster family, AIDS-related stigma, poverty, and exclusion from school are associated with an increased likelihood of emotional and social problems in orphaned children (Cluver et al., 2008; Makame, Ani, & Grantham-McGregor, 2002). In contrast, support from extended family and other orphans may help to mitigate the potential negative effects of parental bereavement (Atwine, Cantor-Graae, & Bajunirwe, 2005). However, little systematic attention has been paid to social and environmental factors that might reduce risk or facilitate resilience in orphaned children and can be targeted in intervention programs.

The present study extends this research by testing a variable-based model of resilience in orphaned adolescents living in a community affected by AIDS in South Africa. It builds on the work of Barber and his colleagues (Barber, Stolz, & Olsen, 2005), who have demonstrated that healthy adolescent development in a variety of cultures is associated with experiences of “connection” (loving, supportive relationships with significant others), “regulation” (supervision, structure, and monitoring), and “respect for individuality” (acknowledging and respecting an adolescent’s individual self by avoiding intrusive, exploitative, or manipulative behaviors). There is evidence that adolescents’ functioning is related to these conditions as experienced in different social contexts (family, school, neighborhood, and peers) and that deficits in experience in one context can be compensated for by experience in other contexts (Barber & Olsen, 1997). This raises the possibility that orphaned adolescents might be protected from experiencing adjustment problems to the extent that they experience connection, regulation, and respect for individuality in their significant relationships and are exposed to regulating forces in their homes and communities.

As the focus of this study was to identify factors that can reduce risk or increase resilience in orphaned adolescents, our first task was to determine the criteria by which to assess relatively good outcomes. The research literature has demonstrated that parental bereavement is statistically associated with increased internalizing problems (Cluver & Gardner, 2007a). Consistent with this literature, results of an earlier study indicated that the orphans who participated in the present study reported more symptoms of depression and anxiety and lower self-esteem than nonorphans from the same communities did (Wild, Flisher, Laas, & Robertson, 2006). In this context, low levels of internalizing problems among orphans imply greater relative resistance to the stresses associated with a parent’s death. Consequently, resilience was operationally defined in terms of low levels of internalizing symptoms and measured using a composite score based on the three measures that distinguished the orphans from the nonorphans. This outcome criterion or “emotional resilience” score was subsequently predicted by selected risk-adversity and asset variables.

Because orphaned adolescents are not necessarily a homogeneous group, the first aim of our study was to identify factors related to the bereavement and the adolescents’ experiences prior to and following the parent’s death that might contribute to the overall level of risk experienced. Our second aim was to investigate the main and interaction effects of family, peer, and community assets in helping to protect adolescents from the potential negative effects of orphanhood and co-occurring risks on their psychological adjustment.

Adolescents were classified as orphans if they had lost one or both parents (UNAIDS/UNICEF/USAID, 2004). We used the World Health Organization (2003) definition of adolescence as those aged between 10 and 19 years. We focused on adolescents because they are expected to form the bulk of the orphan population and because existing services in South Africa concentrate on children aged less than 12 years. In addition, the factors that influence the adjustment of adolescents may differ from those that are important for younger children.

Method

Participants

The sample comprised 159 parentally bereaved adolescents aged between 10 and 19, of whom 87 (55%) were female. The participants were drawn from the Eastern Cape Province of South Africa. In 2005, AIDS accounted for 55% of maternal orphans in this province (Dorrington et al., 2006). The study area comprised a complex of urban and peri-urban settlements characterized by poverty, high levels of unemployment, and a lack of social and economic infrastructure. All but 3 participants (98%) were isiXhosa-speaking, and 95% were enrolled in school at the time of the study.

There are no reliable records of orphans available in South Africa, and recruiting families affected by AIDS is difficult given the stigma and secrecy that surrounds the disease. Consequently, participants were recruited with the assistance of nongovernmental and community-based organizations (NGOS and CBOs) that provide care and support to people living with AIDS. Additional participants were identified by local schools and through word of mouth (“snowballing”). Adolescents whose parents had died less than 6 months previously were excluded from the study in order to reduce the influence of acute bereavement issues. Some researchers describe the 6-month mark as a turning point after which acute grief reactions tend to abate (Cerel et al., 2006).

Measures

Demographic Information and Risk-Adversity Variables

Adolescents completed a brief questionnaire designed to obtain information on their age and sex, past and present living arrangements, and the timing of their parent’s death. Material deprivation was assessed by summing the number of four basic facilities that the household possessed: electricity, a television, a telephone, and a car. A lack of basic household facilities is considered a parsimonious, robust, and child-centered indicator of material deprivation in South Africa (Barnes, Wright, Noble, & Dawes, 2007). In addition, the adolescents’ cumulative stress exposure was assessed by summing their responses to 21 items drawn from the Life Events Questionnaire for Adolescents (LEQ-A; Masten, Neemann, & Andenas, 1994). These items include both discrete and chronic life events and have all been judged to be both negative and independent of the adolescent’s behavior (e.g. “A parent or caregiver had trouble with alcohol or drugs”). For each item, adolescents were asked to indicate whether it was something that had happened to them during the last 6 months (scored 2), something that had happened during their life but more than 6 months ago (scored 1), or something that had never happened to them (scored 0). The LEQ-A has been used previously with urban adolescents in South Africa and Kenya (Seedat, Nyamai, & Vythilingum, 2004).

Family Assets

Connection

The provision of emotional connection and support was measured using the 10-item acceptance subscale from the revised Child Report of Parent Behavior Inventory (CRPBI; Barber et al., 2005; Schaefer, 1965), modified to apply to the adolescents’ current caregiver. The adolescent was asked to respond on a 3-point scale ranging from 0 (not like her/him) to 2 (a lot like her/him) as to how well items (e.g., “gives me a lot of care and attention”) describe this adult. High scores on this measure indicate a relationship characterized by acceptance, affection, and positive evaluation, whereas low scores signify ignoring, neglect, or rejection. Barber et al. have previously used this scale in South Africa and reported a Cronbach’s alpha coefficient of .90 (for mother) and .89 (for father) in an isiXhosa-speaking sample of adolescents. In this sample, alpha was .91.

Regulation

The caregiver’s regulation of the adolescent’s behavior was measured using a 5-item monitoring scale (Barber et al., 2005). The adolescent responded on a 3-point scale ranging from 0 (doesn’t know) to 2 (knows a lot) regarding how much their caregivers “really know” about the following: where you go at night, where you are most afternoons after school, how you spend your money, what you do with your free time, and who your friends are. Barber et al. calculated Cronbach’s alpha coefficients for an isiXhosa translation of the scale to be .71 (for mother) and .82 (for father). In this sample, alpha was .69.

Respect for individuality

Respect for Individuality was measured using the 8-item Psychological Control Scale—Youth Self-Report (PCS-YSR; Barber et al., 2005). The adolescent responded on a 3-point scale ranging from 0 (not like her/him) to 2 (a lot like her/him) as to how well items describe their caregiver. A sample item is, “[My parent/carer/guardian] is a person who . . . is less friendly with me if I do not see things her/his way.” The scale assesses the extent to which adolescents perceive their caregivers as invalidating the adolescent’s feelings, constraining their verbal expressions, personally attacking them, and withdrawing love from them. The response scale was reversed so that high scores indicate more respect for individuality. Barber et al. reported Cronbach’s alpha coefficients for an isiXhosa translation of this scale to be .72 (for mother) and .76 (for father). In this sample, alpha was .61.

Peer Assets

Connection

Following Barber and Olsen (1997), one item was used to index peer connection. Adolescents were asked whether they had a best friend, and those who answered “yes” to this question were then asked, “How much does this person like or love you?” The response scale ranged from 3 (very much) through 0 (little or none). Adolescents who did not have a best friend (n = 38) were given a score of 0 on this measure.

Regulation

Barber and Olsen’s (1997) adaptation of a 11-item measure of peer delinquency developed by Elliot, Huizinga, and Ageton was used as a measure of peer regulation. Adolescents were asked how many of their friends purposely damage or destroy property, run away from home, and so on. The response scale ranged from 4 = all to 0 = none. The scale was reverse coded so that high scores indicate the presence of regulation (nondeviance) in peer relationships. Original alpha coefficients for the isiXhosa translation of this scale are not available. The alpha coefficient in this sample was .85.

Respect for Individuality

One item was used to measure respect for individuality in the adolescent’s peer relationships. With reference to their best friend, adolescents were asked, “How much does this person try to control what you do, think, or say?” The response scale ranged from 0 = none to 3 = very much. The response scale was reversed so that high scores reflect high autonomy.

Community Assets

Connection

Connection to adults in the community was assessed with a 4-item scale measuring how often the adolescent had spent time during the last 6 months with neighbors, parents of friends, community leaders, and church leaders (Barber & Olsen, 1997). The response scale ranged from 0 (not at all) to 6 (every day). Original alpha coefficients for the isiXhosa translation of this scale are not available. In this sample, alpha was .60.

Regulation

Community regulation was assessed by a 5-item scale measuring the presence of social disorganization (Barber & Olsen, 1997). Adolescents were asked how often they had heard in the past few months of things such as the following happening in their neighborhood: a fight in which a weapon was used, violent arguments between neighbors, and youth gang conflicts. The response scale was 0 = never to 3 = often. The scale was reverse coded so that high scores indicate high levels of community regulation. Original alpha coefficients for the isiXhosa translation of this scale are not available. Internal consistency in this sample was .72.

No measure of community respect for individuality was available.

Adolescent Adjustment

Depression

Symptoms of depression were assessed using the 10-item short form of the Children’s Depression Inventory (CDI; Kovacs, 1992). Responses were scored on a scale from 0 to 2, with 2 representing the severe form of a depressive symptom. The short form is based on the 10 best discriminating and most internally consistent items from the longer 27-item form (Kovacs, 1992). Barber et al. (2005) calculated a Cronbach’s alpha of .70 for an isiXhosa translation of this scale in a South African sample. In this sample, alpha was .64.

Anxiety

Symptoms of anxiety were assessed using the Children’s Manifest Anxiety Scale—Revised (RCMAS; Reynolds & Richmond, 1978). Participants were required to answer “yes” or “no” to each of 28 anxiety items; for example, “I worry a lot of the time.” In addition, nine lie-scale items were included to assess socially desirable responding. Support has been provided for the reliability and validity of the RCMAS (Reynolds & Richmond, 1978). Original alpha coefficients for the isiXhosa translation of this scale are not available. In this sample, alpha was .79.

Self-esteem

Self-esteem was assessed using seven items from the global self-worth subscale of the Self-Esteem Questionnaire (SEQ; DuBois, Felner, Brand, Phillips, & Lease, 1996). A sample item is, “I am happy with myself as a person.” Each item was rated on a 4-point scale ranging from strongly disagree to strongly agree. DuBois et al. have provided evidence in support of the validity of the SEQ. The scale has adequate psychometric properties when used with South African youth, including an original Cronbach’s alpha coefficient of .76 (Wild, Flisher, Bhana, & Lombard, 2005). In this sample, alpha was .86.

Procedure

All the questionnaires were translated from English into isiXhosa by translators who spoke isiXhosa as a home language. The accuracy of the translated questionnaires was assured by back-translating them into English and comparing the back-translated version with the original version. Any discrepancies were resolved by negotiation between at least two translators. The translated measures were piloted in Cape Town with a sample of 20 isiXhosa-speaking adolescents orphaned by AIDS.

Three trained and supervised isiXhosa-speaking interviewers who had experience in working with families affected by AIDS acted as fieldworkers. All questionnaires were administered verbally in a one-to-one interview format. Participants were interviewed in private, in their homes, or in an alternative location of their choice.

Ethical approval for this study was obtained from the Health Sciences Faculty of the University of Cape Town. Caregivers and adolescents signed informed assent forms prior to the interview. Participants were advised on the aims and nature of the study, what they would be required to do, who was undertaking the research, and how it would be disseminated and applied on completion. They were informed that they were free to withdraw at any stage of the study, without prejudice, and that their personal information would remain confidential. In order to minimize the possibility of disclosure of the HIV status of the deceased parent where this was not the wish of that person, no direct reference to HIV and AIDS was made in the interview. Participants were given a small token of appreciation (ZAR30, or approximately US$4) for their time, and those in need of assistance were referred to appropriate mental health or social services.

Measuring Emotional Resilience

The three measures of internalizing problems utilized in this study were significantly intercorrelated at .42 for depression and anxiety, -.37 for depression and self-esteem, and -.28 for anxiety and self-esteem (all p values < .001). Following Dutra et al., these variables were, therefore, combined to form a composite score of emotional resilience by standardizing and then averaging the scale scores. The CDI and RCMAS were reverse scored before being converted to a z score so that higher scores indicated greater emotional resilience. A relative, variable-based measure of resilience was chosen in preference to an absolute, person-based approach of dividing the sample into resilient and nonresilient subgroups because the former approach provides a more sensitive outcome measure and increases statistical power (Dutra et al., 2000). Furthermore, the lack of standardization data for the adjustment measures in South Africa meant that there were no established cutoff points for dichotomizing the sample.

Results

Descriptive Statistics

The mean age of the adolescents was 15 years (SD = 2.26). Their mean age when their parent died was 11.73 (SD = 3.23), and the mean length of time since the death was 3.25 years (SD = 2.69). Ninety-seven adolescents (61%) had lost their mother only, 38 (24%) their father only, and 24 (15%) had lost both parents. The relationship between the current caregiver and the adolescent was grandparent in the case of 49% of participants; aunt/uncle in the case of 21%; biological parent, 16%; sibling, 5%; and no relation, 9%. Overall, 96% of adolescents aged between 10 and 14 years and 85% of those aged between 15 and 19 years were in the care of a biological relative. The cause of the parent’s death was HIV and AIDS in the case of 51% of the participants and other illness in the case of 29% (e.g., cancer, heart disease); accident, 9%; and homicide, 5%. The cause of death of the remaining 6% of the parents was unknown or attributed to “betwitchment.” 31 adolescents (19%) lived in households that lacked electricity, and only 11 households (7%) had a car. Detailed descriptive statistics for the main study variables are presented in Table 1, disaggregated by gender and two age groupings.

Table 1.

Descriptive Statistics Disaggregated by Gender and Age Grouping

		Male			Female
	Age	N	M	SD	N	M	SD
Life Events Questionnaire for Adolescents (LEQ-A)	10-14	40	4.85	2.57	44	4.68	2.56
	15-19	32	5.81	3.32	43	6.49	3.11
Family assets
Connection	10-14	40	17.10	4.47	44	17.32	4.21
	15-19	32	17.28	3.55	43	16.07	4.54
Regulation	10-14	40	7.00	2.65	44	7.16	2.85
	15-19	32	6.06	2.53	43	6.12	2.70
Respect for Individuality	10-14	40	10.70	3.37	44	11.23	3.13
	15-19	32	11.41	2.65	43	10.05	3.43
Peer assets
Connection	10-14	40	1.88	1.16	44	1.57	1.13
	15-19	32	1.66	0.94	43	1.67	0.99
Regulation	10-14	40	41.25	4.10	44	41.84	3.21
	15-19	32	40.66	6.51	43	42.07	4.00
Respect for Individuality	10-14	40	1.68	0.75	31	1.23	0.56
	15-19	32	2.12	0.78	33	1.82	0.85
Community assets
Connection	10-14	40	6.95	6.11	44	6.02	5.50
	15-19	32	6.94	6.60	43	6.07	4.81
Regulation	10-14	40	6.50	2.26	43	6.88	2.06
	15-19	32	5.78	2.68	43	6.33	2.68
Outcomes
Depression	10-14	40	2.65	1.85	44	3.34	3.03
	15-19	32	3.41	3.29	43	3.30	2.35
Anxiety	10-14	40	13.45	4.31	44	14.68	5.26
	15-19	32	13.25	5.76	43	15.51	4.95
Self-esteem	10-14	40	22.55	3.15	44	21.32	2.97
	15-19	32	20.69	2.87	42	20.19	2.80

Preliminary Analyses

Preliminary analyses were conducted to examine the relationship between emotional resilience scores and (a) three personal characteristics of children (age, gender, and socially desirable responding), and (b) seven factors related to the bereavement experience itself and the adolescents’ experiences prior to and following the parent’s death that might contribute to the overall level of risk or adversity experienced (time since the parent’s death, child age when the death occurred, whether the death was a result of HIV and AIDS, gender of the deceased [mother, father, or both], whether the child’s current caregiver was a relative or nonrelative, material deprivation, and cumulative stress).The degree of relationship between two continuous variables was assessed using the Pearson product–moment correlation coefficient. The point biserial correlation was used when one of the variables was measured as a dichotomy. Because gender of the deceased had three possible values, a one-way analysis of variance (ANOVA) was used to compare the means for the adolescents who had lost their mother only (M = 0.01, SD = 0.74), father only (M = -0.33, SD = 0.83) or both parents (M = -0.12, SD = 0.66).

Results of the correlational analyses are displayed in Table 2. Emotional resilience scores were significantly negatively correlated with age, material deprivation, and cumulative stress. Males had significantly higher emotional resilience scores than females did, adolescents orphaned due to HIV and AIDS had significantly higher emotional resilience scores than did those orphaned as a result of other causes, and adolescents living with biological relatives had significantly higher emotional resilience scores than did those living with nonrelatives. The effect of gender of deceased was not statistically significant, F(2, 155) = 2.78, p =.07.

Table 2.

Correlations Among Demographic and Risk-Adversity Variables and Emotional Resilience

	1	2	3	4	5	6	7	8	9
1. Child age	—
2. Child gender^a	−.05	—
3. Time since death	.13	.00	—
4. Age when bereaved	.59***	−.04	−.73***	—
5. Cause of death^b	−.11	.13	−.29***	.16	—
6. Current caregiver^c	−.24**	.06	−.04	−.13	.14	—
7. Material deprivation	.13	−.04	−.04	.13	−.28***	−.06	—
8. Cumulative stress	.20*	−.05	−.09	.21**	.03	−.04	.05	—
9. Social desirability	−.02	−.08	−.13	.09	−.13	−.15	.27**	−.02	—
10. Emotional resilience	−.17*	.18*	−.01	−.10	.29***	.22**	−.18*	−.17*	−.13

Dummy-coded variable: 1 = male; 0 = female.

Dummy-coded variable: 1 = HIV and AIDS; 0 = other.

Dummy-coded variable: 1 = relative; 0 = nonrelative.

p < .05. **p < .01. ***p < .001.

Correlations

Table 3 displays the correlations between the asset variables and emotional resilience. All of the correlations except that between peer respect for individuality and emotional resilience were statistically significant. Greater emotional resilience was associated with higher levels of connection and regulation in the family, peer, and community contexts and with greater respect for individuality in the family.

Table 3.

Intercorrelations Among Assets and Emotional Resilience

	1	2	3	4	5	6	7	8
1. Family connection	—
2. Family regulation	.11	—
3. Family respect for individuality	.07	.09	—
4. Peer connection	.18*	.10	−.06	—
5. Peer regulation	−.02	.08	.18*	.01	—
6. Peer respect for individuality	−.26**	−.08	.19*	−.25**	.12	—
7. Community connection	.01	.31***	−.18*	.13	−.18*	−.04	—
8. Community regulation	.32***	−.04	.08	−.11	.18*	−.08	−.20*	—
9. Emotional resilience	.23**	.22**	.23**	.23**	.17*	−.02	.16*	.22**

p < .05. **p < .01. ***p < .001.

Primary Analyses

A hierarchical multiple regression analysis was used to test a main-effects model of resilience by examining the relative and additive associations between risks, assets, and adolescents’ emotional resilience. In addition, an interaction model was tested to determine whether any associations between the assets and emotional resilience varied according to the overall level of stress experienced. Only variables significantly correlated with emotional resilience in the correlational analyses were included in the regression analysis. Variables were entered in blocks, in the following order: Step 1: Age and gender were entered first as control variables. Step 2: The four risk-adversity variables significantly associated with resilience scores in the preliminary analyses (cause of death, current caregiver, material deprivation, and cumulative stress exposure) were entered. Step 3: Family assets (connection, regulation, and respect for individuality) were entered. Step 4: Peer assets (connection and regulation) were entered. Step 5: Community assets (connection and regulation) were entered. Step 6: Interaction (multiplicative) terms were entered. Separate analyses were conducted for the interaction of each asset with cumulative stress. A moderator model is supported if interactions are significant. The variables were entered in the order specified so that those most proximal to the adolescent were entered before more distal variables.

Results are presented in Table 4. Information is provided for the change in R² accounted for by each block of variables and the standardized beta weight (and its standard error) of the individual variables in each block.

Table 4.

Results of Hierarchical Multiple Regression Analysis Predicting Adolescents’ Emotional Resilience

Step	Variable	ΔR²	β	SE β
1.	Demographic controls (set of two)	.06
	Age		−.16*	.08
	Gender^a		.17*	.08
2.	Risk-adversity variables (set of four)	.11***
	Cause of death^b		.22**	.08
	Current caregiver^c		.15*	.08
	Material deprivation		−.09	.08
	Cumulative stress		−.15*	.08
3.	Family assets (set of three)	.07**
	Connection		.14	.08
	Regulation		.16*	.07
	Respect for individuality		.16*	.07
4.	Peer assets (set of two)	.04*
	Connection		.14*	.07
	Regulation		.13	.08
5.	Community assets (set of two)	.05**
	Connection		.19*	.08
	Regulation		.19*	.08
6.	Interactions (alternatives)
	Family Connection × Cumulative Stress	.00	.05	.27
	(Family Regulation × Cumulative Stress)	.00	.17	.23
	(Family Respect for Individuality × Cumulative Stress)	.00	.09	.21
	(Peer Connection × Cumulative Stress)	.01	−.30	.18
	(Peer Regulation × Cumulative Stress)	.00	.21	.55
	(Community Connection × Cumulative Stress)	.01	−.20	.19
	(Community Regulation × Cumulative Stress)	.00	−.13	.16
	R² = .33, Adj. R² = .27***

Dummy-coded variable: 1 = male; 0 = female.

Dummy-coded variable: 1 = HIV and AIDS; 0 = other.

Dummy-coded variable: 1 = relative; 0 = nonrelative.

p ≤ .05. **p < .01. ***p < .001.

After controlling for age and gender, three risk-adversity variables had a significant individual association with the outcome criterion. Adolescents living with a biological relative and who had lost a parent as a result of HIV and AIDS had higher scores on the emotional resilience measure than those living with nonrelatives or whose parents had died of causes other than HIV and AIDS. In contrast, more cumulative stress exposure was associated with lower emotional resilience scores.

Caregiver, peer, and community assets all accounted for a significant increase in the variance after controlling for variables up to and including that step. Greater connection to a friend and to adults in the community, more regulation in the family and community, and greater respect for individuality from the caregiver all had significant unique associations with higher emotional resilience scores. There were no significant interactions between these assets and adolescents’ cumulative stress exposure.

Group Comparison

In order to gain insight into the risk and protective mechanisms that might underlie the regression results, we investigated whether differences in emotional resilience scores associated with the cause of parental death could be accounted for by different levels of community connection. Because the assumption of homogeneity of group variances was not met (indicated by a significant result on Levene’s test), the Mann–Whitney U test was used as a nonparametric alternative to the independent samples t test. Results indicated that the mean community connection rank of adolescents orphaned as a result of HIV and AIDS (88.18) was significantly greater than that of those orphaned as a result of other causes (71.51; z = 2.28, p < .05).

Discussion

The results of this study support the idea that the impact of parental bereavement is likely to vary depending on the broader context in which it occurs, and in particular on the balance that exists between risk and protective factors in adolescents’ families and communities (Rutter, 2000; Ungar, 2008). Although the literature suggests that orphans as a group are at increased risk of internalizing problems, this risk was magnified for adolescents who had lost a parent as a result of a cause other than AIDS, whose primary caregiver was a nonrelative, and who had more cumulative stress exposure. Even when varying levels of risk were taken into account, however, adolescents’ experiences of regulation and respect for individuality in the family context, connection in the peer context, and connection and regulation in the community context were significantly associated with lower levels of internalizing problems.

The results of the regression analyses support previous research on parental bereavement that has found higher rates of internalizing problems in girls than in boys (Dowdney, 2000). In addition, older adolescents were found to be more vulnerable than their younger counterparts. This finding may reflect more adult role–taking among older adolescents (Stein, Riedel, & Rotheram-Borus, 1999), particularly as relatives may be less willing to assume custody for older adolescents than for younger children (Siegel & Gorey, 1994). Older adolescents in the present study were significantly less likely to live with a relative than younger adolescents were (Table 2).

After controlling for age and gender, three risk-adversity variables—adolescents’ cumulative stress exposure, current caregiver, and the cause of the parent’s death—were significantly independently associated with the outcome criterion. The finding that cumulative stress was associated with poorer outcomes is consistent with previous research demonstrating that children exposed to multiple adversities are at greater risk of psychopathology than those experiencing a single major risk (Rutter, 2000). It also supports Richter and Desmond’s (2008) argument that orphanhood is not the only source of vulnerability in communities affected by widespread poverty and HIV and AIDS. The reduced risk of internalizing difficulties associated with living with a biological relative is also not surprising. Orphaned children in South Africa have identified extended family members as a source of practical care, comfort, and consolation (Cluver & Gardner, 2007b; Rochat & Hough, 2007). In addition, Forehand et al. (1999) found that children who had moved into the stable environment of a close relative did not show an increase in difficulties after the AIDS-related death of their mother. However, the observation that adolescents orphaned as a result of AIDS were less at risk than those orphaned as a result of other causes was unexpected, as it contradicts the findings of previous research (Bhargava, 2005; Cluver et al., 2007). The most likely explanation for this finding is that some of the AIDS-orphaned adolescents were receiving a degree of unofficial practical and emotional assistance from HIV and AIDS support groups to which their parent(s) had belonged and which was not available to adolescents orphaned as a result of other causes—perhaps because official government recognition and support for palliative care is limited to AIDS patients (Gwyther, 2002). The AIDS-orphaned adolescents in our sample were less economically disadvantaged than other orphans (see Table 1), and there is evidence that poverty mediates the association between orphanhood and psychological problems (Cluver et al., 2009). In addition, the AIDS-orphaned adolescents experienced more community connection than did those orphaned as a result of other causes. This suggests that we may have found more internalizing symptoms among AIDS-orphaned adolescents in areas where support groups are less active (e.g., in rural areas) or among families where HIV and AIDS was not recognized or acknowledged as the cause of death.

Indeed, the results of this study emphasize the important role that social support plays in facilitating healthy outcomes in orphaned adolescents. More specifically, the findings support a main-effects model of resilience in which the risks and assets experienced in adolescents’ families, peer groups, and communities contribute independently and additively toward explaining their psychological adjustment, without interaction. Previous research has highlighted connection, regulation, and respect for individuality as three central dimensions of parenting influencing adolescent development (Barber et al., 2005). This study suggests that after the death of a parent, adolescents’ psychological well-being is significantly influenced by the extent to which such experiences are provided by alternative or surviving caregivers, friends, and/or adults in the community. Greater connection to a friend and to adults in the community, more regulation in the family and community, and greater respect for individuality from the caregiver were all associated with greater emotional resilience. This supports previous research suggesting that resilience is facilitated by warm, supportive relationships, opportunities to exercise autonomy and decision making, and appropriate supervision and monitoring by parents, teachers, and neighbors that serves to protect adolescents from further environmental risk (Rutter, 2000). These protective effects were evident regardless of the overall level of stress experienced.

In considering these results, there are some limitations of the study that need to be taken into account. First, the reliance on self-report data raises the possibility that the findings may have been inflated by shared method variance. In addition, a lack of access to medical records meant that we were not always able to verify the reported cause of the parent’s death. In contrast to other informants, adolescents have direct access to their own emotional states, and their perceptions of their environment are crucial to understanding its impact on them (Rutter, 2000). Nevertheless, future research would benefit from the inclusion of multiple sources of information and verbal autopsy methods for determining causes of mortality.

A second limitation of this study is the possibility that the nonrandom sampling procedure may have led to an overrepresentation of AIDS-orphaned adolescents whose families were receiving support from NGOS or CBOS, and hence, to an underestimation of internalizing symptoms in this group. Although the apparent protective effect of being known to an NGO or CBO speaks to the central goals of this study, further research is needed to determine whether the results can be generalized to other AIDS-affected communities.

A third caveat is that the resilience construct was limited to low levels of internalizing symptoms. Luthar et al. argue that when a risk factor is more likely to affect one domain of functioning over others, it is logical to prioritize that domain in assessing outcomes. However, the maintenance of competent functioning in this domain does not imply positive adaptation across other important areas such as scholastic functioning or social competence (Luthar et al., 2000; Rutter, 2000). It is also important to note that most of the measures used in this study were developed in White, middle-class populations in North America and Europe. Although particular care was taken to select measures that have been found to be reliable and valid across cultures and that have been used in other studies in South Africa, it is possible that such psychological instruments do not fully capture locally meaningful experiences and understanding of poor mental health. For example, clinical reports suggest trends among Black South Africans to report internalizing psychological distress through somatic complaints rather than emotional symptoms (Swartz, 1998). Furthermore, the study design was cross-sectional and correlational, precluding conclusions about causality. Some of these limitations could be addressed by future research using prospective, longitudinal designs and a wider range of outcome measures.

Conclusion

A large, and rapidly growing, number of children in South Africa are orphaned as a result of the AIDS pandemic, and providing them with care and protection is an increasing concern. There is a growing body of research on the psychosocial problems experienced by orphaned children. However, little attention has been paid to understanding individual differences in adjustment. The current study extends the small body of research on resilience in orphans by using a comprehensive battery of measures of family, peer, and community resources.

The findings suggest that intervention efforts should not be limited to AIDS-orphaned adolescents but that they should rather focus on those who are in the care of nonrelatives and/or have been exposed to multiple adversities before or after the parent’s death. Girls and older adolescents may also be in particular need of support. The research results also suggest that practitioners and policy makers may be able to facilitate resilience in orphaned adolescents by strengthening the capacity of their extended families, friends, and competent adults outside of the family to provide them with emotional support, respect for individuality, and appropriate supervision. Thus, preventive programs to foster good parenting in adults caring for orphans and positive social experiences at school may well be beneficial (Rutter, 2000). Expanding the access of families affected by life-threatening illnesses other than AIDS to community-based palliative care may also have value. The Memory Box program (which includes family interventions and children’s groups) has demonstrated some success in helping South African children orphaned or emotionally affected by HIV and AIDS (Richter, Manegold, & Pather, 2004) and could potentially be expanded to other provinces and to children orphaned as a result of other causes. The task now is to identify additional factors that foster resilience and that can be incorporated in intervention programs.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Financial support for this research was provided by the Bristol-Myers Squibb “Secure the Future” project; the South African Medical Research Council; and the AIDS and Society Research Unit, Faculty of Health Sciences Research Committee and University Research Committee of the University of Cape Town. The authors would like to thank Yvonne Gilbert for coordinating the fieldwork, and Patricia Betsha, Namfazi Ziwele, and Zanele Ngesi for conducting the interviews. We would also like to express our gratitude to Regina Meke, the SA Red Cross Society, NAPWA, the Christelike Maatskaplike Raad, the East London Hospice, and the Sisonkhe support group for assisting with recruitment.

Bios

Lauren G. Wild is a senior lecturer in the Department of Psychology at the University of Cape Town. Her research focuses on family stress, caregiving, and child adjustment and on risk-taking behavior in adolescence.

Late Alan J. Flisher was the Sue Streungmann professor of child and adolescent psychiatry and mental health and head of the Division of Child and Adolescent Psychiatry at the University of Cape Town. His areas of expertise included psychiatric epidemiology, mental health policy and services research, and adolescent health.

Brian A. Robertson is an emeritus professor in the Department of Psychiatry and Mental Health at the University of Cape Town. His current interests are HIV/AIDS and mental health and the development of child and adolescent mental health services in Africa.

References

Atwine

Cantor-Graae

Bajunirwe

(2005). Psychological distress among AIDS orphans in rural Uganda. Social Science and Medicine, 61, 555-564.

Barber

B. K.

Olsen

J. A.

(1997). Socialization in context: Connection, regulation, and autonomy in the family, school, and neighborhood, and with peers. Journal of Adolescent Research, 12, 287-315.

Barber

B. K.

Stolz

H. E.

Olsen

J. A.

(2005). Parental support, psychological control, and behavioral control: Assessing relevance across time, method, and culture. Monographs of the Society for Research in Child Development, 70(4), 1-137.

Barnes

Wright

Noble

Dawes

(2007). The South African index of multiple deprivation for children: Census 2001. Cape Town, South Africa: HSRC Press.

Bhargava

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

Cerel

Fristad

M. A.

Verducci

Weller

R. A.

Weller

E. B.

(2006). Childhood bereavement: Psychopathology in the 2 years postparental death. Journal of the American Academy of Child and Adolescent Psychiatry, 45, 681-690.

Cluver

Gardner

(2007a). The mental health of children orphaned by AIDS: A review of international and Southern African research. Journal of Child and Adolescent Mental Health, 19, 1-17.

Cluver

Gardner

(2007b). Risk and protective factors for psychological well-being of children orphaned by AIDS in Cape Town: A qualitative study of children and caregivers’ perspectives. AIDS Care, 19, 318-325.

Cluver

Gardner

Operario

(2007). Psychological distress amongst AIDS-orphaned children in urban South Africa. Journal of Child Psychology and Psychiatry, 48, 755-763.

10.

Cluver

Gardner

Operario

(2008). Effects of stigma on the mental health of adolescents orphaned by AIDS. Journal of Adolescent Health, 42, 410-417.

11.

Cluver

Gardner

Operario

(2009). Poverty and psychological health among AIDS-orphaned children in Cape Town, South Africa. AIDS Care, 21, 732-741.

12.

Davids

Skinner

(Eds.). (2006). A situational analysis of orphans and vulnerable children in four districts of South Africa. Cape Town, South Africa: HSRC Press.

13.

Dorrington

R. E.

Johnson

L. F.

Bradshaw

Daniel

(2006). The demographic impact of HIV/AIDS in South Africa. National and provincial indicators for 2006. Cape Town, South Africa: Centre for Actuarial Research, South African Medical Research Council and Actuarial Society of South Africa.

14.

Dowdney

(2000). Annotation: Childhood bereavement following parental death. Journal of Child Psychology and Psychiatry, 7, 829-830.

15.

DuBois

D. L.

Felner

Brand

Phillips

R. S. C.

Lease

M. A.

(1996). Early adolescent self-esteem: A developmental-ecological framework and assessment strategy. Journal of Research on Adolescence, 6, 543-579.

16.

Dutra

Forehand

Armistead

Brody

Morse

Simon Morse

Clark

(2000). Child resiliency in inner-city families affected by HIV: the role of family variables. Behaviour Research and Therapy, 38, 471-486.

17.

Forehand

Pelton

Chance

Armistead

Morse

P. S.

Stock

(1999). Orphans of the AIDS epidemic in the United States: Transition-related characteristics and psychosocial adjustment at 6 months after mother’s death. AIDS Care, 11, 715-722.

18.

Gwyther

(2002). South Africa: The status of palliative care. Journal of Pain and Symptom Management, 24, 236-238.

19.

Hough

E. S.

Brumitt

Templin

Saltz

Mood

(2003). A model of mother-child coping and adjustment to HIV. Social Science and Medicine, 56, 643-655.

20.

Johnson

Dorrington

(2001). The impact of AIDS on orphanhood in South Africa: A quantitative analysis (CARE Monograph No. 4). Retrieved from http://www.commerce.uct.ac.za/Research_Units/CARE/Monographs/Monographs/mono04.pdf

21.

Kovacs

(1992). Children’s depression inventory. Niagara Falls, NY: Multi-Health Systems.

22.

Lin

K. K.

Sandler

I. N.

Ayers

T. S.

Wolchik

S. A.

Luecken

L. J.

(2004). Resilience in parentally bereaved children and adolescents seeking preventive services. Journal of Clinical Child and Adolescent Psychology, 33, 673-683.

23.

Luthar

S. S.

Cicchetti

Becker

(2000). The construct of resilience: A critical evaluation and guidelines for future work. Child Development, 71, 543-562.

24.

Madhavan

(2004). Fosterage patterns in the age of AIDS: continuity and change. Social Science and Medicine, 58, 1443-1454.

25.

Makame

Ani

Grantham-McGregor

(2002). Psychological well-being of orphans in Dar El Salaam, Tanzania. Acta Paediatrica, 91, 459-465.

26.

Masten

A. S.

(2001). Ordinary magic: Resilience processes in development. American Psychologist, 56, 227-238.

27.

Masten

A. S.

Neemann

Andenas

(1994). Life events and adjustment in adolescents: The significance of event independence, desirability and chronicity. Journal of Research on Adolescence, 4, 71-97.

28.

Reynolds

C. R.

Richmond

B. O.

(1978). What I think and feel: A revised measure of children’s anxiety. Journal of Abnormal Child Psychology, 6, 271-280.

29.

Richter

L. M.

Desmond

(2008). Targeting AIDS orphans and child-headed households? A perspective from national surveys in South Africa, 1995-2005. AIDS Care, 20, 1019-1028.

30.

Richter

Foster

(2005). Strengthening systems to support children’s healthy development in communities affected by HIV/AIDS. Geneva, Switzerland: Department of Child and Adolescent Health and Development (CAH), World Health Organization.

31.

Richter

Manegold

Pather

(2004). Family and community interventions for children affected by AIDS. Cape Town, South Africa: HSRC Publishers.

32.

Rochat

Hough

(2007). Enhancing resilience in children affected by HIV and AIDS: Children’s views and experiences of resilience enhancing family and community practices. Pretoria, South Africa: Child, Youth, Family and Social Development Research Programme, Human Sciences Research Council.

33.

Rutter

(2000). Resilience reconsidered: Conceptual considerations, empirical findings, and policy implications. In Shonkoff

J. P

Meisels

S. J.

(Eds.), Handbook of early childhood intervention (2nd ed., pp. 651-682). Cambridge, UK: Cambridge University Press.

34.

Schaefer

E. S.

(1965). Children’s reports of parental behaviour: An inventory. Child Development, 36, 413-424.

35.

Seedat

Nyamai

Vythilingum

(2004). Trauma exposure and post-traumatic stress symptoms in urban African schools. British Journal of Psychiatry, 184, 169-175.

36.

Siegel

Gorey

(1994). Childhood bereavement due to parental death from Acquired Immunodeficiency Syndrome. Journal of Developmental and Behavioral Pediatrics, 15 (Suppl.), 66-70.

37.

Stein

J. A.

Riedel

Rotheram-Borus

M. J.

(1999). Parentification and its impact on adolescent children of parents with AIDS. Family Process, 38, 193-208.

38.

Swartz

(1998). Culture and mental health: A Southern African view. Cape Town, South Africa: Oxford University Press.

39.

UNAIDS/UNICEF/USAID. (2004). Children on the brink 2004: A joint report of new orphan estimates and a framework for action. New York, NY: USAID.

40.

Ungar

(2008). Resilience across cultures. British Journal of Social Work, 38, 218-235.

41.

Wild

L. G.

Flisher

A. J.

Bhana

Lombard

(2005). Psychometric properties of the Self-Esteem Questionnaire for use with South African adolescents. South African Journal of Psychology, 35, 195-208.

42.

Wild

L. G.

Flisher

A. J.

Laas

Robertson

B. A.

(2006, July). The psychosocial adjustment of adolescents orphaned in the context of HIV/AIDS. Poster presented at the International Society for the Study of Behavioural Development Biennial Meeting, Melbourne, Australia.

43.

World Health Organization. (2003). Strategic directions for improving the health of children and adolescents. Geneva, Switzerland: Author.