The Global Burden of Disease Study 2015 estimated that there were 1.2 m HIV/AIDS deaths worldwide (95% confidence interval [CI]: 1.1 to 1.3 m), of which 859,000 (95% CI: 804,610 to 912,930) were in sub-Saharan Africa.1 There are a range of studies about the impact of orphanhood in sub-Saharan Africa, particularly in countries where orphanhood increased sharply in the 1990s due to the AIDS epidemic.2–9 These studies have mostly focused on schooling and sometimes health outcomes. Over time, they have evolved from the use of cross-sectional survey data, to more sophisticated analysis with longitudinal data. This has facilitated interpreting results as informing on the causal impact of orphanhood. Generalizing these findings, the impacts that are found are mostly associated with the loss of a child's mother, rather than paternal deaths.
Another strand of the literature on orphanhood in sub-Saharan Africa focuses on psychosocial consequences. The first case-control study investigating this came from Tanzania, where 41 orphans where compared with 41 matched nonorphans from the same neighborhoods.10,11 The study highlighted how orphans were not just socioeconomically disadvantaged, but also had elevated levels of internalizing problems, potentially jeopardizing their long-run mental health.
Similar evidence from other countries soon followed and over the past 15 years, evidence has accumulated from many countries in the region, such as South Africa,12–14 Uganda,15,16 Ethiopia,17 Zimbabwe,18 Namibia,19 and Rwanda.20 By and large, these studies found negative associations between orphanhood and psychosocial outcomes, although some studies stressed that circumstances surrounding the orphanhood experience are at least as important as orphanhood itself. This point is echoed in a recent systematic review focusing on the relation between AIDS-related orphanhood and child mental health.21
Much of the work is cross sectional in nature and compares between orphans and nonorphans. Two recent studies were able to harness the power of panel data to take the analysis a step further. One study from South Africa followed orphans and nonorphans over 4 years and found that the mental health outcomes of orphans worsened over this period.22 A study from Ethiopia, by contrast, found that the negative short-term effects of preteenage maternal death disappeared by the age of 14–15.23
These longitudinal data sets have sharpened insights into the long-run impact of parental death on psychosocial outcomes. However, neither had a sample of nonorphaned children who could then be compared in adulthood to those who had suffered parental loss. This would address concern about preorphanhood confounders. To the best of our knowledge, our study is the first to be able to use that kind of set-up.
We analyzed a sample of 1108 children interviewed at baseline in 1991–1994 in the Kagera region of Tanzania as part of the Kagera Health and Development Survey (KHDS). This region was heavily affected by HIV early on in the epidemic. The baseline included nonorphaned children and by the endline, about 18 years later in 2010, 15% had lost their mother before the age of 19 and double that number had lost their father.
This study focused on self-esteem as the main outcome of interest. During the 2010 follow-up round, we measured the respondent's global self-esteem, which can be described as his or her overall sense of self-worthiness, self-acceptance, and self-respect.24,25 A recent systematic review of the empirical research on the mental health effects of orphanhood by AIDS-related causes categorized studies by outcomes and showed a concentration of studies looking at depression, anxiety, and conduct problems.21 Of all the studies cited above in this introduction, only 1 looked at self-esteem.23 Self-esteem is of interest as a measure of psychosocial well-being; it has been shown that those with positive self-esteem earn more and are more satisfied with life.26–29 Global self-esteem is related to 2 domains of the Big Five personality traits, neuroticism and extraversion. The 10-item Rosenberg Self-Esteem Scale (RSES)30–32 is a widely used measure of global self-esteem (among the alternatives are the Tennessee Self-concept Scale50,51 and the Beck Youth Inventories48).
A study that translated the items into 28 languages and administered it to 16,998 participants across 53 nations found that the internal reliability and factor structure of the Rosenberg scale were psychometrically sound across many languages and cultures, although there is a tendency for people from collectivist cultures to avoid extreme ends of the scale as well as some cross-cultural variability in responses to negatively formulated items.33 Thus, although self-esteem is only 1 aspect of a constellation of psychosocial well-being measures that may be affected by parental death—and indeed low self-esteem may coincide with other adverse personality traits—its importance for earnings and life satisfaction on the one hand and the pleasing psychometric properties of RSES on the other make it an important and feasible trait to study.
Our research strategy was to compare the 2010 RSES of children who became orphaned before the age of 19 to those who did not. Our study sample comprised people who were originally interviewed for the first time between 1991 and 1994 when young (between 0 and 16 years) and not orphaned, who were reinterviewed in 2010 as young adults (15–30 years old). Given the period of study and age of the respondents, we captured impacts of orphanhood in childhood on self-esteem in young adults. We further disaggregated results by whether the child was orphaned in preteenage or teenage years. In a second step, we controlled for a variety of background characteristics.
DATA AND METHODS
The KHDS is a publically available data set.34 The 2010 round of the KHDS attempted to reinterview 6353 individuals first interviewed in 1991–1994 and representative for the Kagera Region at that time. In 2010, all respondents between the ages of 15 and 30 years were eligible to be administered the 10-item Rosenberg Self-Esteem Scale (described below). If there were 2 or more household members ages 15–30, then the survey software on handheld devices randomly selected 1 eligible person per household to be administered the self-esteem questions.
We studied people originally interviewed for the first time between 1991 and 1994 when young (between 0 and 16 years) and not orphaned, who were reinterviewed in 2010 as young adults (15–30 years old). There are 1839 young persons who were not orphaned in the first interview. The 2010 interview teams traced and interviewed 1430 (78%) that are used for the analysis. Of the remaining, 233 (13%) were not found, 151 (8%) had died, and 13 (0.7%) refused to be interviewed. For 12 (0.7%), we could not determine their orphanhood status.
It is a mobile population. Although 94% of these individuals originated from rural villages, by 2010 they were found residing, on average, 158 km from their 1991–1994 locations with 65% still found in villages, but 25% in towns and 10% in cities.
Table 1 shows that of the 1430 children, nonorphaned at baseline and successfully interviewed in 2010, 15% had lost their mother, 30% their father, and 9% both parents before the age of 19 by 2010. The higher rates of paternal orphanhood are typical for the region and exist, despite a slightly higher prevalence of HIV among women in Tanzania—6.2% among women aged 15–49 compared with 3.8% among similarly aged men.36 This pattern has been attributed to sex patterns in age-specific mortality and the age gap between husbands and wives.35 Table 1 further breaks these numbers down by whether the parent died during the child's preteenage (0–12 years old) or teenage years (13–19 years old). The data do not allow us to ascertain the cause of death.
Of the 1430 successfully interviewed individuals, 322 were excluded from the analysis because they were not randomly selected to be administered the self-esteem module, giving a final analytical sample of 1108 people. Table 1 shows that orphanhood rates within this sample were near identical to those in the full sample.
The 10-item self-esteem questionnaire was a Swahili translation of the original Rosenberg30 questionnaire. We ensured that the Swahili wording accurately reflected the original meaning in English through a process of translation and back translation. We first inverted the 4-point Likert response codes of the negatively formulated items. We then standardized each item by subtracting the mean and dividing by the SD. Finally, we constructed the RSES as the average of the standardized items. RSES values range from −1.31 to 1.09, with an average of 0.02 and a SD of 0.38. For the analysis, we divided the RSES by its SD so that we can assess the differences in self-esteem in terms of SDs of RSES, which is more meaningful.
Table 2 shows that in terms of age, sex, weight, and height at baseline, the (future) maternal and paternal orphans in our sample were similar to those who do not lose a parent (minima, maxima and standard deviations are given in Supplemental Digital Content Table A1, http://links.lww.com/QAI/B75). However, paternal orphans came from households where the head was older and had lower education outcomes. At baseline, their households consumed about 1.5% less than others and they were 5.7% points less likely to have been living in a house with good flooring.
Maternal orphans had a different preorphanhood profile. The household in which they resided were consuming about 9% more per capita and were 6.4% points more likely to have a cement floor. Furthermore, maternal orphans had higher educated mothers and fathers.
Clearly, orphans came from somewhat different socioeconomic backgrounds compared with nonorphans, and these differences vary by the sex of the deceased parent. This is not surprising, given that the nonrandom distribution of HIV prevalence across sex and socioeconomic status in Tanzania and highlights the importance of taking into account the potential confounders when comparing orphans to nonorphans.36,37
We started our analysis by regressing SDs of RSES on dummies indicating whether the individual suffered paternal death or maternal death. These were 2 indicator variables when we considered any death at 0–19 years of age and 4 indicator variables when we divide parental death depending on whether it occurred during preteenage (0–12) or teenage (13–19) years. In a second regression model we did the same, but tightened the comparison by adding dummies for the origin village, sex, and age. In a final model, we further added the remaining preorphanhood characteristics from Table 2. This last regression is our preferred specification. These preorphanhood characteristics are included as controls to compare like for like in this observational setting. That is, we want to compare 2 children from similar households, where one of the children later becomes orphaned. These are not static traits. We measured these also in 2010, but we did not include the 2010 values in the regressions because these measures may be affected by the orphanhood experience (if, for example, lower self-esteem affects labor productivity and household income).
We also explored the interaction of parental death with (1) sex of the respondent, (2) the time that has passed since the death, (3) the death of the other parent (double orphanhood), and (4) 3 contemporaneous variables measuring years of schooling, body stature, and wealth.
Figure 1 shows the results from the 3 models described above (no controls, minimal controls, and full controls) for maternal and paternal deaths. The point estimates and CIs are presented in Supplemental Digital Content Table A2 (http://links.lww.com/QAI/B75). The average self-esteem score in adulthood did not differ in any statistically significant manner by maternal orphanhood, regardless of the specification. By contrast, paternal orphans' RSES was on average about 0.204 (95% CI: 0.059 to 0.348) SDs lower for those orphaned between the ages of 0–19 across the 3 alternate specifications. The fact that additional controls did not change the result suggests that cross-sectional data would produce the same findings as longitudinal surveys.
We introduced the age dimension to when a parent passes in Figure 2 and Supplemental Digital Content Table A3, (http://links.lww.com/QAI/B75), by looking at parental deaths when the young person was 0–12 years and when 13–19 years. Again, maternal deaths were not found to impact self-esteem. Paternal orphanhood in preteenage years is associated with a decline in RSES, and the magnitude is consistent across specifications. Referring to the third specification, children who suffer preteen (0–12 years) paternal loss scored 0.172 (95% CI: −0.007 to 0.350) SDs lower on RSES, whereas those who lost their father during teenage years score 0.236 (95% CI: 0.039 to 0.434) SDs lower. These findings mirror the direction of results on the same sample for a measure of locus of control (the extent to which someone believes what happens in life is beyond or in one's control), an adaptation of the Rotter internality/externality scale.38 These results are not presented here.
We interacted the orphan dummies with each other, with a sex dummy, and with the years between the interview and the parental death. The results, presented in Supplemental Digital Content Table A4, (http://links.lww.com/QAI/B75), show no sex differentiated effects (column 1), no multiplicative effects of double orphanhood (column 2), and no evidence that the effects depended on the time since bereavement (column 3).
In a further specification [Supplemental Digital Content Table A4 (http://links.lww.com/QAI/B75) column 4], we controlled for 2 contemporaneous human capital outcomes (the number of formal years of schooling and the attained height of the child in 2010) and 1 contemporaneous wealth indicator (household expenditures per capita in 2010). Including these separately or jointly did not change the coefficients on the orphanhood variables, which indicates that the correlation between paternal orphanhood and adverse self-esteem outcomes holds across young adults of a variety of education, health, and wealth outcomes.
The last Tanzania HIV/AIDS and Malaria Indicator Survey 2011–2012 puts HIV/AIDS prevalence among the 15–49-year-old population at 6.2% and showed that orphanhood is widespread in Tanzania.36 One of every 11 children (9%) younger than the age of 18 has lost either one or both parents and one of every 7 Tanzanian households (15%) hosts an orphan.
Our data come from the Kagera Region in the northwestern part of the country, bordering Uganda. The first 3 Tanzanian AIDS cases were detected in this region.39 Because Kagera has a much longer history of HIV/AIDS than most other areas, it now provides a unique window into intergenerational impacts, with the children of the victims of the early peak of the disease now transitioned into adulthood. Furthermore, the high HIV prevalence in the region prompted a number of data collection activities, which now provide the necessary baseline data to control for preorphanhood baseline confounders. Around the time of the baseline, the Kagera region had elevated prevalence rates of HIV, which were slowly decreasing. In 1990, 22.4% (95% CI: 20.6 to 25.2) of women going for antenatal care in the region's main hospital in Bukoba tested positive for HIV.40 The results should then be interpreted as informing on the plight of orphans in a highly HIV-affected area.
Orphans could suffer from lower self-esteem through a variety of pathways,10 and some studies suggest that AIDS orphans may suffer more severely from social stigma.22,41,42 But empirical research has shown that stigmatized individuals actively protect their self-esteem, such that there is no simple one-to-one inverse relation between social stigma and low self-esteem.43 This highlights the need to study these effects in the long run to assess their persistence.
The debate on orphanhood is often focused on maternal orphans, perhaps in part motivated by previous findings of the importance of maternal status with regard to child schooling and nutrition. Still, there are many good reasons to examine the role of the father. For example, a nascent literature in economics shows how father presence plays a key role in the intergenerational transmission of education and earnings in high-income countries.44–46 A much earlier example, was a line of research on black identity in the United States, initiated in the wake of the 1965 Moynihan Report, which was concerned with the possible contribution of elevated father absence in the black community to poor self-esteem.47
Our study points to the importance of paternal death, whether HIV related or not, for long-run self-esteem of older children. This complements other work that has highlighted the adverse effects of maternal death on children of younger ages, but mostly narrowly focuses on schooling.2–9 Thus, our work broadens the range of outcomes measuring well-being and highlights the role of the father for psychosocial outcomes, with teenage paternal orphans suffering lower self-esteem.
The strength of our study comes from the fact that we could observe children before orphanhood, which allows us to control for preorphanhood characteristics. The long, 18-year time frame gives us some confidence that these are long-term and persistent effects. The 2 most important limitations of our study are, first, that it is focused on just 1 region in Africa and it remains unclear whether the results can be generalized outside this context. Second, our study is focused only on self-esteem, while there may be a host of other outcomes that are relevant.
Our study builds the case for further development and evaluation of psychosocial support programs for children who lose their parents. Evaluations have been conducted of a peer-group support program for AIDS orphans in Uganda48 and a South African project that runs art and education activities to boost self-esteem, self-efficacy, and HIV insight for children affected by HIV.49 Both these programs were found to have positive effects on some psychosocial outcomes, but not on self-esteem. We do find positive effects on self-esteem documented in 2 evaluations of economic empowerment projects for orphans in Uganda.50,51
1. Wang H, Wolock TM, Carter A, et al. Estimates of global, regional, and national incidence, prevalence, and mortality of HIV
, 1980–2015: the Global Burden of Disease Study 2015. Lancet HIV
2. Ardington C, Leibbrandt M. Orphanhood and schooling in South Africa
: trends in the vulnerability of orphans
between 1993 and 2005. Econ Dev Cult Change. 2010;58:507.
3. Beegle K, De Weerdt J, Dercon S. The intergenerational impact of the African orphans
crisis: a cohort study from an HIV
/AIDS Affected Area. Int J Epidemiol. 2009;38:561–568.
4. Case A, Paxson C, Ableidinger J. Orphans
: parental death, poverty, and school enrollment. Demography. 2004;41:483–508.
5. de Walque D. Parental education and children's schooling outcomes: evidence from recomposed families in Rwanda. Econ Dev Cult Change. 2009;57:723–746.
6. Evans D, Miguel E. Orphans
and schooling in Africa
: a longitudinal analysis. Demography. 2007;44:35–57.
7. Senne JN. Death and schooling decisions over the short and long run in rural Madagascar. J Popul Econ. 2014;27:497–528.
8. Thomas KJ. Family contexts and schooling disruption among orphans
in post- genocide Rwanda. Popul Res Policy Rev. 2010;29:819–842.
9. World Bank. Education and HIV
/AIDS, A Window of Hope. Washington, DC: World Bank; 2002.
10. Makame V, Ani C, Grantham-McGregor S. Psychological well-being of orphans
in dar El Salaam, Tanzania. Acta Paediatr. 2002;91:459–465.
11. Foster G. Beyond education and food: psychosocial well-being
. Acta Paediatr. 2002;91:502–504.
12. Cluver L, Gardner F, Operario D. Psychological distress amongst AIDS-orphaned children in urban South Africa
. J Child Psychol Psychiatry. 2007;48:755–763.
13. Govender K, Reardon C, Quinlan T, et al. Children's psychosocial wellbeing in the context of HIV
/AIDS and poverty: a comparative investigation of orphaned and non- orphaned children living in South Africa
. BMC Public Health. 2014;14:615.
14. Reardon C, George G, Mucheuki C, et al. Psychosocial and health risk outcomes among orphans
in mixed households in KwaZulu-Natal, South Africa
. Afr J AIDS Res. 2015;14:323–331.
15. Atwine B, Cantor-Gaae E, Bjunirwe F. Psychological distress among AIDS orphans
in rural Uganda. Social Sci Med. 2005;61:555–564.
16. Nabunya P, Ssewamala F. The effect of parental loss on the psychosocial wellbeing of AIDS-orphaned children living in AIDS-impacted communities: does gender matter? Child Youth Serv Rev. 2014;43:131–137.
17. Bhargava A. AIDS epidemic and the psychological well-being and school participation of Ethiopian orphans
. Psychol Health Med. 2005;10:263–275.
18. Nyamukapa C, Gregson S, Lopman B, et al. HIV
-associated orphanhood and children's psychosocial distress: theoretical framework tested with data from Zimbabwe. Am J Public Health. 2008;98:133–141.
19. Ruiz-Casares M, Thombs B, Rousseau C. The association of single and double orphanhood with symptoms of depression among children and adolescents in Namibia. Eur Child Adolesc Psychiatry. 2009;18:369–376.
20. Tehetna A, Anna-Maija P, Raija-Leena P. Stigma, marginalization and psychosocial well-being
in Rwanda: exploring the mediation role of social support. AIDS Care. 2016;28:736–744.
21. Sharp C, Jardin C, Marais L, et al. Oprhanhood by AIDS-related causes and mental health: a developmental psychopathology approach. J HIV
22. Cluver L, Orkin M, Gardner F, et al. Persisting mental health problems among AIDS-orphaned children in South-Africa
. J Child Psychol Psychiatry. 2012;53:4363–4370.
23. Himaz R. Impact of parental death in middle childhood and Adolescense on child outcomes. J Afr Economies. 2013;22:463–490.
24. Rosenberg M. Conceiving the Self. Malabar, FL: Robert E. Krieger; 1979.
25. Wylie R. The Self-concept: Theory and Research on Selected Topics. Rev ed. Vol. 2. Lincoln, NE: University of Nebraska Press; 1979.
26. Bowles S, Gintis H, Osborne M. The determinants of earnings: a behavioral approach. J Econ Lit. 2001;39:137–176.
27. Almlund M, Duckworth A, Heckman J, et al. Personality psychology and economics. In: Hanushek E, Machin S, Woessman L, eds. Handbook of the Economics of Education. Vol. 4. Amsterdam, the Netherlands: Elsevier; 2011:1–181.
28. Drago F. Self-esteem
and earnings. J Econ Psychol. 2014;32:480–488.
29. Deiner E. Subjective well-being. Psychol Bull. 1984;95:542–575.
30. Rosenberg M. Society and the Adolescent Self-image. Princeton, NJ: Princeton University Press; 1965.
31. Byrne BM. Measuring Self-concept Across the Lifespan: Issues and Instrumentation. Washington, DC: American Psychological Association; 1996.
32. Wylie R. The Self-concept: Theory and Research on Selected Topics. Rev ed. Vol. 1. Lincoln, NE: University of Nebraska Press; 1974.
33. Shmitt D, Allik J. Simultaneous administration of the Rosenberg Self-Esteem
Scale in 53 nations: exploring the universal and culture-specific features of global self-esteem
. J Personal Social Psychol. 2005;89:623–642.
34. De Weerdt J, Beegle K, Bie Lilleør H, et al. Kagera health and development survey basic information document. 2012. Rockwool Foundation Study paper No. 46.
35. Ainsworth M, Filmer D. Inequalities in children's schooling: AIDS, orphanhood, poverty, and gender. World Development. 2006;34:1099–1128.
36. Tanzania Commission for AIDS (TACAIDS), Zanzibar AIDS Commission (ZAC), National Bureau of Statistics (NBS), Office of the Chief Government Statistician (OCGS), and ICF International. Tanzania HIV
/AIDS and Malaria Indicator Survey 2011–2012. Dar es Salaam, Tanzania: TACAIDS, ZAC, NBS, OCGS, and ICF International; 2013.
37. Msisha W, Kapiga S, Earls F, et al. Socioeconomic status and HIV
seroprevalence in Tanzania: a counterintuitive relationship. Int J Epidemiol. 2008;37:1297–1303.
38. Rotter J. Generalized expectancies for internal versus external control of reinforcement. Psychol Monogr. 1966;80:1–28.
39. Kwesigabo G, Killewo J, Urassa W, et al. HIV
-1 Infection prevalence and incidence trends in areas of contrasting levels of infection in the Kagera region, Tanzania, 1987–2000. J Acquir Immune Defic Syndr. 2005;40:585–591.
40. Kwesigabo G, Killewo JZ, Urassa W, et al. Monitoring of HIV
-1 infection prevalence and trends in the general population using pregnant women as a sentinel population: 9 years of experience form the Kagera region of Tanzania. J Acquir Immune Defic Syndr. 2002;23:410–417.
41. Cluver L, Orkin M. Cumulative risk and AIDS-orphanhood: interactions of stigma, bullying and poverty on child mental health in South Africa
. Social Sci Med. 2009;69:1186–1193.
42. Deacon H, Stephney I. HIV
/AIDS, Stigma and Children. Cape Town, South Africa
: HSRC Press; 2007.
43. Corcker J, Major B. Social stigma and self-esteem
: the self-protective properties of stigma. Psychol Rev. 1989;96:608–630.
44. Fertig AR. Is Intergenerational Earnings Mobility Affected by Divorce?” Center for Research on Child Wellbeing Working Paper 2002–2004. Princeton, NJ: Princeton University; 2007.
45. Gould ED, Simhon A. Does Quality Time Produce Quality Children? Evidence on the Intergenerational Transmission of Human Capital Using Parental Deaths. IZA Discussion Paper 5487. Bonn, Germany: Institute for the Study of Labor; 2011.
46. Kalili A, Mogstad M, Rege M, et al. Father presence and the intergenerational transmission of educational Attainment. J Hum Resour. 2016;51:869–899.
47. Porter J, Washington R. Black identity and self-esteem
: a review of studies of black self-concept, 1968–1978. Annu Rev Sociol. 1979;5:53–74.
48. Kumakech E, Cantor-Graae E, Maling S, et al. Peer-group support intervention improves the psychosocial well-being
of AIDS orphans
: cluster randomized control trial. Social Sci Med. 2009;68:1038–1043.
49. Muelle J, Alie C, Jonas B, et al. A quasi-experimental evaluation of a community- based art therapy intervention exploring the psychosocial health of children affected by HIV
. Trop Med Int Health. 2011;16:57–66.
50. Ssewamala F, Han C, Neilands T. Asset ownership and health and mental health functioning among AIDS-orphaned adolescents: findings from a randomized clinical trial in rural Uganda. Social Sci Med. 2009;69:191.
51. Ssewemala F, Karimili L, Torsten N, et al. Applying a family-level economic Strengthening intervention to improve education and health-related outcomes of school- going AIDS-orphaned children: lessons from a randomized experiment in Southern Uganda. Prev Sci. 2016;17:134–143.