Depressive Symptoms in Youth Heads of Household in Rwanda Correlates and Implications for Intervention

ARTICLE Depressive Symptoms in Youth Heads of Household in Rwanda Correlates and Implications for Intervention

Neil W. Boris, MD; Lisanne A. Brown, PhD; Tonya R. Thurman, PhD; Janet C. Rice, PhD; Leslie M. Snider, MD, MPH; Joseph Ntaganira, PhD; Laetitia N. Nyirazinyoye, MPH

Objective: To examine the level of depressive symp- depressive symptoms that exceeded the clinical cutoff were toms and their predictors in youth from one region of associated with having 3 basic household assets or fewer, Rwanda who function as heads of household (ie, those such as a mattress and a spare set of clothes (odds ratio responsible for caring for other children) and care for [OR], 1.69; 95% confidence interval [CI], 1.06-2.70), eat- younger orphans. ing less than 1 meal per day (OR, 1.68; 95% CI, 1.09- 2.60), reporting fair health (OR, 1.32; 95% CI, 0.76- Design: Cross-sectional survey 2.29) or poor health (OR, 2.33; 95% CI, 1.17-4.64), endorsing high levels of grief (OR, 2.67; 95% CI, 1.73- Setting: Four adjoining districts in Gigonkoro, an im- 4.13), having at least 1 parent die in the genocide as op- poverished rural province in southwestern Rwanda. posed to all other causes of parental death (OR, 1.83; 95% CI, 1.10-3.04), and not having a close friend (OR, 1.91; Participants: Trained interviewers met with the eldest 95% CI, 1.17-3.12). There was an interaction between member of each household (n=539) in which a youth marginalization from the community and alcohol use; 24 years old or younger was caring for 1 child or more. youth who were highly marginalized and did not drink alcohol were more than 3 times more likely to report Main Exposure: Serving as a youth head of house- symptoms of depression (OR, 3.07; 95% CI, 1.73-5.42). hold. When models were constructed by grouping theoretically related variables into blocks and controlling for other Main Outcome Measures: Rates and severity of de- blocks, the emotional status block of variables (grief and pressive symptoms using the Center for Epidemiologic marginalization) accounted for the most variance in de- Studies Depression scale; measures of grief, adult sup- pressive symptoms. port, social marginalization, and sociodemographic factors using scales developed for this study. Conclusions: Orphaned youth who head households in rural Rwanda face many challenges and report high rates Results: Of the 539 youth heads of household, 77% were of depressive symptoms. Interventions designed to go be- subsistence farmers and only 7% had attended school for yond improving food security and increasing house- 6 years or more. Almost half (44%) reported eating only hold assets may be needed to reduce social isolation of 1 meal a day in the last week, and 80% rated their health youth heads of household. The effect of head-of- as fair or poor. The mean score on the Center for Epi- household depressive symptoms on other children liv- demiologic Studies Depression scale was 24.4, exceed- ing in youth-headed households is unknown. ing the most conservative published cutoff score for adolescents. Multivariate analysis revealed that reports of Arch Pediatr Adolesc Med. 2008;162(9):836-843

Author Affiliations: Department of Psychiatry and N 2005, IT WAS ESTIMATED THAT overwhelmed, and many orphans either Neurology, Tulane University 290 000 children younger than 18 head households or live on the streets.4-7 School of Medicine (Dr Boris), years in Rwanda were orphans, The demographic makeup of such house- and Departments of one of the highest numbers holds and the psychological status of or- International Health and worldwide of children who have phans caring for other children are not well Development (Drs Brown, lost both parents.1 The combined effects characterized.8 Furthermore, the num- Thurman, and Snider), and I of the 1994 genocide, in which approxi- ber of youth-headed households (de- Biostatistics (Dr Rice), Tulane mately 10% of the population was killed, fined for this study as households headed University School of Public Health and Tropical Medicine, and the human immunodeficiency virus by unmarried individuals younger than 24 New Orleans, Louisiana; and (HIV) epidemic make Rwanda unique. years) will increase with time in areas af- 9-11 Rwanda School of Public Health Most African orphans have been ab- fected by HIV/AIDS. (Dr Ntaganira and sorbed into informal fostering systems.2,3 Depressive symptoms in heads of Ms Nyirazinyoye), Kigali. Such systems, however, are increasingly household are of particular concern be-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 836

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 cause of the association between poor developmental out- propriate instruments.20 Translation and backtranslation pre- comes for children and caregiver depression.12 Across cul- ceded review of survey items by a technical committee of youth tures, depression affects interactive patterns between the and professionals who attended to the context, cultural, and caregiver and the child, disturbing the development of linguistic relevance of the questions. The survey was then pre- children’s emotion regulation.13-15 Defining and measur- tested at a local street-children’s center and then piloted with 79 youth heads of households living adjacent to the survey area. ing depression across cultures, however, requires con- The Center for Epidemiologic Studies Depression (CES-D)21 sideration of how culture and language might affect symp- scale was incorporated into the survey with no substantive changes. 16 tom expression. Local terms equivalent to the CES-D 4-point scale were devel- Studies suggest that valid and reliable surveys for de- oped. The CES-D measures affective, somatic, and cognitive com- pression can be constructed using local language, though ponents of depression and has been widely used as a screening an extensive process of piloting, adapting, and translating tool, including in adolescents and high-risk groups.22,23 The Cron- western surveys for use in nonwestern settings, is neces- bach ␣ for the CES-D in this study was 0.86. sary.17-19 A primary objective of the present study was to Recent findings suggest that the distribution of CES-D scores is different for older adults than for adolescents and young examine the level of depressive symptoms and their pre- 24 dictors in youth from one region of Rwanda who function adults. For this analysis, we used the scale as continuous and also used a cutoff score of 24 to identify those youth with mod- as heads of household and care for younger orphans. erate to severe symptoms.25 The cutoff score of 24 improves case detection in this age group,26 although, to our knowl- METHODS edge, there are no published studies of youth from Rwanda. We developed several scales using our focus group and piloting to inform each scale. Principal factor analysis ensured SURVEY SITES AND SAMPLING PROCEDURES that each scale was unidimensional. Respondents indicated their level of agreement from “strongly agree” to “strongly dis- In 2004, a survey was conducted in the former Gigonkoro prov- agree,” with “don’t know/uncertain” scored in the middle. We ince, a rural province in southwestern Rwanda, selected be- dichotomized each scale at the median to facilitate the estima- cause it is particularly impoverished. Before the survey, insti- tion of odds ratios. tutional review board approval from the health sciences panel An adult support scale was formed from 4 items (Cron- at Tulane University and from a national review panel in Rwanda bach ␣, .85). Items included whether youth had an adult in their was received. Youth-headed households are common in south- life whom they could trust to give them advice and guidance western Rwanda, and in 2001, an international nongovern- or an adult who comforts them. mental organization had begun to target the more than 700 A marginalization scale was formed to assess the degree to households in this region that community leaders identified as which youth felt isolated from the surrounding community being headed by youth; the survey was designed to guide pro- (Cronbach ␣, .77).27 Respondents indicated their level of agree- gram expansion. To be included in this study, youth heads of ment with 6 items including whether they thought that no one household had to have little or no contact with their parents cared about them or that people in the community would rather (ie, be functioning as orphans). hurt than help them. During the course of the survey field work, 41 of the homes Focus group data confirmed that ongoing grief about pa- originally identified were found not classifiable as youth- rental loss was important. A 7-item grief scale (Cronbach ␣, headed households (ie, parent had returned, youth had been .66) was created for this study, including questions such as adopted, or other adult caregiver was present) and, thus, were whether youth often thought about their lost loved one(s) and ineligible; 115 of the households could not be located; most whether the loss of their loved one(s) had led to residual an- youth heads of household were reported to have moved for work, ger, loss of faith in God, or loss of confidence in others. Ͻ marriage, or boarding school. Only 2 eligible respondents ( 1%) We created 2 binary item scales. The first was an assets scale refused to participate. The final sample (N=539) included those that indicates whether the head of household owns a blanket aged 24 years or younger who were caring for 1 or more other and/or a spare set of clothes or shoes and whether the house- household members. hold has a latrine, a light source (oil lamp or better), or a mat- All youth heads of household were interviewed in their homes tress. The second was an exposure to maltreatment scale, de- for approximately 1 hour. A small bag of household items was veloped from our focus group data, that included having been provided as a token of appreciation to interviewees. Thirty- beaten, not being paid for work, and experiencing damage to eight youth younger than 25 years who reported attempting property or attempts to take land or possessions. We used item suicide or always thinking about suicide in the 2 months be- analysis to determine that the scales were unidimensional and fore the survey were immediately referred to local mental health dichotomized these scales using a median split. authorities for further evaluation; these youth were included Other variables empirically or theoretically linked to de- as interview participants. pressive symptoms were included in the analysis, including health status, cause of parental death, consumption of alco- INSTRUMENT AND SCALES hol, experience of forced sex, sexual risk-taking behaviors, and sociodemographic data. Health status consisted of self- The survey was designed to assist with measuring interven- reported health (excellent, good, fair, or poor). The parental tion effect on youth psychosocial functioning. Scales measur- death variable was coded yes if either parent was reported to ing symptoms of grief, depression, adult support, and margin- have died in the genocide. Sociodemographic variables in- alization from the community in addition to questions about cluded age, sex, assets scale, number of meals per day, area of demographic factors, general health and functioning, HIV knowl- residence, and educational achievement level. edge, vulnerability/protection, sexual behaviors, and risk behaviors were included. STATISTICAL ANALYSIS The development of the survey instrument was informed and refined by a multistep process detailed elsewhere.8 Focus Analyses were conducted using commercially available soft- groups and free listing with orphaned youth helped select ap- ware (STATA version 8; StataCorp LP, College Station, Texas).

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 837

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 teracted with contact with relatives in predicting depressive Table 1. Sociodemographic Characteristics symptoms. We included this interaction in all of the models of 539 Respondents that include the social support block of variables. We also included an interaction between marginalization and use of al- Variable No. (%) cohol. Psuedo-R2 estimates are based on the full main effects Sex models only. Male 267 (49.5) Female 272 (50.5) RESULTS Age, y Յ20 272 (50.5) Ն20 267 (49.5) SOCIODEMOGRAPHIC DATA District in former Gigonkoro region Karaba 157 (29.1) Table 1 gives the characteristics of the 539 households Nyamapabe 121 (22.5) and youth heads of households living with 1 or more other Mudosomwa 149 (27.6) household members. There were about equal numbers Nyraguru 112 (20.8) Educational achievement, primary school, y of males and females heading the households; their mean Յ3 234 (43.4) age was 20 years. Seventy-seven percent were subsis- Ͼ3 305 (56.6) tence farmers; 43% had attended primary school for 3 Household assets years or less, and only 7% had completed more than 6 Յ3 229 (42.5) years of primary school. Most respondents reported hav- Ͼ 3 310 (57.5) ing a blanket, a latrine, and a spare set of clothes; only Meals eaten, No./d 1 239 (44.3) 10.8% owned a mattress. Forty-four percent reported eat- 2 or 3 300 (55.7) ing only 1 meal per day, and almost 80% rated their health Health status as fair or poor. Excellent or good 107 (19.9) Youth in 71.4% of households reported that both par- Fair 317 (58.8) ents were dead, 26.2% reported that 1 parent was dead, Poor 115 (21.3) and 2.4% reported that both parents were alive. In cases No. of parents dead 0 13 (2.4) where 1 or both parents were alive, the parents were in 1 141 (26.2) jail or living elsewhere or their whereabouts or status were 2 385 (71.4) unknown. Parent(s) killed in genocide Almost one-fourth of the participants who reported No 411 (76.3) that a parent had died indicated that 1 or both of them Yes 128 (23.7) were killed in the genocide. Causes of death other than genocide included poisoning, and AIDS and other ill- nesses. It was impossible to verify youth reports of pa- The depressive symptoms score from the CES-D scale was the rental cause of death, and reports of poisoning repre- dependent variable. We decided a priori to group potential present a more socially accepted way of reporting death dictors of depressive symptoms into the following blocks: health resulting from AIDS. status, social support, cause of parental death, emotional status, exposure to maltreatment, lifestyle behaviors, and sociodemographic data. Social support included the adult support EMOTIONAL WELL-BEING scale plus a question about contact with relatives and a ques- AND RISK BEHAVIORS tion about whether the head of household believes that he or she has a close friend. Emotional status included the grief and Grief clearly had a role in how youth viewed the world marginalization scales. Exposure to maltreatment included ex- around them; more than half had high levels of grief perience of forced sex added to positive responses on the mal- (Table 2). For example, 64% of participants stated that treatment scale. Lifestyle behaviors included ever having had they had lost confidence in people, and more than 40% en- sex and drinking alcohol. dorsed feeling that life was meaningless or that they had After computing frequency distributions, we explored the lost faith in God since the death(s) of their parent(s). bivariate association of each predictor with depression using ␹2 tests. Multivariate logistic regression was conducted to es- The youth in this sample reported feeling stigma- timate odds ratios, 95% confidence intervals, and P values. tized and socially isolated. For example, 76% of partici- To determine the amount of variance in depression that pants agreed or strongly agreed with the statement that blocks of variables explain and to develop a multivariate model, the community rejects orphans. Only 26% strongly agreed we calculated pseudo-R2 using the MacKelvey and Zavoina mea- that they had at least 1 friend. Nevertheless, the youth sure.28 We determined the amount of variance in depression were not completely without support; 66% named 3 that each block of variables explained alone and after control- sources of adult support, and most of the respondents ling for health status, cause of parental death, and sociodemo- (82%) reported having contact with relatives. graphic data. Most youth (70.5%) reported some form of maltreat- We hypothesized that the relationships between the pre- ment such as being beaten; experiencing attempts to have dictors and depression might differ by sex and initially ran the regression models separately by sex. Sex differences were, how- land or possessions taken, or property damage; and not ever, minimal, and we subsequently combined the sample. Nev- being paid for work. Risk behavior was common, with ertheless, we tested for significant interaction between each pre- 32% reporting ever having had sex and 49% reporting dictor and sex in the bivariate models and in the final model. alcohol use. Fourteen percent of the sample (n=63 fe- There was only 1 significant interaction involving sex: sex in- males and 13 males) reported experiencing forced sex.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 838

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Depressive symptoms were common (Table 2). The mean depression score was 24.4, which is above the most Table 2. Emotional Well-being and Risk Behaviors conservative published CES-D cutoff score of 24 for ado- of 539 Respondents lescents.24 Put another way, 53% of the sample screened positive for depression, though the range of CES-D scores Variable No. (%) (0-55) was considerable. Grief level Low (Յ3) 234 (43.4) High (Ͼ3) 305 (56.6) PREDICTORS OF DEPRESSIVE SYMPTOMS Marginalization from community Low (Յ3) 236 (43.8) At univariate analysis, we found several variables to be High (Ͼ3) 303 (56.2) statistically associated with depressive symptoms Adult support Low (Յ3) 185 (34.3) (Table 3). Multivariate analysis was used to further ex- High (Ͼ3) 354 (65.7) plore these findings, including possible interactions. Contact with relatives For the regression models, we grouped predictors as No 97 (18.0) follows: health status, social support, demographic data, Yes 442 (82.0) cause of parental death, emotional status (grief and mar- Have at least 1 friend ginalization from the community), experience of mal- No 401 (74.4) treatment, and lifestyle (ever had sex or used alcohol). Yes 138 (25.6) Maltreatment experience The groups were entered in the model one at a time. No 159 (29.5) Table 3 also gives the adjusted odds ratios from the final Yes 380 (70.5) regression model. Ever had sex Most of the variables significantly associated with de- No 365 (67.7) pressive symptoms at univariate analysis remained sig- Yes 174 (32.3) nificant in the regression model. Youth with 3 house- Ever had forced sex No 463 (85.9) hold assets or fewer were more likely to score above the Yes 76 (14.1) cutoff for depression, as were those reporting eating fewer Use alcohol than 1 meal per day. Youth with fair or poor health were No 274 (50.8) more likely to be depressed, though the relationship be- Yes 265 (49.2) tween health status and depressive symptoms was some- Depressive symptoms (CES-D cutoff, 24) what diminished in the final model. Յ24 253 (46.9) Ͼ High levels of grief were significantly associated with 24 286 (53.1) depressive symptoms above the screening cutoff, and those Abbreviation: CES-D, Center for Epidemiologic Studies Depression. who reported that 1 or both parents had died in the genocide (as compared with all other causes of parental death) were almost twice as likely to score above the screening school in the primary (pre–high school) years, most ex- cutoff as well. Those who said they did not have a close perience food insecurity as subsistence farmers, and many friend were more likely to meet the criteria for depression. report poor health, endorse feeling stigmatized, and are Two interaction terms were included in the model: see- socially isolated. More than 70% report instances of mal- ing relatives by sex and marginalization by alcohol use. treatment by others in the community, and 1 in 6 fe- Though the interaction of sex and depressive symptoms male respondents report forced sex. Grief caused by pa- did not reach statistical significance, male respondents rental loss further complicates the risk profile for most who did not have contact with their relatives were more respondents. likely to report high levels of depressive symptoms, More than half of the 539 youth we interviewed were whereas for female respondents, not having contact with above the most conservative age-specific screening cut- relatives was protective. Among youth reporting high lev- off on the CES-D, a score in other cultures consistent with 25 els of marginalization from the community, those who moderate to severe symptoms. Our data are consonant did not drink alcohol were more than 3 times more likely with our expectations about youth who head house- to report symptoms of depression, which suggests that holds in postgenocide Rwanda. A 1999 study con- drinking is protective for marginalized youth. In conducted in Rwanda revealed that 15% of a representative trast, there was no significant relationship between de- group of postgenocide adults met strict criteria for de- 18 pressive symptoms and alcohol use for those reporting pression using a locally developed and tested survey, low marginalization. In sum, having few household as- though more representative surveys of adolescents and sets and little food, reporting fair or poor health, having young adults are unavailable. high levels of grief and marginalization, having lost a par- Orphan status in Rwanda, as in many parts of Africa, ent during the genocide, and having no close friend pre- is a marker for risk conditions associated with depres- dicted depressive symptoms in the regression model. sion. In Uganda, for example, orphan status was the only significant predictor of depressive symptoms in youth aged 11 to 14 years, with orphans more than 6 times as likely COMMENT to endorse high levels of depressive symptoms than nonorphaned children.29 Orphan status also independently pre- Rwandan youth aged 12 to 24 years who care for younger dicted internalizing symptoms in Tanzanian children.30 children have severe challenges. Most dropped out of Orphan status remained associated with depression

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 839

Variable Depressed Respondents, % Unadjusted OR (95% CI) Adjusted OR (95% CI) Sex Male 49.1 1 [Reference] . . . Female 57.0 1.38 (0.98-1.93) . . . Contact with relatives Yes 50.5 1 [Reference] . . . No 65.0 1.82 (1.15-2.88) . . . Male Yes ...... 1 [Reference] No ...... 2.17 (0.84-5.63) Female Yes ...... 1 [Reference] No ...... 0.55 (0.27-1.13) Age, y Յ20 50.4 1 [Reference] 1 [Reference] Ͼ20 55.8 1.24 (0.89-1.75) 1.19 (0.78-1.81) District Nyaraguru 56.3 1 [Reference] 1 [Reference] Karaba 55.4 0.97 (0.59-1.58) 0.77 (0.42-1.39) Nyamapabe 59.4 1.56 (0.73-2.06) 0.94 (0.50-1.77) Mudosomwa 41.6 0.55 (0.34-.91) 0.63 (0.35-1.14) Educational achievement, primary school, y Յ3 50.4 1 [Reference] 1 [Reference] Ͼ3 55.1 1.21 (0.86-1.69) 1.35 (0.88-2.09) No. of household assets Յ3 59.4 1.56 (1.10-2.22) 1.69 (1.06-2.70) Ͼ3 48.4 1 [Reference] 1 [Reference] Meals eaten, No./d Յ1 63.6 2.17 (1.54-3.03) 1.68 (1.09-2.60) 2 or 3 44.7 1 [Reference] 1 [Reference] Health status Excellent or good 35.5 1 [Reference] 1 [Reference] Fair 51.7 1.95 (1.24-3.06) 1.32 (0.76-2.29) Poor 73.0 4.92 (2.78-8.72) 2.33 (1.17-4.64) Parent(s) killed in genocide No 48.9 1 [Reference] 1 [Reference] Yes 66.4 2.06 (1.36-3.13) 1.83 (1.10-3.04) Grief level Low (Յ3) 36.3 1 [Reference] 1 [Reference] High (Ͼ3) 65.9 3.39 (2.37-4.84) 2.67 (1.73-4.13) Marginalization from community Low (Յ3) 32.2 1 [Reference] . . . High (Ͼ3) 69.3 4.75 (3.30-6.86) . . . Adult support Low (Յ3) 69.2 2.78 (1.92-4.00) . . . High (Ͼ3) 44.6 1 [Reference] . . . Contact with relatives Yes 50.5 1 [Reference] . . . No 65.0 1.82 (1.15-2.88) . . . Male Yes 45.5 ...... No 72.2 3.12 (1.44-6.77) . . . Female Yes 55.9 ...... No 60.7 1.22 (0.68-2.18) . . . Have at least 1 friend Yes 39.1 1 [Reference] . . . No 57.9 2.14 (1.44-3.17) 1.91 (1.17-3.12) Maltreatment experience No 37.7 1 [Reference] . . . Yes 59.5 2.42 (1.66-3.54) 1.46 (0.91-2.34) Ever had sex No 51.2 1 [Reference] . . . Yes 56.9 1.26 (0.87-1.81) 1.43 (0.86-2.37) Ever had forced sex No 51.6 1 [Reference] . . . Yes 61.8 1.52 (0.92-2.50) 0.76 (0.38-1.54) Use alcohol No 62.0 2.08 (1.49-2.94) . . . Yes 43.8 1 [Reference] . . . Low marginalization Use alcohol 29.6 1 [Reference] 1 [Reference] Do not use alcohol 35.6 1.32 (0.76-2.28) 3.07 (1.73-5.42) High marginalization Use alcohol 57.9 1 [Reference] 1 [Reference] Do not use alcohol 78.2 2.61 (1.58-4.31) 1.05 (0.56-1.98)

Abbreviations: CI, confidence interval; OR, odds ratio; ellipses, not applicable.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 840

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 after controlling for other factors in a large study com- reported high levels of marginalization and no alcohol paring orphans with other vulnerable children in Zim- use also reported high levels of depressive symptoms. babwe.31 In sum, compared with other disadvantaged Much of the available alcohol in rural Rwanda is locally- children in Africa, orphans are particularly likely to re- produced and consumed in village celebrations and cer- port symptoms of anxiety and depression.32 emonies; orphans with few resources are unlikely to be The predictors of depressive symptoms in the or- able to afford to drink unless they are included in such phaned youth we interviewed were complex. We ex- events. The most marginalized youth may be those most pected health status to be associated with depressive likely to be excluded from village events, which would symptoms; in the final model, those with fair or poor mean that these youth would have few opportunities to health reported higher levels of depressive symptoms. drink. Social alcohol use in Rwanda is a proxy for social Our finding linking depressive symptoms and having inclusion. We had only 1 general question about alco- only 1 meal per day or having few household assets is hol use, however, and so captured both common social less well established. Hunger has, however, been linked use and more uncommon problem use. These margin- to depression and anxiety in children in less impover- alized youth may also either avoid social contact (com- ished settings,33 and going to bed hungry was associated mon when depressed) or be shunned and, therefore, ex- with depressive symptoms in orphaned children in perience depressive symptoms. Furthermore, depressed Tanzania.30 youth may drink to self-medicate, and more data on the In our sample, grief and social marginalization ac- relationship between alcohol use and depressive symp- counted for the largest percentage of the variance in de- toms in similar contexts are necessary. pressive symptoms both when sociodemographic data Neither sex nor maltreatment experiences were signifi- were controlled for and when they were not. Grief is con- cantly associated with reporting high levels of depressive founded with orphan status. Our data suggest that many symptoms. It may be that males, who are culturally less likely orphans experience ongoing grief symptoms years after to have to care for other children, experience such respon- the loss of loved ones. In the Rwandan context, unre- sibility as highly stressful. As for maltreatment, our sur- solved grief may be linked to exposure to trauma, which vey included questions ranging from more common ha- interferes with the process of grieving. For Rwandan or- rassment (eg, threats to steal property and not being paid phans, traumatic grief, marked by an admixture of grief for work) to being beaten to uncommon experiences such and posttraumatic symptoms,34 is likely prevalent. Symp- as sexual assault. Relative to grief, poor health, lack of food, toms such as intrusive re-experiencing and hyper- and not having a close friend, however, it seems that these arousal make it difficult for the individual with trau- kinds of experiences do not significantly contribute to de- matic grief to traverse the typical grieving process.35 One pressive symptoms in the harsh context in which these clue that the interplay between trauma and grief was im- youth heads of household live. portant in this sample is that youth who reported paren- Our survey captured a significant amount of the vari- tal death during the genocide were almost twice as likely ance in depressive symptoms among youth heads of to report more depressive symptoms, even when account- household; other factors, however, seem to have a role, ing for other factors. Children may have particular prob- and both more qualitative and survey research is neces- lems making sense of violent parental death.36,37 Be- sary. We focused on youth who cared for younger chil- cause our survey was conducted just before the 10th dren (primarily siblings and offspring) because of data anniversary of the genocide, for which many communi- suggesting that caregiver depressive symptoms have im- ties in Rwanda had large-scale memorial services, it is pos- portant negative developmental implications for those sible that the timing of the survey may have elicited an- children.12 Our data suggest that younger children liv- niversary reactions and heightened grief, especially in ing in homes headed by orphaned youth are at grave risk. youth who lost a parent during the genocide. There is Potential genetic effects (most of the heads of house- evidence that distressed individuals flood local clinics dur- hold are first-degree relatives) may compound the effect ing April when yearly public memorials occur (J.N.; per- of the interactive effects of being cared for by a person sonal observation; February 7, 2007). with depressive symptoms.14 Given the psychosocial con- In addition to grief, not having a close friend was text, it seems unlikely that depressive symptoms would strongly associated with depressive symptoms in this remit (at least not without recurring) in most of those study. The link between social support and depression already affected. Furthermore, the kind of social mar- in collectivist societies is important,38 and orphaned ado- ginalization the heads of household endorsed might also lescents and young adults are particularly affected by so- directly affect younger members of the family. cial isolation.39 The few comparative studies in Africa sug- There are several important limitations to this study. gest that orphan status is a predictor of perceptions of Our sample cannot be considered representative of or- marginalization; orphans tend to report either fewer or phaned youth in other areas or countries. Compound- more negative interactions with both peers40 and adults.41,42 ing economic and social risks combined with high lev- Given the cross-sectional nature of our data, however, it els of exposure to trauma make it difficult to generalize is not possible to determine whether not having a close from data gathered in postgenocide Rwanda. Further- friend precipitated depressive symptoms or resulted from more, though youth-headed households are an increas- them. ingly common phenomenon as the HIV/AIDS epidemic When we looked beyond close peer relationships, we results in more children being orphaned, both broad cul- found that perceptions of marginalization from the wider tural factors and unique local factors influence how or- community interacted with alcohol use. Those youth who phans are or are not absorbed into communities.8,9

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 841

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 The cross-sectional nature of our data is another im- tistical analysis: Rice. Obtained funding: Boris, Brown, and portant limitation. Depressive symptoms are important, Snider. Administrative, technical, and material support: but whether, for example, they result from marginaliza- Boris, Brown, Thurman, Rice, Snider, Ntaganira, and tion or make it more likely for youth to perceive that they Nyirazinyoye. Study supervision: Brown, Snider, Ntaganira, are marginalized cannot be determined. Our reliance on and Nyirazinyoye. self-report further complicates matters. Recall bias may Financial Disclosure: None reported. have increased the likelihood that respondents en- Funding/Support: This study was supported by the Popu- dorsed grief, poor health, depressive symptoms, or mar- lation Council Horizons Program, which was funded by ginalization. On the other hand, it has been argued that the US Agency for International Development. the cultural proscription on open displays of emotion in Role of the Sponsor: The funder had a role in the de- Rwanda might actually limit endorsement of symptoms sign and conduct of the study; collection, management, of psychological distress among interviewees.43 analysis, and interpretation of the data; and prepara- Another major limitation of this study is the lack of a tion, review, and approval of the manuscript. comparison group. Without a comparison group, we can- Additional Contributions: Our partners at World Vi- not be sure whether nonorphaned youth living in this sion International, especially Eleazar Mugirira, were es- social context are any less distressed. A challenge for re- pecially helpful in data gathering. We thank the youth search with orphans is to identify a group that is com- who graciously participated in the initial focus groups parable. There is mounting evidence to suggest that or- that informed our survey and in the survey itself. phans are uniquely at risk for adverse outcomes. Several of the scales we used in our survey (those cap- turing grief, degree of adult support, and perceived mar- REFERENCES ginalization) were developed and piloted for this study and, thus, need additional validation in other popula- 1. United Nation’s Childrens’ Fund. Africa’s Orphaned and Vulnerable Children: Chil- ␣ dren Affected by AIDS. New York, NY: United Nation’s Childrens’ Fund; 2006. tions. Some of the scales had a Cronbach reliability of 2. Masmas TN, Jensen H, da Silva D, Høj L, Sandström A, Aaby P. The social situ- less than 0.80, and refinement of these scales in other set- ation of motherless children in rural and urban areas of Guinea-Bissau. Soc Sci tings is indicated. Med. 2004;59(6):1231-1239. 3. Monasch R, Boerma JT. Orphanhood and childcare patterns in sub-Saharan Africa: an analysis of national surveys from 40 countries. AIDS. 2004;18(suppl 2): CONCLUSIONS S55-S65. 4. Deininger K, Garcia M, Subbarao K. AIDS-induced orphanhood as a systemic Interviews of a large sample of impoverished and or- shock: magnitude, impact and program interventions in Africa. World Dev. 2003; 31(7):1201-1220. doi:10.1016/S0305-750X(03)00061-5. phaned youth younger than 24 years who care for de- 5. Madhavan S. Fosterage patterns in the age of AIDS: continuity and change. Soc pendents in rural Rwanda revealed high levels of depres- Sci Med. 2004;58(7):1443-1454. sive symptoms. Hunger, grief, few assets, poor health 6. Veale A, Quigley P, Ndibeshye T, Nyirimihigo C. Struggling to survive: orphan status, and indices of social marginalization were asso- and community dependent children in Rwanda. Rwanda Ministry of Local Gov- ernment and Social Affairs and UNICEF report, 2001. http://www.unicef.org ciated with more depressive symptoms in this sample. /evaldatabase/files/RWA_2000_001.pdf. Accessed March 19, 2007. Ten years after the Rwandan genocide and in the midst 7. Watts H, Lopman B, Nyamukapa C, Gregson S. Rising incidence and prevalence of the HIV/AIDS epidemic, the effects of poverty and so- of orphanhood in Manicaland, Zimbabwe, 1998 to 2003. AIDS. 2005;19(7): cial disruption on the most vulnerable youth in Rwanda 717-725. are evident. The effect of caregiver depression on younger 8. Boris NW, Thurman TR, Snider L, et al. Infants and young children living in youth- headed households in Rwanda: implications of emerging data. Infant Ment Health children living in youth-headed households is not yet J. 2006;27(6):584-602. doi:10.1002/imhj.20116. known. Further study of orphans and vulnerable chil- 9. Oleke C, Blystad A, Rekdal OB. “When the obvious brother is not there”: political dren in countries such as Rwanda, in particular, studies and cultural contexts of the orphan challenge in northern Uganda. Soc Sci Med. that inform large-scale interventions, are necessary if the 2005;61(12):2628-2638. 10. Heuveline P. Impact of the HIV epidemic on population and household struc- next generation of youth is to thrive. ture: the dynamics and evidence to date. AIDS. 2004;18(suppl 2):S45-S53. 11. Foster G. The capacity of the extended family safety net for orphans in Africa. Psychol Health Med. 2000;5(1):55-63. doi:10.1080/135485000106007. Accepted for Publication: February 1, 2008. 12. Goodman SH, Gotlib IH. Risk for psychopathology in the children of depressed Correspondence: Neil W. Boris, MD, Department of Psy- mothers: a developmental model for understanding mechanisms of transmission. chiatry and Neurology, Tulane University School of Medi- Psychol Rev. 1999;106(3):458-490. cine, 1440 Canal St, Box TB-52, New Orleans, LA 70112 13. Beardslee WR, Gladstone TR, Wright EJ, Cooper AB. A family-based approach to the prevention of depressive symptoms in children at risk: evidence of paren- ([email protected]). tal and child change. Pediatrics. 2003;112(2):e119-e131. Author Contributions: Dr Brown had full access to all 14. Kim-Cohen J, Moffitt TE, Taylor A, Pawlby SJ, Caspi A. Maternal depression and of the data in the study and takes responsibility for the children’s antisocial behavior: nature and nurture effects. Arch Gen Psychiatry. integrity of the data and the accuracy of the data analy- 2005;62(2):173-181. sis. Study concept and design: Boris, Brown, Thurman, 15. Richter LM. The Importance of Caregiver-Child Interactions for the Survival and Healthy Development of Young Children: A Review. Geneva, Switzerland: World Snider, and Ntaganira. Acquisition of data: Boris, Brown, Health Organization; 2004. Thurman, Snider, Ntaganira, and Nyirazinyoye. Analy- 16. Bolton P. Cross-cultural validity and reliability testing of a standard psychiatric sis and interpretation of data: Boris, Brown, Thurman, Rice, assessment instrument without a gold standard. J Nerv Ment Dis. 2001;189 Snider, Ntaganira, and Nyirazinyoye. Drafting of the manu- (4):238-242. 17. Bolton P, Neugebauer R, Ndogoni L. Prevalence of depression in rural Rwanda based script: Boris, Brown, and Snider. Critical revision of the on symptom and functional criteria. J Nerv Ment Dis. 2002;190(9):631-637. manuscript for important intellectual content: Boris, Brown, 18. Bolton P. Local perceptions of the mental health effects of the Rwandan genocide. Thurman, Rice, Snider, Ntaganira, and Nyirazinyoye. Sta- J Nerv Ment Dis. 2001;189(4):243-248.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 842

©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 19. Saint Arnault D, Sakamoto S, Moriwaki A. The association between negative self- 32. Snider L, Dawes A. Psychosocial vulnerability and resilience measures for national- descriptions and depressive symptomology: does culture make a difference? Arch level monitoring of orphans and other vulnerable children: recommendations for Psychiatr Nurs. 2005;19(2):93-100. the revision of the UNICEF psychological indicator. http://www.childinfo.org/files 20. Bolton P, Tang AM. An alternative approach to cross-cultural function assessment. /PsychosocialMeasures_Final06.pdf. Accessed June 10, 2008. Soc Psychiatry Psychiatr Epidemiol. 2002;37(11):537-543. 33. Weinreb L, Wehler C, Perloff J, et al. Hunger: its impact on children’s health and 21. Radloff LS. The CES-D scale: a self-report depression scale for research in mentalhealth.Pediatrics.2002;110(4):e41-e50.http://pediatrics.aappublications.org the general population. Appl Psychol Meas. 1977;1(3):385-401. doi:10.1177 /cgi/content/full/110/4/e41. Accessed February 7, 2007. /014662167700100306. 34. Jacobs S, Mazure C, Prigerson H. Diagnostic criteria for traumatic grief. Death 22. Colletta ND. At risk for depression: a study of young mothers. J Genet Psychol. Stud. 2000;24(3):185-199. 1983;142(second half):301-310. 35. Cohen JA, Mannarino AP, Knudsen K. Treating childhood traumatic grief: a pilot 23. Lyketsos CG, Hoover DR, Guccione M, et al. Changes in depressive symptoms study. J Am Acad Child Adolesc Psychiatry. 2004;43(10):1225-1233. as AIDS develops. The Multicenter AIDS Cohort Study. Am J Psychiatry. 1996; 36. Currier JM, Holland JM, Neimeyer RA. Sense-making, grief, and the experience 153(11):1430-1437. of violent loss: toward a mediational model. Death Stud. 2006;30(5): 24. Chabrol H, Rodgers R, Rousseau A. Relations between suicidal ideation and di- 403-428. mensions of depressive symptoms in high-school students. J Adolesc. 2007; 37. Vigil GJ, Clements PT. Child and adolescent homicide survivors. Complicated 30(4):587-600. grief and altered worldviews. J Psychosoc Nurs Ment Health Serv. 2003;41 25. Roberts RE, Lewinsohn PM, Seeley JR. Screening for adolescent depression: a (1):30-39. comparison of depression scales. J Am Acad Child Adolesc Psychiatry. 1991; 38. Pretorius TB. Problem-solving appraisal in the association of life stress and de- 30(1):58-66. doi:10.1097/00004583-199101000-00009. pression: a South African study. Psychol Rep. 1992;71(3, pt 1):855-862. 26. Rushton JL, Forcier M, Schectman RM. Epidemiology of depressive symptoms in the National Longitudinal Study of Adolescent Health. J Am Acad Child Ado- 39. Wild L, Flisher A, Laas S, Robertson B. The psychosocial adjustment of adoles- lesc Psychiatry. 2002;41(2):199-205. cents orphaned in the context of HIV/AIDS. J Child Adolesc Ment Health. In press. 27. Thurman TR, Snider L, Boris N, et al. Psychosocial support and marginalization 40. Cluver L, Gardner F. The psychological well-being of children orphaned by AIDS of youth-headed households in Rwanda. AIDS Care. 2006;18(3):220-229. in Cape Town, South Africa. Ann Gen Psychiatry. 2006;5:8-17. 28. Mittlböck M, Schemper M. Explained variation for logistic regression. Stat Med. 41. Manuel P. Assessment of orphans and their caregivers’ psychological well- 1996;15(19):1987-1997. being in a rural community in central Mozambique [master’s thesis]. London, 29. Atwine B, Cantor-Graae E, Bajunirwe F. Psychological distress among AIDS or- England: Institute of Child Health; 2002. phans in rural Uganda. Soc Sci Med. 2005;61(3):555-564. 42. Gilborn L, Apicella L, Brakarsh J, et al. Orphans and Vulnerable Youth in Bula- 30. Makame V, Ani C, Grantham-McGregor S. Psychological well-being of orphans wayo, Zimbabwe: an Exploratory Study of Psychosocial Well-being and Psycho- in Dar El Salaam, Tanzania. Acta Paediatr. 2002;91(4):459-465. social Support Programs. Washington, DC: Population Council; 2006. 31. Nyamukapa C, Gregson S. Extended family’s and women’s roles in safeguard- 43. Pham PN, Weinstein HM, Longman T. Trauma and PTSD symptoms in Rwanda: ing orphans’ education in AIDS-afflicted rural Zimbabwe. Soc Sci Med. 2005; implications for attitudes toward justice and reconciliation. JAMA. 2004;292 60(10):2155-2167. (5):602-612.

Announcement

Sign Up for Alerts—It’s Free! Archives of Pediatrics & Adolescent Medicine offers the ability to automatically re- ceive the table of contents of Archives via e-mail when it is published online. This also allows you to link to individual articles and view the abstract. It makes keep- ing up-to-date even easier! Go to http://pubs.ama-assn .org/misc/alerts.dtl to sign up for this free service.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 162 (NO. 9), SEP 2008 WWW.ARCHPEDIATRICS.COM 843