Family Contextual Factors are Differentially Associated with Depressive Symptoms among Boys and Girls with Perinatally Acquired HIV

Abstract

Youth with perinatally acquired HIV (PHIV) are at risk for depressive symptoms, which are associated with a range of adverse outcomes. Although family contextual factors associated with depressive symptoms differ among boys and girls without PHIV, it is unclear whether this is also the case among youth with PHIV. Participants included 314 youth with PHIV (M = 12.88, SD = 3.08 years old; 51% male; 85% Black/Latinx) and their caregivers. Higher levels of caregivers’ own depressive symptoms, caregiver-child detachment, and family conflict were associated with higher levels of caregiver-reported youth depressive symptoms. Less consistent discipline was associated with higher levels of youth-reported depressive symptoms. Higher youth-reported depressive symptoms were associated with greater family cohesion among boys and greater caregiver detachment among girls. Consideration of contextual variables is essential for interventions for depressive symptoms among youth with PHIV, but attention to sex differences with family contextual factors is also important.

INTRODUCTION

Advances in antiretroviral therapies have led to increased overall health and survival rates among youth with perinatally acquired human immunodeficiency virus (PHIV); nevertheless, concerns regarding mental health outcomes among these youth have grown [1, 2]. Issues regarding biological risk associated with HIV have been raised, given the virus’ crossing of the blood-brain barrier and early seeding of viral reservoirs [3], associated inflammation [4], and potentially neurotoxic nature of antiretroviral medications [5]. Youth with PHIV are also more likely to experience family-related stressors (e.g., parental illness or death, transitions among caregivers, responsibility for care of other family members) [2, 6]. Unique psychosocial issues include frequent visits to doctors, delayed puberty, missed academic and social opportunities, and potential mortality [2, 7]. These experiences may influence the timing and achievement of normative developmental milestones [8], such as school completion, employment, relationships, and independent daily functioning, all of which may contribute to the elevated risk for depressive symptoms among youth with PHIV [2, 9]. Although youth with PHIV are more likely to receive behavioral or pharmacological interventions than youth without PHIV, many do not receive active treatment [4, 10]. This increased risk is particularly worrisome given that psychiatric problems among youth are generally associated with academic and social impairment, unsafe sex, and medication nonadherence [1, 11].

Youth mental health is influenced by multiple risk factors in the home that entail various levels of family processes, including caregiver psychological functioning, caregiver-child interactions, and family-level factors [8, 10, 12, 13], which can be jointly considered as “family contextual factors.” These family contextual factors are consistently associated with depressive symptoms among youth without PHIV [14, 15] and youth with HIV positive mothers [16, 17]. Nevertheless, there is a dearth of research concurrently considering family contextual factors or examining the relations among these factors and youth depressive symptoms among families of youth with PHIV. Caregivers of youth with PHIV may experience increased stress as a function of their own health, as well as concerns about the well-being of youth in their care, their financial status, disclosure of the youth’s HIV status, and the desire to normalize experiences associated with PHIV and shield youth from stigma and discrimination [18]. All of these stressors may affect caregiver-child interactions and discipline strategies, family functioning, and caregiver mental health [19, 20]. In addition, parental HIV infection is associated with poverty, substance use, and parental psychiatric symptoms, which may further contribute to problematic parent-child interactions [10]. Few studies have focused on protective factors for mental health among youth with PHIV despite the possibility that facing these challenges as a family may result in more attentive parenting and family cohesion [18] and contribute to greater communication and caregiver involvement [2, 9], all of which are associated with positive outcomes among youth with PHIV. Although evidence-based interventions for improving mental health among youth with PHIV are limited [2, 19], existing findings support the need for integration of mental health services to address youth psychological symptoms and contextual processes into PHIV treatment and care. Moreover, a fuller understanding of family contextual factors that may distinguish youth with PHIV most vulnerable to developing depressive symptoms may be critical to such intervention efforts.

It is well-established that there are differences in prevalence rates and risk factors for depressive symptoms among boys and girls without PHIV [e.g., 21, 22, 23], which suggest that the pathways to depression among youth with PHIV may vary on the basis of youth sex. For example, by puberty, females in the general population are two to three times more likely to suffer from depression compared with males [23, 24]. Psychological and social changes during adolescence may additionally contribute to the emergence of differences in rates of depression among boys and girls. Increasing sexual maturity may differentially affect social roles of females vs. males, and interpersonal changes and expectations (e.g., focus on autonomy vs. interpersonal relationships) may result in differential exposure to stressful life events [25]. Girls are also more likely to use ruminative coping styles, which are linked to increased risk for depression [26]. Unfortunately, few studies of depression among youth with PHIV have explored differences in risk factors for depressive symptoms among boys and girls. Boys and girls without PHIV differ in associations among family environment, caregiver-child interactions, and caregiver psychological symptoms with youth depressive symptoms [e.g., 27]. Girls are more vulnerable than boys to family conflict [25, 27, 28] and maternal depressive symptoms [29], as well as more likely to worry that family conflict will negatively impact the quality of the caregiver-child relationship [30]. Indeed, the associations between family conflict and depressed mood have been shown to be stronger among girls than boys [27, 31], though results are mixed, with other work suggesting similar levels of association for boys and girls [32]. These findings warrant further investigation as to whether these differences are evidenced among youth with PHIV, who are at increased risk for family-related stressors compared to youth without PHIV and youth with other chronic illnesses [2, 9, 18].

One important methodological issue to consider is the low rates of agreement among caregivers and youth in reporting youth depressive symptoms [33], which are also evident among HIV-infected youth and their caregivers [34]. The discrepancy in ratings is likely attributable to various factors, including differences in knowledge of, access to, or classification of youth’s depressive symptoms as distressing, as well as differential presentations and youth’s ability to regulate their behaviors given context-specific demands [33, 35]. For example, girls may be rated as exhibiting higher levels of depressive symptoms because some observable depressive symptoms are socioculturally discouraged among boys [36]. Among HIV-infected youth, informant discrepancies are also related to unique psychosocial and HIV-related factors [34]. The current study examined the associations of youth depressive symptoms with family contextual factors drawn from multiple levels, including family cohesion, family conflict, caregiver-child detachment, consistent enforcement of discipline, and caregivers’ own depressive symptoms. To address potential informant variations, youth depressive symptoms were examined for caregiver- and youth self-reports separately. We hypothesized that girls would evidence higher rates of depressive symptoms according to both youth and caregiver reports. Based on prior research suggesting that girls are more vulnerable to family contextual factors [e.g., 29], we predicted that higher levels of youth depressive symptoms would be associated with less family cohesion and caregiver consistency in discipline, and greater family conflict, more severe caregiver depressive symptoms, and greater caregiver-child detachment, and that these factors would be related to higher levels of depressive symptoms among girls than boys.

METHOD

Participants

Participants with PHIV (N = 314) were recruited from 29 clinics in the United States and Puerto Rico involved in an NIH-funded research initiative, the International Maternal Pediatric Adolescent AIDS Clinical Trials Group (IMPAACT) P1055 study, a non-treatment, observational study of youth ages 6-17 years old who were perinatally infected with HIV or HIV-affected, and their families. The study was designed to examine prevalence rates and severity of emotional and behavioral symptoms among a cohort of perinatally HIV-infected children and adolescents and a demographically matched control group. The initial screening occurred in 2005 and 2006; among 1162 youth who were eligible for participation, 50% (n = 582) enrolled. Participation rates were higher among youth identified as Latinx, but did not differ significantly by sex or HIV status [37]. The P1055 study was completed in 2008. Given biological (e.g., virus’ crossing of the blood-brain barrier, inflammation, effects of medications [3, 4, 5]), developmental (e.g., challenges achieving milestones and developmental expectations [2, 7, 8]), and contextual (e.g., family stressors [2, 6]) challenges associated with PHIV, the present study considered only youth with PHIV. Additional inclusion criteria required that youth had lived with the same caregiver over the past year and had an IQ of at least 70.

Procedure

Prior to participant recruitment, each participating clinic submitted a site implementation plan to the study chairs for review and approval. Plans were required to delineate procedures for making psychiatric referrals; managing unintended HIV disclosure, recruiting, and retaining participants; incentives; and maintaining quality control. Site coordinators were instructed to ask participants whether they had mental health concerns at scheduled visits and take appropriate action for participants who became upset, concerned, or even curious about questions in the assessment battery. Study chairs conducted monthly reviews of referrals and related outcomes. Consent procedures assured youth that their responses would be confidential with the exception of any information indicating abuse, neglect, or harm to self or others. Disclosure of child abuse or neglect was reported to child protective services. To obtain a representative sample balanced for age and sex, lists of all eligible youth with PHIV and peer comparisons within the designated age range were generated by the study team for each of the 29 participating sites. Lists were sorted into blocks of 8 youth each, balanced for age (older [≥ 12 years old] vs. younger [< 12 years old]) and youth sex. Sites were required to contact each participant in a block before moving onto the next block and continued enrollment until 400 participants in each group were entered or enrollment was closed.

At study entry, youth and caregivers completed an extensive battery of questionnaires and rating scales including information about demographic, youth, caregiver, and family characteristics. For youth with PHIV, lifetime history of antiretroviral medications and major HIV-related diagnoses were obtained. The present study considers data that were derived from this baseline assessment. All measures were administered in a hospital outpatient clinic. All participants were encouraged to provide their own answers on self-report instruments, with staff available to read the questions as needed. Incentives varied by site because of differences related to local IRB requirements and included small cash amounts, gift cards, or vouchers for attending the study visit [10, 37, 38, 39].

Measures

Youth Depressive Symptoms.

Caregivers reported on youth depressive symptoms using the Child and Adolescent Symptom Inventory-4R (CASI-4R) [40], and youth reported on their own depressive symptoms using the Youth Inventory-4R (YI-4R) [41], both of which are DSM-referenced rating scales. Items are rated from 0 (never) to 3 (very often). Non-redundant items from the major depressive episode and dysthymic disorder subscales were summed to create a depression symptom severity score. Evidence supports satisfactory reliability (e.g., test-retest, internal consistency) and validity (e.g., convergent, discriminant) of the YI-4R (current sample α = .76) [42, 43, 44] and CASI-4R (current sample α = .77) in clinic and community samples [48] and in studies of many pediatric samples [5].

Caregiver-Youth Interactions.

Caregivers reported on the quality of the caregiver-youth relationship using the Parent’s Report (PR) [45], which contains 20 behavioral items pertaining to how caregivers interact with their children. We examined two subscales (4 items each): detachment (e.g., distance or withdrawal from the child; “I am unaware of what he thinks or feels;” α = .61), and consistent enforcement of discipline (e.g., caregiver’s commitment to established rules and procedures; “I punish him for disobeying;” α = .53). Items are rated on a scale from 0 (never) to 6 (always). The PR demonstrates good test-retest reliability and external validity [45] and has demonstrated predictive and concurrent validity and good internal consistency in studies of youth with internalizing problems [35, 46, 47]. The internal consistency of the PR subscales in the present study is lower than what is typically found, possibly related to heterogeneity among primary caregivers, duration of youth’s residing with the primary caregiver, and/or effects on family context based on factors associated with PHIV and/or parental HIV.

Family Environment.

Caregivers reported on the family environment using the Family Environment Scale (FES) [48]. The present study examined two subscales (9 items each): family cohesion (e.g., commitment and support among family members; “help/support one another;” α = .54), and family conflict (e.g., overt discord among family members; “fight a lot” and “get angry-throw things;” α = .57). Items are rated as true (1) or false (0). The cohesion and conflict scales have demonstrated good internal consistency among community-based samples and high convergent validity [47, 49, 50]. The FES subscales were less psychometrically sound than expected in the current sample, which could be a function of variability within the sample as noted with regard to the PR.

Caregiver Depressive Symptoms.

Caregivers reported on their own depressive symptoms using the DSM-referenced Adult Self-Report Inventory-4 (ASRI-4) [51]. The major depressive disorder subscale is comprised of the DSM-IV-TR symptoms for major depressive disorder (e.g., “feels worthless or inferior”). Items are rated from 0 (never) to 3 (very often) and summed to create a symptom severity score (α = .84). This subscale has exhibited satisfactory reliability (e.g., test-retest, internal consistency) and validity (e.g., convergent, discriminant) with comparable scales of adult psychopathology and differentiates non-clinic and clinic samples [51].

Analytic Approach

Descriptive statistics and bivariate correlations were conducted to examine relations among study variables, and independent samples t-tests were used to determine whether boys and girls differed on study variables, in SPSS Version 26.0. Two multiple regression analyses were conducted in Mplus Version 8 [52] with caregiver-reported youth depressive symptoms or youth self-reported depressive symptoms as the outcome variable. Mplus uses Full Information Maximum Likelihood (FIML) estimation to address missing data, given that omitting participants with missing data may bias an analytic sample [53]. All available data are used to estimate model parameters, which allows for smaller errors in parameter estimates and standard errors compared to other missing data strategies [53, 54]. Some PHIV youth were drawn from the same family and consequently observations are not independent (M = 1.08, SD = .27; range = 1-2 youth from each family; n = 301 youth drawn from independent families); thus, nesting within families was controlled in the analyses using the “cluster” option in Mplus.

All continuous predictor variables were z-scored (M = 0, SD = 1) before inclusion in the regression equations, and interaction terms were created using these z-scored variables. Step 1 included age, sex, and the 5 contextual variables (i.e., caregivers’ own depressive symptoms, cohesion, conflict, detachment, and consistency). Step 2 included all Step 1 variables and interaction terms between youth sex and the family contextual variables (i.e., sex × caregivers’ own depressive symptoms, sex × cohesion, sex × conflict, sex × detachment, and sex × consistency) [55, 56]. Post-hoc probing was conducted for significant interaction terms using procedures described by Holmbeck [56]. Specifically, we computed conditional moderator variables in which each group (i.e., boys or girls) was assigned a value of zero, then created interaction terms using the conditional group and z-scored family contextual variable. We then ran regressions with the conditional group variable, z-scored family contextual variable, and the conditional group × family contextual variable interaction term. From these analyses, we obtained the slope (unstandardized beta) and constant (intercept), which were used to graph the interactions.

RESULTS

Sample Characteristics

Youth with PHIV (N = 314) were on average 12.88 years old (SD = 3.08 years), and caregivers identified 50.80% as male and 49.20% as female, 85.30% Black or Latinx, and 14.70% as White (Table I). Caregivers ranged in age from 22-80 years old (M = 47.56, SD = 11.12), and youth lived with their respective caregiver for 1-18 years (M = 10.40, SD = 4.03). Primary caregivers who participated included biological mothers (34%), biological fathers (8%), biological grandparents (11%), adoptive mothers (24%), adoptive fathers (4%), other biological relatives (14%), and non-biological relatives (e.g., foster or stepparents, 4%). Thirty-nine percent (n = 122) of youth currently resided with a biological parent who was HIV positive (32% with biological mother only, 3% with biological father only, and 4% with both biological parents). The majority (59%) of youth had undetectable viral load at study entry (< 400 copies per mL), 74% had entry CD4% > 25%, and 78% did not have a prior AIDS-defining diagnosis. About two-thirds (66%) were receiving highly active antiretroviral therapy (ART) with protease inhibitors (Pis), and an additional 15% were receiving ART without Pis. Median duration of ART was 6.5 years [4, 10].

Table I.

Youth, Family, and HIV-related Characteristics at Study Entry of Youth with PHIV

Variable	n (% of available data)
Youth Characteristics
Ethnicity (n = 313)
African-American, Non-Hispanic	173 (55.27)
White, Non-Hispanic	46 (14.70)
Latinx, all ethnicities	94 (30.03)
Family Characteristics
Lives with biological caregiver (N = 314)	132 (42.04)
Primary caregiver education (n = 311)
≤ 8th grade	21 (6.75)
Some high school (≥ 1 year)	70 (22.51)
High school graduate or GED	75 (24.12)
Some college (≥ 1 year)	86 (27.65)
College degree	47 (15.11)
Master’s/doctoral degree	12 (3.86)
Primary caregiver income (n = 278)
< $10,000	51 (18.35)
$10,000 – $20,000	83 (29.86)
$20,001 – $40,000	73 (26.26)
$40,001 – $70,000	48 (17.27)
≥ $70,001	23 (8.27)
HIV-related Characteristics
Taking antiretroviral drugs (N = 314)	289 (92.04)
Entry CD4% (n = 313)
0-24	82 (26.20)
≥25	231 (73.80)
Nadir CD4% (n = 306)
0-14	127 (41.50)
15-24	100 (32.68)
≥25	79 (25.82)
Entry HIV viral load (n = 313)
0-400 copies/mL	185 (59.10)
401-10,000 copies/mL	64 (20.45)
>10,000 copies/mL	64 (20.45)
Peak HIV-1 RNA viral load (n = 306)
0-10,000 copies/mL	29 (9.48)
10,001 – 100,000 copies/mL	101 (33.01)
>100,000 copies/mL	176 (57.52)

Compared to boys in the community-based normative sample, boys in the present sample had lower CASI-4R caregiver-reported youth depression symptom severity scores (t(157) = 5.00, p < .01, Hedge’s g = .41) and youth-reported YI-4R depression symptom severity scores (t(145) = 5.18, p < .01, Hedge’s g = .47). Compared to girls in the community-based normative sample, girls with PHIV had lower youth-reported depression symptom severity scores (t(144) = 5.61, p < .01, Hedge’s g = .47) but similar levels of caregiver-reported youth depression symptom severity scores (t(150) = 0.30, p = .77, Hedge’s g = .03).

Descriptive statistics (Ms and SDs for the sample and among boys and girls, ranges, ns) and bivariate correlations among continuous variables are presented in Table II. Caregiver-reported youth depressive symptoms were associated with youth self-reported depressive symptoms and age, though the magnitude of these correlations was small; positively and moderately associated with caregivers’ own depressive symptoms, caregiver-youth detachment, and family conflict; and negatively associated with family cohesion. Correlations among family contextual variables were in expected directions and of low to moderate magnitude. Independent samples t-tests indicated that girls exhibited higher levels of caregiver-reported youth depressive symptoms: t(276.34) = 2.47, p = .014, Cohen’s d = .14; and youth self-reported depressive symptoms: t(285.73) = 3.16, p = .002, Cohen’s d = .18. No other differences between boys and girls emerged among study variables (ts range from .05-1.68, ps range from .09 to .96, ds range from .00 to .09).

Table II.

Bivariate Correlations (p values), Means, SDs, Ranges, and ns for Study Variables

Variable	1	2	3	4	5	6	7	8
1. Age	-
2. Caregiver-reported youth depressive symptoms	.18 (<01)	-
3. Youth self-reported depressive symptoms	−.13 (.022)	.17 (<.01)	-
4. Caregivers’ own depressive symptoms	.05 (.419)	.27 (<.01)	.06 (.276)	-
5. Cohesion	−.16 (<.01)	−.27 (<.01)	−.02 (.711)	−.22 (<.01)	-
6. Conflict	.05 (.384)	.26 (<.01)	.06 (.326)	.17 (<.01)	−.43 (<.01)	-
7. Detachment	.14 (.015)	.29 (<.01)	.08 (.180)	.28 (<.01)	−.26 (<.01)	.23 (<.01)	-
8. Consistency	−.11 (.053)	−.15 (<.01)	.09 (.119)	−.09 (.114)	.17 (<.01)	−.19 (<.01)	−.31 (<.01)	-
Sample M	12.91	1.84	7.26	8.16	7.44	2.39	4.30	19.37
Sample SD	3.08	2.47	4.07	4.86	1.53	1.75	3.96	3.37
Sample Range	6.25-18	0-14	0-27	1-30	2-9	0-9	0-18	9-24
Sample n	314	314	301	308	314	314	312	313
Boys M	12.92	1.50a	6.41b	7.94	7.58	2.28	4.13	19.20
Boys SD	3.15	2.04	4.13	4.64	1.47	1.75	3.66	3.48
Boys n	158	158	150	156	158	158	158	158
(% sample)	(50.3%)	(50.3%)	(49.8%)	(50.6%)	(50.3%)	(50.3%)	(50.6%)	(50.5%)
Girls M	12.91	2.19a	8.10b	8.38	7.29	2.49	4.51	19.54
Girls SD	3.02	2.82	5.10	5.09	1.60	1.76	4.24	3.27
Girls n	156	156	151	152	156	156	154	155
(% sample)	(49.7%)	(49.7%)	(50.2%)	(49.4%)	(49.7%)	(49.7%)	(49.4%)	(49.5%)

^aResults of independent samples t-tests indicate girls > boys (p = .014).

^bResults of independent samples t-tests indicate girls > boys (p = .002).

Primary Analyses

The regression analyses jointly considered z-scored family contextual variables and interactions between youth sex and these contextual variables as predictors of caregiver-reported youth depressive symptoms (Table III) and youth self-reported depressive symptoms (Table IV). Older age and female sex predicted caregiver-reported and youth self-reported depressive symptoms. Caregivers’ own depressive symptoms, caregiver-youth detachment, and family conflict predicted caregiver-reported youth depressive symptoms (Table III), and consistency in enforcement of discipline predicted youth self-reported depressive symptoms (Table IV). Step 2 results indicated significant interaction terms for sex × detachment in predicting caregiver-reported youth depressive symptoms (Table III), and sex × cohesion in predicting youth self-reported depressive symptoms (Table IV). Post-hoc probing of the sex × detachment interaction term indicated that the slope was significantly different from zero among girls (B = 1.19, p < .01), but not boys (B = .14, p = .439), for caregiver-reported youth depressive symptoms (Figure 1). Boys exhibited similar (and low) levels of depressive symptoms regardless of level of detachment. Among girls, higher levels of detachment were associated with higher depressive symptoms. Post-hoc probing of the sex × cohesion interaction term for youth self-reported depressive symptoms indicated that the slope was significantly different from zero among boys (B = .64, p = .023), but not girls (B = −.49, p = .192; Figure 2). Girls reported similar levels of depressive symptoms regardless of the levels of family cohesion, whereas boys reported higher levels of depressive symptoms in the context of higher family cohesion. Consistent with the results of the independent samples t-test analyses, the figures also indicate that girls’ symptom levels were higher than boys’.

Table III.

Multiple Regression Analyses with Caregiver-Reported Youth Depressive Symptoms (N = 314) as the Outcome

Variable	B	SE B	β	p
Step 1
Age	.34	.13	.13	.008
Sex	−.55	.24	−.11	.024
Caregivers’ own depressive symptoms	.45	.14	.18	.001
Cohesion	−.22	.15	−.09	.141
Conflict	.34	.17	.14	.047
Detachment	.41	.15	.16	.007
Consistency	−.08	.14	−.03	.589
Step 2
Sex × caregivers’ own depressive symptoms	−.37	.29	−.10	.210
Sex × cohesion	−.21	.30	−.06	.500
Sex × conflict	.05	.35	.01	.881
Sex × detachment	−.80	.26	−.21	.002
Sex × consistency	.16	.28	.05	.571

Table IV.

Multiple Regression Analyses with Youth Self-Reported Depressive Symptoms (n = 301) as the Outcome

Variable	B	SE B	β	p
Step 1
Age	−.64	.30	−.13	.032
Sex	−1.54	.52	−.16	.002
Caregivers’ own depressive symptoms	.19	.25	.04	.445
Cohesion	.07	.27	.02	.797
Conflict	.24	.31	.05	.438
Detachment	.49	.32	.10	.128
Consistency	.53	.24	.12	.023
Step 2
Sex × caregivers’ own depressive symptoms	.03	.50	.01	.946
Sex × cohesion	1.37	.54	.20	.013
Sex × conflict	.70	.60	.10	.247
Sex × detachment	.21	.58	.03	.717
Sex × consistency	−.02	.50	.00	.975

Figure 1.

Relation between detachment and caregiver-reported youth depressive symptoms among boys and girls.

Figure 2.

Relation between cohesion and youth self-reported depressive symptoms among boys and girls.

DISCUSSION

There is strong evidence indicating a greater risk for depressive symptoms among older girls vs. boys [21, 22, 23]; however, this finding is not well-established among youth with PHIV, a population impacted by immunologic, developmental, and social vulnerabilities that may confer additional risk for depressive symptoms. As expected, for both caregiver- and youth self-report measures, girls with PHIV evidenced higher levels of depressive symptoms compared to boys, supporting the extension of this consistent finding from samples without PHIV to this vulnerable population of youth [21, 22, 23]. Girls with PHIV may use more ruminative coping styles and experience higher levels of social and psychological stressors that confer risk for depressive symptoms as they approach puberty as compared to boys with PHIV, consistent with research among youth without PHIV [25, 26]. Higher levels of caregivers’ own depressive symptoms, family conflict, and caregiver-child detachment predicted higher levels of caregiver-reported youth depressive symptoms, and inconsistency in discipline predicted higher levels of youth self-reported depressive symptoms. The main effect of detachment was qualified by differences between boys and girls. Specifically, girls demonstrated higher levels of depressive symptoms in the context of higher levels of caregiver detachment, consistent with hypotheses and previous research suggesting that girls may be more strongly influenced by family relationship dynamics than boys [56].

The present study uniquely considers both child sex and key family contextual factors associated with depressive symptoms among youth with PHIV, who may be particularly vulnerable, both biologically and emotionally, to problems in family- or caregiver-level relationships. Youth with PHIV and their families often face psychosocial and medical challenges not typically encountered by youth without PHIV [2, 8, 18]. Chronic illness has the potential to disrupt typical caregiver-child relationships, possibly leading to a ripple effect for youth mental health across development [57]. In combination with normative autonomy-seeking from caregivers around the time of adolescence, this process may explain elevated depressive symptoms among girls in the context of caregiver detachment.

Nevertheless, caregivers of youth with PHIV may react to these stressors with more focused, active, and positive parenting to maximize the youth’s quality of life [18]. This may be especially true when the youth appears to exhibit depressive symptoms, which may account for the finding that higher levels of depressive symptoms among boys are associated with greater family cohesion. The significant main effect of caregiver depressive symptoms is consistent with findings that caregiver depressive symptoms may affect caregiver-child and family interactions and thereby contribute to youth psychological problems [58]. It should also be noted that youth participating in the study received close attention, care, and flexibility from the research staff, many from a very young age, which may have also served as a protective factor. Indeed, boys and girls had lower or similar levels of depressive symptoms compared to CASI-4R and YI-4R community-based normative samples, which may speak to the protective nature of study participation.

The present study has a number of strengths. The large, geographically representative sample of PHIV youth, as well as their broad age range, allowed for assessment across developmental periods, which increases the generalizability of findings. Likewise, the present study’s assessment battery of risk factors drawn from multiple contextual levels captures several aspects of the family context, which is a particular strength given the limited research about family contextual risk factors among youth with PHIV [9]. In addition, the use of multiple informants to assess youth depressive symptoms reduces concerns about mono-rater biases and accounts for potential contextual and informant variations in youth’s depressive symptoms. Of note, and consistent with previous research among youth with PHIV and their caregivers [34], relatively low agreement was observed between caregiver- and youth self-reported youth depressive symptoms. Such informant discrepancies between caregivers and youth for depressive symptoms may reveal meaningful variability that can inform treatment planning and implementation [33, 34].

Despite these strengths, there are several limitations. The analyses are cross-sectional and as such preclude inferences about causality or direction of effects. In addition, some of the indices of family contextual factors in the present study had low internal consistency, suggesting potential issues of reliability and construct validity, which may have influenced results. Symptom levels were generally lower in the present sample compared to a community-based normative sample [42, 43]; thus, future research should aim to replicate these findings in a different clinical or pediatric sample. Another limitation is the possibility that the association between caregivers’ own depressive symptoms and caregiver-reported youth depressive symptoms may be attributed to caregivers’ negative attentional bias stemming from their own depressive symptoms [59]. Using alternative rating scales or methods to measure constructs of interest could address this potential confound, as well as determine whether findings generalize to other informants and indices. Given increasing survival rates, the prevalence of psychological problems among youth with PHIV is likely to grow as these youth age. Research that considers family contextual factors and youth psychological functioning in other samples of youth with PHIV will be necessary to evaluate the generalizability of these findings from a US sample and inform prevention and treatment approaches for psychological symptoms among youth with PHIV.

These findings support the importance of integrating mental health services into care for youth with PHIV, as has been recommended for individuals across the globe living with HIV/AIDS [60]. To maximize affordability, acceptability, and availability of mental healthcare for HIV positive individuals, it has been recommended that existing primary healthcare workers and counselors receive additional training (ideally in collaboration with local mental health service providers) to provide basic assessment and management of mental health problems as part of the typical HIV/AIDS service [60]. Further, mounting evidence suggests that family involvement in intervention for youth with PHIV is critical to not only effectively increase medication adherence and reduce sexual risk behavior, but also manage youth mental health symptoms [2]. One example that is consistent with this possibility is the Collaborative HIV-Prevention and Adolescent Mental Health Family Program (CHAMP+), a family-based, culturally sensitive psychosocial intervention for youth with PHIV [61]. Importantly, the intervention may be offered in a hospital setting, where youth are already receiving their HIV primary care. The curriculum is delivered through interactive activities, discussions, role plays, and games, and is focused on improving communication, relationship maintenance skills, social support, and HIV education in a manner developmentally appropriate for preadolescent HIV positive children and their caregivers [61]. Evaluations of this family-centered intervention indicate that it leads to improved medication adherence, treatment knowledge, and youth-caregiver communication around HIV [6]. Despite promising preliminary results, more work is needed to support feasibility of implementation. The results of the present study moreover suggest that interventions may benefit from increased attention to differences in family contextual factors among boys and girls with PHIV, and that screening children with PHIV and their families for these factors and depressive symptoms could facilitate identification of youth at risk for depressive symptoms and inform prevention efforts to mitigate such risk. Though puberty is a time when youth begin to assert greater independence from caregivers, the present study suggests that caregiver detachment is a risk factor for depressive symptoms among girls with PHIV, and may therefore be a key intervention target that has not been explicitly considered among youth with PHIV. Practitioners working with girls with PHIV and/or their family members may see benefits of increased vigilance for early signs of youth depressive symptoms and focus on strengthening caregiver-child connections.

Future research should seek to examine mechanisms that might account for the unexpected finding that boys’ depressive symptoms are associated with greater family cohesion; indeed, the direction of the effect of cohesion on boys’ depressive symptoms opens up further questions around risk and protective factors among families with a male youth with PHIV. Based on the relatively low agreement between caregiver- and youth self-reported depressive symptoms, practitioners working with families of youth with PHIV may be mindful of potential biases related to child sex in assessment of depression. Finally, future research should utilize a prospective design to explore the temporal relation between family contextual factors and depressive symptoms, as well as address whether the influence of these caregiver- and family-related factors are specific to depressive symptoms or also associated with other psychological outcomes (e.g., behavior problems, anxiety) observed among youth with PHIV [1].

CONCLUSIONS

In summary, there is much to consider as a caregiver of a youth with PHIV that may negatively impact caregiver mental health symptoms, caregiver-child interactions, and broader family-level functioning, each of which may in turn influence psychosocial functioning of youth with PHIV. The present study joins others in highlighting the importance of family involvement in interventions among youth with PHIV, and uniquely illuminates differences among boys and girls in both levels of and risk factors for depressive symptoms among this high-risk population. These risk factors (e.g., caregiver depressive symptoms, conflict, consistency), as well as differences among boys and girls (detachment and cohesion), should be further investigated and considered in the development and dissemination of evidence-based psychosocial interventions for youth with PHIV. Through a better understanding of how to effectively target and intervene in the development of depressive symptoms, youth with PHIV and their families may be better equipped to be resilient in the face of challenges and thereby attenuate risk for depressive symptoms and related negative sequelae.