Introduction
As increasing numbers of heterosexuals, particularly women, become infected with HIV, their families are affected by AIDS [1]. By the year 2000, 52 000-125 000 children and adolescents in the US will be orphaned by AIDS, reflecting annual rates of death that are equal to or higher than the national estimates of parental death caused by cancer or motor vehicle accidents [1]. Parents living with AIDS are confronted with a series of decisions regarding their children, including the disclosure of their illness status to their children and who will care for their children if they die. Women disclose personal information significantly more often than men [2]; thus, mothers with AIDS may disclose to their children more often than fathers.
Parents often regulate their behaviour to their child's developmental capacities; therefore, we also anticipated that the decisions of parents living with AIDS would differ, depending upon the child's age.
Parental gender and the age of the child were anticipated to influence disclosure, as well as custody arrangements. Children are likely to be at higher risk for long-term negative outcomes if parents do not make custody plans [3]. Children bereaved by sudden, unexpected parental loss demonstrate more negative outcomes than children who are prepared for such a loss [4] and the legal complications are greater [3]. As there are no data on these patterns among parents living with AIDS, our first goal is to describe how these parents cope with the issue of disclosure of their illness and custody for their children.
The disclosure and custody choices of parents living with AIDS present challenges to their children, who must cope with prejudice and stigmatization from others [5], with anticipatory grief due to their parent's illness, and they may experience increased anxiety for their own security and welfare. Most of the existing data on the adjustment of bereaved children comes from middle-class, white children who have a surviving parent and who were recruited from large university medical centres [6]. In contrast, parents living with AIDS are predominantly disenfranchised single-parent minority group members who have been involved in injecting drug use subcultures in inner cities [7] and have a highly stigmatizing illness [5] that leads to social isolation [8]. Bereaved children are at increased risk for depression, anxiety, conduct disturbance, academic difficulty, decreased self-esteem, somatic complaints and suicidal acts [8]. Thus, our second goal is to report how the practice of disclosure and custody arrangements of parents living with AIDS influence their adolescent's immediate and short-term social and mental health adjustment over 3 to 6 months.
Methods
Participants
One hundred and fifty-one parents living with AIDS, aged 25-60 years, were recruited from a consecutive series of names of persons with AIDS accepted for case management services at the New York City (NYC) Division of AIDS Services. Eligibility included having a diagnosis of AIDS (mean time = 15 months) and at least one non-infected adolescent child aged 12-18 years. Of 232 English- and Spanish-speaking (3%) families who were referred, voluntary informed consent was obtained from 65% of the parents (n = 151); 81% of the participating parents allowed their adolescents aged 12 years or older to be interviewed (n = 171) and 70% were reassessed at a follow-up interview 3-6 months later.
Procedures
Interviews were conducted by predominantly (62%) African-American or Latino two-person teams in the homes of the parents living with AIDS. Assessment domains covered parental disclosure and parental custody plans.
a. Parental disclosure: the parents living with AIDS indicated, for each child, the type of disclosure made regarding their illness using the following categories: no disclosure, ill, chronically ill, HIV-infected, has AIDS, or is going to die.
b. Parental custody plans: the parents living with AIDS were asked about each child regarding: discussions with potential guardians about custody; contacts with social service agencies; wills; and standby agreements.
In the adolescent assessments the following domains were assessed at recruitment and 3 to 6 months later. The assessments at recruitment included a lifetime and current (previous 3 months) time frame:
a. Sexual risk indicators included abstinence, the number of sexual partners for vaginal or anal sex, and the number of unprotected encounters of vaginal or anal sex.
b. Four dichotomous indicators of substance use and a weighted substance-use index were calculated. Use of cigarettes, alcohol, and marijuana in the past 3 months versus abstention were dummy-coded. The fourth indicator dummy-coded any 'harder' drug (stimulants, inhalants, hormones, cocaine/crack, heroin). Specific substances were weighted by seriousness [9] as: 0.25, cigarettes/day; 0.50, alcohol; 1, marijuana; 2, amphetamines; 3, inhalants, hallucinogens, and steroids; 4, crack and cocaine; and 5, heroin.
c. Emotional distress. The Brief Symptom Inventory [10], provides the average degree of distress across 53 symptoms during the previous week on a scale from 0 (not at all) to 4 (extremely) and a summary count of symptoms of distress [Cronbach's alpha (α) = 0.94].
d. Delinquency. Stealing, fighting, and vandalism events during the past 3 months were counted for a summary of 12 behaviours.
Results
Description of the samples
The mean age of the parents living with AIDS was 37.6 years [standard deviation (SD), 5.2; range, 25-60]; 41% were Latino and 39% were African-American. Mothers represented 86.8% of the sample. About half (46%) of the parents living with AIDS had graduated from high school. Household compositions varied: 94% included children living in the household, the remainder had children temporarily in foster care placements, group homes, or incarcerated; 23% included an adult partner; 9% lived with a parent, 3% with other relatives, and 24% were the single head of a household.
About one-third of the parents had injected drugs although, currently, only 3.3% injected drugs and fathers (15%) were significantly more likely to be injecting drugs than mothers (1.3%). The lifetime number of sexual partners was high (mean, 121; SD, 620; median, eight); 2.6% always used condoms. Other high-risk behaviours included sex with partners who bartered sex (17%), injected drugs (66%), were HIV-seropositive (57%), or were bisexual (21%). In the pre-vious 3 months, 54% abstained from sexual activity and among the sexually active (n = 69), most had one sexual partner (mean, 1.2; SD, 1.1; median, 1.0), and 68% of their sexual encounters were protected by condoms. Current health status was poor, with 45% reporting a major illness onset in the previous year. A mean of 15 (SD, 5.6; range, 0-23) physical symptoms was reported over the previous 3 months; these were symptoms that resulted in moderate distress (mean, 1.9; SD, 1.0; range 0-5).
Disclosure and custody planning were reported for each child. On average, parents reported 2.6 children (mean age 13.2 years, ranging from a few months to 18 years old) and totalling 386 children for the sample of 151 parents. Half of the children were female (49.5%). The age of the children was distributed as follows: 10.1% were aged 5 years or younger; 20.4% were aged 6-10 years; 31.6% were aged 11-14 years; 31.3% were aged 15-18 years; and 6.5% were over 18. The 171 adolescents (53% female) were in eighth grade (usually consisting of 13-year olds) on average (mean age 15.2 years; SD, 1.9). More than half (58%) had had to retake a school year; 2% had graduated from high school (with 74% too young to have graduated), and 9% were currently out of school.
Disclosure
Figure 1 summarizes the frequency and type of disclosure grouped by the gender of the parents living with AIDS. Typically, children were told that their parents were HIV-positive (74%). Overall, only 11% disclosed nothing (neither illness nor HIV status). These patterns are similar for fathers and mothers. However, the patterns vary significantly by the age of the child (χ2 [degrees of freedom (df) = 4] = 110.5, P < 0.001), as demonstrated in Fig. 2. As expected, children to whom HIV-positive-serostatus was disclosed were significantly older [F statistics (df = 1, 384), 108.2; P < 0.001; mean, 14.5; SD 3.6] than children who were uninformed (mean, 10.0; SD, 5.0). Within families, 11.3% of parents living with AIDS told none of their children about their illness, seropositivity, or AIDS diagnosis, whereas 44% told all of their children about their HIV serostatus. Parents disclosed similar information to male and female children.
Custody plans
Almost all the parents initiated a discussion of potential guardianship for their children within a few months of being diagnosed with AIDS. Mothers (81%) and fathers (75%) initiated custody discussions with members of their social support network at similar rates. Parents discussed potential guardianship with their own parents (29.5%), a sibling (27.8%), children over 18 years old (10.6%), extended family members (2.0%), a romantic partner (10.6%), and a friend (3.3%). Far less common were discussions with staff in social service agencies (24%) or with the other parent (16%). Legal actions were initiated by 30% of the parents living with AIDS (31% mothers, 25% fathers). The age of the child significantly influenced the type of action taken [χ2 (df = 3) = 8.7, P < 0.03], but the child's gender did not influence custody arrangements. Each type of action was more likely to have occurred for children aged 11 years or younger, compared with older children. Parents consulted a potential guardian for 75.9% of the children and 99% of those consulted agreed to be the guardian. However, informing the child of custody arrangements varied by the child's age. The mean age of children who had been informed of the parents' custody arrangements (mean, 13.0 years; SD, 4.3) was significantly older than those uninformed [mean, 10.2 years; SD, 5.1; F (df = 1, 342), 23.1; P < 0.001]. Wills were made for younger children (mean, 11.0 years; SD, 4.7), whereas children without such arrangements were older [mean, 12.8 years; SD, 4.7; F (df = 1, 379), 9.7; P < 0.002]. Overall, parents initiated 2.1 (SD, 1.1; range, 1-5) custody discussions about their children.
Relationship of disclosure to custody arrangements
The parents' disclosure of serostatus was not significantly associated with custody arrangements, even when patterns for adolescents [χ2 (df = 1) = 0.3; non-significant (NS)] and younger children [χ2 (df = 1) = 0.4, NS] were examined separately. Among the 73% who disclosed their serostatus to their adolescent children, 23% had also made legal custody arrangements. Among the 27% who did not disclose, 26% made legal arrangements. These patterns were not related to the adolescent's gender, ethnicity, or age.
Adolescent adjustment associated with the choices made by parents living with AIDS regarding disclosure of HIV infection and legal custody arrangements.
Because 49 parents living with AIDS had more than one adolescent child enrolled in the study, we evaluated the family-level effect of parents living with AIDS' disclosure by fitting hierarchical models for continuously-scaled dependent variables using SAS PROC MIXED [11]. Independent variables in the model included age, gender, disclosure, and legal arrangements. The magnitude of the intra-class correlations, capturing the degree to which outcomes are correlated across children in the same families, ranged between 0.003 (P = 0.99) and 0.136 (P = 0.37) for mental health and delinquency. It was not necessary to run PROC MIXED for the sexual risk behaviours because youths who were sexually active did not have sexually active siblings; therefore, intra-class correlations would be artefactual. We examined a subsample consisting of one adolescent, randomly selected from each family (n = 122 at recruitment; n = 90 at follow-up). The overall pattern of results was similar, indicating no family effect.
Bivariate tests of significance were conducted to examine differences in sexual risk, substance use, emotional distress, and delinquency between adolescents who were informed that their parent was HIV-positive or had AIDS. Similar analyses were conducted for adolescents covered by legal custody plans. Analyses of sexual risk behaviour were conducted with a subsample of 54 adolescents at recruitment and 42 adolescents at follow-up who reported at least one sexual partner in the 3 months preceding the interview.
Mean scores of each measure for current behaviours are summarized in Table 1, grouped by informed and uninformed adolescents. Adolescents informed of their parents' HIV serostatus were significantly less likely to be sexually abstinent in their lifetime [48.5% versus 67.6%; χ2 (df = 1) = 4.2, P < 0.039]. There were no differences in the number of current sexual partners at recruitment among those who were sexually active. However, among the sexually active, informed adolescents reported significantly more unprotected sexual risk acts [Student's t test (df = 47), 2.2; P < 0.03] than uninformed adolescents. At the follow-up assessment, the informed adolescents were abstinent less often [χ2 (df = 1) = 5.2 P = 0.02]; but adolescents had similar numbers of partners and unprotected acts.
At recruitment, cigarette use was more likely among informed youths [χ2 (df = 1) = 7.9, P < 0.005], but alcohol and marijuana use were similar. At the follow-up interview, informed adolescents were significantly more likely to use cigarettes [χ2 (df = 1)= 5.8; P = 0.016] but were similar in alcohol and marijuana use. Abstinence varied by substance, both at recruitment and follow-up. However, significant differences were found between informed and uninformed adolescents on the weighted substance use index, which takes into account the level and severity of use at both recruitment [Student's t test (df = 82), -3.1; P = 0.003] and follow-up [t test (df = 80), -2.93; P = 0.004].
Informed and uninformed youths at recruitment were similar in the mean number of symptoms and the level of emotional distress. At follow-up, informed adolescents had more symptoms of emotional distress, [t test (df = 116), -2.6; P = 0.01], and a higher level of distress, [t test (df = 62), -2.9; P = 0.005].
Although adolescent behaviour patterns were related to their parents' disclosure of illness, patterns were not associated with legal custody arrangements (data available from M.J.R-B.). At recruitment, the lifetime reports of adolescents with legal custody arrangements indicated significantly higher rates of sexual abstinence [65% versus 48%; χ2 (df = 1) = 4.0, P < 0.05] and significantly fewer sexual partners [mean, 4.9; SD, 13.7 versus mean, 1.4; SD, 2.9; t test (df = 149), 2.7; P < 0.008]. At recruitment, adolescents with legal custody arrangements reported significantly higher rates of sexual abstinence [64.0% versus 80.4%; χ2 (df = 1) = 4.2; P = 0.04], but the sexually active were similar in the numbers of sexual partners and risk acts. At follow-up, rates of sexual abstinence were similar across youths with and without legal custody plans. Overall, the presence of legal custody arrangements was not associated with significant differences in alcohol or drug use, emotional distress, or delinquency.
Discussion
NYC is unique in providing access to 30% of the nation's parents living with AIDS [1]. In addition, NYC is the only site where a centralized roster exists of all parents living with AIDS, allowing us to assess a sample similar in ethnicity, gender, and age to those acquiring HIV heterosexually and through substance use nationally [12]. Our sample was disenfranchised Latinos or African-Americans living with AIDS who were single parents with a history of substance use, multiple sexual partners, and sex in high-risk contexts.
Unexpectedly, mothers and fathers demonstrated very similar behaviour patterns. However, as expected, parents adjusted their behaviour based on their children's developmental age. It may be that parenthood is more salient than gender in determining these behaviour patterns, since the behaviour of mothers and fathers was similar. Younger children were not informed about their parent's AIDS diagnosis, whereas older siblings were. Differential disclosure to different children created a system in which 'secrets' exist among one-third of the families, a pattern consistently associated with unhealthy adjustment and communication in other research [13]. Longitudinal analysis may identify whether children who were not informed of their parent's serostatus learn from their siblings or parents over time. We must also develop ways of assessing children's awareness of a 'secret' illness. However, the behaviour of parents living with AIDS regarding disclosure was dissimilar to those for custody arrangements, suggesting different determinants of these processes.
Surprisingly, the adolescents' behaviour within a family was unrelated to each other, but was related to their parent's behaviours. Adolescents informed of their parent's illness were more likely to engage in unprotected sex, substance use, and be emotionally distressed. The higher prevalence of smoking among informed adolescents and the greater severity of substance use (frequency and level) is cause for concern, given the well-established finding that the use of cigarettes and alcohol often precede the use of marijuana and harder drugs [9,14]. It also appears that other problems may emerge 3 to 6 months later where earlier problems did not exist (e.g., emotional distress). This finding is particularly disturbing, given the tendency of AIDS care providers to consistently encourage disclosure of illness [15,16]. If these findings persist over time, it may be that disclosure to an adolescent is not positive for the youth's long-term adjustment. However, the current findings cover only a 6-month period. It may be that disclosure enhances short-term problems, but provides the parent and the adolescent the opportunity to resolve interpersonal issues surrounding grief and bereavement. Although coping with these issues is difficult and has a short-term negative impact (e.g., over 6 months), the youths' long-term adjustment following the parent's death may be better than the adjustment of youths who were unprepared for their parent's death or who pretended that they were not aware of the parent's diagnosis, when, in fact, they had learned of the parent's diagnosis without a formal disclosure by the parent. Secrets within families have been seen as generating problems, rather than helping family members toward better adjustment [17]. The data provided in this article are important in challenging our assumptions regarding the positive benefits of disclosure. However, the data may indicate that the adjustment of parents and adolescents may be a complex, developmental process. If these patterns hold over time, individually tailored intervention programs may be needed to help parents living with AIDS cope with how and when to disclose their serostatus to their children, and help them decide what disclosure patterns are appropriate, given the developmental and adjustment needs of their individual children.
As with the disclosure findings, the custody plans of parents living with AIDS vary according to their children's age, and are similar for mothers and fathers. Contacting social service agencies or making legal arrangements were not common. Half as many families make custody plans for older children as for younger children, highlighting the need for further research to investigate factors associated with legal custody arrangements for older children. It may be that parents living with AIDS who initiate legal custody arrangements are a self-selected group based on other factors. They may have children who are 'easy' in temperament or have personality styles that may assist in making custody arrangements easier.
Although we have known for at least 5 years that an epidemic of AIDS orphans is emerging [3], our ability to anticipate the short- and long-term consequences for these children has not been clear. The unique stigmatization associated with AIDS, socioeconomic and geographic clustering of cases, and the paucity of longitudinal prospective research on bereaved children make it difficult for policymakers, program developers, and researchers to design interventions for these children. We can anticipate that there will be increased stress and negative outcomes for many of these children [3], out-comes that will be associated with increased costs to society. Our challenge is to plan how to improve the outcomes and help guide parents living with AIDS to face the challenges of when, how, and to whom to disclose their illness and how to effectively plan for their children's care following their death.
Acknowledgements
We thank the parents and adolescents who participated in this study, as well as those who assisted in it, including T. Belin, C. Cantwell, B. Gardner-Williams, J. Hudis, K. Langabeer, J. Lehane, P. Lewis and M. Riedel.
References
1. Michaels D, Levine C: Estimates of the number of motherless youth orphaned by AIDS in the United States. JAMA 1992, 268:3456-3461.
2. Mark E, Alper TG: Women, men, and intimacy maturation. Psychol Women Q 1985, 9:81-88.
3. Levine C: The new orphans and grieving in the time of AIDS. In AIDS and the New Orphans. Edited by Dane BO, Levine C. Westport: Auburn House; 1994:1-11.
4. West SG, Sandler I, Pillow DR, Baca L, Gersten JC: The use of structural equation modeling in generative research: Toward the design of a preventive intervention for bereaved children. Am J Community Psychol 1991, 19:459-480.
5. Herek GM, Capitanio JP: Public reactions to AIDS in the United States: A second decade of stigma. Am J Public Health 1993, 83:574-577.
6. Sandler IN, West SG, Baca L, et al.: Linking empirically based theory and evaluation: The family bereavement program. Am J Community Psychol 1992, 20:491-521.
7. Zayas LH, Romano K: Adolescents and parental death from AIDS. In AIDS and the New Orphans. Edited by Dane BO, Levine C. Westport: Auburn House; 1994:59-76.
8. Rotheram-Borus MJ, Miller S, Murphy DA, Draimin BH: An intervention for adolescents whose parents are living with AIDS. Clin Child Psychol Psychiatry 1997, 2:201-209.
9. Kandel DB, Yamaguchi K: From beer to crack: Developmental patterns of drug involvement. Am J Public Health 1993, 83:851-855.
10. Derogatis LR: Brief symptom inventory. Baltimore: Clinical Psychometric Research; 1992.
11. SAS Institute Inc.: SAS Technical Report P 229, SAS/STAT Software: Changes and Enhancements, Release 6.07. Cary: SAS Institute Inc., 1992:289-366.
12. Centers for Disease Control and Prevention: HIV/AIDS Surveillance Report: US HIV and AIDS Cases Reported Through June 1996. Atlanta: CDC; 1996:8.
13. Wright LS, Garrison J, Wright NB, Stimmel DT: Childhood unhappiness and family stressors recalled by adult children of substance abusers. Alcohol Treatment Q 1991, 8:67-80.
14. Kandel DB, Yamaguchi K, Chen K: Stages of progression in drug involvement from adolescence to adulthood: Further evidence for the gateway theory. J Stud Alcohol 1992, 53:447-457.
15. Draimin BH: Everything you need to know when a parent has AIDS. New York: Rosen Publishing Group; 1994.
16. Dane B: Mourning in secret: how youngsters experience a family death from AIDS. In Orphans of the HIV Epidemic. Edited by Levine C. New York: United Hospital Fund of New York; 1993:60-68.
17. Cooklin A, Barnes GG: Taboos and social order: new encounters for family and therapist. In Secrets in Families and Family Therapy. Edited by Imber-Black E. New York: WW Norton; 1993:292-328.
© Lippincott-Raven Publishers.