The effect of parental chronic illness on quality of attachment of adolescents P. Vitali
Student Number: s1757946 Master Thesis Clinical Psychology
Supervisor: Dr. D. S. Sieh Institute of Psychology
Leiden University
Abstract
The aim of this study was to investigate the link between parental chronic physical illness (CPI) or chronic mental illness (CMI) and quality of attachment of adolescents. Additionally, the moderating effect of adolescent age and depression levels of the parent with CPI in this link were examined. Two data sets were used with each a target and control group, the first of which included 130 adolescents from families with one parent suffering from CPI and their 160 parents, and 112 adolescents from families with two healthy parents and their 138 parents. The second data set included 26 adolescents with a parent suffering from CMI and 122 adolescents with two healthy parents. Quality of attachment of adolescents was measured using the Inventory of Parent and Peer Attachment (IPPA) which consists of three subscales measuring communication, trust and alienation from their parents. Quality of attachment of adolescents with a parent suffering from CPI or CMI did not differ from that of adolescents with two healthy parents. Results remained the same after controlling for family income, gender of the ill parent and gender of adolescents. No moderating effect of adolescent age or depression levels of parents with CPI was observed. These findings follow the direction of recent research supporting that chronic illness of a parent is not always associated with adverse outcomes for adolescents.
The effect of parental chronic illness on quality of attachment of adolescents “Sometimes I’ll just feel alone, like I’m the only one dealing with this and it’s like having the weight of the world on my shoulders”. This is how a 13-year-old adolescent puts to words how it feels like having a parent suffering from chronic illness (Phillips & Lewis, 2015). It is estimated that 15% of people suffering from chronic conditions are parents (Barkmann, Romer, Watson, & Schulte-Markwort, 2007). This percentage is bound to rise, as chronic illness rates are increasing globally (Chan, 2011). Chronic illness seems to pose a threat to quality of parent-child
relationship (Pedersen & Revenson, 2005; Rolland, 1999) and children’s psychosocial adjustment (Sieh, Viser-Meily, & Meijer, 2013). However, literature in this field remains limited (Pedersen & Revenson, 2005). This study focuses on the effect of chronic physical illness (CPI) or chronic mental illness (CMI) of the parent on quality of attachment of their adolescent children.
The term “chronic illness” is used to define health conditions that persist over time, requiring continuous health care provision (Martin, 2007). Those conditions can be treated but not cured, and they are associated with impairments in physical, emotional and social functioning (Wikman, Wardle & Steptoe, 2011). Although chronic illness is often considered to be a physical condition as in case of multiple sclerosis, psychopathology that persists over time, such as depression, psychosis, and substance dependence, falls under the chronic illness classification as well
(Martin, 2007). Chronic physical illness (CPI) or chronic mental illness (CMI) of a parent affects all family members, especially children, as family roles become re-adjusted (Mehta, Cohen & Chan, 2008). Children experience long periods of absence of parents, as the ill parent is often hospitalized, bed-bound and weak (Pedersen & Revenson, 2005; Rolland, 1999). Additionally, the healthy spouse faces increased family responsibilities, as he or she is the first to take over the
familial duties of the ill parent (Rolland, 1990). As a result, both parents become physically and emotionally unavailable and they are often unresponsive towards their children’s needs (Bowlby, 1983). Children may also face increased familial responsibilities, such as helping the disabled parent or taking care of younger siblings (Ireland & Pakenham, 2010a). This role shift may result in parentification, meaning that children adopt responsibilities beyond age appropriateness (Hooper, 2007). Parentification is a transition where parents lose their role as caregivers who cover the physical needs of their children, as well as their role as supporters who fulfill the psychological needs of their children. Indeed, as Ireland and Pakenham (2010a) indicated, in case of chronic illness of a parent, children adopt tasks both as caregivers (i.e. cooking the family meals) and supporters of the family (i.e. ensuring the psychological well-being of the ill or healthy parent). Children of parents suffering from CMI have additional challenges to deal with: As symptoms of mental illness can lead to rapid mood swings, distractibility, and irritability (Hipwell, Goossens, Melhuis, & Kumar, 2000), parents with CMI often appear to have
disorganized, inconsistent and unstructured interactions with their children (Oyserman, Bybee, Mowbray, & Hart-Johnson, 2005; Oyserman, Mowbray, Meares, & Firminger, 2000). A parent diagnosed with depression or schizophrenia can display withdrawn or violent behavior and hostile emotions which create distance to the child (Hartney &Barnard, 2015; Wai Wan & Green, 2009). Consequently, the emotional closeness that usually characterizes parent-child relationships can be partially lost or reduced. Finally, CPI and CMI are in many cases combined. Sufferers from CPI are more likely than non-sufferers to experience co-morbid CMI, usually depression (Liew, 2011). Co-morbidity of CPI and CMI is associated with cumulative negative effects on children’s adjustment, such as increased stress levels (Sieh, Meijer, & Viser-Meily, 2010).
Prior research in the field of parental illness has concluded that, in comparison to adolescents with two healthy parents, adolescents with parents suffering from CPI faced an increased risk of internalizing problems (Sieh et al., 2013), including depression (Morley, Selai, Schrag,
Jahanshahi & Thomson, 2011), anxiety, and somatic complaints (Barkman et al., 2007). CMI of a parent has been linked to an increased risk for children to manifest internalizing psychopathology (Dean et al., 2010), such as anxiety and depression (Larsson, Knutsson-Medin, Sundelin & Trost von Werder, 2000). Internalizing problems during adolescence, in turn, have been associated with low quality of parent-child relationship (Buist et al., 2016).
Introducing the concept of attachment, Bowlby (1983) referred to the child’s disposition to seek contact with a specific figure that acts as a safe haven, especially in times of distress.
Accordingly, children possess an innate tendency to form close relationships with their care-givers, in order to fulfill not only physical, but also emotional needs of safety and consolation in times of anguish. However, negative life events such as chronic illness of a parent can mitigate a transition from secure to insecure attachment (Waters, Hamilton, & Winfield, 2000). Bowlby (1983) described CPI or CMI of the parent as a condition that can lead to lower quality of attachment. Although the theoretical construct of attachment was at first limited to infants’ and young children’s relationship with their caregivers, during the years, it has evolved to an ongoing process that expands to adolescence (Ainsworth, 1991; Waters et al., 2000). During adolescence, parents continue to be perceived by their children as the most important source of support and security (Nickerson and Nagle, 2005).
Research evidence on the effect of chronic illness of a parent on quality of attachment is equivocal. On the one hand, in case of a parent with CPI, Pakenham and Cox (2012) as well as
Sieh et al. (2013), did not find differences in quality of attachment between adolescents with a parent diagnosed with CPI and adolescents with two healthy parents. Further, Ireland and Pakenham (2010b) did not observe significant differences in quality of attachment among adolescents with parents suffering from CPI and those with parents suffering from CMI, as their quality of attachment was similar to that of adolescents with two healthy parents. On the other hand, numerous studies indicated that infants with mothers suffering from CMI had lower quality of attachment than infants with healthy mothers (Coyl, Roggman, & Newland, 2002; Hipwell et al., 2000). In one of the few studies including adolescents, Weinfield, Sroufe, and Egeland (2000) implemented a prospective design that examined quality of attachment from infancy to early adulthood and found that CMI of the mother was a common factor for children whose quality of attachment declined from childhood to adolescence.
Review of previous research reveals the issues that have no received sufficient attention. Studies on quality of attachment when parents suffer from CPI or CMI remain scarce, especially in comparison to the amount of studies on quality of attachment in case of a chronically ill child (i.e. Berant, Mikulincer & Shaver, 2008). Further, the different effect of CPI or CMI on quality of attachment is understudied. Research on parents with CMI is focused on infancy, and only one study has examined the different effect of CPI or CMI of the parent on quality of attachment during adolescence (Ireland & Pakenham, 2010b). Finally, despite indications that CPI with co-morbid CMI has cumulative negative effect on adolescents’ adjustment (Sieh et al., 2010), no study has examined the effect of CPI with co-morbid CMI of the parent on quality of attachment of adolescents. Adolescents may be especially vulnerable to attachment changes parental chronic illness elicits. First, the greater cognitive capacity of their age allows greater understanding of consequences of chronic illness, causing greater distress in comparison to younger children
(Pedersen & Ravensen, 2005). Second, increased maturity of adolescents leads to the adoption of more demanding responsibilities during redistribution of family roles, because care giving duties increase as children get older (Pedersen & Ravensen, 2005). Lastly, while adolescence is a period of yearning for autonomy, chronic illness of a parent keeps adolescents family-bound, creating a developmental mismatch between their need for independence and the family’s need for
commitment (Rolland, 1999). Adolescents may respond to the above with irritability towards their parents or withdrawal from the parent-child relationship (Quinn-Beers, 2001).
This study aims to investigate the following issues: First, to examine the differences in quality of attachment between adolescents with chronically ill and healthy parents, hypothesizing that adolescents of parents with CPI or CMI will report lower quality of attachment than adolescents with two healthy parents. Second, in case that our first hypothesis is supported, we will
investigate whether the type of chronic condition of the parent affects quality of attachment of adolescents differently. The moderating effect of depression level of parents suffering from CPI will be taken into account to gain insight into cumulative negative effect of CPI with co-morbid CMI on quality of attachment of adolescents. Further, the moderating effect of age will be investigated; hypothesizing that quality of attachment of older adolescents with parents suffering from CMI or CPI is lower compared to younger adolescents.
Method
Participants
This study used two data sets of participants (Data set 1 and 2). Adolescents had to be from 10 to 20 years old, living with their parents. Adolescents, who lived outside the Netherlands, were not
fluent in Dutch or suffered from cognitive disabilities or severe somatic conditions were excluded. Adolescents with mild conditions, such as asthma or dyslexia, were eligible for participation.
Data set 1 included cross-sectional data from 161 adolescents with a parent diagnosed with CPI, their ill parent, (n = 87), and the ill parent's healthy spouse, (n = 84), as well as 112 adolescents with 138 healthy parents. For this study, only participants with data available for the whole family were included: The CPI target group consisted of adolescents between 10 and 20 years old, their parent suffering from CPI and their healthy spouses from 80 families. The majority of parents with CPI were mothers (62.2%, n = 51). The most common CPI reported was multiple sclerosis (30.5%), paraplegia (6.1%), Parkinson’s disease (4.9%) and rheumatoid arthritis (4.9%). The control group included adolescents between 11 and 19 years old and their two healthy
parents from 69 families (see Figure 1). All participants were Dutch.
Figure 1. Flow chart of participants of Data set 1. Data Set 1
Parents with CPI (n = 87) Healthy spouses (n = 84) Adolescents (n = 161)
Healthy parents (n =138) Adolescents (n = 112)
Parents with CPI (n = 80) Healthy spouses (n = 80) Adolescents (n = 130)
Healthy parents (n = 138) Adolescents (n = 112)
CPI target CPI control Excluded participants
(no data for all family members):
7 parents with CPI
4 healthy spouses from the CPI target group
31 adolescents from the CPI target group
Data set 2 included a cross-sectional sample of adolescents with a parent suffering from CMI for three months minimum and a sample of adolescents with two healthy parents. This data set included information provided by children only, hence no parental demographic information in spite of diagnosis and duration of illness was available. We excluded participants who provided no data at all, cases of fictitious data and cases that only provided little initial demographic information. The final sample consisted of 148 adolescents from 145 families. Although recruitment of the target group addressed adolescents with parents suffering from CMI, 16 adolescents, corresponding to 61.5% of the target group, reported having a parent with CPI. Age of adolescents in the target group ranged from 10 to 22 years. The most common CMI reported was depression (40%) and alcohol abuse (20%). The control group consisted of adolescents between 10 and 23 years old with two healthy parents (see Figure 2). All adolescents were Dutch except for one Spanish participant belonging to the target group.
Figure 2. Flow chart of participants of the Data set 2. CMI target group
(n = 67)
CMI control group (n = 131)
Deletion of 40 cases (Remaining: n = 27)
Deletion of five cases (Remaining: n = 126)
Deletion of one case (Remaining: n = 26)
Deletion of four cases (Remaining: n = 122) Data Set 2 (N = 230)
Deletion of 32 cases that provided no data
Deletion of 45 cases of fictitious data
Deletion of five cases that provided only few demographic data
Instruments
Demographic variables. Adolescents and their parents answered questions about their gender,
age and nationality. Adolescents answered questions about their education and school performance and frequency of contact with their parents. Parents were asked about their education, living situation and income. Parents with CPI answered additional questions
concerning the nature and duration of their condition. Adolescents of parents with CMI answered additional open questions about their parents' diagnosis and duration of illness.
Depression level. Depression level of parents with CPI were measured using the Beck
Depression Inventory (BDI), a self-report questionnaire consisting of 21 items with four possible responses that range in severity of depressive symptoms (Beck, Steer, & Carbin. 1988). Those items cover behavioral symptoms (“I can work as well as before/ It takes extra effort to start something/ I have to push myself hard to do anything/ I am no more able to do something”), somatic symptoms (“I have not lost weight/ I have lost weight because I am on a diet/I have lost more than two kilos/I have lost more than four kilos/I have lost more than six kilos”) and
cognitive-affective symptoms (“I do not feel sad./I feel sad/I am sad all the time and I can’t get over it/ I am so sad that I can’t stand it”). Higher scores indicate more severe depressive
symptoms: scores lower than 10 indicate none or minimal depression, scores between 10 and 18 mild to moderate depression, scores between 19 to 29 moderate to severe depression, and scores above 30 severe depression. Chronbach’s alpha indicated good reliability for the total scale (α=.83).
Quality of attachment of adolescents. Quality of attachment of adolescents was measured using the revised Inventory of Parent and Peer Attachment (IPPA), a self-report instrument
created by Armsden and Greenberg (1987). Although the IPPA contains measures concerning attachment towards parents and peers, for this study only the parental subscales were used. The subscale trust measured the shared understanding and respect between parents and children, with items as “My mother/father respects my feelings”. The subscale communication measured quality of spoken communication, with items as “If my mother/father knows I have something, she/he asks me what”. Lastly, the subscale alienation measured the degree of perceived emotional isolation from parents or anger towards them, with items as “I am angry at my mother/father”. In the present study, adolescents answered 12 items evaluating their attachment to each parent. Each item was rated on a 4-point Likert scale ranging from 1 (almost never) to 4 (almost always), with higher scores indicating higher quality of attachment. Quality of attachment towards parents was calculated by the sum of the paternal and maternal subscales. Chronbach’s alpha for the IPPA was satisfactory for the mother scale (α =.80), father scale (α =.80) and total scale (α =.85).
Procedure
Data set 1 (CPI target and control group) had already been collected by Sieh et al. (2012) for a cross-sectional study concerning the impact of parental CPI on adolescents’ functioning. Areas of recruitment included rehabilitation centers, hospitals, schools, community centers, general health practitioner’s offices, websites of patients’ organizations and public libraries. Potential
participants and recruiters were informed about the study by posters and brochures, as well as orally by professionals of the institutions that participated in the study. Research assistants, trained by Sieh according to the research protocol, visited families who volunteered to participate and met the inclusion criteria. After informed consent had been signed, self-report questionnaires were administered to adolescents and their parents. As a reward for their participation,
approved by the ethical commission of the research Institute of Child Development and Education of the University of Amsterdam on December 2nd, 2008.
Data set 2 (CMI target and control group) was collected by master and bachelor psychology students of Leiden University. Flyers stating the aim and design of the study were distributed in mental health institutions, schools and university locations, hospitals and public places such as libraries. Families interested in participating contacted the researchers via e-mail or phone. Data collection was conducted electronically. After signing written consent, adolescents filled in online questionnaires. As a reward for their participation, adolescents entered a lottery, having a chance to win one of four gift cards of 25€ value each (see Figure 3). The second study was approved by the ethics committee of the Institute of Psychology at Leiden University on February 10th, 2014.
Figure 3. Procedure of recruitment.
Data Set
Data set 1 (CPI target and control group) Data set 2 (CMI target and control group)
Families contacted the researchers
Administration of questionnaires Online questionnaires
Reward for adolescents: lottery Participants informed by announcements Participants informed by announcements
Trained research assistants visited families
Informed consent signed
Reward for all adolescents
Data analysis
Preliminary analyses. Missing data analysis was conducted for all groups of participating
parents and adolescents. The percentage of missing data was below 5%, so no statistical correction was needed (Buhi, Goodson, & Neilands, 2008), except for IPPA scores (6.6%). Multiple imputation was conducted for IPPA scores, using IPPA items and total score as predictors (Janssen et al., 2010).
Z- Scores were calculated as a function of the raw total scores of quality of attachment for target groups of adolescents, using the mean of the corresponding control group as a reference (DeShea & Toothaker, 2015).
In order to examine pre-existing differences between target and control groups of parents and adolescents, age, gender, GPA and education of adolescents and living situation, years living with their spouse, number of children living at home, income, education, depression levels and age of parents were compared. Analyses were conducted using IBM SPSS version 20. All significance tests were two tailed, using an alpha level of .05.
Main analyses. To determine whether quality of attachment differed between target and
control groups, Kruskall-Wallis test with planned contrasts was conducted for the total IPPA scale as well as the three subscales of trust, communication, and alienation. To control for the effect of economic status and gender of parent with CPI on quality of attachment, (Gorrese & Ruggieri, 2012; Rawatlal, Pillay & Kliewer, 2014), an additional ANCOVA was conducted for Data set 1 (CMI target and control group) with family income and gender of parent as covariates (Mertler & Vannatta, 2002). Family income was coded from 1 (less than 1000 euro) to 8 (4.000 and more). A parametric test was used as the Central Limit Theorem indicates that for large
sample sizes, as in case of Data set 1, data tend to be normally distributed (Field, 2009).
Bootstrap was used as an additional way to ensure robust analysis (Wilcox, 2009). To control for the effect of gender of adolescents on quality of attachment (Gorrese & Ruggieri, 2012), an ANCOVA was conducted for Data sets 1 and 2 using gender of adolescents as a covariate. Bootstrap was used to ensure robust analysis (Wilcox, 2009).
Moderation analysis was conducted to investigate the effect of age on the relationship between parental health condition and quality of attachment of adolescents (Bauman, Sallis,
Dzewaltowski, & Owen, 2002; Stone-Romero & Liakhovitski, 2002). Furthermore, the
moderating effect of depression level of parents with CPI was investigated by taking into account their BDI scores (see Figure 4). Parental health condition was coded as 1 (parent with CPI or CMI accordingly) or 0 (two healthy parents). Bootstrap was used to ensure robust analysis (Wilcox, 2009). Moderation analyses were conducted using the PROCESS (version 2.15) macro for SPSS created by Hayes (2016).
Moderation 1: Moderating effect of age (Data set 1).
Moderation 2: Moderating effect
Age of adolescents (years)
Parental health condition:
Parent with CPI
Two healthy parents
Quality of attachment of adolescents (IPPA scores)
Moderation 2: Moderating effect of depression level of parents with CPI (Data set 1).
Moderation 3: Moderating effect of age
Moderation 3: Moderating effect of age (Data set 2).
Figure 4. Moderation analyses
Results
Differences in demographic characteristics between groups of parents
Parents of the CPI target group (ill parents and their healthy spouses) and control group (two healthy parents) did not differ in age, education, living situation, years living with their spouse and number of children living with them. There were no differences in gender composition
Depression levels of parents with CPI (BDI scores)
Parental health condition: • Parent with CPI • Two healthy parents
Quality of attachment of adolescents (IPPA scores)
Parental health condition: • Parent with CMI • Two healthy parents
Quality of attachment of adolescents (IPPA scores)
between the CPI target and control group. Within the CPI target group the majority of ill parents were mothers, p < .01. Families of the CPI target group had lower income than families of the CPI control group, p < .05. Post hoc comparisons revealed that mean depression level of parents with CPI was higher than mean depression level of their healthy spouses and parents of the CPI control group, p < .01. Mean depression level of healthy spouses of the CPI target group was higher than mean depression level of parents of the CPI control group, p < .01. For the majority of parents depression levels ranged from no depression to mild depression. Parents suffering from CPI exhibited severe depression levels more frequently than their healthy spouses or healthy parents of the control group (see Table 1).
Differences in demographic characteristics between groups of adolescents.
Adolescents of the CMI target and control group were older than adolescents of the CPI target and control group, p < .01. Adolescents of the CMI target group were older than adolescents of the CMI control group, p < .01. The CMI target and control group included more girls than the CPI target and control group, p < .01. Adolescents of the CPI target and control group had lower GPA than adolescents of the CMI target and control group, p < .01. The majority of adolescents in the CPI target and control group were following secondary education, whereas the majority of adolescents in the CMI target and control group were following university education.
Differences in quality of attachment between groups of adolescents. Z-Scores for quality of
attachment of adolescents of the CPI (z = -.06) and CMI target group (z = -.01) did not deviate from the control group's standard deviation, indicating that quality of attachment scores were similar to those of the control groups. Contrary to our hypothesis, quality of attachment of adolescents did not differ between adolescents of parents with CPI, CMI or healthy parents, for
the IPPA total score as well as for the subscales trust, communication and alienation, p >.05. Differences in quality of attachment of adolescents remained non-significant after controlling for family income and gender of the ill parent in the CPI sample as well as for gender of adolescents in the CPI and CMI sample, both for the IPPA total score and the three subscales.
Table 1.
Differences between groups of parents and adolescents
CPI target CPI control CMI target CMI control Parents Ill parent Healthy spouse
Mean age in years (SD) 46.84 (5.55) 47.48 (5.80) 47.30 (5.07) - - Mean children at home (SD) 2.09 (1.3) 2.03 (.97)
Mean depression level (SD) 12.65 (7.99) 7.05 (6.50) 4.11 (3.94) - -
None/Minimal 50.2% 78.6% 87.6% - -
Mild 30.1% 15.1% 5.7% - -
Moderate 16.5% 3.8% 0.7% - -
Severe 3.8% 1.3% - - -
Gender (female) 62.2% 31.3% 50% - -
Living with partner/ married 92.6% 93.7% 94.9% - - Mean years with spouse (SD) 20.93 (4.99) 20.29 (6.30) - -
Family income (euro) -
1000 to 1999 14.1% 7.2% - - 2000 to 2999 44.9% 30.4% - 3000 to 3999 21.8% 35.5% - - 4000 or more 19.2% 24.6% - - Education Primary education 9.9% 14% 6.6% - - Secondary education 39.5% 38% 34.7% - - (Pre)university education 45.7% 46.9% 55.5% - - Adolescents
Mean age in years (SD) 14.58 (2.39) 14.50 (2.25) 18.75 (2.05) 18.02 (2.78) Mean GPA (SD) 6.92 (0.86) 7.23 (0.77) 6.85 (0.64) 6.94 (0.78) Mean quality of attachment (SD) 78.24 (10.69) 78.77 (8.77) 77.81 (14.30) 77.62 (11.59) Mean communication (SD) 21.98 (5.23) 22.31 (4.75) 22.92 (5.53) 21.83 (5.27) Mean trust (SD) 27.55 (3.99) 27.93 (3.22) 27.19 (5.34) 27.47 (4.25) Mean alienation (SD) 28.67 (3.30) 28.71 (2.75) 27.69 (4.82) 28.05 (3.36) Gender (female) 53.8% 50.8% 84.6% 79.5% Education Primary education 12.6% 12.3% 2.9% 5.3% Secondary education 71.6% 72% 42.7% 13.7% (Pre)university education 11.5% 13.1% 45.8% 74.8% Note: SD= Standard Deviation
Moderators on quality of attachment in case of a parent with CPI.
The model investigating the effect of parental condition on quality of attachment while taking
into account depression level of the CPI parent as a moderator did not contain a significant predictor, F (3, 229) = .29, p > .05.
The overall model investigating parental health condition on quality of attachment of
adolescents while taking into account the age of adolescents as a moderator was significant, F (3, 236) = 11.41, p < .01, meaning that the overall model had predictive ability on quality of
attachment of adolescents. However, it explained only 11% of the variance in quality of
attachment of adolescents, R2 = .11. Parental health condition was not a significant predictor of quality of attachment of adolescents and no moderating effect was observed. There was a direct effect of adolescent age on quality of attachment, b = -1.40, t = - 5.52, p < .01, with quality of attachment decreasing as adolescents got older (see Table 2).
Table 2.
Moderating effect of depression levels of parents and adolescent age on quality of attachment of adolescents. b Coefficient
(Standardized)
Standard Error
95% Confidence Intervals Lower Bound Upper Bound Depression level
Constant 78.8 .69 77.3 80.1
Depression levels .01 .09 -.17 .18
Parental health condition -.83 1.37 -.60 .54
Parental health condition*depression level .30 .18 -.04 .66 Adolescent age
Constant 78.5 .61 77.3 79.7
Adolescent age -1.40* .25 -1.90 -.90
Parental health condition -.45 1.19 -2.80 1.90
Adolescent age*parental health condition .43 .50 -.56 1.42
Moderators on quality of attachment in case of a parent with CMI.
The overall model investigating parental health condition on quality of attachment of adolescents while taking into account adolescent age as a moderator was significant, F (3, 139) = 4.86, p < .01, meaning that the overall model has a predictive ability on quality of attachment of
adolescents. However, it explained only 4% of the variance in quality of attachment of adolescents, R2 = .04. Parental health condition was not a significant predictor of quality of attachment of adolescent and no moderating effect was observed. There was a direct effect of adolescent age on quality of attachment of adolescents, b = -.95, t = -2.93, p < .01, with quality of attachment of adolescents declining with increasing age (see Table 3).
Table 3.
Moderating effect of adolescent age on quality of attachment of adolescents b Coefficient
(Standardized)
Standard Error
95% Confidence Intervals Lower Bound Upper Bound
Constant 77.63 .99 75.70 79.59
Adolescent age -0.95* .32 -1.59 -.31
Parental health condition -2.11 2.93 -7.91 3.69
Adolescent age*parental health condition 1.52 .90 -.26 3.30
Note: *p < .01
Discussion
The first aim of this study was to investigate the effects of chronic physical illness (CPI) and chronic mental illness (CMI) of the parent on quality of attachment of adolescents. Results suggested that there were no differences between adolescents with chronically ill and healthy parents in (subscales of) quality of attachment. These findings remained after controlling for the effect of family income and gender of the ill parent in case of parents with CPI, as well as after
controlling for the effect of gender of adolescents in case of parents with CPI and CMI. The second aim of this study was to investigate the moderating effect of depression level of parents suffering from CPI, as well as the moderating effect of adolescent age on quality of attachment of adolescents with parents suffering from CPI or CMI. Depression level of parents with CPI did not moderate the relationship between parental health condition and quality of attachment of
adolescents. Similarly, no moderating effect of age was observed for adolescents with parents suffering from CPI or CMI. However, age significantly predicted quality of attachment of adolescents regardless of parental health condition, with quality of attachment decreasing as adolescents got older.
Contrary to our hypothesis as well as Bowlby's (1986) theory, adolescents of parents who suffer from CPI or CMI do not differ in their quality of attachment and level of communication, trust, and alienation from adolescents with two healthy parents. Adolescence is an age of increased cognitive capacity and maturity (Pedersen & Revenson, 2005), which leads to different attachment relationships with caregivers (Waters & Cummings, 2000). Contrary to childhood, during which physical proximity is necessary to ensure high quality of attachment, quality of attachment of adolescents is based on perceived support of the parental figure irrespective of physical availability (Armsden & Greenberg, 1987). Thus, chronic illness may limit physical availability of the parents (for example, due to hospitalization of ill parents or increased duties of their healthy spouses), but not their ability to be perceived as a source of support in case of need by their adolescent children. On the one hand, the finding that CPI of the parent does not affect quality of attachment of adolescents is in line with prior research of Ireland and Pakenham (2013, 2010b). This finding points towards the direction of research supporting that having a parent
suffering from chronic illness is not always associated with adverse outcomes for adolescents. Ireland and Pakenham (2012) noted that CMI or CPI of the parent can also have positive outcomes for adolescents, for instance, strengthened interpersonal relationships and increased maturity. It is possible that quality of attachment mediates the relationship between chronic illness of the parent and positive outcomes for adolescents. To support this view, Gerber (2010) considered high quality of attachment to be one of the possible mechanisms of the development of resilience in adolescence. However, this study expected that chronic illness of a parent would negatively affect quality of attachment of adolescents; this is why resilience was not examined. On the other hand, in case of parental CMI, our results contradict previous research of Coyl, Roggman and Newland (2002) and Hipwell et al. (2000), who concluded that CMI of the mother negatively affected quality of attachment of children. However, those studies measured
attachment towards mentally ill mothers in a sample of infants. Participating children in this study were adolescents, an age group significantly different from infants in terms of
developmental and attachment needs (Waters & Cummings, 2000). Also, quality of attachment referred to mothers and fathers as well as both the mentally ill and healthy parent. Although attachment in infancy is focused on the primary caregiver, usually the mother, as children grow older, attachment bonds are formed with both parents (Ainsworth, 1991). In case of adolescents with chronically ill parents, attachment towards the healthy parent may outweigh experiences that could negatively affect quality of attachment (for instance, inability of the ill parents to become actively involved with their children because of their symptoms).
Moderation analysis suggested that depression levels of parents with CPI did not affect quality of attachment of adolescents. This finding supports the results already presented, as CMI of the
parents does not appear to predict quality of attachment of adolescents. However, this finding contradicts previous research indicating that parental CPI with co-morbid mental illness has cumulative negative effects on adjustment of their adolescent children (Sieh et al., 2012). It should be noted that the majority of parents with CPI exhibited mild to moderate depression level. Low score range in depression level may have made the moderating effect undetectable, as, according to Stone-Romero and Liakhovitski (2002), restricted range of a moderator greatly obscures its observed effect due to low statistical power. Present findings indicate that low to moderate levels of depression in parents with CPI do not affect quality of attachment of adolescents, but it remains to be investigated whether severe depression level in combination with CPI of the parent have cumulative negative effect on quality of attachment of adolescents. Concerning the moderating effect of age, contrary to our hypothesis, age of adolescents did not moderate the relationship between parental health condition and quality of attachment. This finding contradicts Rollands’ theory (1999), according to which increased responsibilities and familial needs that chronic illness of a parent brings along conflict adolescents' need for
independence, negatively affecting quality of attachment with their parents. Research by Sieh et al. (2012) and Pakenham and Cox (2010b) confirmed that adolescents of parents suffering from chronic illness face greater familial duties and increased care giving responsibilities in
comparison to adolescents with two healthy parents. Yet, Pakenham and Cox (2012) noted that increased care giving responsibilities of adolescents with parents suffering from chronic illness favored their quality of attachment, possibly because of the formation of close family bonds. Hence, it is possible that this positive aspect of care giving counterbalances experiences that may clash with adolescents’ need for independence, helping them to maintain high quality of
attachment in case of parental chronic illness may have been more pronounced if both children and adolescents were included (Romero & Liakhovitski, 2002), as it is possible that quality of attachment in case of parental chronic illness varies across different stages of children’s development.
Although no moderating effect of the age of adolescents in the relationship between parental health condition and quality of attachment of adolescents was observed, age significantly predicted quality of attachment of adolescents regardless of parental condition, with quality of attachment decreasing as adolescents got older. This corresponds to prior research of Nickerson and Nagle (2005), who reported decreasing attachment to parents for pre-adolescents, as
communication and trust towards parents is progressively substituted by communication and trust towards peers. The findings of this study indicate that quality of attachment to parents also
decreases for children in early and late adolescence. Probably, this is a function of the
replacement of various aspects of parental attachment by peer attachment (Nickerson & Nagle, 2005). This finding may seem to contradict our results indicating that adolescents of parents who suffer from CPI or CMI maintain high quality of attachment with them. However, quality of attachment is differently defined according to children’s age and developmental stage (Waters & Cummings, 2000). During adolescence, a decrease in attachment to parents is expected as part of normal psychosocial development (Ainsworth, 1991). Adolescents of parents suffering from chronic illness seem to follow the attachment pattern that their age and developmental stage predict, similar to that of their peers with two healthy parents (Ainsworth, 1991).
with parents suffering from a CMI, the majority of participating adolescents of this group reported a CPI diagnosis of their parent. Possibly, adolescents misinterpreted the term
“diagnosis” of the questionnaire, referring to their parents’ physical symptoms although CMI may also have been present or predominant. This led to a heterogeneous sample of children whose parents suffered from CMI, CPI, or maybe both. The heterogeneity of the target group of
adolescents with parents suffering from CMI in combination with its small sample size may have obscured the impact of parental CMI on quality of attachment of adolescents, resulting in a Type II error (Field, 2009). Secondly, as more than one adolescent from each family participated, dependence of subjects resulting to family clustering is another factor that limits the conclusions of this study. However, it should be noted that children of the same family often differ in their quality of attachment with their parents, as attachment is a result of parent-child interaction that can vary significantly among siblings (Rauer & Volling, 2007). Finally, the sample consisted of parents with high socioeconomic status. As higher socioeconomic status has been shown to predict better quality of attachment (Rawatlal, Pillay, & Kliewer, 2014), the generalizability of our findings is limited. Parents of low socioeconomic status suffering from chronic illness may have limited access to medical care and support (Sieh et al., 2013). Thus, their adolescents may be exposed to more intense parental symptoms leading to lower quality of attachment.
In the future, the present study could be replicated using a more homogeneous sample of adolescents whose parents suffer exclusively from CMI. Further, children ranging from school-age to late adolescence could be compared in order to investigate the effect of school-age on quality of attachment using a moderator with adequate levels to ensure statistical power (Stone-Romero & Liakhovitski, 2002), or a longitudinal design could be implemented in order to examine the
attachment patterns of children with parents suffering from chronic illness across their developmental stages until they reach late adolescence. Further, moderators that research has shown to be relevant for children’s attachment but were not part of this study could be examined, such as quality of attachment to the healthy parent, duration of parental symptoms or care giving experiences of children (Ireland & Pakenham, 2010; Rawatlal, Pillay & Kliewer, 2014).
Multilevel analysis should be used to deal with dependence of siblings’ scores due to family clustering (Sieh et al., 2010). Alternatively, only one child per family could be used in the analyses. Using a more diverse sample of parents regarding education and income could help investigate quality of attachment of adolescents coming from families where chronic illness is combined with low socioeconomic status, increasing the generalizability of the findings. Finally, in line with research of Ireland and Pakenham (2012), the mediating role of high quality of attachment when parents suffer from CPI or CMI on positive outcomes for adolescents could be investigated.
In conclusion, this study increases our understanding concerning quality of attachment of adolescents who have parents suffering from CPI or CMI. Additionally, it is one of the few studies providing information about the moderating effect on quality of attachment of adolescents of adolescent age in case of CPI or CMI and depression level in case of CPI. Families should be informed that CMI or CPI of a parent does not necessarily result in low quality of attachment of adolescents, and that adolescents with a parent suffering from chronic illness that exhibit lower levels of communication and trust and higher levels of alienation (Armsden & Greenberg, 1987) than their peers with healthy parents are not the rule, but the exception.
References
Ainsworth, M. D. S., (1991). Attachments and other affectional bonds across the life cycle. In C. M. Parkes, S. Stevenson-Hinde, & S. Marris (Eds), Attachment across the life cycle (33-52). New York: Routledge.
Allen, J. P., Marsh, P., McFarland, C., McElhaney, K. B., Land, D. J., Jodl, K. M., & Peck, S. (2002). Attachment and autonomy as predictors of the development of social skills and delinquency during midadolescence. Journal of Consulting and Clinical Psychology, 70, 56-66. doi: 10.1037/0022-006X.70.1.56
Allen, J. P., McElhaney, K. B., Land, D. J., Kuperminc, G. P., Moore, C. W., O'Beirne–Kelly, H., & Kilmer, S. L. (2003). A secure base in adolescence: Markers of attachment security in the mother–adolescent relationship. Child Development, 74, 292-307.
doi:10.1111/1467-8624.t01-1-00536
Armsden, G. C., & Greenberg, M. T. (1987). The inventory of parent and peer attachment: Individual differences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence, 16, 427-454. doi: 10.1007/BF02202939
Barkmann, C., Romer, G., Watson, M., & Schulte-Markwort, M. (2007). Parental physical illness as a risk for psychosocial maladjustment in children and adolescents: epidemiological findings from a national survey in Germany. Psychosomatics, 48, 476-481.
doi:10.1176/appi.psy.48.6.476
Bauman, A. E., Sallis, J. F., Dzewaltowski, D. A., & Owen, N. (2002). Towards a better understanding of the influences on physical activity: The role of determinants, correlates,
causal variables, mediators, moderators, and confounders. American Journal of Preventive Medicine, 23, 5-14. doi:10.1016/S0749-3797(02)00469-5
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: twenty-five years of evaluation. Clinical Psychology Review, 8, 77-100. doi:10.1016/0272-7358(88)90050-5
Berant, E., Mikulincer, M., & Shaver, P. R. (2008). Mothers' attachment style, their mental health, and their children's emotional vulnerabilities: A 7 year study of children with congenital heart disease. Journal of Personality, 76, 31-66. doi:10.1111/j.1467-6494.2007.00479.x
Bowlby, J. (1983). Attachment and loss (2nded.). New York: Basic Books.
Buhi, E. R., Goodson, P., & Neilands, T. B. (2008). Out of sight, not out of mind: Strategies for handling missing data. American Journal of Health Behavior, 32, 83-92. doi:
10.5993/AJHB.32.1.8
Buist, K. L., Verhoeven, M., Hoksbergen, R., ter Laak, J., Watve, S., & Paranjpe, A. (2016). Associations of perceived sibling and parent-child relationship quality with internalizing and externalizing problems: Comparing Indian and Dutch early adolescents. The Journal of Early
Adolescence. Advance online publication. doi: 10.1177/0272431616653473
Chan, M. (2011, April 27).The rise of the noncommunicable diseases: An impending disaster. Opening remarks at the WHO global forum, Moscow, Russian Federation. Retrieved from http://www.who.int/dg/speeches/2011/global_forum_ncd_20110427/en/
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159. doi: 10.1037/0033-2909.112.1.155
Coyl, D. D., Roggman, L. A., & Newland, L. A. (2002). Stress, maternal depression, and negative mother–infant interactions in relation to infant attachment. Infant Mental Health Journal, 23, 145-163. doi:10.1002/imhj.10009
Dean, K., Stevens, H., Mortensen, P. B., Murray, R. M., Walsh, E., & Pedersen, C. B. (2010). Full spectrum of psychiatric outcomes among offspring with parental history of mental disorder. Archives of General Psychiatry, 67, 822-829.
doi:10.1001/archgenpsychiatry.2010.86
DeShea, L., & Toothaker, L. E. (2015). Introductory statistics for the health sciences. New York: CRC Press.
Field, A. (2009). Discovering statistics using SPSS (3rd Ed). London: Sage.
Gorrese, A., & Ruggieri, R. (2012). Peer attachment: A meta-analytic review of gender and age differences and associations with parent attachment. Journal of Youth and
Adolescence, 41(5), 650-672.
Hartney, E., & Barnard, D. K. (2015). A framework for the prevention and mitigation of injury from family violence in children of parents with mental illness and substance use
problems. Aggression and Violent Behavior, 25, 354-362. doi:10.1016/j.avb.2015.09.004
Hayes, A. F. (2016). The PROCESS macro for SPSS and SAS. Retrieved from http://www.processmacro.org/index.html
Hipwell, A. E., Goossens, F. A., Melhuish, E. C., & Kumar, R. (2000). Severe maternal
psychopathology and infant–mother attachment. Development and Psychopathology, 12(02), 157-175.
Hooper, L. M. (2007). The application of attachment theory and family systems theory to the phenomena of parentification. The Family Journal, 15, 217-223.
doi:10.1177/1066480707301290
Ireland, M. J., & Pakenham, K. I. (2010a). The nature of youth care tasks in families
experiencing chronic illness/disability: Development of the Youth Activities of Caregiving Scale (YACS). Psychology and Health, 25, 713-731. doi:10.1080/08870440902893724
Ireland, M. J., & Pakenham, K. I. (2010b). Youth adjustment to parental illness or disability: the role of illness characteristics, caregiving, and attachment. Psychology, Health &
Medicine, 15, 632-645. doi:10.1080/13548506.2010.498891
Larsson, B., Knutsson-Medin, L., Sundelin, C., & von Werder, A. T. (2000). Social competence and emotional/behavioural problems in children of psychiatric inpatients. European Child &
Adolescent Psychiatry, 9, 122-128. doi:10.1007/PL00010708
Liew, H. P. (2011). Depression and chronic illness: A test of competing hypotheses. Journal of Health Psychology, 17, 100-109. doi:1359105311409788.
Martin, C. M. (2007). Chronic disease and illness care: adding principles of family medicine to address ongoing health system redesign. Canadian Family Physician, 53(12), 2086-2091.
Meeus, W., Oosterwegel, A., & Vollebergh, W. (2002). Parental and peer attachment and identity development in adolescence. Journal of Adolescence, 25, 93-106.
doi:10.1006/jado.2001.0451
Mehta, A., Cohen, S. R., & Chan, L. S. (2009). Palliative care: a need for family systems approach. Palliative and Supportive Care, 7, 235-243. doi:10.1017/S1478951509000303
Morley, D., Selai, C., Schrag, A., Jahanshahi, M., & Thompson, A. (2011). Adolescent and adult children of parents with Parkinson's disease: incorporating their needs in clinical
guidelines. Parkinson’s Disease, 2011. doi: 10.4061/2011/951874
Mertler, C. A., & Vannatta, R. A. (2002). Advanced and multivariate statistical methods. Los Angeles, CA: Pyrczak.
Nickerson, A. B., & Nagle, R. J. (2005). Parent and peer attachment in late childhood and early adolescence. The Journal of Early Adolescence, 25, 223-249.
doi: 10.1177/0272431604274174
Oyserman, D., Bybee, D., Mowbray, C., & Hart-Johnson, T. (2005). When mothers have serious mental health problems: Parenting as a proximal mediator. Journal of Adolescence, 28, 443-463. doi: 10.1016/j.adolescence.2004.11.004
Oyserman, D., Mowbray, C. T., Meares, P. A., &Firminger, K. B. (2000). Parenting among mothers with a serious mental illness. American Journal of Orthopsychiatry, 70, 296-315. doi:10.1037/h0087733
Pakenham, K. I., & Cox, S. (2012). The nature of caregiving in children of a parent with multiple sclerosis from multiple sources and the associations between caregiving activities and youth adjustment overtime. Psychology & Health, 27, 324-346. doi:
10.1080/08870446.2011.563853
Pakenham, K. I., & Cox, S. (2014). The effects of parental illness and other ill family members on the adjustment of children. Annals of Behavioral Medicine, 48, 424-437.
doi:10.1007/s12160-014-9622-y
Parsa, N., Yaacob, S. N., Redzuan, M. R., Parsa, P., & Esmaeili, N. S. (2014). Parental attachment, inter-parental conflict and late adolescent's self-efficacy. Asian Social Science, 10, 123-131. doi: 10.5539/ass.v10n8p123
Pedersen, S., & Revenson, T. A. (2005). Parental illness, family functioning, and adolescent well-being: A family ecology framework to guide research. Journal of Family Psychology, 19, 404-419. doi: 10.1037/0893-3200.19.3.404
Phillips, F., & Lewis, F. M. (2015). The adolescent’s experience when a parent has advanced cancer: Qualitative inquiry. Palliative Medicine,29, 851-858. doi:
10.1177/0269216315578989
Quinn-Beers, J. (2001). Attachment needs of adolescent daughters of women with cancer. Journal of Psychosocial Oncology, 19, 35-48. doi:10.1300/J077v19n01_03
Rauer, A. J., & Volling, B. L. (2007). Differential parenting and sibling jealousy: Developmental correlates of young adults’ romantic relationships. Personal Relationships, 14, 495-511. doi:10.1111/j.1475-6811.2007.00168.x
Rawatlal, N., Pillay, B. J., & Kliewer, W. (2015). Socioeconomic status, family-related variables, and caregiver–adolescent attachment. South African Journal of Psychology, 45, 551-563. doi:10.1177/0081246315588906
Rolland, J. S. (1999). Parental illness and disability: A family systems framework. Journal of
Family Therapy, 21, 242-266. doi:10.1111/1467-6427.00118
Sheehan, G., & Noller, P. (2002). Adolescent’s perceptions of differential parenting: Links with attachment style and adolescent adjustment. Personal Relationships, 9, 173-190.
doi:10.1111/1475-6811.00011
Sieh, D. S., Visser-Meily, J. M. A., & Meijer, A. M. (2013). Differential outcomes of
adolescents with chronically ill and healthy parents. Journal of Child and Family Studies, 22, 209-218. doi: 10.1007/s10826-012-9570-8
Sieh, D. S., Visser-Meily, J. M. A., Oort, F. J., & Meijer, A. M. (2012). Risk factors for problem behavior in adolescents of parents with a chronic medical condition. European Child &
Adolescent Psychiatry, 21, 459-471. doi:10.1007/s00787-012-0279-4
Stone-Romero, E. F., & Liakhovitski, D. (2002). Strategies for detecting moderator variables: A review of conceptual and empirical issues. Research in Personnel and Human Resources Management, 21, 333-372.
Wan, M. W., & Green, J. (2009). The impact of maternal psychopathology on child–mother attachment. Archives of Women's Mental Health, 12, 123-134. doi: 10.1007/s00737-009-0066-5
Waters, E., Hamilton, C. E., &Weinfield, N. S. (2000). The stability of attachment security from infancy to adolescence and early adulthood: General introduction. Child Development, 71, 678-683.doi:10.1111/1467-8624.00175
Weinfield, N.S., Sroufe, L.A., & Egeland, B. (2000). Attachment from infancy to early adulthood in a high risk sample: Continuity, discontinuity, and their correlates. Child Development, 71, 695-702. doi: 10.1111/1467-8624.00178
Wikman, A., Wardle, J., & Steptoe, A. (2011). Quality of life and affective well-being in middle-aged and older people with chronic medical illnesses: A cross-sectional population based study. PLoS One, 6, 1. doi:10.1371/journal.pone.0018952
Wilcox, R. R. (2009). Robust ANCOVA using a smoother with bootstrap bagging. British Journal of Mathematical and Statistical Psychology, 62, 427-437. doi:
10.1348/000711008X325300
World Health Organization (2014). Non-communicable diseases global monitoring framework: indicator definitions and specifications. Retrieved from
