Impact of Family Caregiving by Youth on Their Psychological Well-Being: A Latent Trait Analysis
- First Online:
- Cite this article as:
- Cohen, D., Greene, J.A., Toyinbo, P.A. et al. J Behav Health Serv Res (2012) 39: 245. doi:10.1007/s11414-011-9264-9
- 297 Views
Secondary data analyses were conducted on a survey dataset from 1,281 middle school students to analyze the impact of family caregiving on self-reports of psychological well-being using the Multiple Indicators, Multiple Causes (MIMIC) model. Factor analysis resulted in four latent factors underlying psychological functioning, and the MIMIC model revealed significant caregiver effects on three factors: anxiety/depression, engaged coping, and disengaged coping, but not life satisfaction. Youth caregivers, especially those living with the care recipient, reported significantly higher anxiety/depression and a greater use of both coping styles compared to non-caregivers. Caregiving has a negative influence on the emotional well-being of youth with dual student–caregiver roles. The utilization of more coping strategies may reflect needing to try many approaches to school/family stressors because supports and experience are limited. Research to clarify how caregiving mediates the behavioral health and academic success of youth and also impacts care recipients and the family is warranted.
Family caregiving has become a normative experience in the twenty-first century. People are living longer, the challenges of managing relatives with dementias and other chronic illnesses at home are increasing, the length of hospital stays is declining, the number of families caring for injured and disabled military personnel is growing, and economic pressures have eroded the disposable income of families to hire supportive services. However, in spite of over 40 years of research and increasingly higher public awareness about adult family caregivers,1,2 a large group of informal care providers has gone unnoticed in the USA: youth caregivers who are children and adolescents under 18 years of age.3,4 A large share of family caregiving responsibilities has shifted onto this age group because of several demographic trends, including decreasing family size, geographic dispersion of families, high divorce rates, an increasing number of single parents, as well as multiple marriages and reconstituted families.5,6 All these trends have created care gaps which are being met by youth who are often the only ones available.5
The USA lags behind research and policy initiatives in the United Kingdom (UK) and Australia, where several surveys of youth in caregiving programs have revealed several consistent findings: children with an average age of 12–13 years provide substantial unpaid family caregiving, including health and personal care, emotional support, household tasks, and handling emergencies.7 Those with extensive involvement experience adverse effects on their health, family relationships, connections with peers and the broader community, educational performance, and future career options.7–10
The exact prevalence of youth caregivers in the USA is unknown. The only national survey conducted in 2004 estimated there were at least 1.4 million youth caregivers between the ages of 8 and 18 years who provided substantial caregiving support to family members, mostly grandparents and parents.11 The results of several other US studies suggest that the number of caregiving youth is much higher.12,13 A 2009 survey estimated there were 37 million households with a caregiver in which 37% (14 million) had children in the household, but no information was given about the roles children played.13
The only other large-scale US study of youth caregivers was a 2002 survey of 11,029 students in 54 middle and high schools in Palm Beach County, Florida who answered questions about school work and family health.3 A total of 6,210 (56.3%) students reported that they provided hands-on care for a family member with a health condition, and more than two thirds (4,166) reported that assisting with their relative’s care had a negative impact on their academic performance. More than 35% reported that it interrupted their ability to study, and 25% said they were unable to complete homework assignments. Over 18% of student caregivers reported more than one negative effect on their academic performance.
The consequences of family caregiving for child development and adjustment are not well understood. Some studies show that youth caregivers may acquire increased confidence, maturity, empathy, self-esteem, and resilience.8,14,15 Being involved in caregiving also appears to promote positive perceptions of self-identity, self-worth, and create a sense of belonging.16 However, most study results emphasize that caregiving appears to have a negative impact on many youth, affecting emotional and physical health, educational achievement, psychosocial adjustment, and overall well-being.3–10 The long-term effect on adjustment and development has not been studied.
The 2004 national survey10 and the Palm Beach County survey3 only reported descriptive statistics about prevalence of youth caregivers and impact based on age, gender, race, and caregiver status. No multivariate quantitative analyses were done comparing caregiving and non-caregiving populations, risk factors for negative consequences of caring, mediators of the relationship of level of responsibility to health, educational problems, and psychosocial impact. Likewise, most research in the UK and Australia has been qualitative studies with youth caregiver groups, and the few quantitative studies of caregiving youth and non-caregiving peers have used relatively small samples.17,18
The objective of this study was to conduct a secondary data analysis of a school-based survey to examine the reported psychological and behavioral functioning in a sample of middle school students who were caregivers for relatives at home and outside their home compared to non-caregiving peers. There were two hypotheses: (1) caregiving would have a negative impact on youth caregivers’ emotional health with respect to three self-reported outcome variables—symptoms of anxiety and depression, coping styles in response to stress, and life satisfaction; and (2) youth caring for relatives in the home would have greater adverse effects on these outcome measures compared to youth caring for relatives outside the home.
This research involved secondary data analysis of a dataset with no personal identifiers. The study was reviewed by the University of South Florida Institutional Review Board and approved as an exempt study.
The dataset included items administered in a survey of 1,281 students, ages 10 to 14 years, from two middle schools in one Florida school district. The two schools were selected by staff from the school district’s Department of Intervention Services and the American Association of Caregiving Youth to be the first schools to participate in a school-based caregiver intervention program, the Caregiving Youth Project (CYP).3 The two schools were chosen because of the high prevalence of caregiving observed in the 2002 Palm Beach County survey,3 organizational culture, and general impression of student needs. The Director of Student Intervention Services and the CYP Project Director met with school principals who then convened a meeting seeking the approval of school staff members responsible for providing resources or referrals to students who demonstrate academic, social, or behavioral needs.
The school district’s Department of Research and Accountability sent letters to parents/guardians describing the CYP, requesting permission for their child to participate in a computer-based survey to determine eligibility for CYP services, and giving parents the choice of opting out if they disapproved. Students were asked to assent to the computer survey. Parent/guardians of the caregiving youth who were invited to participate in the CYP signed a consent form permitting participation in CYP services, but student participation in all CYP activities was voluntary.
The survey included demographic characteristics (three items) and caregiver variables (four items) as well as measures of anxiety/depression (six items), coping (six items), and life satisfaction (five items).
Items included grade, gender, and race/ethnicity.
Students were asked whether they provided assistance to a family member who needed special care because of a medical condition, advanced age, or a disability, and answers could be one of three possibilities: yes, and I live with that person; yes, but I do not live with that person; or no, I am not a caregiver. Those who self-identified as providing assistance were then asked to specify their relationship to the care recipient as well as types of assistance with activities of daily living and instrumental activities of daily living.
Students completed the anxious/depressed subscale of the Behavior Problems Index (BPI), which is one of six subscales that comprise the 28-item rating scale for children ages 4 to 17 years.19 This subscale consists of six questions about sudden mood changes, worries, fearfulness, sadness, not being loved, and feelings of worthlessness, and higher scores indicate higher levels of anxiety and depression. Although the full BPI was standardized using parents’ reports about problem behaviors, students were the respondents in this survey. Each scale item is rated as often true, sometimes true, or not true over the past three months, and each item is scored as 1 if the response is often true or sometimes true or as 0 if not true.19 The items are summed, and a higher score reflects more anxiety and depression. The total BPI score and the anxious/depressed subscale when based on parents’ reports have an alpha coefficient of .86 and .74, respectively.19,20 Although there are no published psychometric data about the BPI with child respondents, the internal consistency of the anxious/depressed subscale in this sample (Cronbach’s alpha = .76) was comparable to the .74 reported for parent respondents.20
Students completed six items from the Social Stress Version of the Response to Stress Questionnaire (RSQ).21 The full 57-item RSQ, developed for parent respondents, measures a range of coping mechanisms in reference to age-appropriate social stressors experienced by adolescents. The RSQ divides coping along two broad dimensions: voluntary vs. involuntary processes and engagement and disengagement strategies. Based on these dimensions, the RSQ has five factors, and within each factor, it is comprised of different sets of individual coping strategies. The RSQ has good reliability and validity, including internal consistency with alphas ranging from .73 to .85.21 There are no psychometric data for the RSQ answered directly by adolescents.
Three of the six RSQ items used in the student survey measured voluntary engaged coping strategies that reflected responses to deal directly with social stressors (i.e., think of ways to deal with or fix the problem, ask others for help or ideas, and try to fix the situation). These items comprised all problem-solving items on one factor of the RSQ, primary control engaged coping, comprised of three categories with nine total items. Another three items measured voluntary disengagement coping to distance from the problem (i.e., wishing to be smarter, wishing the problem to go away, and wishing someone would take the problem away). These items included all wishful thinking items on another RSQ factor, secondary control disengagement coping, which had four categories with 12 total items. Items were rated on the use of a particular coping strategy, from 1 (not at all) to 4 (a lot), and higher total and subscale scores reflect a greater propensity to use more coping strategies. Both sets of coping items had an alpha of .68, which given only three items in each set, shows good reliability.
Life satisfaction was measured using the complete five-item Satisfaction with Life Scale (SWLS) which measures global life satisfaction.22 Each item is a statement scored from 1 (strongly disagree) to 7 (strongly agree), with total scores ranging from 5 (low satisfaction) to 35 (high satisfaction). Previous studies demonstrate that the SWLS total score has a two-month test–retest reliability of .82 and an internal consistency reliability of 0.87.22 Internal consistency was .77 in the current study.
A latent variable approach was used to assess the effects of caregiver status on three outcome measures simultaneously by comparing three student groups (non-caregivers, caregivers living with care recipients, and caregivers not living with care recipients) while controlling for background covariates, including grade, gender, and race/ethnicity. Since two of the three outcome variables were subscales of instruments that did not have known psychometric properties with child respondents, measurement errors and potential differential item functioning (DIF) were concerns. Therefore, a latent variable analysis approach was preferred to conventional regression models. The Multiple Indicators Multiple Causes (MIMIC) model was used to examine differences in the means of error-free latent factor outcomes across student groups.23–26 The analysis was performed in three steps: a preliminary exploratory factor analysis (EFA) to review measurement characteristics and identify the optimal number of underlying factors; a confirmatory factor analysis to determine how well the results of the EFA fit the data and yield a valid factor structure; and after obtaining a satisfactory fit, MIMIC was applied by adding the caregiver variable and covariates to the model.
Each observed item was analyzed as an ordered categorical variable, and all analyses were conducted in Mplus Version 626 using the default means and variance adjusted by a weighted least squared estimator. This is a robust method that does not assume a multivariate normal distribution for items and is appropriate for ordered categorical variables.26,27 Several model fit indices were used, including chi-square, comparative fit index (CFI), and root mean square error of approximation (RMSEA). CFI is the relative position of the model accurately representing the data between two opposites. In extreme cases, an independence model that assumes no relationships in the data yields the worst fit (CFI = 0), while a saturated model provides a perfect fit (CFI = 1). CFI values greater than 0.9 indicate good fit. RMSEA is one of the most informative criteria and describes the average size of the residuals between the expected model estimated covariance and actual sample data covariance. RMSEA values less than 0.05 are preferred.28
Basic demographic characteristics of middle school student sample by caregiver status (N = 1,281)
Not a caregiver
Caregiver living with care recipient
Caregiver not living with care recipient
Based on the EFA results, including a Scree plot, the four-factor solution was the preferred choice (X2 = 240.05, df = 74, RMSEA = .042). The four factors included anxiety/depression, life satisfaction, and two coping factors. As expected, the two extracted anxiety/depression and life satisfaction factors matched the anxiety/depression and life satisfaction scales item-for-item. One coping factor corresponded to three items measuring primary control engaged efforts to cope and the second factor consisted of three items measuring secondary control disengaged efforts to cope.
Item-level means by student caregiver status
Items and scale/subscale
Status: caregiver/living with care recipient
Yes (not with)
Primary control engagement coping:
When I have a problem
I try to think of different ways to deal with or change the problem or fix the situation.
1 to 4
I ask other people for help or for ideas about how to make the problem better.
1 to 4
I do something to try to fix the problem or take action to change things.
1 to 4
When I have a problem
I wish that I were stronger, smarter, or more popular so that things would be different.
1 to 4
I deal with the problem by wishing it would just go away, that everything would work itself out.
1 to 4
I wish that someone would just come and get me out of the mess.
1 to 4
In most ways my life is close to perfect
1 to 7
The conditions of my life are excellent.
1 to 7
I am satisfied with my life.
1 to 7
So far, I have gotten the important things I want in life.
1 to 7
If I could live my life over, I would change almost nothing.
1 to 7
In the last 3 months,
I have worried too much.
1 to 3
I have had sudden changes in moods or feelings.
1 to 3
I have felt unhappy, sad, or depressed.
1 to 3
I have been fearful or anxious
1 to 3
I have felt worthless or inferior.
1 to 3
I have felt like or complained that no one loves me.
1 to 3
MIMIC model results: standardized effects of caregiver status on latent outcomes adjusted for gender, race/ethnicity, and school grade as well as relationship between latent outcomes
Caregiver status (ref: no)
Yes, live with care recipient
Yes, but does not live with care recipient
Gender (ref: male)
Race/ethnicity (ref: Caucasian (White))
Asian, multiracial/multiethnic (mixed) and others
Correlations between latent factors
For all other adjusted estimates, females reported significantly higher levels of anxiety/depressive symptoms than males (p < .001). They scored significantly lower than males in their use of coping styles where they actively tried to deal with a problem and higher in their use of disengaged wishful thinking coping styles (p < .001). Compared to Caucasians, minority groups reported significantly lower utilization of engaged coping strategies (p < .001), and with the exception of Latino/Hispanic students, other minority groups reported significantly lower life satisfaction (p < .001). There were no significant direct effects of caregiver status, gender, or racial group on any item score (i.e., no group-related DIF was observed).
Prevalence of caregiving
The prevalence of caregiving youth in this large sample of students from two middle schools, 36.0%, was lower than the prevalence of 56% reported in the Palm Beach County survey of 54 middle and high schools.3 There are no other school-based surveys of caregiving youth to compare these findings, but the figures are striking. The prevalence results in this study likely reflect the demographic characteristics of Palm Beach County which has large populations of older adults and persons with HIV/AIDS as well as diversity in racial, economic, and household characteristics.3
Caregiving impact on youth mental health
The present study is one of the first to quantitatively analyze differences between youth caregivers, including those who lived with care recipients and those who did not, and non-caregivers in a large school sample. The results are not only an important step in affirming that a significant proportion of children and adolescents have caregiving responsibilities for family members, primarily older relatives, but that this also significantly affects their mental health, similar to adult caregivers.29 An important research priority is to clarify the taxonomy and clinical significance of youth caregiver depression and anxiety.
As hypothesized, youth caregivers were at significantly higher risk for anxiety and depression, with caregivers living with care recipients at the highest risk. However, other variables that could potentially contribute to their heightened vulnerability in addition to caregiving stress, such as the quality of attachment to parents or other family members, academic difficulties, and lack of available supports, were not measured in the dataset.30 Future research should clarify whether the multiple interacting psychological, behavioral, social, school, and family factors that mediate emotional distress in youth caregivers are different from non-caregiving youth. A large body of research describes the role of family functioning, such as marital conflict, on the development of psychological and behavioral problems in children.30,31 Family caregiving demands are likely to increase family dysfunction and contribute to increased emotional distress in youth caregivers associated with caregiving and school pressures.
Coping responses to stress
Although youth caregivers used more strategies in both coping style categories compared to non-caregivers, the frame of reference students used to formulate their responses is unknown because they answered each item in response to the introductory statement “When I have a problem.” It is possible that the greater use of all coping processes by youth caregivers reflected several factors: their perception of having more problems and increased stress associated with family and school pressures; and/or their need to try different coping strategies because they had no guidance about what to do.
Although these students are trying many approaches to caregiving, there is no literature describing how young caregivers acquire caregiving skills. Research is needed to characterize the process whereby these early adolescents learn what needs to be done (e.g., are enabled by family members to provide help, watch others provide care, or learn what needs to be done on their own) as well as how they cope with success and failure in the learning process. The learning process is likely to vary depending on their age and developmental level, living arrangements, family structure and dynamics, the impact of the experience on their perceptions of their competence, and the availability of information and support. Indeed, the coping with social stress results suggest they are trying many strategies precisely because they have limited experience and supports to guide them.
Life satisfaction is a broad construct, defined as a person’s cognitive appraisal of the quality of their overall life or with specific domains, and it is considered a key component of overall well-being.32,33 Research on adolescent life satisfaction is scant, and although studies have yielded conflicting results about factors related to optimal adolescent life satisfaction, family relationships appear to be the most important factor in perceived life well-being.34
Both youth caregivers and non-caregivers reported scores indicating that overall they slightly agreed with being satisfied about life as a whole. The lack of a significant difference between caregiver and non-caregiver youth may reflect the fact that the life satisfaction measure used in this study elicited a global perception of the student’s world view and did not target specific domains, such as satisfaction with family, school, peers, or caregiving. The available literature has consistently emphasized that general measures of life satisfaction mask perceptions and distinctions adolescents make about aspects of their world.32,34
The major limitation of the study is reliance on a secondary data set that used subscales of psychological test measures to minimize the testing burden on middle school students. Two of the three outcome measures were subscales originally developed and used with parents, not children. However, the internal consistencies, i.e., alpha, obtained for the students in this study are generally similar with those reported for adults, and no differential item functioning was detected across the groups tested in the model.
Implications for Behavioral Health
This study is an initial empirical step to understand the well-being of youth caregivers challenged to care for family and to operate successfully in their educational and social worlds. The results have implications for behavioral health care services for adolescents in schools, health care settings, and home. It is essential that youth caregivers be recognized as a vulnerable population by several constituencies who are as yet largely unaware of their existence: school professionals, including teachers, principals, guidance counselors, and school nurses; health care professionals, including pediatricians, primary care physicians, nurses, and psychologists; community agencies in the home care, mental health, family, and aging networks; as well as caregiving advocacy and support organizations.1–5
Adolescence is a stressful developmental period without the added dimensions of caring for family members with health problems, and much needs to be learned about the short-term and long-term effects of caregiving on the personal, family, and academic life of young adolescents. Since the youth caregiving population will continue to grow, development of evidence-based interventions to support them and prevent negative health, psychosocial, and educational outcomes must be a priority. All the currently available informational caregiving resources and evidence-based interventions have been developed for spouses and adult children.1,35
Future studies should incorporate tests and measures that identify the severity of the health care/psychosocial care needs of the person(s) receiving care in relation to the emotional distress of youth caregivers; the relationship of youth caregivers to the family, school, and friendship social support systems; changes in the nature of the attachment to parents and other family members; the impact of caregiving on obesity and adolescent identity; the impact of caregiving on risky health behaviors, including violence, suicide, and self-injury; and the impact of caregiving on educational performance and academic success over time.
Finally, in addition to the research agenda, it is important to emphasize that many other countries have recognized the needs of youth caregivers. The UK countries have passed several major national laws that provide for a range of services and programs.7 These include financial allocations to assess vulnerable children, provide community and home-based services for care recipients, families, and youth, as well as support programs and resources for youth caregivers. Australia, Canada, and a number of European and African countries are also working to address the needs of youth caregivers through new policies and practices.36,37 Therefore, perhaps the most important implication of the results of this study is the need for policies and practices in the USA that support the behavioral health and overall well-being of youth with dual caregiver–school roles and their families.
Conflict of Interest
None of the authors have any conflict of interest to disclose.