Caregiver burden

Caregiver burden is a multifactorial construct. In this paper, caregiver burden experienced by aging parents of adults with ASD was conceptualized to comprise four domains- emotional, developmental, time dependence, and financial burden, based on adult caregiving literature [23,24]. Parents experience emotional burden, which includes negative emotions such as sadness, grief, depression, anxiety, guilt, and blame [25–27]. The developmental aspect of burden pertains to parents’ perceptions of being “off-time” in their development when compared to their same-aged peers. Extant evidence indicated that parental caregivers were more likely to report missing out on a normal way of life, experience social disconnectedness from relatives and friends, and feel unsupported by systems and society at large [28,29]. Time dependence is the perception of burden secondary to the time restrictions faced by parents because of the enduring and lifelong demands of caregiving for adult children [24,30]. Time devoted to caregiving can reduce opportunities to associate with friends, family members, and spouses, and can restrict parent caregivers’ potential for pursuing career, leisure, and community interaction opportunities [7,31]. Financial burden encompasses ongoing economic costs that families contend with to ensure that all family members, including the adult child with ASD, have access to age-appropriate health care, educational, and career opportunities. Parents often have to reduce working hours, forgo promotions, and higher educational pursuits because of caregiving demands that can affect family income and retirement benefits [31,32].

Predictors of caregiver burden

Caregiver burden is influenced by individual- and family-level factors. Individual-level factors include the clinical and socio-demographic characteristics of the individual with ASD and their caregivers. The core clinical traits of ASD (e.g., ability to communicate and socialize) influences an individual’s functional independence, placing differential demands on parental caregivers [13,30,33]. Many socio-demographic characteristics of parents also can influence their perceptions of the burden associated with caregiving. Aging parents, who are themselves experiencing age-related health challenges while supporting their adult children with ASD, are likely to experience greater caregiver burden, than younger parents [6,34]. In addition, many aging parents also may be providing care to other family members such as another child, spouse, or parent [35]. Recent estimates indicated that close to a quarter of informal family caregivers in the United States are providing care for more than one family member [36]. Compound caregiving for multiple individuals in aging families can also increase a parent’s perceived caregiver burden [37,38].

Study purpose

The purpose of this study was to examine the relationship between the ability of adults with ASD to perform ADLs and parental perceptions of caregiver burden. It was hypothesized that higher dependence in ADLs would be associated with a greater perception of caregiver burden after adjusting for factors in the caregiving context.

Research design

Using a web-based survey, cross-sectional data was collected electronically from a convenience sample of parents of adult children with ASD from multiple states in the United States.

Participants

Participants had to be at least 50 years of age, able to read and comprehend English, have access to a computer for the web-based survey, and self-report being a parent of at least one adult (aged 18 or above) with ASD to be included in the study. Eligible participants were asked to read the information sheet and if they were willing to participate in the study, they were directed to complete the online survey. A total of 353 self-administered surveys were submitted to the website. This number was reduced by 33 participants, including 23 who did not provide a sufficient number of responses and 10 who did not meet the inclusion criteria because the child with ASD had not reached the age of 18 years or the parent was younger than 50 years of age. The final sample of older parents included 320 respondents who met all criteria for inclusion.

The majority of the participants (n = 212, 66.3%) were aged 50–59 years, while the remaining (n = 108, 33.7%) were aged 60-years or older. Most of the participants were female (n = 259, 82.0%), Caucasian (n = 289, 90.9%), reported being married (n = 254, 79.2%), and had higher levels of education (n = 223, 69.7% with a bachelor’s degree or higher). The majority (n = 204, 63.8%) were working either full-time, part-time, or were self-employed, and more than half (n = 185, 59.1%) reported an annual family income greater than $60,000 (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0244844.t001

Variables of interest

Data analysis

First, frequency distributions and descriptive statistics were used to summarize all variables of interest (Table 1). Second, prior to running regression models, pairwise correlation analyses were used to examine the bivariate relationships between the independent and dependent variables. Correlation plots were generated (S1 Fig) using the R software package [61,62]. To characterize the form of the association between the primary independent variable (dependence in ADLs of the adult child with ASD) and each of the dependent variables (caregiver burden and the four domains) linear, quadratic, and non-parametric Lowess functions [63] were fitted and plotted (S2 Fig). Based on these estimates it was determined that the linear form provides optimal fit (also see sensitivity analyses below). As such, linear associations were adopted and presented as primary findings in this study.

For each dependent variable three linear regression models were fit to examine (1) crude, (2) caregiver characteristics, and (3) adult with ASD characteristics adjusted models. The aim of these incremental adjustments was to determine the extent to which the associations between caregiver burden and dependence in ADLs would be attenuated by these adjustments. Table 2 includes the unstandardized beta coefficients derived from these models. To maintain parsimony, variables with bivariate associations significant at p < 0.05 were retained for use as covariates in the regression models (however, see sensitivity analyses below). S3 Fig summarizes the marginal means plots for each of the caregiver burden outcomes over the dependence in ADLs continuum, and their 95% confidence intervals. These plots help visualize the reported associations and the reductions in magnitudes resulting from covariates adjustments [64].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Linear associations between dependence in ADLs and caregiver burden¹.

https://doi.org/10.1371/journal.pone.0244844.t002

Several sensitivity analyses were conducted to ensure robustness of the primary findings. Sensitivity results are included as supplementary material to the manuscript. First, several power transformations of the primary exposure (dependence in ADLs), including higher polynomial (quadratic, and cubic) and natural-log functions were considered and tested (S1 Table). AIC and BIC fit statistics showed no consistent evidence for improvement in fit (relative to linear) as a result of these transformations. However, the covariates adjusted models for the quadratic and cubic transformations were used to calculate and plot the marginal means derived from these models, and their 95% confidence intervals, to allow interested readers to compare the estimates across linear, quadratic and cubic fit (S2 Table and S4 Fig). Second, to test for any possible attenuations in the primary results derived from the linear regression described above (regression model 3) a fourth model was fitted to adjust for the full set of covariates described in Table 1. The estimated marginal means derived from these models plotted in S5 Fig allow readers to visualize and compare these estimates relative to the primary reported findings. The data used for these analyses and the associated codebook are available for review in S1 and S2 Data respectively.

Description of adult children with ASD

The majority (80.3%) of adults with ASD represented in this study were male, with a mean age of 25.1 years (SD = 7.09). Most (77.5%) were living with their parents, and about 42% were working for pay (full-time, part-time, with a job coach, or in a sheltered workshop). About 23.4% of the adult children were attending public schools, 16.6% were attending a university or community college, 4.1% were attending a vocational training program. Most adults with ASD (73.8%) could communicate on topics in a meaningful way. About half (50.3%) required minimal prompting, direction, or redirection for their behavior, and 4.4% had typical age-appropriate adult behavior, while 11% of parents reported serious challenging behaviors. About 42% had friends in the community or were part of a social group, while the remaining had no friends outside of their family (Table 1).

The mean scores for the dependence in ADLs in descending order (higher scores indicate greater dependence on parents) were: money management (M = 2.92), SD = 1.13), taking public transportation (M = 2.80, SD = 1.26), cooking simple meals (M = 2.31, SD = 1.14), community engagement (M = 2.11, SD = 1.07), simple cleaning e.g. bed making (M = 2.10, SD = 1.0), grooming (M = 1.77, SD = 0.93), bathing (M = 1.51, SD = 0.91), dressing (M = 1.33, SD = 0.67), eating (M = 1.25, SD = 0.57), functional mobility (M = 1.26, SD = 0.65), toileting (M = 1.28, SD = 0.65), and with an overall mean of 1.86 (SD = 0.68).

Parental caregiver burden description

The mean for total burden as reported by the participants was 2.67 (SD = 0.63), when measured on a 5-point rating scale, with higher scores indicating greater caregiver burden. The domain level scores indicated that the highest type of burden was financial (M = 3.05, SD = 0.65), followed by developmental (M = 2.92, SD = 1.05), followed by time dependence (M = 2.88, SD = 1.01), and emotional (M = 1.84, SD = 0.76).

Correlation analysis

The results of pairwise correlation analysis summarized in S1 Fig indicate the direction and magnitude of the associations between the variables of interest. Total caregiver burden was positively associated with dependence in ADLs (r = 0.50, p <0.001) and compound caregiving (r = 0.13, p = 0.02). The inverse associations of burden with adult child’s ASD characteristics indicated that parents of adult children with fewer behavioral (r = -0.48, p <0.001), and social challenges (r = -0.18, p = .002) reported less caregiver burden. The inverse association of burden with health (r = -0.27, p <0.001) imply that parents with lower quality of health perceived higher levels of caregiver burden. The magnitude of these associations was small for emotional and financial burden.

Regression analyses

First, unadjusted bivariate linear regression models showed that the associations between dependence in ADLs and caregiver burden (with the exception of emotional burden) were all statistically significant and in the expected direction, indicating that parents of adult children who needed to provide more assistance for ADLs were more likely to perceive greater caregiver burden. The estimated unstandardized coefficients were b = 0.47 (p < 0.001), b = 0.48 (p < 0.001), b = 1.2 (p < 0.001), and b = 0.12 (p = 0.023) for total, developmental, time dependence, and financial burden, respectively. The explained variances shown by the estimated R² were 25%, 10%, 65%, and 2%, respectively (Table 2).

Second, adjusting for parent characteristics attenuated the associations (unstandardized beta coefficients) between dependence in ADLs and total caregiver burden (b = 0.43; [p < 0.001]), developmental burden (b = 0.42; [p < 0.001]), and time dependence burden (b = 1.18; [p <0.001]) by 9%, 13%, and 2%, respectively. Adjustment for parent characteristics led to an additional 5%, 6%, and 4% absolute increase in level of explained variance for these types of burden, respectively. The association between dependence in ADLs and financial burden became statistically insignificant at this step.

Finally, adjusting for behavioral challenges further attenuated, but did not completely explain, the magnitudes of the associations (unstandardized beta coefficients) between dependence in ADLs and total burden (b = 0.24; [p <0.001]), developmental burden (b = 0.27; [p < 0.05]), and time dependence burden (b = 0.98; [p < 0.001]) by an additional 30%, 36%, and 17%, respectively. Adjustment for characteristics of the adult child with ASD led to an additional 8%, 5%, and 4% absolute increase in the levels of explained variance for these burdens, respectively. Plots of the marginal means derived from these models to help visualize the associations are presented in S3 Fig. Additionally, as expected, the sensitivity analyses (S5 Fig) showed no evidence for attenuations or change in the associations between the caregiver burden outcomes and ADLs from additional controls for the covariates detailed in Table 1.

Practice implications

The association of dependence in ADLs of adult children with ASD and parental perceptions of caregiver burden is of practical relevance to family members, practitioners, and policy makers. The findings are relevant to practitioners, such as occupational therapists, social workers, and psychologists, who are committed to improving the quality of life of individuals with ASD and their family members. For example, occupational therapy researchers found that parents often identify independence in ADLs as a top area of intervention for their children with ASD [70]. Gaining mastery in ADLs also is associated with positive quality of life outcomes for adults with ASD, such as post-secondary education, employment, and independent living [15,40,71]. Two practical ways of empowering aging parents of adult children with ASD in dealing with their daily challenges related to ADLs are described below.

First, therapists should consider providing direct services to adolescents and adults with ASD to improve their functional independence. Doing so can decrease the time constraints faced by parents in helping their adult children and thus free up the time and energy required for engaging with same age peers in leisure and community engagement. It is recommended that older children with ASD be provided with non-traditional direct services to increase affordability and cost-effectiveness. Practitioners should create lifelong, sustainable community-based programs to empower adults with ASD who are at risk for decline in their ADL performance after adolescence [72].

Second, therapists should develop trainings focused on ADLs for families of adults with ASD as has been demonstrated with other caregiver populations [57,73]. For example, aging parents can be trained to help adults with ASD gain and/or maintain mastery in ADL skills into adulthood. Such parental training is important in light of longitudinal research indicating that adults with ASD are at risk to lose mastery over their daily living skills [12,72]. Including siblings of children with ASD in ADL-focused interventions can reduce a culture of dependence in the home environment from childhood and promote independence in ADLs. Further, including siblings in therapy sessions can enhance emotional connections with the child with ASD leading to stronger relationships in adulthood [74,75].

This study also yields important policy implications. In particular, there is a critical need to devise programs focused on addressing the challenges faced by families of adults with ASD who often struggle to find support for their adult children as they transition out of the public-school system [26,41]. Emerging research points to the need for ADL therapies to improve family quality of life of individuals with ASD [66]. The increased incidence of ASD, along with the chasm in service needs for adults with ASD, has created an emerging niche for practitioners such as occupational therapists [55,76]. Therapists should consider community-based models of service delivery because they tend to be family-centered, user-friendly, and cost-effective when compared to the traditional medical-model of service delivery that can lead to long-term dependence on professionals [77,78].

Limitations

The limitations of this study pertain to the homogeneity of the sample and challenges associated with standardized measurements of ASD traits. The non-probability sampling techniques used to recruit participants resulted in sample homogeneity (i.e., the participants were predominantly well-educated, affluent, married, Caucasian mothers who could access the Internet to participate in this web-based survey). The survey response rate could not be determined, as there was no way of determining how many potential participants heard about the opportunity to participate given the wide spread methods of disseminating the survey link. The overrepresentation of non-Latino white families from a higher socioeconomic status limits the generalizability of the findings. Parents were asked to self-report their adult child’s diagnosis of ASD instead of using an autism checklist or a specialized scale to measure the severity of the condition and the level of dependence in ADL. The response categories used for determining the core characteristics of ASD were based on salient criteria available in the literature. The survey design, allowed respondents to choose only one criterion. However, it is possible for individuals to have a range of traits that overlap several responses.

Future directions

Future research should include a broader array of aging parents of adults with ASD, particularly by targeting the recruitment of racial/ ethnic minorities and those living in low resource, disadvantaged households, and communities. Future efforts should be invested in increasing the accuracy of measurements of the strengths and challenges faced by adults with ASD. This could be done in two ways. First, integrate standardized scales to measure the severity of ASD including social communication skills and adaptive functioning. Second, researcher-developed survey items included in this study could be refined by asking the participants to rank (on a scale of 0–100) the challenges of their adult child with ASD on each of the response categories within each of the domains (communication, social skills, and behavioral challenges). Doing so, would allow for a more precise characterization and adjustment to these challenges in statistical models. Finally, it is recommended that the CBI and the CRA be validated for aging caregivers of adults with ASD.