Child Health Status and Health Care Use in Grandparent- Versus Parent-Led Households

The National Survey of Children’s Health (NSCH) is a cross-sectional survey funded and directed by the Maternal and Child Health Bureau of the United States Health Resources and Services Administration and fielded by the United States Census Bureau. The survey provides representative estimates of key indicators of child health at the state and national levels for noninstitutionalized United States children aged 0 through 17 years. The NSCH measures indicators such as the child’s physical, mental, emotional, and developmental health; health care access, utilization, and quality; and family factors that affect the child’s health. Data are collected annually from parents or other caregivers through web- or paper-based questionnaires. More information about the survey, including its design and administration, is available elsewhere.^12–17 

For the current cross-sectional analysis, we combined NSCH data across 4 years (2016 to 2019) to increase the sample size. Overall, response rates ranged from 37% to 43% depending on year, and interview completion rates (or the proportion of households with children who completed a questionnaire) ranged from 70% to 80%. The analytic sample included 117 371 children ages 0 to 17 years. We applied Rapoport and colleagues’ operationalization of grandparent- versus parent-led households.²  Specifically, “grandparent-led households” were defined as households in which the caregiver respondent was a grandparent, either with no other caregiver or with another grandparent as caregiver. “Parent-led households” were defined as households in which the caregiver respondent was a biological or adoptive parent, either with no other caregiver or with another biological or adoptive parent or stepparent as caregiver. Households with other caregiver arrangements were excluded.

Caregiver respondents answered questions about whether the child currently had certain physical, emotional, mental, and developmental conditions as well as special health care needs.¹⁸  Respondents also answered questions on their experiences as a caregiver, including emotional support, family resilience, handling caregiving demands, frustration getting health care services for the child, help with care coordination, and impact of child health on employment (Supplemental Table 5). In addition, caregivers provided information about the child’s health care access and utilization, including usual source of sick care, preventive check-ups, preventive dental care, mental health treatment, specialty care, forgone care, medical expenses, and problems paying for medical expenses.

We compared grandparent-led households with parent-led households with respect to the sociodemographic characteristics of the caregiver and child, household characteristics, current health conditions of the child, caregiving experiences, and the child’s access to and utilization of health care. We estimated frequency distributions with weighted percentages and 95% confidence intervals (CIs) and tested bivariate associations between each of the measures of interest and household type using design-based Pearson χ² tests of independence.

For the measures of children’s health status and health care access and utilization, we used logistic regressions to model outcomes of interest as a function of household type. Regressions were adjusted for potential confounders, which were selected based on observed sociodemographic differences between household types. We adjusted for the following caregiver characteristics: sex, physical health status, and mental health status. We also adjusted for the following child characteristics: sex, age category, race/ethnicity, and insurance status/type. Additionally, models were adjusted for federal poverty level, highest household education, and number of caregivers. The health care access/utilization models were further adjusted for child general health status to account for differential need for health care. Model results are shown as adjusted prevalence ratios, indicating the increased risk of a particular outcome in grandparent-led households compared with parent-led households, while holding all other covariates constant.

All analyses were weighted based on the survey’s sampling design to produce estimates that were representative of children nationally and were conducted using StataSE 15 (StataCorp, 2017; Stata Statistical Software: Release 15; College Station, TX: StataCorp LLC). The threshold for statistical significance was set at α = .05. Observations with missing or unknown data were dropped from the analysis. Sex (0.1% missing), race (0.4% missing), and ethnicity (0.6% missing) were imputed using hot-deck imputation, and household poverty (17.2%) was multiply imputed using regression methods. Six poverty implicates were used to calculate estimates with correct standard errors. More information is available elsewhere on imputation methods.¹⁹  This study was exempt from institutional review board review because it used publicly available, deidentified data.

The sample for this analysis included 3464 children from grandparent-led households and 113 907 children from parent-led households (Table 1). Among grandparent-led households, the majority of caregivers were female (Supplemental Figure 1). Compared with parent caregivers who responded to the survey, a greater proportion of grandparent caregivers were female (75.7% vs 69.2%, P < .001). The mean caregiver age was 59.9 years in grandparent-led households and 41.6 years in parent-led households, and a larger proportion of grandparent caregivers were in the older age categories (50+ years) compared with parent caregivers (P < .001). A smaller proportion of grandparent caregivers reported excellent/very good physical health (51.1% vs 69.8%, P < .001) and mental health (72.7% vs 77.9%, P = .02) compared with parent caregivers.

Child sex and age did not differ by household type. The distribution of the child’s race and ethnicity differed by household type (P < .001). Specifically, a larger proportion of children in grandparent-led households were non-Hispanic Black, compared with parent-led households (30.0% vs 10.9%). Additionally, fewer children in grandparent-led compared with parent-led households were reported to be in excellent/very good health status (81.9% vs 91.9%, P < .001). In terms of insurance status, more children in grandparent-led households had public insurance (64.1% vs 26.4%), whereas more children in parent-led households had private insurance (P < .001).

A larger proportion of grandparent-led households than parent-led households had lower income levels, lower education levels, received government benefits, experienced food insufficiency, were unemployed, had English as their primary language, and had only 1 caregiver in the household (P < .001 for all).

The observed (unadjusted) prevalence of several health conditions varied according to household type (Table 2). Namely, children in grandparent-led households had higher rates of several physical health conditions including asthma (10.3% vs 7.5%, P = .002), oral health problems (18.9% vs 13.1%, P < .001), and overweight/obesity (40.3% vs 29.7%, P < .001), compared with children in parent-led households. Children in grandparent-led households also had higher rates of emotional, mental, or developmental conditions including ADHD (16.3% vs 8.0%, P < .001), behavioral/conduct problems (13.9% vs 6.1%, P < .001), depression (6.6% vs 3.1%, P < .001), developmental delays (10.7% vs 4.9%, P < .001), learning disabilities (13.9% vs 6.2%, P < .001), and speech and language disorders (8.2% vs 5.4%, P = .03). A greater proportion of children in grandparent-led households had a qualifying special health care need compared with parent-led households (P < .001).

For caregivers’ experiences, grandparents reported lower rates of feeling emotionally supported (70.4% vs 76.3%, P = .001) and experiencing high family resilience (76.1% vs 82.6%, P < .001), compared with parents (Table 2). Among employed caregivers, a greater proportion of grandparents than parents reported that the child’s health affected their employment (4.2% vs 2.8%, P = .04). However, there were no differences between grandparents and parents with respect to reports of handling the demands of raising a child, frustration in getting health care services for the child, and receiving care coordination help when needed.

The observed prevalences for children’s health care access and utilization also differed by household type (Table 3). Children in grandparent-led households had lower rates of having a usual source of sick care (65.7% vs 79.5%, P < .001), receiving a preventive checkup in the past year (64.6% vs 77.1%, P < .001), receiving a preventive dental visit in the past year (73.8% vs 80.6%, P < .001), and receiving specialty care when needed (78.6% vs 90.2%, P < .001) compared with children in parent-led households. In addition, children in grandparent-led households had higher rates of forgone care (5.9% vs 2.8%, P = .002) compared with children in parent-led households. Children in grandparent-led households were less likely to have any medical expenses (33.7% vs 70.1%, P < .001) compared with parent-led households; however, among those with any expenses, a greater proportion of grandparents than parents reported problems paying for health care (23.0% vs 15.2%, P = .01).

After adjusting for potential confounders, several of the associations between household type and children’s health conditions and health care access/utilization remained statistically significant (Table 4). Specifically, relative to parent-led households, there was an increased prevalence among grandparent-led households of children with ADHD (adjusted prevalence ratio [aPR] = 1.60; 95% CI, 1.30-1.91), behavioral/conduct problems (aPR = 1.54; 95% CI, 1.25-1.83), depression (aPR = 1.73; 95% CI, 1.24-2.23), learning disability (aPR = 1.48; 95% CI, 1.17-1.79), and any special health care needs (aPR = 1.34; 95% CI, 1.18-1.49); and a decreased prevalence of having a usual source of sick care (aPR =0.93; 95% CI, 0.88-0.97), preventive checkups (aPR = 0.94; 95% CI, 0.89-0.99), and any medical expenses (aPR =0.90; 95% CI, 0.86-0.95). Associations between household type and the following child health conditions and health care access/utilization indicators were no longer significant after adjustment: asthma, oral health problems, overweight/obesity, developmental delay, preventive dental visits, specialty care, forgone care, and problems paying medical expenses.

In this analysis of a national sample, the study findings corroborate and build on previous research on the health and well-being of children living in grandparent-led households. In terms of the health status of grandparent caregivers, this study upholds previous findings that grandparent caregivers report worse physical and mental health compared with parent caregivers.^6,8  In regard to child health outcomes, previous literature has reported mixed findings on the prevalence of common conditions among children in grandparent-led compared with parent-led households; however, the majority of these studies were not based on nationally representative samples.^2,11  The current study found children in grandparent-led households had higher observed rates of several physical, emotional, mental, and developmental health conditions. Although having a disproportionately higher burden of health conditions could have contributed to the circumstances rendering parents unable to care for their child, it also underscores the importance of children in grandparent-led households receiving essential health care services.

The health conditions and health care utilization patterns of custodial grandparents have been well studied; however, no studies to date have examined utilization of essential health care services among children in grandparent-led households using a nationally representative sample. Consequently, this study addresses an important research gap, indicating that children in grandparent-led households are less likely to use certain essential health care services compared with children in parent-led households.

Further investigation is necessary to elucidate why children in grandparent-led households may not receive necessary health care services. This study found that a larger proportion of grandparent-led households than parent-led households are affected by sociodemographic factors such as reliance on public insurance for the child, lower household education levels, and living in a single caregiver household. After adjusting for these possible confounders, we found that only a few health care access/utilization outcomes remained significant, underscoring the important role of sociodemographic factors in accessing needed health care. Previous research suggests custodial grandparents can be difficult for service providers to reach, may not seek help because they lack enabling resources (eg, transportation) to obtain services, or have been disappointed with available services.^11,20  Furthermore, research suggests that whether the grandparent became a custodian through informal or formal mechanisms can affect service utilization, with grandparents providing care through formal mechanisms accessing and using more services.²¹  Other possible contributing factors could include being unaware of current pediatric recommendations for frequency of obtaining certain health care services, prioritizing other responsibilities associated with caring for a child, or the grandparents’ own physical and emotional health obstacles. Additionally, the current study findings indicate that grandparents disproportionately face challenges such as being a single caregiver, having lower income, and possessing lower levels of education, which may affect their ability to secure needed health care for themselves and the children in their care. The majority of children in grandparent-led households in this study did not have any medical expenses, making cost a less likely barrier. However, for the children with any expenses, a greater proportion of grandparents than parents reported problems paying for health care, which could partially explain reduced health care utilization among this subpopulation. Interestingly, in this study, grandparents did not report increased difficulties with several other factors relevant to obtaining health care for their grandchild. Namely, grandparent caregivers reported no differences in handling the demands of raising a child, frustration in getting health care services for the child, and receiving help with care coordination compared with parent-led households. However, compared with parents, grandparents did report lower rates of emotional support and high family resilience.

Previous studies have documented that grandparent caregivers can experience isolation from peers and several emotional challenges.^11,22  Custodial grandparents must contend with the changing circumstances in their own life, the events in their child’s life that require the grandparent to become a primary caregiver, as well as the trauma their grandchild may be experiencing around their changing guardianship.¹¹  These emotional challenges compounded with physical vulnerabilities of older caregivers may make it more difficult for custodial grandparents to provide all the necessary care for their grandchildren. One study found that resilience can mediate these negative stressors and consequently interventions to promote resiliency among grandparents may be valuable.^11,23  A few studies have shown that interventions aiming to strengthen supportive factors, such as health management, and reduce risk factors, such as isolation, can be beneficial to grandparents’ well-being.^11,24  For social support, in particular, grandparents who participate in support groups have been shown to report increased social support compared with those who do not participate.²⁵ 

Despite the numerous stressors facing custodial grandparents, they also possess numerous strengths including resourcefulness, benefit finding, and positive caregiving.¹¹  To promote grandparents’ strengths, service sectors interacting with custodial grandparents should be mindful of their complex needs. This study indicates that grandparent caregivers often intersect with or could benefit from the services of other sectors, evidenced by their increased utilization of government benefits and increased rates of food insecurity and unemployment relative to parent caregivers. Indeed, large proportions of grandparent-led households were found to interact with service systems providing financial supports through Temporary Assistance for Needy Families/welfare, Women, Infants, and Children program, food stamps, and free and reduced cost school meals. Therefore, service systems supporting grandparents, including the health care sector, should aim for service coordination and cross-promotion of other sectors’ benefits to optimize the care and well-being of the grandparent–grandchild dyad.

There are several study limitations to consider. First, given that data are cross-sectional, causality cannot be asserted. Second, the survey items only capture caregiver-reported information and may be subject to recall bias if a caregiver cannot accurately remember which health care services the child has used. Third, the survey did not capture the reason why the child is under the care of a grandparent nor how long the child has been in the caregiving arrangement. Additional information surrounding the circumstances of grandparent caregiving arrangements could explain, at least in part, some of the differences found between grandparent- and parent-led households. Finally, if survey respondents were systematically different from nonresponders, nonresponse bias could lead to over- or underestimates of the outcomes presented in this analysis. However, previous analyses of nonresponse bias in the NSCH have found that although nonresponse is generally higher in areas with smaller college-educated populations, lower incomes, less homeownership, and larger non-White populations, no strong or consistent evidence of nonresponse bias is evident after survey weights are applied.²⁶  Despite these limitations, this study can help inform stakeholders working to improve the health of grandparent caregivers and their grandchildren by highlighting areas where increased supportive services may be necessary. Of note, although previous studies of custodial grandparents have frequently been limited by small sample sizes, the current study leverages the ability to pool multiple years of nationally representative data to yield a large sample of more than 3300 children living in grandparent-led households.

Children living in grandparent-led households have worse physical, emotional, mental, and developmental health than children in parent-led households, and lower utilization of essential health care services. Grandparent caregivers may benefit from additional support to ensure that they can access primary, preventive, and specialty health care for their grandchildren in a timely manner. Health care systems, as well as other social services systems, have the potential to address the complex needs of grandparent-led households by providing support services that address the needs of both grandparents and their grandchildren.

ADHD

attention deficit hyperactivity disorder

aPR

adjusted prevalence ratio

CI

confidence interval

NSCH

National Survey of Children’s Health