Family Well-Being in Grandparent- Versus Parent-Headed Households Eli Rapoport, BS,A Nallammai Muthiah, BS,A Sarah A

Home , Grandparent

Family Well-being in Grandparent- Versus Parent-Headed Households Eli Rapoport, BS,a Nallammai Muthiah, BS,a Sarah A. Keim, PhD, MA, MS,b,c Andrew Adesman, MDa,d

BACKGROUND AND OBJECTIVES: Little is known about the 2% of US children being raised by their abstract grandparents. We sought to characterize and compare grandparent- and parent-headed households with respect to adverse childhood experiences (ACEs), child temperament, attention-deﬁcit/hyperactivity disorder (ADHD), and caregiver aggravation and coping. METHODS: Using a combined data set of children ages 3 to 17 from the 2016, 2017, and 2018 National Survey of Children’s Health, we applied survey regression procedures, adjusted for sociodemographic confounders, to compare grandparent- and parent-headed households on composite and single-item outcome measures of ACEs; ADHD; preschool inattention and restlessness; child temperament; and caregiver aggravation, coping, support, and interactions with children. RESULTS: Among 80 646 households (2407 grandparent-headed, 78 239 parent-headed), children in grandparent-headed households experienced more ACEs (b = 1.22, 95% conﬁdence interval [CI]: 1.07 to 1.38). Preschool-aged and school-aged children in grandparent-headed households were more likely to have ADHD (adjusted odds ratio = 4.29, 95% CI: 2.22 to 8.28; adjusted odds ratio = 1.72, 95% CI: 1.34 to 2.20). School-aged children in these households had poorer temperament (badj = .25, 95% CI: 20.63 to 1.14), and their caregivers experienced greater aggravation (badj = .29, 95% CI: 0.08 to 0.49). However, these differences were not detected after excluding children with ADHD from the sample. No differences were noted between grandparent- and parent-headed households for caregiver coping, emotional support, or interactions with children. CONCLUSIONS: Despite caring for children with greater developmental problems and poorer temperaments, grandparent caregivers seem to cope with parenting about as well as parents.

’ aDivision of Developmental and Behavioral Pediatrics, Steven and Alexandra Cohen Children’s Medical Center of WHAT S KNOWN ON THIS SUBJECT: Nearly 3 million New York, Lake Success, New York; bCenter for Biobehavioral Health, The Research Institute at Nationwide children today are raised by their grandparents, often Children’s Hospital, Columbus, Ohio; cDepartment of Pediatrics, College of Medicine and Department of because of social adversity. Research to date has Epidemiology, College of Public Health, The Ohio State University, Columbus, Ohio; and dDepartment of Pediatrics, primarily demonstrated negative social and health Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, Hempstead, New York outcomes for caregivers and children in grandparent- Mr Rapoport conceptualized and designed the study and conducted the statistical analyses; headed households. Ms Muthiah conceptualized and designed the study and drafted the initial manuscript; Dr Keim WHAT THIS STUDY ADDS: In a large, nationally conducted the statistical analyses; Dr Adesman conceptualized and designed the study; and all representative US sample, attention-deﬁcit/hyperactivity authors reviewed and revised the manuscript and approved the ﬁnal manuscript as submitted and disorder and childhood adversity appear to be agree to be accountable for all aspects of the work. responsible for some of the behavioral and DOI: https://doi.org/10.1542/peds.2020-0115 developmental disparities observed between Accepted for publication Jun 22, 2020 grandparent- and parent-headed households. No differences in caregiver coping and emotional support Address correspondence to Andrew Adesman, MD, Division of Developmental and Behavioral were found. Pediatrics, Steven and Alexandra Cohen Children’s Medical Center of New York, 1983 Marcus Ave, Suite 130, Lake Success, NY 11042. E-mail: [email protected] To cite: Rapoport E, Muthiah N, Keim SA, et al. Family Well- being in Grandparent- Versus Parent-Headed Households. Pediatrics. 2020;146(3):e20200115

Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 146, number 3, September 2020:e20200115 ARTICLE The number of children being raised The literature regarding the health The National Survey of Children’s by their grandparents has grown and developmental outcomes of Health (NSCH), a cross-sectional considerably in recent years, from 2.5 children raised by grandparents annual survey of households in the million in 2005 to 2.9 million in has yielded mixed findings. United States with children 2015.1 Although grandparents can Researchers of previous studies have ,18 years old, offers the unique provide support and stability in demonstrated that, in addition to opportunity to compare grandparent- families, the increase in custodial having experienced adversity early in and parent-headed households with grandparenting in the United States life, these children tend to have fewer respect to both children and has primarily been driven by the coping resources because they cannot caregivers. In this study, we aimed to inability of some parents to care for turn to their parents for support.22 assess, using this large nationally their children,2 and up to 72% of Some studies indicate that children representative data set, differences children raised by grandparents have being raised by grandparents have between grandparent- and parent- been exposed to at least one adverse, a higher prevalence of developmental headed households in terms of traumatic event.3 In light of rising delays,23 behavioral issues,24 and sociodemographics, caregiver–child incarceration rates,4,5 the current academic difficulties,25 suggesting interactions, adverse childhood opioid crisis,6 and the recent that the combination of higher experiences (ACEs), and other economic recession,7 children who traumatic event exposure and poorer caregiver and child variables, enter nonparental kinship care face coping skills in children in controlling for underlying a unique living environment and nonparental care may hinder positive sociodemographic differences. complex relationships that can impact social development. However, their long-term development. after adjusting for selection bias METHODS caused by child and family The demographic and health background factors, other studies correlates of grandparents assuming Sample have shown that nonparental care is the caregiving role have been well- The Maternal and Child Health not associated with poorer cognitive characterized by previous research. Bureau of the US Health Resources skills or behavioral problems.26 Most custodial grandparents are aged and Services Administration Reverse causation is a possibility 50 to 59 years8 and, compared with examined the physical and emotional as well because poorer child health parents, tend to have poorer health of noninstitutionalized – – may introduce disruption and physical9 12 and mental13 15 health children ages 0 to 17 through the instability to caregiving before taking on the demanding role nationally representative NSCH.31 The arrangements.27 of parenting a child. Custodial NSCH used a 2-phase multimode grandparents may also feel isolated survey approach based on the Census Previous researchers have further from peers because the demands of Address Master File, and data were explored grandparent and grandchild caregiving can be time-consuming.16 weighted to account for nonresponse outcomes among subpopulations of To that end, grandparents raising and sociodemographics. The 2016, grandparent caregivers. Analyses their grandchildren often report 2017, and 2018 NSCH data sets were have individually been focused on receiving inadequate support from combined for cross-sectional analysis grandmothers15,28,29 and those around them, and evidence per the NSCH Guide to Multi-Year grandfathers11 raising grandchildren, suggests they are less likely to receive Estimates.32 Children ages 3 to 17 racial differences among support resources.17,18 were included. grandfamilies,10,29 the diverse There has historically been less cultural attitudes and outcomes of Households in which the respondent attention on the positive outcomes grandparenting,14,30 and how single- was a grandparent and the other of the grandfamily household grandparent caregivers compare with primary caregiver in the household structure. Although it is true that single-parent caregivers.11,12 These was a grandparent, or there was no caregiving is particularly taxing for studies individually provide key other primary caregiver in the older adults,19 evidence suggests that insights into select components of the household, were categorized as even when faced with unique grandfamily. However, because of “grandparent-headed households.” financial and health burdens, differences in samples and analytic Households in which a primary custodial grandparents and their methods, results are difficult to caregiver was a biological or adoptive grandchildren can thrive.20 In compare between studies. No recent parent and the other primary fact, many grandparents raising studies have investigated both child caregiver in the household was grandchildren report that they and caregiver measures using a biological or adoptive parent or would perform the same role again a single, large, nationally stepparent, or there was no other if given the chance.21 representative sample. primary caregiver in the household,

Downloaded from www.aappublications.org/news by guest on September 29, 2021 2 RAPOPORT et al were categorized as “parent-headed validity of the derived scale. In separately in 1- and 2-caregiver households.” Households with other addition to these composite scales, households. structures were excluded. count variables for the number of For all analyses, P values were ACEs experienced were created for derived from 2-sided statistical Single-Item Outcome Measures children with complete responses for tests, and associations with P values Caregivers answered questions about all ACEs. ,.05 were considered to be whether the child had ever statistically significant. All analyses experienced each of 7 individual Statistical Analysis were conducted in R, version 4.0.0, ACEs (binary) and whether the child Grandparent-headed households by using package survey, version had a current medical diagnosis of were compared against parent- 4.0, and all analyses accounted for attention-deficit/hyperactivity headed households on caregiver and the complex survey design of the disorder (ADHD) (binary). child sociodemographics and child NSCH combined data set. This Availability of emotional support was health by using second-order study was exempt from institutional assessed by using the question, Rao–Scott adjusted x2 tests. Logistic review board review because it “during the past 12 months, was and linear regressions were used to used publicly available, there someone that you could turn to model outcomes of interest as deidentified data. for day-to-day emotional support functions of household structure with parenting or raising children” (grandparent- versus parent-headed (binary; “no” or “yes”). Caregiver RESULTS household). Regressions were coping (binary) was measured with adjusted for potential confounders, The eligible sample included 2407 the question, “how well do you think which were selected on the basis of grandparent households (631 single- you are handling the day-to-day observed sociodemographic grandparent households and demands of raising children?” differences between household 1776 two-grandparent households) Responses were dichotomized as structures and anticipated and 78 239 parent households “very well” versus “somewhat well,” associations with the behavioral (10 115 single-parent households and “not very well,” or “not at all.” outcomes of interest based on the 68 124 two-parent households). Composite Outcomes literature. Models were adjusted for Grandparent caregivers achieved caregiver sex, caregiver education, lower levels of education (F = 34.7, To facilitate analysis of inattention or household poverty level, 1- vs 2- P , .001) and had lower household restlessness, child temperament, caregiver household, and child age, incomes (F = 51.2, P , .001). They parental aggravation, frequency of sex, race and ethnicity, and health were also more likely to be female quality family interactions, and status, with the exception of models (F = 8.5, P = .004) and to be in a one- neighborhood support, 6 composite for ADHD diagnosis, which did not caregiver household (F = 77.8, P , scales were derived by aggregating control for child health status because .001) (Table 2). responses to individual Likert items ADHD is a component of child health. in the NSCH, as noted in Table 1. Child sex and age were not associated Additionally, because ADHD has been Responses to individual component with household structure, but the associated with ACEs,33 an additional items were weighted such that all distribution of child race and logistic regression was conducted items contributed equally to the ethnicity differed with household controlling for the occurrence of each composite scales. Inattention or structure, especially in the proportion individual ACE. restlessness and the frequency of of grandparents compared with quality family interactions were only Statistically significant associations parents who cared for non-Hispanic assessed for children ages 3 to 5, between household structure and Black children (30.5% vs 11.4%, temperament and parental child temperament and parental respectively). Children in aggravation were separately assessed aggravation were reexamined in grandparent-headed households were for children ages 3 to 5 and ages 6 to a sample excluding children with also less likely to be in excellent or 17, and neighborhood support was ADHD. Additionally, the association very good health (78.2% vs 90.4%; assessed for all children in the between household structure and F = 47.4, P , .001). sample. Internal consistency, as inattention and restlessness was measured by Cronbach a, was assessed among 3- to 5-year-old ACEs calculated for each scale. Also, the children without a diagnosis of ADHD Children in grandparent-headed association between the inattention to determine if subthreshold ADHD households were more likely to have or restlessness scale and ADHD phenotypes were associated with experienced each of the ACEs diagnosis was examined by using household structure. Availability of (Table 3); on average, children in a linear regression to evaluate the emotional support was assessed grandparent-headed households

Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 146, number 3, September 2020 3 TABLE 1 Composite Measure Definitions and Component Items From the 2016, 2017, and 2018 NSCH ADHD Diagnosis and Symptoms a Scale Component Items (Range) Caregivers in grandparent-headed Inattention and restlessness, “How often is this child easily distracted?” (0 [none of the time] to 3 households were more likely to have b ages 3–5 [all of the time]) children with ADHD than those in “ Compared to other children his or her age, how often is this child parent-headed households for able to sit still?” (0 [all of the time] to 3 [none of the time])c Range: 0–12 “How often does this child keep working at something until he or children ages 3 to 5 (7.8% vs 1.5%, she is finished?” (0 [all of the time] to 3 [none of the time])c adjusted odds ratio [aOR] = 4.29, 95% Cronbach a: 0.69 “When he or she is paying attention, how often can this child follow CI: 2.22 to 8.28) and ages 6 to 17 instructions to complete a simple task?” (0 [all of the time] to 3 (18.0% vs 9.9%, aOR = 1.72, 95% CI: c [none of the time]) 1.34 to 2.20). After controlling for Temperament, ages 3–5 “How often does this child play well with others?” (0 [all of the time] to 3 [none of the time].)c ACEs, ADHD was still more common “How often does this child become angry or anxious when going in grandparent-headed households from one activity to another?” (0 [none of the time] to 3 [all of the for children ages 3 to 5 (aOR = 3.27, b time]) 95% CI: 1.52 to 7.02) but not children “ When excited or all wound up, how often can this child calm down ages 6 to 17 (aOR = 1.17, 95% CI: quickly?” (0 [all of the time] to 3 [none of the time])c Range: 0–15 “How often does this child lose control of his or her temper when 0.91 to 1.50). things do not go his or her way?” (0 [none of the time] to 3 [all of The inattention and restlessness scale the time])b Cronbach a: 0.62 “This child bounces back quickly when things do not go his or her was associated with ADHD diagnoses way.” (0 [definitely true] to 3 [not true])d in children ages 3 to 5; preschool-aged Temperament, ages 6–17 “This child stays calm and in control when faced with a challenge.” children with ADHD scored an average (0 [definitely true] to 3 [not true])d of 3.80 (95% CI: 3.43 to 4.16) points e Range: 0–6 “This child argues too much.” (0 [not true] to 3 [definitely true]) higher on the scale than those without Cronbach a: 0.56 Parental aggravation, ages 3–5 “During the past month, how often have you felt that this child is ADHD. In the sample of 3- to 5-year- and ages 6–17 much harder to care for than most children his or her age?” (0 old children without ADHD, household [Never] to 4 [Always]) structure was not associated with Range: 0–12 “During the past month, how often have you felt that this child does inattention and restlessness (adjusted things that really bother you a lot?” (0 [never] to 4 [always]) beta [badj] = 0.11, 95% CI:20.27 to Cronbach a: 0.76 and 0.79 “During the past month, how often have you felt angry with this a child?” (0 [never] to 4 [always]) 0.49). Cronbach for inattention and Quality family interaction, ages “During the past week, how many days did you or other family restlessness and other composite 3–5 members tell stories or sing songs to this child?” (0 [every day] outcome measures is reported in to 3 [0 days]) Table 1. Range: 0–6 “During the past week, how many days did you or other family members read to this child?” (0 [every day] to 3 [0 days]) Child Temperament Cronbach a: 0.76 “During the past week, how many days did you or other family members read to this child?” (0 [every day] to 3 [0 days]) Children ages 3 to 5 did not differ in Neighborhood support, ages “People in this neighborhood help each other out.” (0 [definitely temperament between grandparent- 3–17 disagree] to 3 [definitely agree]) headed and parent-headed households – “ ’ ” Range: 0 9 We watch out for each others children in this neighborhood. (0 (b = .25, 95% CI: 20.63 to 1.14), [definitely disagree] to 3 [definitely agree]) adj Cronbach a: 0.81 “When we encounter difficulties, we know where to go for help in whereas children ages 6 to 17 in our community.” (0 [definitely disagree] to 3 [definitely agree]) grandparent-headed households had b a Phrasing of several component items changed slightly between NSCH versions. The table reports the phrasing from the poorer temperament ( adj =.23,95% 2016 NSCH. CI: 0.07 to 0.40). However, this b Response options differed between NSCH versions. 2016: (0) none of the time, (1) some of the time, (2) most of the time, association was not robust to the and (3) all of the time. 2017 and 2018: (0) never, (1) sometimes, (1) approximately half the time, (2) most of the time, and (3) always. removal of children with ADHD from c Response options differed between NSCH versions. 2016: (0) all of the time, (1) most of the time, (2) some of the time, (3) the sample (badj = .19, 95% CI: 20.01 none of the time. 2017 and 2018: (0) always, (1) most of the time, (2) approximately half the time, (2) sometimes, and to 0.38). (3) never. d Response options differed between NSCH versions. 2016 and 2017: (0) definitely true, (1.5) somewhat true, and (3) not true. 2018: (0) always, (1.5) usually, (1.5) sometimes, and (3) never. Caregiver Aggravation e fi Response options differed between NSCH versions. 2016 and 2017: (0) not true, (1.5) somewhat true, and (3) de nitely Similarly, no association was noted true. 2018: (0) never, (1.5) sometimes, (1.5) usually, and (3) always. between household type and aggravation among caregivers of experienced 1.22 (95% confidence confounders, children in grandparent- children ages 3 to 5 (badj =.17,95%CI: interval [CI]: 1.07 to 1.38) more ACEs headed households experienced 20.23 to 0.57), whereas grandparent than children in parent-headed significantly more ACEs overall caregivers of children ages 6 to 17 households. Even after adjusting for (Table 4). were more likely to experience

Downloaded from www.aappublications.org/news by guest on September 29, 2021 4 RAPOPORT et al TABLE 2 Sample Characteristics and Demographics of Grandparent Households and Parent but this difference was not robust to – – Households of Children Ages 3 17, 2016 2018 NSCH (N = 80 646) adjustment for confounders (badj = Characteristics Grandparent Parent Rao–Scott P .10, 95% CI: 20.26 to 0.46). Households Householdsa Adjusted Caregivers in grandparent-headed (n = 2407) (n = 78 239) F-statistic households had slightly more n %b n %b supportive neighborhoods (badj = .32, Primary caregiver sex 8.5 .004 95% CI: 0.04 to 0.59). No differences Male 674 24.9 25 347 31.1 were noted in caregiver coping (aOR Female 1697 75.1 52 550 68.9 = 0.96, 95% CI: 0.77 to 1.19). Primary caregiver 34.7 ,.001 Additionally, in 2-caregiver education fi Less than or equal to 48 7.3 694 4.2 households, no statistically signi cant eighth grade differences were noted in caregiver Ninth to 12th grade, 202 20.9 1535 7.5 likelihood of having someone to turn no diploma to for day-to-day emotional support High school 687 28.6 8047 14.5 with parenting and raising children graduate or GED Vocational or trade 209 6.8 3614 6.1 (70.2% grandparents versus 76.1% program parents, aOR = 1.00, 95% CI: 0.75 to Some college 481 14.3 11 395 13.9 1.35). Among 1-caregiver households, Associate degree 269 10.1 8173 8.5 fewer grandparents reported having Bachelor’s degree 281 6.1 24 898 25.4 someone for day-to-day emotional Master’s degree 153 4.3 14 362 14.9 Doctorate or 37 1.6 5040 5.0 support, but no differences were professional noted in the adjusted model (59.4% degree grandparents versus 69.0% parents, Household income, % of 51.2 ,.001 aOR = 0.77, 95% CI: 0.53 to 1.10). federal poverty level 0–99 563 33.2 7169 17.9 100–199 585 31.8 11 577 21.0 DISCUSSION 200–399 770 21.8 24 168 27.7 $400 489 13.1 35 325 33.4 In this cross-sectional analysis of No. caregivers 77.8 ,.001 a large nationally representative 1 caregiver 631 30.6 10 115 15.0 sample of children ages 3 to 17, 2 caregivers 1776 69.4 68 124 85.0 children being raised by grandparents Child age, y 2.7 .07 were more likely to have had adverse 3–8 406 21.7 13 544 19.1 9–12 958 43.5 27 022 40.2 experiences and a diagnosis of ADHD. 13–17 1043 34.9 37 673 40.7 Additionally, school-aged children in Child sex 0.0 .98 grandparent-headed households had Male 1240 51.0 40 300 51.0 poorer temperaments, and their Female 1167 49.0 37 939 49.0 caregivers experienced greater Child race and ethnicity 46.0 ,.001 Hispanic 303 20.2 8471 24.7 aggravation from parenting. White, non-Hispanic 1406 40.7 55 970 53.6 Importantly, after excluding children Black, non-Hispanic 376 30.5 4003 11.4 with ADHD from our analyses, Multiracial or other, 322 8.6 9795 10.4 differences in child temperament and non-Hispanic caregiver aggravation were no longer , Child health status 47.4 .001 fi Excellent or very 1985 78.2 72 078 90.4 statistically signi cant. Additionally, good health although ADHD was more prevalent Good, fair, or poor 415 21.8 5958 9.6 among children in grandparent- health headed households, we did not find GED, general equivalency diploma. differences in inattention and a Parent households included 2 biological or adoptive parents, 1 biological or adoptive parent plus 1 stepparent, or 1 restlessness among young children biological or adoptive parent. b Prevalence figures weighted to be nationally representative. without an ADHD diagnosis. The results from our analyses, in b elevated aggravation ( adj = .29, 95% Additional Caregiver Measures many ways, are similar to what is CI: 0.08 to 0.49). This association was Caregivers in grandparent-headed currently known about grandparents not robust to the removal of children households had more frequent raising grandchildren. Compared with b with ADHD from the sample ( adj = .16, quality family interactions with their parent caregivers, custodial 95% CI: 20.06 to 0.38). child (b = .54, 95% CI: 0.19 to 0.90), grandparents had lower educational

Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 146, number 3, September 2020 5 TABLE 3 ACEs Among Grandparent and Parent Households, 2016–2018 NSCH (N = 80 646) accounting for ACE exposure, ACEs Grandparent Parent ORb (95% CI) aORb,c (95% CI) although our effect estimate was Households Householdsa attenuated, grandparent-headed (n = 2407) (n = 78 239) households remained more likely to b b n % n % have children with ADHD. A possible Child experienced parent explanation may be the heritability of or guardian divorcing ADHD.36 Mothers with ADHD are or separating No 852 43.6 61 268 77.5 Reference Reference more likely to experience unplanned 37 Yes 1478 56.4 16 251 22.5 4.47 (3.64 to 5.49) 4.34 (3.28 to 5.73) pregnancies, which are a common Child experienced parent precipitating factor for grandparents or guardian dying raising their grandchildren; because No 1953 87.0 75 496 97.3 Reference Reference Yes 359 13.0 1867 2.7 5.40 (4.09 to 7.14) 3.84 (2.63 to 5.59) the children of mothers with ADHD Child experienced parent are more likely to also have ADHD, or guardian serving this pathway may explain elevated time in jail rates of ADHD among children in No 1331 64.4 73 728 94.2 Reference Reference Yes 986 35.6 3534 5.8 8.92 (7.34 to 10.84) 6.24 (4.92 to 7.93) grandparent-headed households. Child saw or heard Furthermore, there is a higher parents or adults prevalence of substance abuse among slapping, hitting, adults with ADHD38; substance abuse, kicking, or punching one another in the as well as the elevated incarceration home rates associated with it, further No 1687 79.1 74 131 95.3 Reference Reference contributes to grandparent caregiving Yes 604 20.9 3086 4.7 5.38 (4.35 to 6.66) 4.32 (3.35 to 5.57) ’ Child was a victim of because of parents inability to violence or witnessed effectively care for their children. violence in the Importantly, we did not find evidence neighborhood of differences in inattention and No 1986 88.8 74 774 96.4 Reference Reference Yes 318 11.2 2440 3.6 3.34 (2.59 to 4.31) 2.27 (1.67 to 3.09) restlessness among preschool-aged Child lived with anyone children without an ADHD diagnosis who was mentally ill, between grandparent-headed and suicidal, or severely parent-headed households. depressed No 1828 85.9 70 759 92.6 Reference Reference In our sample, grandparent-headed Yes 476 14.1 6341 7.4 2.08 (1.68 to 2.56) 2.00 (1.57 to 2.54) Child lived with anyone households were much more likely to who had a problem have children with ADHD. Because with alcohol or drugs children with ADHD tend to exhibit No 1375 70.3 70 850 92.6 Reference Reference more externalizing behaviors and are Yes 932 29.7 6362 7.4 5.26 (4.35 to 6.35) 5.20 (4.17 to 6.47) often perceived as harder to care for OR, odds ratio. by their caregivers, it is not surprising a Parent households included 2 biological or adoptive parents, 1 biological or adoptive parent plus 1 stepparent, or 1 biological or adoptive parent. that we found poorer child b Weighted to be nationally representative. temperament and elevated parental c Adjusted for caregiver sex, caregiver education, household poverty level, number of caregivers, child age, child sex, child aggravation in grandparent-headed race and ethnicity, and child health status. households. These findings are also consistent with the current attainment and household income. opioid crisis,6 our findings are similar understanding of behavioral and Children raised by grandparents were to previous research about the social characteristics of children also more likely to have experienced precipitating factors of the grandfamily raised by grandparents.3,8,25,39 a variety of ACEs than children raised household structure, as well as its by parents, reinforcing past findings financial and health correlates.3,22 The fact that significant differences in about children in nonparental care.34 child temperament and parental Adverse experiences have been shown Given the established association aggravation disappeared when we to have cumulative associations with between ACE exposure and ADHD,33 excluded children with ADHD from behavioral problems, developmental it is unsurprising that we identified analyses suggests that ADHD itself delays, and difficulties in school and elevated rates of ADHD among both may be responsible for many adult outcomes like substance abuse preschool and school-aged children between-group differences in child and depression.35 In light of rising raised in grandparent-headed behaviors and social characteristics. incarceration rates4,5 and the current households. However, after Children in nonparental care are at

Downloaded from www.aappublications.org/news by guest on September 29, 2021 6 RAPOPORT et al TABLE 4 ACE Composite Measures Among Grandparent and Parent Households, 2016–2018 NSCH controlling for sociodemographic (N = 77 281) confounders, we found that ACE Composite Grandparent Parent ORb (95% CI) aORb,c (95% CI) grandparents report greater support Measures Households Householdsa from their neighborhoods. However, (n = 2148) (n = 75 133) this difference in neighborhood n %b n %b support between grandparents and Child experienced parents was fairly small and unlikely $1 ACE to have substantial implications. No 447 28.4 52 567 68.8 Reference Reference Yes 1701 71.6 22 566 31.2 5.56 (4.27 to 7.24) 5.20 (3.71 to 7.29) Given grandparent caregivers’ limited Child experienced $2 ACEs access to emotional support, it has No 956 56.6 66 326 87.8 Reference Reference been suggested that grandparent Yes 1192 43.4 8807 12.2 5.49 (4.50 to 6.71) 4.88 (3.79 to 6.30) households may be particularly in Child experienced need of social support services to $ 3 ACEs cope with the difﬁculties associated No 1310 70.0 71 014 94.2 Reference Reference 21 Yes 838 30.0 4119 5.8 6.99 (5.70 to 8.57) 6.19 (4.78 to 8.01) with raising grandchildren. In Child experienced conjunction with previous evidence $4 ACEs that grandparents who serve as the No 1605 82.5 73 238 97.2 Reference Reference primary caregiver for a child were Yes 543 17.5 1895 2.8 7.50 (6.04 to 9.33) 6.41 (4.86 to 8.45) twice as likely to develop symptoms OR, odds ratio. of depression than noncaregiving a Parent households included 2 biological or adoptive parents, 1 biological or adoptive parent plus 1 stepparent, or 1 41 biological or adoptive parent. grandparents, it is vital that b Weighted to be nationally representative. grandparents raising grandchildren c Adjusted for caregiver sex, caregiver education, household poverty level, number of caregivers, child age, child sex, child take advantage of support groups in race and ethnicity, and child health status. their community and on-line. Policies to create local grandparent-raising- higher risk for living under unstable than differences inherent to grandchildren support programs can caregiving arrangements,35 which grandparent and parent caregivers, may provide ways to cope, informational puts these children at greater risk for explain this small disparity. Additionally, support, social support, and resource externalizing behavior problems.40 although caregivers in grandparent- connections to caregivers.20,42,43 In However, many other factors may also headed households were more likely to a recent randomized controlled trial, be involved. Although our analyses experience aggravation from parenting, Pandey et al43 (2018) compared the controlled for many key we did not identify differences in effectiveness of traditional child sociodemographic variables, residual caregivers’ reported ability to cope welfare services with 3 community- confounding related to other risk with the daily demands of caregiving. based forms of support for custodial factors (eg, prenatal alcohol exposure, We found that grandparent and parent grandmothers. They concluded that lead exposure, or family history of caregivers in both 1- and 2-caregiver traditional child welfare is better ADHD) may have impacted our households did not differ in their odds suited for the needs of parents and analyses. Large-scale longitudinal of having someone to turn to for day- foster parents and that peer-based studies examining children in to-day emotional support with community programs provide greater nonparental care would be necessary parenting or raising children, after informational and emotional support to determine the extent to which controlling for confounders. to grandmothers raising their externalizing behavior contributes to Of concern, 24% of parent caregivers grandchildren. Given the difﬁculties child placement in nonparental care. and 30% of grandparent caregivers in many grandparent caregivers face Although the differences between 2-caregiver households did not have with respect to emotional support grandparent- and parent-headed someone to turn to for day-to-day with parenting, pediatricians should households have important implications emotional support. Moreover, 31% of refer these caregivers to community- for children and caregivers, so too do single-parent caregivers and 41% of based organizations oriented toward the similarities between these 2 groups. single-grandparent caregivers lacked supporting grandparents raising For example, although we found that this type of support. Given the grandchildren; in particular, grandparent-headed households have demographic characteristics of our pediatricians should be mindful of the a higher frequency of quality sample, it is possible that custodial additional support needed by interactions between caregivers and grandparents may not have as many grandparents in one-caregiver children, our analyses indicate that friends and family to rely on for households. Organizations such as sociodemographic differences, rather parenting support. Notably, after Grandfamilies.org provide a directory

Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 146, number 3, September 2020 7 of national and state-specific caregiver variables and child caregivers, the NSCH did not include resources and support groups, which variables using the same large, questions evaluating important pediatricians can use to guide and nationally representative sample. characteristics of caregivers and the counsel custodial grandparents.44 This methodology allows for caregiver–child relationship, such as fi Although past studies have revealed consistent interpretations of ndings caregiver race or the duration of time that children raised in grandparent- pertaining to both children and that the child has been in the care of headed households may have poorer caregivers. Additionally, whereas their parent or grandparent. Our outcomes throughout adolescence and researchers of many studies analyzed inability to account for these adulthood, our findings suggest that individual Likert items when underlying household characteristics fi efforts to identify children who would examining child and caregiver may have impacted our ndings. benefit from medical or mental health outcomes, in our study, we used interventions would be best served composite measures, reducing the CONCLUSIONS impact of random variation and through screening that identifies ACEs In this study, we highlight many measurement error on our findings. and ADHD. The American Academy of profound differences between Pediatrics has suggested that However, the use of these composite grandparent- and parent-headed pediatricians screen their patients for measures also introduced notable households. Even after adjusting for early childhood adversity to identify limitations to this study. In particular, potential confounders, children in children at high risk for toxic stress.45 with the exception of the inattention grandparent-headed households were Given that pediatricians tend to under- and restlessness scale, we could not much more likely to have experienced identify risk factors such as ACEs and evaluate the construct validity of our psychosocial adversity. Additionally, unmet social needs,46 and given the composite scales. Another potential school-aged children in grandparent- elevated prevalence of ACEs among limitation of these measures was headed households had poorer children in grandparent-headed caused by minor variation between temperaments and their caregivers households, pediatricians should be individual components in different reported greater aggravation. particularly mindful of the importance iterations of the NSCH, which may However, no differences were noted of early childhood screening in this have introduced some inconsistency with respect to how well caregivers population. Continued research into the to our composite measures. However, were handling the day-to-day complex interplay between childhood the majority of our composite demands of parenting. With nearly 3 adversity, ADHD, and physical and measures had strong internal million children now being raised by emotional health is essential for the consistency, indicating that the one or both grandparents, development and refinement of individual component items were pediatricians must be mindful of the effective screening and interventions in closely related to each other. demographic, psychosocial, and this high-risk population. In addition to the limitations parenting challenges that characterize One strength of our study is its large, introduced by our composite many grandparent-headed nationally representative sample of measures, our study was also limited households. In addition to screening 80 646 caregivers of children, by the reliance on caregiver report. It children in these families for including 2407 grandparents raising is possible that grandparents are adversity and heightened stress, their grandchildren, making this the more critical about or have higher pediatricians should refer these largest study to date examining expectations for the behavior of their families to appropriate support childhood adversity, caregiver–child grandchildren. However, this type of groups and other resources relationships, and other related bias is less likely to apply to reports committed to meeting the needs of measures in grandparent-headed about ACEs or medical diagnosis of parenting grandparents. households using a nationally ADHD. Additionally, the NSCH representative sample. The sample questionnaire’s focus on lifetime size allowed analytical models to exposure to adversity did not allow ABBREVIATIONS control for many key confounders. us to determine if the ACEs occurred ACE: adverse childhood experience Whereas most previous studies of before or after the child’s placement ADHD: attention-deficit/ grandparent households have focused with their caregiver. Finally, although hyperactivity disorder on psychological, behavioral, and we had the ability to control for aOR: adjusted odds ratio health measures among either the demographic differences between CI: confidence interval caregivers or the children, in this groups, as with any retrospective NSCH: National Survey of study, we directly compare cross-sectional analysis, residual Children’s Health grandparent households and parent confounding remains a possibility. For badj: adjusted beta households with respect to both example, beyond the number of

Downloaded from www.aappublications.org/news by guest on September 29, 2021 8 RAPOPORT et al PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). Copyright © 2020 by the American Academy of Pediatrics FINANCIAL DISCLOSURE: The authors have indicated they have no ﬁnancial relationships relevant to this article to disclose. FUNDING: No external funding. POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conﬂicts of interest to disclose.

REFERENCES 1. Wiltz T. Why More Grandparents Are 11. Whitley DM, Fuller-Thomson E. The service providers working with Raising Children. Philadelphia, PA: The health of the nation’s custodial grandparents raising grandchildren. Pew Charitable Trusts; 2016 grandfathers and older single fathers: J Appl Gerontol. 2015;34(2):138–157 fi 2. Arditti JA. Family Problems: Stress, ndings from the behavior risk factor 20. Lent JP, Otto A. Grandparents, Risk, and Resilience. Hoboken, NJ: John surveillance system. Am J Men Health. grandchildren, and caregiving: the – Wiley & Sons; 2014 2017;11(6):1614 1626 impacts of America’s substance use – 3. Sprang G, Choi M, Eslinger JG, Whitt- 12. Whitley DM, Fuller-Thomson E, crisis. Generations. 2018;42(3):15 22 Woosley AL. The pathway to Brennenstuhl S. Health characteristics 21. Hayslip B Jr., Kaminski PL. grandparenting stress: trauma, of solo grandparent caregivers and Grandparents raising their fi relational conflict, and emotional well- single parents: a comparative pro le grandchildren: a review of the being. Aging Ment Health. 2015;19(4): using the behavior risk factor literature and suggestions for practice. 315–324 surveillance survey. Curr Gerontol Gerontologist. 2005;45(2):262–269 Geriatr Res. 2015;2015:630717 4. Bloom B, Phillips S. In whose best 22. Hayslip B Jr., Fruhauf CA, Dolbin- interest? The impact of changing public 13. Baker LA, Silverstein M. Depressive MacNab ML. Grandparents raising policy on relatives caring for children symptoms among grandparents raising grandchildren: what have we learned with incarcerated parents. In: Children grandchildren: the impact of over the past decade? Gerontologist. with Parents in Prison. Abingdon, participation in multiple roles. 2017;57(6):1196 – J Intergener Relatsh. 2008;6(3):285–304 United Kingdom: Routledge; 2017:63 74 23. Nanthamongkolchai S, Munsawaengsub 5. Travis J, Western B, Redburn FS. The 14. Tang F, Xu L, Chi I, Dong X. Psychological C, Nanthamongkolchai C. Comparison of Growth of Incarceration in the United well-being of older Chinese-American the health status of children aged States: Exploring Causes and grandparents caring for grandchildren. between 6 and 12 years reared by Consequences. Washington, DC: J Am Geriatr Soc. 2016;64(11): grandparents and parents. Asia Pac National Academies Press; 2014 2356–2361 J Public Health. 2011;23(5):766–773 6. Manchikanti L, Helm S II, Fellows B, 15. Yalcin BM, Pirdal H, Karakoc EV, Sahin 24. Goulette NW, Evans SZ, King D. Exploring et al. Opioid epidemic in the United EM, Ozturk O, Unal M. General health the behavior of juveniles and young States. Pain Physician. 2012;15(3, perception, depression and quality of adults raised by custodial suppl):ES9-ES38 life in geriatric grandmothers providing grandmothers. Child Youth Serv Rev. care for grandchildren. Arch Gerontol – 7. Livingston G, Parker K. Since the Start 2016;70:349 356 Geriatr. 2018;79:108–115 of the Great Recession, More Children 25. Edwards OW, Daire AP. School-age Raised by Grandparents. Washington, 16. Hayslip B Jr., Glover RJ. Custodial children raised by their grandparents: DC: Pew Research Center; 2010 grandparenting: perceptions of loss by problems and solutions. Journal of non-custodial grandparent peers. 8. Ellis RR, Simmons T. Coresident Instructional Psychology. 2006;33(2): Omega (Westport). 2008;58(3):163–175 – Grandparents and Their Grandchildren: 113 119 2012. Washington, DC: US Department of 17. Taylor MF, Marquis R, Coall DA, Batten R, 26. Berger LM, Bruch SK, Johnson EI, James Commerce; 2014 Werner J. The physical health dilemmas S, Rubin D. Estimating the “impact” of facing custodial grandparent 9. Hughes ME, Waite LJ, LaPierre TA, Luo Y. out-of-home placement on child well- caregivers: policy considerations. All in the family: the impact of caring being: approaching the problem of Cogent Med. 2017;4(1):1292594 for grandchildren on grandparents’ selection bias. Child Dev. 2009;80(6): – health. J Gerontol B Psychol Sci Soc Sci. 18. Gerard JM, Landry-Meyer L, Roe JG. 1856 1876 2007;62(2):S108–S119 Grandparents raising grandchildren: 27. Bramlett MD, Blumberg SJ. Family the role of social support in coping ’ 10. Whitley DM, Fuller-Thomson E. African- structure and childrens physical and with caregiving challenges. Int J Aging American solo grandparents raising mental health. Health Aff (Millwood). Hum Dev. 2006;62(4):359–383 – grandchildren: a representative profile 2007;26(2):549 558 of their health status. J Community 19. Fruhauf CA, Pevney B, Bundy-Fazioli K. 28. Smith GC, Palmieri PA. Risk of Health. 2017;42(2):312–323 The needs and use of programs by psychological difficulties among

Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 146, number 3, September 2020 9 children raised by custodial 35. Beal SJ, Greiner MV. Children in 42. Hayslip B Jr., Blumenthal H, Garner A. grandparents. Psychiatr Serv. 2007; nonparental care: health and social Social support and grandparent 58(10):1303–1310 risks. Pediatr Res. 2016;79(1–2): caregiver health: one-year longitudinal 184–190 ﬁndings for grandparents raising 29. Goodman C, Silverstein M. their grandchildren. J Gerontol B Grandmothers raising grandchildren: 36. Sprich S, Biederman J, Crawford MH, Psychol Sci Soc Sci. 2015;70(5): family structure and well-being in Mundy E, Faraone SV. Adoptive and 804–812 culturally diverse families. biological families of children and Gerontologist. 2002;42(5):676–689 adolescents with ADHD. J Am Acad Child 43. Pandey A, Littlewood K, Cooper L, et al. Adolesc Psychiatry. 2000;39(11): Connecting older grandmothers raising 30. Wang CD, Hayslip B Jr., Sun Q, Zhu W. 1432–1437 grandchildren with community Grandparents as the primary care resources improves family resiliency, providers for their grandchildren: 37. Owens EB, Hinshaw SP. Adolescent social support, and caregiver self- a cross-cultural comparison of Chinese mediators of unplanned pregnancy efﬁcacy. J Women Aging. 2019;31(3): and U.S. Samples. Int J Aging Hum Dev. among women with and without 269–283 2019;89(4):331–355 childhood ADHD. J Clin Child Adolesc – Psychol. 2020;49(2):229 238 44. Grandfamilies.org. GrandFacts: state 31. U.S. Census Bureau. 2016 National 38. Kalbag AS, Levin FR. Adult ADHD and fact sheets for grandparents and other Survey of Children’s Health substance abuse: diagnostic and relatives raising children. Available at: Methodology Report. Washington, DC: treatment issues. Subst Use Misuse. www.grandfamilies.org/state-fact- U.S. Census Bureau; 2018 2005;40(13–14):1955–1981, 2043–2048 sheets. Accessed June 3, 2020 32. US Census Bureau. 2018 National 39. Radel L, Bramlett M, Chow K, Waters A. 45. Garner AS, Shonkoff JP; Committee on Survey of Children’s Health Guide to Children Living Apart From Their Psychosocial Aspects of Child and Multi-Year Estimates. Washington, DC: Parents: Highlights From the National Family Health; Committee on Early US Census Bureau; 2019 Survey of Children in Nonparental Care. Childhood, Adoption, and Dependent 33. Brown NM, Brown SN, Briggs RD, Washington, DC: US Department of Care; Section on Developmental and German M, Belamarich PF, Oyeku SO. Health and Human Services; 2016 Behavioral Pediatrics. Early childhood Associations between adverse 40. Newton RR, Litrownik AJ, Landsverk JA. adversity, toxic stress, and the role of childhood experiences and ADHD Children and youth in foster care: the pediatrician: translating diagnosis and severity. In: Proceedings distangling the relationship between developmental science into lifelong from the 2014 Pediatric Academic problem behaviors and number of health. Pediatrics. 2012;129(1). Societies Meeting; May 3–6, 2014; placements. Child Abuse Negl. 2000; Available at: www.pediatrics.org/cgi/ Vancouver, Canada 24(10):1363–1374 content/full/129/1/e224 34. Radel LF, Bramlett MD. Children in 41. Minkler M, Fuller-Thomson E, Miller D, 46. Kerker BD, Storfer-Isser A, Szilagyi M, Nonparental Care: Findings from the Driver D. Depression in grandparents et al. Do pediatricians ask about 2011–2012 National Survey of Children’s raising grandchildren: results of adverse childhood experiences in Health. Washington, DC: US Department a national longitudinal study. Arch Fam pediatric primary care? Acad Pediatr. of Health and Human Services; 2014 Med. 1997;6(5):445–452 2016;16(2):154–160

Downloaded from www.aappublications.org/news by guest on September 29, 2021 10 RAPOPORT et al Family Well-being in Grandparent- Versus Parent-Headed Households Eli Rapoport, Nallammai Muthiah, Sarah A. Keim and Andrew Adesman Pediatrics originally published online August 3, 2020;

Updated Information & including high resolution figures, can be found at: Services http://pediatrics.aappublications.org/content/early/2020/07/30/peds.2 020-0115 References This article cites 33 articles, 1 of which you can access for free at: http://pediatrics.aappublications.org/content/early/2020/07/30/peds.2 020-0115#BIBL Subspecialty Collections This article, along with others on similar topics, appears in the following collection(s): Developmental/Behavioral Pediatrics http://www.aappublications.org/cgi/collection/development:behavior al_issues_sub Psychosocial Issues http://www.aappublications.org/cgi/collection/psychosocial_issues_s ub Adoption & Foster Care http://www.aappublications.org/cgi/collection/adoption_-_foster_car e_sub Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.aappublications.org/site/misc/Permissions.xhtml Reprints Information about ordering reprints can be found online: http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on September 29, 2021 Family Well-being in Grandparent- Versus Parent-Headed Households Eli Rapoport, Nallammai Muthiah, Sarah A. Keim and Andrew Adesman Pediatrics originally published online August 3, 2020;

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://pediatrics.aappublications.org/content/early/2020/07/30/peds.2020-0115

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2020 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on September 29, 2021