Caregiver Expressed Emotion in Families of Youth With Spina Bifida: Demographic, Medical, and Family Correlates

Abstract

Objective

Caregiver expressed emotion (EE), an interview-based measure of emotional valence within an interpersonal relationship, is associated with psychosocial outcomes across multiple conditions. Guided by a model implicating a bidirectional role of “Chronic Family Stress” in the unfolding of EE in family environments, the current study examined demographic, medical, and family-level variables in association with EE in caregivers of children with spina bifida (SB).

Methods

Data were combined from 2 distinct studies of families with a child with SB, resulting in a sample of 174 (ages 8–17). Linear regressions examined the family stressors and child variables in association with maternal and paternal warmth and criticism, as coded from EE interviews.

Results

Higher levels of family stress were associated with paternal criticism (p = .03), while having non-Hispanic White children was associated with both maternal and paternal criticism (ps < .005). Having children younger in age (ps < .01) and without a shunt (ps < .01) was associated with higher warmth.

Conclusions

Family stressors, absence of the negative impacts of systemic racism, shunt status, and age appear to be associated with the expression of EE in caregivers of a child with SB. Findings highlight multiple assessment considerations, including assessing EE when children are younger to engage caregivers with children with SB when they are more likely to be expressing more warmth. Pinpointing factors associated with caregiver EE in SB will help to better identify families at risk for high levels of criticism and also aid in the development of targeted prevention and intervention programs.

Introduction

Spina bifida (SB) is a congenital birth defect that affects nearly 1 in 1,000 births each year (Copp et al., 2015) and is associated with significant impacts to physical, psychological, and intellectual development (Stiles-Shields et al., 2019). SB develops in utero when the neural tube fails to properly close, creating a lesion. The location of this lesion is related to the severity of SB (i.e., sacral, lumbar, thoracic; Copp et al., 2015). Common challenges associated with SB include paralysis of the lower extremities, complicated bladder and bowel programs, neurological problems (e.g., hydrocephalus), and cognitive deficits (e.g., IQ, executive functioning, attention; Copp et al., 2015). In addition, youth with SB are at an increased risk for experiencing several psychosocial difficulties, including depression, anxiety, and poorer social adjustment. Indeed, youth with SB tend to have different social development trajectories and experiences than their typically developing peers, such as having fewer friendships and less contact with peers outside of school (Holmbeck et al., 2010). Considering the numerous, complex needs of youth with SB, studies that identify potential correlates—particularly those that may be targeted with intervention—are of the utmost importance.

The family environment of individuals with SB is a particularly relevant and potentially influential factor. Past research has documented links between the overall family environment, specific parenting practices, and adjustment among youth with SB (Essner & Holmbeck, 2010; Holmbeck & Devine, 2010). To better understand the unique interaction patterns within families of children with SB, research has also begun exploring the impact of expressed emotion (EE; Kelly et al., 2011). EE, which is a measure of the emotional valence within an interpersonal relationship as reported by one individual in an interview format, was first evaluated in families of individuals with schizophrenia and measured five dimensions of the family relationship: emotional over-involvement, hostility, criticism, warmth, and positive comments (Brown & Rutter', 1966). Through this research, EE was identified as a significant predictor of symptom relapse (Hooley, 1985; Ma et al., 2021). Subsequent studies have provided additional support for the measurement of EE, documenting associations between EE and mood disorders, eating disorders, and non-suicidal self-injury across the life span (Duclos et al., 2018; James & Gibb, 2019; Weintraub et al., 2021).

Considering that individuals with SB may rely more heavily on caregivers for support due to their medical needs and social isolation from peers, they may be particularly susceptible to the effects of criticism within the home. Conversely, warmth may be beneficial and help youth compensate for psychosocial difficulties. While EE has been noted in caregivers of children with developmental disabilities (Thompson et al., 2018), to our knowledge, only one study to date has examined EE in families who have a child with SB. Specifically, Kelly and colleagues (2011), based on a portion of the data that were included in the current report, found that less warmth and more criticism were associated with higher depressive symptoms in youth with SB. These findings are an important first step in expanding our understanding of caregiver EE in this population and its contributions to child outcomes. However, additional research examining correlates associated with EE is warranted. Indeed, such correlates likely precede the potential expression of EE and therefore may help identify families who are at risk for displaying such behaviors as well as the negative effects associated with this important aspect of the family environment.

Pinpointing factors associated with caregiver EE in SB can not only help to better identify families at risk for high levels of criticism but also aid in the development of targeted prevention and intervention programs. To ground an examination of correlates of the variable presentation and experience of SB with caregiver EE, a model of Toxic Family Stress was used to drive measure selection (Peris & Miklowitz, 2015). Specifically, Toxic Family Stress refers to chronic or frequent stressors, such that familial protective factors are not well supported. Toxic Family Stress has been proposed as a means of understanding the onset and maintenance of psychiatric symptoms via bidirectional effects among caregiver, child, and family-level variables, including EE. For example, in the face of chronic or frequent stressors (e.g., systemic, family, and/or condition-related) a caregiver may begin to respond maladaptively via expressing criticism. This EE likely contributes to further stress in the family system and, therefore, perpetuates the likelihood of continued or increased criticism, which then increases the likelihood of the onset or exacerbation of a child psychiatric disorder, which places further chronic stress on the family environment, and so on in a perpetuating cycle (Peris & Miklowitz, 2015). Conversely, caregiver warmth may serve to buffer some of the impacts of chronic stress, and therefore decrease the likelihood of caregiver criticism and the perpetuating cycle described above. To avoid conflating the experience of having a child with SB with the term “toxic,” the framework will hereafter be referred to as “Chronic” Family Stress.

The application of the model for the current study is shown in Figure 1 and focuses on stressors that likely precede or are correlates of EE. Specifically, measure selection was driven by likely experiences of five types of chronic stress for families with a child with SB. The first variable selected was condition-related stressors. As previously noted, there is wide variability in the presentation of SB. Most children demonstrate significant cognitive and physical challenges, whereas some do not (Stiles-Shields et al., 2019). However, caregivers of children with SB have been found to demonstrate more negative parenting behaviors (e.g., psychological control, intrusiveness) than caregivers of typically developing children (Holmbeck & Devine, 2010). In addition, emerging evidence suggests that lower cognitive ability in children with SB is related to lower levels of adaptive parenting practices (Winning et al., 2020). Given that SB and its severity appear to affect various aspects of the family dynamic, these factors may also impact caregiver EE. For instance, caregivers of children with SB who have a more severe presentation based on type (i.e., myelomeningocele is the most severe form of SB), ambulation status, lesion level, or shunt status may respond to their child in a way that is less warm and/or more critical. Indeed, mothers of children with intellectual disabilities (which has an overlapping symptom profile with SB; Stiles-Shields et al., 2019) have been found to exhibit more negative behaviors toward their children and demonstrate higher levels of criticism (Beck et al., 2004).

Figure 1.

Chronic (toxic) stress model to predict caregiver expressed emotion

Family-level stressors, caregiver mental health, experiences of systemic racism, and socioeconomic status (SES) were selected as the remaining variables. Family-level stressors (the second stressor variable) include factors ranging from relational strains to financial struggles, and are an important component of the overall stressors involved in caring for a child with a chronic medical condition (Pinquart, 2018). Relatedly, the third variable selected was caregiver mental health. Indeed, having a child with SB can have consequences for caregivers’ psychological well-being, leading to greater stress, anxiety, and depression (Holmbeck & Devine, 2010; Vermaes et al., 2005). In turn, it is possible that this elevated distress may heighten caregiver EE behaviors, as associations between these variables have been found in other pediatric populations (Baker et al., 2000; Hastings & Lloyd, 2007). Fourth, the experience of systemic racism was selected as a source of chronic family stress. Indeed, the pervasive and multi-level impacts of systemic racism on individuals and families with minoritized identities cannot be understated as a life-long source of stress (Jones et al., 2020). While EE has been hypothesized to be a culturally specific construct (Bhugra & McKenzie, 2003; López et al., 2009), the literature appears to have overlooked broader contextual experiences with systemic inequities and injustices. Finally, SES was selected as the final source of chronic family stress. The increased level of stress associated with disparities inflicted upon families with fewer resources (i.e., lower SES) has been associated with negative family environments and even increased levels of inflammation in children with chronic conditions (Farrell et al., 2018).

The purpose of the current study was to extend the existing literature on EE by applying a model of chronic family stress to examine the association of five types of chronic family stress: (a) condition-related stressors; (b) family-level stress; (c) caregiver mental health symptoms; (d) the experience of systemic racism; and (e) SES, with caregiver EE in families of youth with SB (Peris & Miklowitz, 2015). It was hypothesized that greater levels of these types of chronic family stress would be associated with higher caregiver criticism, whereas lower levels of chronic family stress would be associated with caregiver warmth. Given inconclusive findings regarding the effects of child age and gender on EE (Beck et al., 2004; Cartwright et al., 2011; Han & Shaffer, 2014), analyses examining associations among these variables and caregiver EE were viewed as exploratory.

Methods

Participants

Participants were from two separate archival samples, based on two larger, longitudinal (non-intervention) investigations of youth with SB and their families (e.g., Devine et al., 2012; Holmbeck et al., 2003). Both studies recruited youth with SB from Midwest hospitals and through a statewide SB association. For study 1, which started in the early 1990s, 68 youth with SB began the study at eight to nine years of age and were assessed every two years (e.g., 8–9 years old at time 1 and 10–11 years old at time 2). Participants were recruited from four sources (a children’s hospital, a children’s hospital caring exclusively for children with physical disabilities, a university-based medical center, and a statewide SB association) and inclusion criteria were: (a) a diagnosis of SB; (b) aged 8–9 years at time 1; (c) ability to speak/read in English; and (d) live within 120 miles of Chicago, IL. For study 2, 140 youth with SB began the study at 8–15 years of age and were also assessed every 2 years (e.g., 8–15 years old at time 1 and 10–17 years old at time 2). Inclusion criteria were: (a) a diagnosis of SB; (b) aged 8–15 years at time 1; (c) ability to speak/read in English or Spanish; (d) the involvement of at least one primary caregiver; (e) cognitive ability to complete study activities (e.g., questionnaires, neuropsychological testing); and (f) lives within 300 miles of Chicago, IL.

To create a sample for the current study that would include a diverse range of ages (i.e., 8–17), data from time 5 of study 1 (n = 52, 76% of original sample at time 1; when youth were 16–17 years old) and data from time 1 of study 2 (n = 140; when youth were 8–15 years old) were combined. From these samples, 18 participants were excluded due to: (a) missing EE-related data and/or (b) being 18 years of age (i.e., for participants aged 18 years and older, caregivers did not participate in the assessments). Thus, the final sample included 42 youth with SB and their caregivers from study 1 and 132 youth and their caregivers from study 2, for a total N of 174. Participants ranged in age from eight to 17 (M = 12.60 ± 3.05) and 64.4% were female (n = 112). Most of the sample was non-Hispanic White (60.3%; this lack of racial and ethnic diversity was driven by study 1), had myelomeningocele (the most severe form of SB; 85.1%), and a lumbar lesion level (61.5%). Table I displays the sample demographic and SB characteristics.

Table I.

Participant Demographic, Medical, and Psychosocial Characteristics (n = 174)

Demographic and medical characteristics
	M/SD (range); n (%)
Age	12.60/3.05 (8–18)
Hollingshead SES	40.38/14.86 (8–66)
Sex
Male	62 (35.6)
Female	112 (64.5)
Race
Asian	2 (1.1)
Black or African American	21 (12.1)
White	145 (83.3)
Othera	6 (3.3)
Ethnicity
Hispanic	40 (23.0)
Non-Hispanic	134 (77.0)
SB type
Myelomeningocele	148 (85.06)
Other	23 (13.22)
Missing	3 (1.72)
Lesion level
Sacral	35 (20.11)
Lumbar	107 (61.49)
Thoracic	25 (14.37)
Unsure	2 (1.15)
Missing	5 (2.87)
SB severity	6.81/1.68 (2–10)

Psychosocial characteristics
	M/SD (range)
Maternal criticism	13.11/2.97 (10–24.50)
Maternal warmth	20.44/5.15 (11–34.50)
Paternal criticism	12.69/3.11 (10–24)
Paternal warmth	20.13/5.57 (10.50–32.50)
Maternal FILE	3.01/0.93 (1–5)
Paternal FILE	2.68/0.85 (1–5)
Maternal FSS	1.99/.59 (1–5)
Paternal FSS	1.89/.61 (1–5)
Maternal SCL	0.35/0.36 (0–1.89)
Paternal SCL	0.26/0.27 (0–1.47)

Note. FILE = Family Inventory of Life Events; FSS = Family Stress Scale; M = mean; NIH = National Institutes of Health; SB = spina bifida; SCL = Symptom Checklist; SD = standard deviation; SES = socioeconomic status.

^aThe “Other” category is comprised of individuals who self-identified with racial groups outside of the NIH-specific categories at the time of assessment, including individuals who self-identified as more than one racial identity.

Procedure

Studies 1 and 2 were approved by university and hospital Institutional Review Boards (IRB). In compliance with IRB approval, informed child assent and guardian consent were obtained prior to all data collection for both studies. Trained graduate and undergraduate students conducted assessments for both studies in the families’ homes. For time 5 of study 1, youth with SB and caregivers independently completed self-report questionnaire packets. Following this, caregivers were audio-recorded as they privately responded to open-ended interview questions about their child’s functioning. Participants were compensated $100 for their time. For time 1 of study 2, data were collected across two in-home assessment sessions, which involved self-report questionnaire and interview completion. Families were compensated for their time with gifts (e.g., reusable water bottle) and $150.

Measures

In both studies, caregivers (i.e., mothers and fathers) individually completed self-report questionnaires about their child and/or family’s functioning that are described below.

Demographics

Caregivers of the children with SB completed a questionnaire detailing the child’s demographic information (e.g., age, race, ethnicity). Of note, both mothers and fathers were recruited to participate; however, gender identity, sexual orientation of the caregiver, and non-parent caregivers (e.g., grandparents) were not assessed. Families with one parent (i.e., a single mother or single father) were included. In addition, race and ethnicity were assessed based on National Institutes of Health categories of the time (e.g., time 1 of study 1 occurred in the early 1990s); the assessments of these variables do not reflect inclusive categories and options (e.g., “Other” being the only option for those who identified with more than one racial identity).

Caregiver Expressed Emotion

The methodology used to assess caregiver EE from studies 1 and 2 has been described in full detail elsewhere (Kelly et al., 2011). Briefly, for both studies, caregivers privately responded to 16 questions assessing their child’s functioning (e.g., future functioning, peer interactions, academic performance, autonomy). Each response was rated separately for warmth and criticism by trained coders, who rated items independently and in duplicate (i.e., each item was rated by two coders). Specifically, each response was rated for warmth based on the expressed fondness for and interest in the child (i.e., 1 = not at all warm, 3 = moderately warm, 5 = extremely warm); each response was rated for criticism based on the expressed distaste for or anger toward the child (i.e., 1 = neither critical nor hostile, 3 = moderately critical or hostile, 5 = severely critical or hostile). Content, emotional expressions (e.g., laughing), and tone of speech were accounted for when coding each response. Each response was also considered separately (as opposed to coding “overall” EE for the interview) and coders were instructed to avoid assumptions about what caregivers “meant” to convey, nor feelings that may have motivated certain answers. All coders received at least three hours of training and the requirement of reaching at least 90% coding agreement with an expert coder. Warmth and criticism totals were calculated by collapsing scores across the 16 items and across the two coders. Non-significant correlations emerged for warmth and criticism for both maternal (r[156] = −0.14, p = .09) and paternal EE (r[117] = −0.10, p = .3), supporting separate investigations of these distinct EE constructs. Interrater reliability was computed with intraclass reliability correlations (ICCs), with 0.60 or above considered as adequate (Kieffer et al., 2004). The ICCs for maternal warmth (0.90) and criticism (0.88) and paternal warmth (0.91) and criticism (0.78) were adequate.

Chronic Stress

Chronic Stress 1: Condition-Related Stress

SB-related stress was measured using SB characteristics and SB-specific items from the Family Stress Scale (FSS; described below in chronic stress 2). Participants’ SB characteristics were: (a) type (myelomeningocele [most severe form of SB] or other); (b) shunt status (no or yes); (c) lesion level (sacral, lumbar, or thoracic); and (d) ambulation status (an ambulation score of 1–3 for study 1 or gross motor function classification of 1–4 for study 2 (coded using the Gross Motor Function Classification System for SB; Wilson et al., 2006). Given the difference in range for ambulation/gross motor function classification, mobility total scores were converted to z-scores in each study sample separately and then combined across samples to create a single mobility score. Higher scores for each SB characteristic indicated higher likelihood for condition-related stress.

Chronic Stress 2: Family-Level Stress

Family-level stress was measured using the Family Inventory of Life Events (FILE) and the FSS items not specific to SB. The FILE is a 90-item self-report measure designed to assess the accumulation of normative and non-normative life events and stresses (McCubbin, 1981). Respondents indicate whether their family has experienced events across nine domains: intra-family strains, marital strains, pregnancy and childbearing strains, finance and business strains, work-family transitions and strains, illness and family care strains, losses, transition in and out, and family legal violations. Norms for the FILE vary by seven family stages (i.e., couple [I] through retirement [VII]) and for families with a child with a chronic illness (e.g., cardiac illness, diabetes); however, broadly, higher scores indicate the experience of more potentially stressful family events. The FILE has demonstrated acceptable test–retest reliability (0.72, 0.77) and internal consistency ranges from 0.79 to 0.82 (McCubbin et al., 1996). The FILE demonstrated acceptable internal consistency in the current sample for mother (α = 0.86) and father report (α = 0.85).

The FSS is a 19-item self-report measure designed to assess common stressors in families with a child with SB (Quittner et al., 1990). Using a 5-point Likert scale, 13 items not specific to SB (e.g., going out into the community) and six condition-specific items (e.g., catheterization) are rated. The FSS total score is the mean rating of all scores (i.e., 1 = not at all stressful to 5 = extremely stressful). The FSS has demonstrated adequate internal consistency with chronically ill populations (Quittner et al., 1998). To better understand findings within this study, FSS items not specific to SB and condition-specific items were also independently explored for associations to EE; these subscales demonstrated acceptable internal consistency for mother (non-SB α = 0.90, SB α = 0.84) and father report (non-SB α = 0.98, SB α = 0.97).

Chronic Stress 3: Caregiver Mental Health Symptoms

The Symptom Checklist-Revised (SCL-90-R) is a 90-item self-report measure designed to assess psychological symptoms experienced over the past seven days (Derogatis et al., 1976). Rating on a 5-point Likert scale (e.g., 0 = not at all distressed to 4 = extremely distressed), parents completed questions that comprise nine subscales: somatization, obsessive–compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. The mean severity response across all items is used to create a Global Severity Index, with higher scores indicating greater psychological distress. The SCL-90-R has demonstrated acceptable reliability and has been used in previous studies of families with a child with SB (e.g., Friedman et al., 2004) and demonstrated acceptable internal consistency in the current sample for mother and father report (αs = 0.99).

Chronic Stress 4: Experience of Systemic Racism

The stress of experiencing the impacts of systemic racism was measured based on the participant with SB’s self-reported racial and ethnic identity.

Chronic Stress 5: SES

SES was measured using the Hollingshead Four Factor Index of SES (Hollingshead, 1975). Specifically, mother and father education and occupation scores were averaged to calculate family SES; for single-parent or families with only one employed caregiver, individual scores were used to represent the SES of the family. Scores ranged from 8 to 66, with higher scores indicating greater SES.

Data Analysis

Independent-samples Mann–Whitney U tests were conducted to compare distributions of key demographic variables between the included (n = 174) and excluded families (n = 18), as well as between studies 1 and 2 to assess for historical threats to validity. Preliminary analyses examined descriptive information, including: means, standard deviations, ranges, and frequencies, for demographic and medical characteristics, as well as caregiver-reported variables of interest.

To estimate variability in caregiver EE that could be explained by mother-reported variables, father-reported variables, and demographic and medical characteristics, linear regressions were used with either (a) maternal criticism, (b) maternal warmth, (c) paternal criticism, or (d) paternal warmth as the dependent variable (DV). Independent variables were specific to the reporter on EE. For example, mother-reported FILE, FSS, and SCL were the IV’s used to estimate maternal EE. Conversely, father-reported FILE, FSS, and SCL were the IV’s used to estimate paternal EE. Child age, sex, experience of systemic racism, and SB-related variables were also examined as IV’s estimating the four DVs.

Results

Included versus Excluded Families

Compared to the 174 families included in the current study, excluded families (n = 18) had children that were significantly older (U = 884.50, p < .005), and had a greater relative proportion of male children than did the included families (χ²(1) = 5.64, p = .02). There were no significant differences between included vs. excluded families with regards to reported race/ethnicity (χ[1] = 0.50, p = .48) or SB severity (U = 2,267.5, p = .89). As expected, there were differences between the samples with respect to age (i.e., study 1 youth were 16–17 years old and study 2 youth were 8–15 years old; p < .001) and the likelihood of experiencing systemic racism (i.e., study 1 was primarily non-Hispanic White participants; p = .002), but there were no differences between the samples on SB type (p = .2), shunt status (p = .9), lesion level (p = .9), mobility (p = .1), FILE (ps > .3), FSS (ps > .6) or SCL-90 scores (ps > .8).

Associations with EE in Mothers of Children with SB

Table II displays linear regression results estimating maternal criticism and warmth from (a) condition-related stressors; (b) family-level stress; (c) caregiver mental health symptoms; (d) the experience of systemic racism; (e) SES; and (f) exploratory variables (age, gender). Of these models, one yielded statistically significant results for maternal criticism. Specifically, the model including the experience of systemic racism as the IV yielded a positive regression coefficient (B = 0.33) and was statistically significant, F(1, 154) = 19.18, p < .005 (f² = 0.12 [small to medium effect size; Cohen, 2013]), suggesting that mothers of non-Hispanic White children (i.e., those unlikely to be negatively impacted by systemic racism) were more likely to express criticism. Two models yielded statistically significant results for maternal warmth. The model including age as the IV yielded a negative regression coefficient (B = −0.32), F(1, 153) = 17.00, p < .005 (f² = 0.11), suggesting mothers expressed more warmth for younger children. In addition, the model with shunt status yielded a negative regression coefficient (B = −0.20), F (1, 153) = 6.28, p = .01 (f² = 0.04), suggesting that mothers expressed more warmth for children without a shunt.

Table II.

Linear Regressions Estimating Maternal Expressed Emotion

Variables	Maternal criticism				Maternal warmth
	B	SE	β	F	B	SE	β	F
SB type (condition-related stress)				0.00				3.70
Constant	13.10	0.67			22.43	1.12
SB type	−0.00	0.71	0.00		−2.31	1.20	−0.15
Shunt status (condition-related stress)				0.02				6.28**
Constant	13.06	0.51			22.23	0.85
Shunt status	0.09	0.57	0.01		−2.41	0.96	−0.20
Lesion level (condition-related stress)				0.08				0.17
Constant	13.29	0.51			20.62	0.87
Lesion level	−0.06	0.21	−0.02		−0.15	0.36	−0.03
Ambulation z-score (condition-related stress)				2.92				2.93
Constant	13.29	0.28			20.24	0.48
Ambulation	0.48	0.28	0.16		−0.83	0.48	−0.16
Maternal FILE (family-level stress)				0.70				0.06
Constant	13.96	0.86			20.60	1.50
FILE	−0.23	0.27	−0.07		−0.12	0.48	−0.02
Maternal FSS (condition, family-level stress)				0.03				0.10
Constant	13.15	0.69			20.73	1.16
FSS	0.06	0.31	0.02		−0.16	0.53	−0.03
Maternal SCL (caregiver mental health symptoms)				0.07				1.04
Constant	13.20	0.35			20.81	0.59
SCL	0.18	0.70	0.02		−1.19	1.16	−0.09
Experience of systemic racisma				19.18***				0.03
Constant	11.86	0.36			20.41	0.67
Race	2.03	0.46	0.33		0.05	0.85	<0.01
Hollingshead SES				2.59				1.32
Constant	12.09	0.73			19.15	1.26
Hollingshead SES	0.03	0.02	0.13		0.03	0.03	0.09
Child age (exploratory)				1.23				17.00***
Constant	11.99	1.04			27.30	1.70
Age	0.09	0.08	0.09		−0.55	0.13	−0.32
Child gender (exploratory)				2.20				0.24
Constant	13.58	0.40			20.71	0.70
Child sex	−0.74	0.50	−0.12		−0.42	0.87	−0.04

Note. FILE = Family Inventory of Life Events; FSS = Family Stress Scale; SCL = Symptom Checklist; SB = spina bifida; SES = socioeconomic status.

^aChild-reported identity of “non-Hispanic White” was used as a proxy measure of likelihood of the negative impacts of systemic racism.

**p < .01. ***p < .005.

Associations with EE in Fathers of Children with SB

Table III displays linear regression results estimating paternal criticism and warmth from: (a) condition-related stressors; (b) family-level stress; (c) caregiver mental health symptoms; (d) the experience of systemic racism; (e) SES; and (f) exploratory variables (age, gender). Of these models, two yielded statistically significant results for paternal criticism. First, the model including father-reported FSS as the IV yielded a positive regression coefficient (B = 0.21) and was statistically significant, F(1, 110) = 4.82, p = .03 (f² = 0.04), suggesting that fathers expressed more criticism when they endorsed higher overall family stress. To better understand this finding, additional models isolating the FSS items assessing family stress not specific to SB and SB-specific family stress were examined. FSS items not specific to SB yielded a positive regression coefficient (B = 0.12) and was statistically significant, F(1, 110) = 6.32, p = .01. However, there was no evidence to suggest a significant association with condition-specific family stress. Second, the model including impacts of systemic racism as the IV yielded a positive regression coefficient (B = 0.31) and was statistically significant, F(1, 119) = 12.53, p < .005 (f² = 0.10), suggesting that fathers of non-Hispanic White children were more likely to express criticism. Two models yielded statistically significant results for paternal warmth. Specifically, the model including age as the IV yielded a negative regression coefficient (B = −0.26), F(1, 114) = 7.93, p < .01 (f² = 0.07), suggesting that fathers express more warmth for younger children. Finally, the model including shunt status as the IV also yielded a negative regression coefficient (B = −0.25), F(1, 115) = 7.74, p = .006 (f² = 0.07), suggesting that fathers expressed more warmth for children with SB without a shunt.

Table III.

Linear Regressions Estimating Paternal Expressed Emotion

Variables	Paternal criticism				Paternal warmth
	B	SE	β	F	B	SE	β	F
SB type (condition-related stress)				1.18				3.73
Constant	11.83	0.80			22.96	1.52
SB type	0.92	0.85	0.10		−3.12	1.62	−0.18
Shunt status (condition-related stress)				1.81				7.74**
Constant	11.96	0.61			22.80	1.08
Shunt status	0.92	0.69	0.12		−3.40	1.22	−0.25
Lesion level (condition-related stress)				0.03				2.25
Constant	12.62	0.74			21.81	1.31
Lesion level	0.06	0.34	0.02		−0.89	0.59	−0.14
Ambulation z-score (condition-related stress)				2.30				2.59
Constant	13.03	0.36			19.52	0.62
Ambulation	0.54	0.35	0.16		−0.99	0.62	−0.17
Paternal FILE (family-level stress)				3.93				0.04
Constant	10.90	1.04			19.94	1.90
FILE	0.69	0.37	0.18		0.14	0.68	0.02
Paternal FSS (condition, family-level stress)				4.82*				1.33
Constant	10.53	1.01			22.27	1.86
FSS	1.14	0.52	0.21		−1.09	0.95	−0.11
Paternal SCL (caregiver mental health symptoms)				0.64				2.16
Constant	12.44	0.41			19.42	0.74
SCL	0.87	1.09	0.08		2.87	1.95	0.14
Experience of systemic racisma				12.53***				0.04
Constant	11.33	0.47			20.28	0.90
Race	2.03	0.57	0.31		−0.23	1.10	−0.02
Hollingshead SES				1.98				0.57
Constant	11.59	0.84			18.99	1.53
Hollingshead SES	0.03	0.02	0.13		0.03	0.04	0.07
Child age (exploratory)				1.63				7.93**
Constant	11.23	1.19			25.90	2.09
Age	0.12	0.09	0.12		−0.46	0.16	−0.26
Child gender (exploratory)				1.10				1.50
Constant	13.12	0.50			19.20	0.91
Child sex	−0.63	0.61	−0.10		1.35	1.11	0.11

Note. FILE = Family Inventory of Life Events; FSS = Family Stress Scale; SB = spina bifida; SCL = Symptom Checklist; SES = socioeconomic status.

^aChild-reported identity of “non-Hispanic White” was used as a proxy measure of likelihood of the negative impacts of systemic racism.

^* p < .05. **p < .01. ***p < .005.

Discussion

Guided by a Chronic Family Stress Model, the current study adds to the literature by identifying associations between stressful experiences and caregiver EE in families of youth with SB. Our hypotheses were partially confirmed. The hypothesis that higher stress would be associated with more caregiver criticism was confirmed for family-level stress (FSS) for fathers. However, counter to this hypothesis, those less likely to have been impacted by the experience of systemic racism were more likely to express more criticism. Warmth was associated with the exploratory variable of age and the condition-related stress of shunt status, such that both mothers and fathers were more likely to express more warmth toward younger children with SB and those without a shunt. Stressors associated with EE in the current study: family stress, shunt status, and absence of the stressor of systemic racism, are discussed below, followed by implications for future research.

Family stressors appeared to influence the expression of EE for caregivers of youth with SB. Indeed, fathers of children with SB were more likely to display criticism when family stress was higher. Based on the Chronic Family Stress model and a diathesis-stress formulation, caregivers who express criticism will be more likely to do so when placed under high stress situations (Hooley & Gotlib, 2000; Peris & Miklowitz, 2015). However, this vulnerability appears to be associated with the general experience of stress for fathers, as opposed to condition-specific stressors or the severity of the condition itself. Condition-related stressors were implicated in the expression of more warmth for both parents (i.e., more likely to be warm with children who did not have a shunt), but there was no evidence to suggest an association between severity and criticism. Instead, fathers were more likely to express criticism when they also endorsed higher stress relating to the family’s management of general stress. This finding is in line with the broader EE literature (Rienecke, 2018), indicating that caregiver warmth and criticism are not necessarily inversely related but may have different factors that promote their expression. This finding also highlights an opportunity, as stress management is a teachable skill that can decrease the experience of stress, whereas SB severity (e.g., type, lesion level) is not a modifiable factor.

Differences in EE emerged based on whether the children with SB identified as non-Hispanic White and, as such, were unlikely to have experienced negative impacts from systemic racism. These youth were more likely to have caregivers who expressed criticism. This finding should be interpreted with caution, given that this was a proxy measure for the negative impacts of systemic racism, which was used in the place of a more accurate and direct measure of the impacts of systemic racism due to the nature of when these data were collected (e.g., the baseline of the first longitudinal study occurred in the early 1990s and recruited a primarily White sample). Further, while examined through the lens of systemic racism, another way to view these analyses is as a sample dichotomized based on minoritized status. This is problematic for multiple reasons, including but not limited to: (a) the assumption that the non-Hispanic White experience is the “norm” by which to evaluate the experience of other groups and (b) the blending of diverse racial and ethnic experiences into a single category (Galán et al., 2021). We therefore cautiously propose two potential meanings from this finding. First, this finding may support the importance of other types of stressors being associated with the expression of EE in caregivers of youth with SB. Indeed, as non-Hispanic White youth are less likely to experience negative stressors and impacts from systemic racism, this finding may suggest that the higher levels of criticism in non-Hispanic White parents are more likely tied to different stressors. Further, the nuances of these associations are likely more complex than the main effects explored in the current study. Second, previously proposed theories posit that EE is likely a culturally specific construct (Bhugra & McKenzie, 2003; López et al., 2009). These findings may therefore add to a growing body of evidence suggesting that families’ cultural and daily lived experiences likely establish different expectations, supports, and norms relating to supporting their children with SB (including differences in collectivist and individualistic communities; Liptak et al., 2010; Ohanian et al., 2018; Papadakis et al., 2018). These differences may promote specific strengths in a family system, including caregivers who are less likely to express criticism toward their children with SB.

While we are hesitant to over interpret the findings of the current study, multiple directions for future research are implicated in assessing EE in families with a child with SB. First, assessment would ideally first occur when children with SB are young, a time at which caregivers are more likely to be expressing higher levels of warmth, as indicated in the current study. Establishing a baseline with caregivers at a younger age could provide the opportunity to build on existing strengths (i.e., warmth) and provide psychoeducation around means to manage potential situations or cognitions that might promote the expression of criticism as their children age. Future research could explore implementation methods to embed these assessments most effectively and sustainably—and referrals for prevention and/or intervention services—into current behavioral healthcare appointment workflows. Additionally, as EE interviews may not be feasible in such appointments due to time, future research on alternative measures of EE (e.g., perceived criticism, Family Attitude Scale; Hooley & Parker, 2006) are warranted. Second, EE assessment should be considered at times of higher stress (e.g., family reporting to their physician that they are currently having relationship strains within the family or financial difficulties). Future research is needed to better identify critical windows of opportunity for intervention considering specific stressors. Indeed, EE is not an immutable construct, as demonstrated by parent training to build warmth and reduce negative behaviors (e.g., Wood et al., 2021). Finally, assessing all caregivers for EE would likely be beneficial. Indeed, the current findings demonstrate different associations for mothers vs. fathers, suggesting that different experiences may promote the likelihood of EE expression for one caregiver compared to another. These assessment considerations will promote the likelihood of EE being discussed with caregivers, as well as provide opportunities to expand their awareness and skills in expressing more warmth toward their children.

The current study has multiple strengths, including the analysis of psychosocial data from multiple samples of families with SB, the inclusion of both mothers and fathers, the use of interview methods with objective coding, and the use of the Chronic Family Stress Model to drive the selection of correlates (Peris & Miklowitz, 2015). However, the findings should be considered in light of specific limitations. First, in combining data from different studies, the data collected were from youth living in different decades. Despite no evidence to suggest differences between the samples on the variables of interest in this study, it is likely that historical impacts affected these two samples’ lived experiences. Additionally, the number of variables that could be examined were limited. For example, it is possible that the association between shunt status and caregiver warmth may have been influenced by other variables (e.g., intelligence quotient; Pollenus et al., 2020). Future analyses should include associations with EE that focus on multiple condition-related factors and stressful constructs that extend beyond differences in race and/or ethnicity (Kaufman & Thurston, 2018). It is also of note that the internal consistency of the FSS was high; future research should include more items to assess a broader range of family stressors. Furthermore, this sample included mothers and fathers as single parents or dyads. It is unclear how differences in EE extend to same sex dyads, non-binary parents, or caregivers who are not parents (e.g., grandparents). Second, this sample consisted of research participants. It is possible that the expression of EE may be different in clinical populations of caregivers of children with SB. Finally, while there were significant associations identified based on the absence of negative impacts from systemic racism, this sample was 60.3% non-Hispanic White. Future research must involve more targeted recruitment and sampling efforts to include more diverse samples (Galán et al., 2021). In addition, stakeholder input should be sought to ensure that culturally appropriate assessment tools and methods are included so as to better identify variables that promote warmth vs. criticism in caregivers of youth with SB.

EE has a well-established role in clinical outcomes across a variety of conditions, including youth with SB (Kelly et al., 2011). The current study adds to this literature by identifying correlates of caregiver criticism and warmth in families with SB grounded within the Chronic Family Stress Model. Family stressors and the absence of the negative impacts of systemic racism are implicated in the expression of criticism, whereas child age and lack of a shunt are implicated in the expression of warmth in caregivers of a child with SB. While future research is necessary to assess additional correlates of EE with more diverse samples, assessing EE in caregivers of young children with SB may prove promising in supporting the development of targeted prevention and intervention programs for such caregivers.

Funding

This research was supported in part by grants from the National Institute of Nursing Research and the Office of Behavioral and Social Sciences Research (R01 NR016235), National Institute of Child Health and Human Development (R01 HD048629), and the March of Dimes Birth Defects Foundation (12-FY13-271). This study is part of ongoing, longitudinal studies. Dr. Stiles-Shields is supported by a fellowship from the National Institute of Mental Health (K08 MH125069) and Dr. Kritikos is supported by a Drotar-Crawford Research Grant from the Society of Pediatric Psychology (APA Division 54).

Conflicts of interest: None declared.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.