Parent Stress and Coping Trajectories in Hispanic and non-Hispanic Families of Children at Risk for Autism Spectrum Disorder

Abstract

Significant disparities exist between Hispanic and non-Hispanic families in time-to-diagnosis among children identified as at risk for Autism Spectrum Disorder (ASD); yet, little is known about parent experiences throughout the diagnostic process that may contribute to or help explain these disparities. The current study examined longitudinal trajectories of parenting stress, coping, and perceived family impact during the ASD diagnostic process among an ethnically and racially diverse low income, urban sample, allowing for comparisons between Hispanic and non-Hispanic families. Hispanic families reported lower levels of parenting stress (χ²=5.35[1], p=0.021), avoidant (χ²=9.66[1], p=0.002) and approach (χ²=8.61[1], p=0.003) coping, and negative family impact (χ²=5.39[1], p=0.020) across time (main effects) compared to non-Hispanic families. Further, there were differences in the change in use (time period x ethnicity interaction effects) of both avoidant (χ²=10.80[3], p=0.013) and approach (χ²=11.57[3], p=0.009) coping, as well as negative family impact (χ²=9.81[3], p=0.020), between Hispanic and non-Hispanic parents over time. These differences begin to shed light on the unique experiences of Hispanic parents. Because of the strengths that Hispanic families demonstrate, interventions that leverage those strengths and focus on education, empowerment, and resiliency might be particularly beneficial for Hispanic families and may also better inform work to increase resiliency in non-Hispanic families.

Introduction.

Despite advances in early diagnosis for Autism Spectrum Disorders (ASD; Pierce et al., 2016; Zwaigenbaum, Bauman, Choueiri, Kasari, et al., 2015), the time between parent/provider concern and diagnostic referral and receipt of a diagnosis can take an average of 7 to 19 months (Carbone et al., 2020; Constantino et al., 2020; Penner, Anagnostou, & Ungar, 2018; Zwaigenbaum et al., 2019). Currently, 1 in 54 children in the US is diagnosed with ASD (Maenner, Shaw, & Baio, 2020), continuing the trend of increasing prevalence over the past decade (Baio et al., 2018; CDC, 2007; Rice et al., 2010; Van Naarden Braun et al., 2015). Although no evidence suggests that having ASD is dependent upon sociodemographic characteristics such as race, ethnicity, income, or education (Dyches, Wilder, Sudweeks, Obiakor, & Algozzine, 2004), ASD prevalence estimates are lower for Hispanic children than non-Hispanic White, non-Hispanic Black, and Asian/Pacific Islander children (Maenner et al., 2020).¹ Because of these differences, it is particularly important to also investigate potential differences in parent and family outcomes in Hispanic, compared to non-Hispanic, families over the course of the ASD diagnostic process in order to highlight family risks and resiliencies that might provide targets for intervention and health policy changes that will facilitate timely diagnostic evaluation for Hispanic families.

Literature suggests that Hispanic children, in particular, are often diagnosed later than their non-Hispanic peers (Christensen et al., 2016; Magana, Lopez, Aguinaga, & Morton, 2013; Yingling, Hock, & Bell, 2018). Hispanic families face social injustices (Lopez, 2014) and poorer quality and access to health care (Liptak et al., 2008; Magaña, Parish, Rose, Timberlake, & Swaine, 2012). They also face unique barriers to care such as language differences, unmet service needs such as not having a personal doctor, and difficulty receiving referrals (Blanche, Diaz, Barretto, & Cermak, 2015; Magaña et al., 2013; Smith, Gehricke, Iadarola, Wolfe, & Kuhlthau, 2020). Little is known about parent experiences before or during the ASD diagnostic odyssey, after ASD concern has been raised, but before diagnostic ascertainment, and even less is known about the potentially differing experiences between Hispanic and non-Hispanic families. In the one study to date that examined parenting stress during the diagnostic period (DesChamps, Ibañez, Edmunds, Dick, & Stone, 2020), parents of children who eventually received an ASD diagnosis had higher levels of parenting stress throughout the diagnostic odyssey than parents of children who did not receive an ASD diagnosis. However, the sample did not allow for examinations of ethnicity-based differences.

It is well documented that parents of children with ASD report higher parenting stress than parents of typically developing children (Dabrowska & Pisula, 2010; DesChamps et al., 2020; Estes et al., 2013; Hayes & Watson, 2013; Pisula & Porębowicz-Dörsmann, 2017) or children with other types of neurodevelopmental disabilities (Dabrowska & Pisula, 2010; DesChamps et al., 2020; Hayes & Watson, 2013; Valicenti-Mcdermott et al., 2015). Less is known about the real-time trajectories of parenting stress, coping, and well-being of parents during the period before a child has reached diagnostic ascertainment, which may be particularly important considering the disparities in access to timely diagnostic and treatment services that Hispanic children and their families experience (Christensen et al., 2016; Magaña et al., 2013; Yingling et al., 2018). Studies examining parenting stress and well-being prior to a child’s diagnosis are often retrospective qualitative (e.g., Braiden, Bothwell, & Duffy, 2010; Siklos & Kerns, 2007) or cross-sectional accounts of parents (e.g., Chamak et al., 2011; Crane et al., 2016; Tait, Fung, Hu, Sweller, & Wang, 2016); that is, they include samples of parents of children who already have completed the diagnostic process and have a documented ASD diagnosis (e.g., Konstantareas & Papageorgiou, 2006). Despite limitations, this research indicated that the period from identification of ASD risk through diagnosis to services is typically lengthy and stressful for families, regardless of the child’s eventual diagnosis (Braiden et al., 2010; Chamak et al., 2011; Crane et al., 2016; Siklos & Kerns, 2007; Wong, Yu, Keyes, & McGrew, 2017; Yingling et al., 2018).

Many factors and processes may contribute to differences in parent outcomes and experiences during the diagnostic process. The Stress, Appraisal, and Coping model (Lazarus & Folkman, 1984) provides a framework to examine such factors. For example, Zuckerman et al. (2017) found that while non-Hispanic White families were more likely to report stress of diagnostic process as a barrier to care, Hispanic families were conversely more likely to report experiencing barriers related to knowledge about ASD and trust in providers. Further, Hispanic mothers are more likely to report early concerns about temperament, but not about social or communication skills (Ratto, Reznick, & Turner-Brown, 2016), which are more closely associated with ASD. Differences in contextual experiences and appraisals immediately before, after, and in the few months following diagnostic ascertainment, when parent knowledge about development and ASD, feelings of stigma, and communication with healthcare providers are likely in flux (Zuckerman et al., 2014a; Zuckerman et al., 2014b), may have important consequences for parent and child outcomes.

Active coping strategies, including planning and positive reframing, have been shown to improve adjustment in families of children with ASD (Bekhet, Johnson, & Zauszniewski, 2012; Feinberg et al., 2014; Feinberg, Silverstein, & Ferreira-Cesar, 2013; Whitehead, Dorstyn, & Ward, 2015); yet, parents of children with ASD (Lai, Goh, Oei, & Sung, 2015) and other developmental disabilities (Craig et al., 2020) report more frequent use of active avoidance coping, typically considered maladaptive, than parents of typically developing children. Although most research has focused on non-Hispanic, White families of children with ASD, studies suggest that culture significantly interacts with coping strategies. For example, because of the strong family values and more traditional gender roles in Hispanic families, some research has found that Hispanic individuals tend to rely mostly on religious coping, emotional support, and denial (Epstein-Ngo, Maurizi, Bregman, & Ceballo, 2013), which have been linked to higher stress and depression among non-Hispanic (Hastings et al., 2005) and Hispanic parents of children with ASD (Willis et al., 2016).

Parental experiences during the diagnostic process are salient for addressing parent as well as child outcomes, as these experiences may impact engagement in pre- and post-diagnostic services, and in turn, the likelihood of receiving a timely diagnosis. In post-diagnosis samples, parents of children with ASD who experience elevated stress are less likely to access services to reduce their child’s symptoms of ASD (Karp et al., 2018). Parenting stress has also been found to be associated with reduced effectiveness of early intervention (Osborne, McHugh, Saunders, & Reed, 2008; Reid, Fesalbon, Mendoza, Alvord, & Rich, 2019; Shalev, Lavine, & Di Martino, 2020; Weiss, Viecili, & Bohr, 2015).

The Current Study

Given the unique child symptoms and behaviors related to ASD that contribute to parent well-being (Hayes & Watson, 2013), as well as the bidirectional and linked relationships between parent and child outcomes over time (Rodriguez, Hartley, & Bolt, 2019), continued investigation of parenting stress, coping, and well-being within the context of the ASD phenotype (prior to ASD diagnosis) is crucial to efforts to improve parent and child outcomes (Nordahl-Hansen, Hart, & Øien, 2018) and may represent a critical avenue for enhancing longterm family outcomes.

The current study focused on the trajectory of parent outcomes in Hispanic and non-Hispanic families starting from when their children were first identified as at risk for ASD, and extending beyond a diagnostic evaluation, up to one year post enrollment in the study. While the authors also recognize the pervasive differences in experiences that exist in parent experiences based on race, we have chosen to focus on ethnicity, which is recognized as an independent factor affecting diagnostic outcomes. ASD prevalence estimates have been “approximately identical” for non-Hispanic White, non-Hispanic Black, and non-Hispanic Asian/Pacific Islander children (18.5, 18.3, and 17.9 per 1000 children, respectively); however, ASD prevalence is significantly lower for Hispanic children (15.4 per 1000 children; Maenner et al., 2020). The present research prospectively examined longitudinal trajectories of parenting stress, coping, and perceived family impact during the ASD diagnostic process among Hispanic and non-Hispanic subgroups. It aims to advance understanding of the experiences of Hispanic families, with the goal of decreasing disparities in ASD diagnosis and engagement in services.

In our sample of low income, urban families of children who were recently referred for a diagnostic evaluation due to concerns for ASD, our research questions were: 1) are there differences in parent and family outcomes (overall parenting stress, autism-specific parenting stress, avoidant and approach coping, and perceived negative impact on the family) between Hispanic and non-Hispanic families across four time points throughout the diagnostic process, and 2) are there differences in patterns of parent and family outcomes (overall parenting stress, autism-specific parenting stress, avoidant and approach coping, and perceived negative impact on the family) between Hispanic and non-Hispanic families during this time. It is particularly important to investigate the interaction between ethnicity and time in order to understand how the pattern of parent stress, coping, and family impact may differ between the two groups, shape parent experiences, and contribute to longer-term parent and child outcomes such as parent engagement in care, receipt of services, and child behavioral outcomes.

Methods

Participants

The current study was based on 339 parents/guardians and their children who were enrolled in a multisite randomized controlled trial of Family Navigation (compared to traditional Care Management), a structured intervention consisting of core components (Broder-Fingert et al., 2020) designed to improve access to diagnostic and treatment services for historically underserved children at risk for ASD. In the current sample, 89.7% identified as mothers (n = 304), 7.1% as fathers (n = 24), and 3.2% (n = 11) as legal guardians of the child involved in the study. The term “parent” will hereafter be used to refer to all parents/guardians in the parent-child dyads.

Procedure

Parent-child dyads were recruited from 14 urban pediatric clinics in 3 northeastern US cities. Children, aged 15–27 months, were referred to the study based on risk for ASD identified in pediatric primary care; ASD risk was confirmed by study staff with a parent interview that included administration of the Modified Checklist for Autism in Toddlers, Revised with Follow-Up (M-CHAT-R/F; Robins et al., 2014). Families were excluded if their child had a prior diagnosis of ASD or if there was no identified ASD risk according to parent/provider concern or M-CHAT-R/F screening results. Data were collected in person by research staff who spoke English, Spanish, Haitian Creole, or Vietnamese; telephone interpreter services were used, as needed, for all other languages. Data were collected longitudinally at four time points (Baseline, Time 1: after diagnostic ascertainment or 150 days after enrollment if diagnostic ascertainment not completed, Time 2: 100 days after diagnostic ascertainment or an additional 100 days after enrollment if diagnostic ascertainment not completed, and Time 3: one year after enrollment).

Measures

Family Socio-demographics

Parents reported on sociodemographic factors at Baseline, including their relationship to the child, parent and child age based on date of birth, sex, race, ethnicity, preferred language, and insurance type (public, private, or other). Parents were also asked about their level of education, country of origin, employment status, and whether the child was receiving EI services.

Parenting Stress

Parenting stress was measured at each time point (Baseline, Time 1, Time 2, and Time 3) using the Parenting Stress Index-Short Form, 4^th Edition (PSI-SF; Abidin, 2012), a 36-item parent-report questionnaire in which parents rate their agreement on a 5-point Likert scale from 1 (Strongly Agree) to 5 (Strongly Disagree). Parents are asked to complete all items based on the child that “causes [them] the most stress.” A higher score indicates greater parenting stress (Abidin, 2012). The PSI-SF has been shown to have high internal consistency and adequate test–retest reliability (Abidin & Brunner, 1995), and has been widely used and validated in studies of parents of children with ASD (e.g., Zaidman-Zait et al., 2011) and, specifically, in studies of Spanish-speaking parents of children with ASD (Nieto, López, & Gandía, 2017). As part of this measure, parents were instructed to complete the PSI-SF about the child currently causing them the most stress. Only parents who completed the questionnaire about the child enrolled in the larger study (n=235, 217, 218, and 204 at Baseline, Time 1, Time 2, and Time 3, respectively), as opposed to a sibling, were included in analyses that included the PSI-SF. Continuous total raw scores were used in analyses. Cronbach’s α in the current sample was 0.93.

Autism-specific Parenting Stress

Autism-specific parenting stress was measured at each time point using the Autism Parenting Stress Index (APSI; Nieto, López, & Gandía, 2017), a 13-item Likert-style questionnaire designed to measure the impact of specific core and co-morbid symptoms associated with ASD on parent levels of stress. Items require the parent to rate the degree of stress that they experience as a result of such issues from 0 (not stressful) to 5 (so stressful sometimes we feel we can’t cope). Initial validation of the APSI found a satisfactory internal consistency (Silva & Schalock, 2012). Cronbach’s α in the current sample was 0.88.

Coping

Coping was measured at each time point using the Brief COPE (Carver, 1997), consisting of 28 items. Based on the recommendations of Carver, Scheier, and Weintraub (1989), coping responses were grouped into approach and avoidant coping. The avoidant coping subscale of the Brief COPE includes 12 items that tap into use of denial, substance use, venting, behavioral disengagement, self-distraction, and self-blame. The approach coping subscale of the Brief COPE includes 12 items that tap into active coping, positive reframing, planning, acceptance, seeking emotional support, and seeking informational support. Two ambiguous items that loaded on both the approach and avoidant coping components, humor and religion, were excluded from further analyses (as recommended by Carver et al., 1989). Participants evaluated how likely they were to engage in each coping response ranging from 1 (I don’t do this at all) to 4 (I do this a lot). Total scores of approach and avoidant coping were used in analyses, with higher scores indicating a stronger tendency to use those coping behaviors. The Brief COPE has been shown to demonstrate sufficient internal consistency among samples of parents of children with ASD and DD (Adams, Rose, Jackson, Karakatsani, & Oliver, 2018; Benson, 2010; Hastings et al., 2005). Cronbach’s α’s on the avoidant and approach coping subscales in the current sample were 0.81 and 0.92, respectively.

Family Impact

Family impact was measured at each time point using the Family Impact Questionnaire (FIQ; Donenberg & Baker, 1993). This tool measures parent’s perceptions of their child’s impact on the family on a four-point Likert scale ranging from 1 (not at all) to 4 (very much). The current study used a composite 20-item negative impact score (including the negative feelings about parenting and the impact on parent’s social life subscales), which has been validated previously in samples of families with and without children with developmental disabilities (Blacher & Baker, 2019; McIntyre, 2008; Neece, Kraemer, & Blacher, 2009). Cronbach’s α in the current sample was 0.88.

Language Translation

The PSI-SF and the Brief COPE were available in Spanish. Research staff translated the APSI and FIQ directly for parents into Spanish and used phone interpretation services as needed if an English/Spanish bilingual research staff member was not available. Similarly, bilingual English/Haitian Creole and English/Vietnamese research staff translated all materials for parents who preferred to speak Haitian Creole or Vietnamese during their interactions with research staff. All other language translations were done using phone interpreter services.

Data Analysis Plan

Descriptive statistics were calculated for all variables in SPSS, Version 26 (IBM Corp, 2018). Pearson correlations were used to assess the associations between sociodemographic variables and variables of interest. Chi-squared analyses were performed with socio-demographic categorical variables to determine differences based on ethnicity (Hispanic vs. non-Hispanic). Independent samples t-tests were performed to determine if ethnic groups differed on continuous socio-demographic variables.

Five generalized estimating equation (GEE) models were fit (i.e., one model for each outcome variable: parenting stress, autism parenting stress, avoidant coping, approach coping, and negative family impact) within SAS 9.4 (SAS Institute Inc, 2014) to estimate the difference in the trajectories between ethnic groups over time. First, each model was tested to determine the appropriate working correlation matrix (independent, exchangeable, unstructured), with the models with the lowest quasi-likelihood under the independence model criterion (QIC; IBM Corp, 2018) being selected as the final models (exchangeable for the parenting stress outcome model; unstructured for the remaining four outcome models). Least-squares means were calculated with follow-up Tukey-Kramer post-hoc comparisons. Models included main effects of time period (Baseline, Time 1, Time 2, Time 3), parent ethnicity (Hispanic, non-Hispanic), and a time-varying diagnostic status term. Time-varying diagnostic status was created as a dummy variable at each time point (0 = no diagnosis; 1 = other developmental or behavioral diagnosis; 2 = ASD at each time point) in order to control for which diagnosis a child received, as well as when they reached diagnostic ascertainment, an important factor often overlooked in similar analyses. Models also included a three-way interaction term (time period x ethnicity x time-varying diagnostic status) along with all component two-way interaction terms (time period x ethnicity, time period x time-varying diagnostic status, and ethnicity x time-varying diagnostic status). Finally, treatment arm (Family Navigation or Care Management) was included as a main effect a priori covariate in all models. If any model showed a non-significant interaction effect (p>0.05), the model was fit again with the interaction terms removed, starting with the three-way interaction term and followed by component two-way interaction terms, until only significant interaction terms remained or we arrived at the base main effects model.

Statement of Community Involvement

The research team included several interdisciplinary clinicians or service providers that worked directly with the communities from which the current sample was drawn. Further, two family members of individuals with developmental disabilities were significantly involved in the development of the current study’s research questions and outcome measures, and the interpretation and dissemination of the findings. Finally, Family Navigators and other research staff, many of whom were members of the cultural communities of the individuals participating in the current study, were involved in the design, implementation, and interpretation of the study results.

Results

Sample Demographics

Table 1 displays sample baseline characteristics, overall and by ethnicity. Of the 339 parents in the sample, 96 (28.3%) were Hispanic. Parent mean age was 30.9 (SD=7.0) years. Over half of parents (n=192, 60.4%) reported that they were Black or African American, 40 (12.6%) White, 19 (6.0%) Asian, 2 (0.6%) American Indian/Alaska Native, 15 (4.7%) more than one race, and 50 (15.7%) did not report a race. Child mean age was 21.6 (SD=3.5) months and 239 (70.5%) children were male.

Table 1:

Baseline characteristics of participants by ethnicity

	Total (n=339)	Hispanic (n=96)	Non-Hispanic (n=243)
Variable	Mean (SD)	Mean (SD)	Mean (SD)	F	df	p-value
Parent Age (years)	30.87 (6.99)	29.00 (6.00)	31.62 (7.22)	1.36	1, 333	.245
Child Age (months)	21.55 (3.49)	21.22 (3.50)	21.67 (3.48)	0.03	1, 337	.872
	Frequency (%)	Frequency (%)	Frequency (%)	χ2		p-value
Participating Parent/Guardian				2.41	2	.300
Mother	304 (89.7%)	90 (93.8%)	214 (88.1%)
Father	24 (7.1%)	4 (4.2%)	20 (8.2%)
Other/Legal Guardian	11 (3.2%)	2 (2.1%)	9 (3.7%)
Preferred Language				115.25	2	<.001
English	278 (82.0%)	54 (56.3%)	224 (92.2%)
Spanish	44 (13.0%)	42 (43.8%)	2 (0.8%)
Other	17 (5.0%)	0 (0%)	19 (7.8%)
Parent Race				195.46	5	<.001
Black or African American	192 (60.4%)	7 (8.2%)	185 (79.4%)
White	40 (12.6%)	21 (24.7%)	19 (8.2%)
Asian	19 (6.0%)	0 (0.0%)	19 (8.2%)
American Indian/Alaska Native	2 (0.6%)	2 (2.4%)	0 (0.0%)
More than one race	15 (4.7%)	9 (2.3%)	6 (2.6%)
Not reported	50 (15.7%)	46 (54.1%)	4 (1.7%)
Parent Born outside the US	211 (62.4%)	56 (58.3%)	72 (29.6%)	23.26	1	<.001
Currently Working Outside Home	180 (53.1%)	43 (44.8%)	137 (56.4%)	3.71	1	.054
Parent Education				12.05	3	.007
Less than High School/no GED	53 (15.7%)	24 (25.0%)	29 (12.0%)
High School Degree or GED	222 (65.7%)	60 (62.5%)	162 (66.9%)
Bachelors/Associates degree	53 (15.7%)	11 (11.5%)	41 (16.9%)
Graduate Degree	9 (2.7%)	0 (0.0%)	9 (3.7%)
Child Sex (male)	239 (70.5%)	68 (70.8%)	171 (70.4%)	0.01	1	.933
Insurance (Child)				0.17	1	.684
Public Insurance (Medicaid)	311 (91.7%)	89 (92.7%)	222 (91.4%)
Private or Other	28 (8.3%)	7 (7.3%)	21 (8.6%)
Child Receiving EI	152 (45.1%)	46 (47.9%)	106 (44.0%)	0.43	1	.513
Child Gestational Age <37 weeks	49 (14.5%)	9 (9.4%)	40 (16.5%)	2.79	1	.095

Note. Missing data: Parent Age (n=335); Parent Race/Ethnicity (n=318); Parent Born in US (n=338); Highest Education Completed (n=337); Child Receiving EI (n=337).

As expected, a larger percentage of Hispanic families interacted with study personnel in a language other than English (43.8% vs. 8.6%; p<.001) and were born outside the U.S. (58.3% vs 29.6%; p<.001), compared to non-Hispanic families. Further, more Hispanic parents did not have a high school degree or GED (25.0% vs. 12.0%; p=0.007) than Non-Hispanic parents. In spite of these sociodemographic differences, only a priori covariates were included in models in order to avoid collinearity with ethnicity, the predictor of interest.

Table 2 presents the observed means and standard deviations of the five outcome variables at each time point for the full sample and stratified by parent ethnicity.

Table 2:

Baseline, Time 1, Time 2, and Time 3 observed means and standard deviations of parent and family variables

	Baseline	Time 1	Time 2	Time 3
Variable	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Parenting Stress (PSI-SF raw score)	86.76 (25.63)	87.96 (22.46)	86.98 (23.46)	88.48 (23.50)
Hispanic	83.34 (23.78)	81.54 (21.87)	81.36 (23.67)	86.66 (24.03)
Non-Hispanic	88.43 (26.39)	91.57 (22.06)	89.76 (22.93)	89.28 (23.30)
Autism Parenting Stress Index (APSI)	11.64 (9.06)	15.32 (11.18)	16.13 (11.16)	17.17 (11.41)
Hispanic	11.35 (8.88)	12.85 (10.24)	14.75 (10.21)	15.32 (10.12)
Non-Hispanic	11.75 (9.15)	16.40 (11.43)	16.69 (11.50)	17.95 (11.85)
Avoidant Coping (Brief COPE)	1.51 (0.49)	1.64 (0.50)	1.58 (0.54)	1.63 (0.53)
Hispanic	1.48 (0.40)	1.48 (0.41)	1.46 (0.47)	1.53 (0.50)
Non-Hispanic	1.52 (0.53)	1.71 (0.53)	1.63 (0.56)	1.67 (0.54)
Approach Coping (Brief COPE)	2.37 (0.92)	2.94 (0.77)	2.84 (0.84)	2.86 (0.78)
Hispanic	2.39 (0.86)	2.66 (0.83)	2.62 (0.94)	2.63 (0.85)
Non-Hispanic	2.36 (0.94)	3.06 (0.72)	2.93 (0.79)	2.96 (0.73)
Negative Family Impact (FIQ)	16.34 (11.51)	17.84 (12.03)	17.20 (11.56)	18.21 (11.67)
Hispanic	14.97 (9.17)	14.73 (9.72)	16.16 (9.52)	16.83 (10.61)
Non-Hispanic	16.88 (12.28)	19.21 (12.69)	17.61 (12.29)	18.79 (12.07)

Note. Time 1: after diagnostic ascertainment or 150 days after enrollment if diagnostic ascertainment not completed; Time 2: 100 days after diagnostic ascertainment or an additional 100 days after enrollment; Time 3: one year after enrollment.

Generalized Estimating Equation Models

Table 3 includes results from the five final GEE models. For each outcome we present figures depicting the least squares means by ethnicity for each time point.

Table 3:

Generalized estimating equation estimates.

	Wald χ2	df	p
Model 1: Parenting stress
Time	2.57	3	0.463
Ethnicity	5.35	1	0.021
Diagnostic status (time-varying)	2.93	2	0.232
Treatment	1.91	1	0.167
Model 2: Autism Parenting Stress Index
Time	9.74	3	0.021
Ethnicity	1.80	1	0.180
Diagnostic status (time-varying)	19.62	2	<0.001
Treatment	0.76	1	0.385
Model 3: Avoidant coping
Time	5.40	3	0.145
Ethnicity	9.66	1	0.002
Diagnostic status (time-varying)	2.25	2	0.325
Treatment	0.67	1	0.412
Time × Ethnicity	10.80	3	0.013
Model 4: Approach coping
Time	19.67	3	<0.001
Ethnicity	8.61	1	0.003
Diagnostic status (time-varying)	18.62	2	<0.001
Treatment	4.91	1	0.027
Time × Ethnicity	11.57	3	0.009
Model 5: Negative family impact
Time	3.13	3	0.372
Ethnicity	5.39	1	0.020
Diagnostic status (time-varying)	3.64	2	0.162
Treatment	0.95	1	0.330
Time × Ethnicity	9.81	3	0.020

Parenting stress (Figure 1).

Figure 1. Parenting stress least squares means, by parent ethnicity.

Note. Sample sizes at Baseline, Time 1, 2, and 3: Hispanic: n=77, 78, 72, 62; Non-Hispanic: n=158, 139, 146, 142.

No three-way interaction effect (time period x ethnicity x diagnosis) was found (χ²=5.13[4], p=0.274). In the refitted model (without three-way interaction), no significant two-way interactions were found between time period and ethnicity (χ²=7.05[3], p=0.070), time period and diagnosis (χ²=0.59[4], p=0.964), or ethnicity and diagnosis (χ²=3.23[2], p=0.199). With all interactions removed from the model, there was a significant main effect of ethnicity on parenting stress (χ²=5.35[1], p=0.021).

Autism-specific parenting stress (Figure 2).

Figure 2. Autism parenting stress least squares means, by parent ethnicity.

Note. Sample sizes at Baseline, Time 1, 2, and 3: Hispanic: n=95, 89, 84, 81; Non-Hispanic: n=243, 203, 206, 192.

No three-way interaction effect (time period x ethnicity x diagnosis) was found (χ²=2.44[4], p=0.655). In the refitted model (without three-way interaction), no significant two-way interactions were found between time period and ethnicity (χ²=4.40[3], p=0.221), time period and diagnosis (χ²=5.91[4], p=0.206), or ethnicity and diagnosis (χ²=0.65[2], p=0.724). With all interactions removed from the model, there was not a significant main effect of ethnicity on autism-specific parenting stress.

Avoidant coping (Figure 3).

Figure 3. Avoidant coping least squares means, by parent ethnicity.

Note. Sample sizes at Baseline, Time 1, 2, and 3: Hispanic: n=95, 88, 84, 81; Non-Hispanic: n=241, 203, 207, 192.

No three-way interaction effect (time period x ethnicity x diagnosis) was found (χ²=1.64[4], p=0.801). In the refitted model (without three-way interaction), no significant two-way interactions were found between time period and diagnosis (χ²=1.06[4], p=0.900), or ethnicity and diagnosis (χ²=1.66[2], p=0.435). In the final model, time period x ethnicity interaction effects (χ²=10.80[3], p=0.013) and main effects of ethnicity (χ²=9.66[1], p=0.002) were significant. Post-hoc comparisons of the least squares means revealed significant differences between Hispanic and non-Hispanic parents at Time 1 (β=−0.22, p=0.001) and Time 2 (β=−0.20, p=0.034).

Approach coping (Figure 4).

Figure 4. Approach coping least squares means, by parent ethnicity.

Note. Sample sizes at Baseline, Time 1, 2, and 3: Hispanic: n=95, 89, 84, 80; Non-Hispanic: n=241, 202, 206, 192.

No three-way interaction effect (time period x ethnicity x diagnosis) was found (χ²=3.81[4], p=0.432). In the refitted model (without three-way interaction), no significant two-way interactions were found between time period and diagnosis (χ²=4.57[4], p=0.334), or ethnicity and diagnosis (χ²=0.31[2], p=0.855). In the final model, time period x ethnicity interaction effects (χ²=11.57[3], p=0.009) and main effects of ethnicity (χ²=8.61[1], p=0.003) were significant. Post-hoc comparisons of the least squares means revealed significant differences between Hispanic and non-Hispanic parents at Time 1 (β=−0.35, p=0.008) and Time 3 (β=−0.35, p=0.016).

Negative family impact (Figure 5).

Figure 5. Negative family impact least squares means, by parent ethnicity.

Note. Sample sizes at Baseline, Time 1, 2, and 3: Hispanic: n=95, 89, 84, 81; Non-Hispanic: n=242, 203, 206, 192.

No three-way interaction effect (time period x ethnicity x diagnosis) was found (χ²=8.46[4], p=0.076). In the refitted model (without three-way interaction), no significant two-way interactions were found between time period and diagnosis (χ²=3.94[4], p=0.414), or ethnicity and diagnosis (χ²=0.56[2], p=0.755). In the final model, time period x ethnicity interaction effects (χ²=9.81[3], p=0.020) and main effects of ethnicity (χ²=5.39[1], p=0.020) were significant. Post-hoc comparisons of the least squares means revealed significant differences between Hispanic and non-Hispanic parents at Time 1 (β=−4.49, p=0.015).

Discussion

The current study provides the first investigation of the longitudinal pre-diagnosis experiences of low-income, ethnically underrepresented parents of child at risk for ASD. This study found differences (main effects) between the parenting stress, coping, and family impact trajectories among Hispanic versus non-Hispanic parents during this unique time period. Specifically, these findings indicated that, on average, Hispanic families in our sample reported lower parenting stress, avoidant coping, approach coping, and negative family impact across the four time points compared to non-Hispanic families. Interestingly, there were no such differences seen in autism-specific parenting stress. Further differences emerged in avoidant and approach coping and negative family impact over time between Hispanic and non-Hispanic families (time period x ethnicity interaction effects). Specifically, these findings indicated that non-Hispanic families report more positive change in both avoidant and approach coping, and negative family impact across the four time points of the current study than Hispanic families. These differences begin to shed light on the unique experiences of Hispanic and non-Hispanic parents.

Parenting Stress

Hispanic families reported parenting stress raw score means on the PSI-SF of 81.36-86.66 (61^st-68^th percentile) and non-Hispanic families reported parenting stress raw score means of 88.43-91.57 (69^th-73^rd percentile). This seems to fall within the range of values reported from previous findings of parenting stress levels during the diagnostic process in a predominantly White, non-Hispanic sample of parents with autism-specific and non-ASD developmental concerns about their child prior to a diagnostic evaluation. DesChamps and colleagues (2020) reported means over time in the 62^nd-66^th percentile for parents who had non-ASD developmental concerns about their child and in the 86^th-94^th percentile for parents with autism-specific concerns about their child. Thus, it appears that our overall sample, in general, reported lower levels of parenting stress prior to a diagnostic evaluation for their child than their White, non-Hispanic peers in the study done by DesChamps et al., (2020). It is possible in this previous study that parents who were more stressed were more likely to also have autism-specific concerns about their child prior to a diagnostic evaluation. In the current study, sample inclusion was based on M-CHAT screening results; a positive screening result is not necessarily associated with parenting stress or concern. Future research should focus specifically on the intersections of ethnicity, parent concern, and parenting stress within samples of families going through the diagnostic process for their children at risk for ASD.

Within the current sample, Hispanic parents may experience lower levels of parenting stress compared to non-Hispanic parents during the time period from when their child is identified as at risk for ASD, and extending beyond a diagnostic evaluation, due to differences in perceptions about their child’s development. For example, previous studies involving Hispanic parents of children with ASD have noted optimism as an important factor in parent experiences (Ekas et al., 2016; Willis et al., 2016). This finding is also in line with previous research that Hispanic mothers were more likely to report early concerns about temperament, but not about social or communication skills (Ratto et al., 2016). Hispanic mothers may have a different understanding of the developmental milestones that have been identified and value-marked, or considered “stressful,” within Western culture and research. Alternatively, there may simply be a mismatch between how the PSI-SF appraises parenting stress and how Hispanic parents appraise stress. Further investigation and inclusion of the priorities and cultural understanding of child development within Hispanic families is necessary, as well as assessments about how well the PSI-SF measures these cultural constructs/perceptions of parenting stress is necessary.

Autism-specific parenting stress means (range over time = 11.35-17.95) were slightly lower in the overall sample compared to previous research. Although it is difficult to make comparisons across studies with differing study populations, in a sample of parents whose children had already received a diagnostic evaluation and differential diagnosis, aged 2-6 years, Silva and Schalock (2012) found that parents of children with ASD reported a mean autism-specific parenting stress on the APSI of 22.93 (SD=10.43); while parents of children with another developmental disability reported a mean of 11.75 (SD=6.73). This comparison should be interpreted with caution, yet may suggest that parents experience more autism-specific parenting stress following a diagnostic evaluation and diagnostic label as children age and particularly if they receive an ASD diagnosis.

We found no difference by ethnicity in autism-specific parenting stress prior to a diagnostic evaluation. The difference between findings in the current study related to general parenting stress and autism-specific parenting stress may have to do with Hispanic parents placing less importance/notice to behavior problems associated with ASD (Ratto et al., 2016). Based on our findings, it seems that Hispanic parents experience comparable autism-specific parenting stress to non-Hispanic parents, but that stress does not spill over to create general parenting stress or dysfunction within the parent-child relationship. Parents may recognize their child’s difficulties but remain resilient due to factors such as optimism (Ekas et al., 2016; Willis et al., 2016), faith (Blanche et al., 2015), and a larger kinship network for raising their child (García-Coll & Vázquez García, 1995).

Coping

Although we found main effects indicating Hispanic parents were using less approach coping (generally considered adaptive), we also found main effects showing that Hispanic parents were using less avoidant coping (generally considered maladaptive). This finding again calls into question how we are measuring and valuing coping strategies. It is possible that certain, more specific, types of coping are used by Hispanics and/or are differentially assigned value (i.e., considered adaptive) and are associated with positive outcomes. As stated above, the Hispanic group within our sample reported lower levels of parenting stress; thus, it is not surprising that they are also reporting less use of coping strategies. Consistent with the Stress, Appraisal, Coping framework, we would not necessarily expect parents to actively cope with something that they do not perceive as problematic or stressful (Lazarus & Folkman, 1984).

Although previous research has shown that Hispanic parents are more likely to use religious coping (Abraído-Lanza et al., 2004; Connell & Gibson, 1997), and thus may be more likely to adopt an attitude to simply leave matters up to God, a follow-up analysis showed that, at Baseline, Hispanic parents did not report more use of religious coping than non-Hispanic parents in our sample (t=−0.82, p=0.411) using the sum of two “religious coping” items on the Brief COPE, which are not included in the subscales described in the methods section. In our sample, then, it may simply be that parents are choosing not to engage in coping behaviors until they learn more about ASD and whether their child has it. We found a significant time period x ethnicity interaction, meaning that there were different patterns that emerged across time in the use of avoidant and approach coping strategies in Hispanic and non-Hispanic parents. Additional research about other factors that contribute to these differences is necessary. For example, future research could assess whether increased ASD/child development knowledge/learning over time mediates the relation between parenting stress and coping.

Family Impact

In the current study, Hispanic parents reported less negative impact on the family. These findings are in line with previous research that has shown that Hispanic families show strengths related to family cohesion (Rayan & Ahmad, 2016; Willis et al., 2016). Hispanic parents may be more likely to seek out social support, in line with “familismo,” a concept referring to the commitment of Hispanic families to maintaining strong kinship ties (García-Coll & Vázquez García, 1995). The negative family impact subscale used in the current study focuses specifically on the impact the child has on negative feelings about parenting and the parent’s social life. The finding that Hispanic families reported a lower negative impact on the family suggest major protective factors in those areas. Collectivistic cultures such as those with which most Hispanic families associate, may place more value and opportunity for caregiving help outside of nuclear families, which in turn could lead to less stress, different coping, and less negative impact on parenting feelings and the parent’s social life. These parents may be able to effectively share the burden or offload stress to grandparents/aunts/uncles/cousins and more informal family supports, providing the opportunity for them to take the time to engage with their friends and partners/spouses for emotional support. Additional research is necessary to further understand the people involved in the care of Hispanic children at risk for ASD, as well as how Hispanic parents experience the diagnostic process, assign meaning, and cope as necessary.

Strengths and Limitations

This study adds to a very limited body of work focused on the outcome patterns of parents whose children are currently going through the diagnostic process for ASD. One strength of this study was the study’s sample. To our knowledge, it is the first to focus on a racially and ethnically underrepresented sample of families. Another strength of this study was that we looked at parent experiences over time, both leading up to, during, and immediately following diagnostic ascertainment. Further, we included a time-varying covariate to account for when and what diagnosis the child received. This distinction is often ignored in longitudinal research using similar samples.

This study also had limitations. Because we conducted separate models for each outcome, we are only able to look at time ordered effects of ethnicity on each outcome. We are not able to determine the pathways in which stress, coping, and family impact are related. Further, we are not able to assess causation, as there are many factors that could be contributing to these actual differences by ethnicity or to our interpretations of the differences. Additional longitudinal path modeling may further elucidate the relationships between these variables over time and causal mechanisms. We included a time-varying covariate to account for when the child received a diagnosis and what the diagnosis was (ASD vs. non-ASD), and looked more specifically at the interactions between time, ethnicity, and diagnosis, to better understand differences by ethnicity. Future research can build on these findings by focusing specifically on differences by diagnosis and how ethnicity and diagnosis interact over time following a diagnostic evaluation and differential diagnosis. For example, if a Hispanic parent is told that their child does not have ASD, their autism-specific stress may decrease more than a non-Hispanic parent’s.

We also recognize the extensive diversity that exists among Hispanic parents and families, as noted in the introduction, and within our sample in terms of language preference, country of origin, being born in the US, and education level. Stress, coping, and family impact may look different amongst different Hispanic groups. Of equal importance, is the vast diversity within our non-Hispanic group. This is a major limitation of the current study. Although the overall sample was similar in that they were urban and low-income, the non-Hispanic group represents Black non-Hispanic, White non-Hispanic, and Asian non-Hispanic families, who are culturally and contextually different. While our lower numbers of White non-Hispanic and Asian non-Hispanic families made it difficult to make comparisons in our outcome measures by race, we encourage future research to look closely at the intersecting impact of ethnicity, race, SES, language, education, and immigration status on parent outcomes during the ASD diagnostic process. Further, it is likely that there are similarities among other subgroups within our sample, across Hispanic and non-Hispanic groups. For example, it may be that Hispanic immigrants have more in common with non-Hispanic immigrants than other Hispanic parents born in the US. We recognize and similarly encourage future research to build upon our findings and look closely at multiple indicators of culture and experience.

We similarly recognize that almost 80% of our non-Hispanic sample was Black, whereas only 8% of our Hispanic sample was Black. Due to the structural racism and history of anti-Blackness at the center of social and health disparities within the United States (Bailey et al., 2017; Haley, 2020), our non-Hispanic subgroup of families may have inherently shown higher levels of stress and negative impact on the family because of the compounding nature of having a child with a developmental concern on top of experiencing racism (Nomaguchi & House, 2013). Research on the unique experiences of Black families of children with ASD that goes beyond the simple fact that Black families face significant barriers to accessing high-quality ASD care services (Magaña et al., 2012; Montes & Halterman, 2011), is emerging (Burkett, Morris, Manning-Courtney, Anthony, & Shambley-Ebron, 2015; Hannon, 2017). Future research should continue to document and build upon current understanding about the intersectionality of being a Black parent of a child with ASD.

A final limitation and conclusion of the current study is the uncertainty about the cultural validity of the outcome measures used. Although many measures studying parenting stress, coping, and family impacts and experiences have been used in diverse samples, it is unclear if these measures are truly picking up on the nuances of different cultural perspectives. For example, while the PSI might be a valid way to capture parenting stress, as defined by the current body of literature, it fails to capture alternative perspectives and experiences of stress. It is clear that most of the measures available are based on Western ideologies of what is stressful, what coping means, and what constitutes “impact.” We look forward to future research that helps to redefine and develop alternative frameworks for measuring family outcomes across cultures within ASD, DD, and broader parenting and family research.

Conclusions and Implications

In the current study, Hispanic families reported lower levels of parenting stress and negative family impact than non-Hispanic families. We have presented some hypotheses about why these differences in outcomes may be occurring; however, without causal data, it is difficult to pinpoint specific intervention and support targets. Still, our findings highlight clear strengths among Hispanic families, making it likely that they will benefit from intervention strategies that build upon strengths instead of the alleviation of stressors. Multi-family support groups that focus on empowerment are one example of potentially effective intervention (Banach, Iudice, Conway, & Couse, 2010) and could be tailored to the pre-diagnosis period. Similarly, although it is difficult to link our findings about outcomes to specific intervention targets, because Hispanic families also reported lower use of both avoidant and approach coping compared to non-Hispanic families, it is likely that Hispanic families may benefit from interventions that target parent education and provide clear action steps. In other words, interventions that leverage the strengths of Hispanic families into adaptive and effective action may be helpful. Mindfulness and information group interventions, for example, are promising for parents of children with ASD and other developmental disabilities to reduce stress (Zuckerman et al., 2017) and increase positive reappraisal coping and parent quality of life (Lunsky et al., 2017; Seltzer, Floyd, Song, Greenberg, & Hong, 2011). These types of interventions that focus on resiliency might be particularly beneficial for Hispanic families and may also better inform work to increase resiliency in non-Hispanic families.

It is also possible that, because Hispanic families are less stressed by parenting and experience less of a negative family impact, they may not be as motivated to pursue a diagnostic evaluation as their non-Hispanic peers. In part, low income Hispanic parents, who were more likely in our sample to be immigrants and speak Spanish, may not feel the benefits of developmental screening for their children would outweigh the potential risks of engaging with health services. Health policy changes, such as ensuring a culturally competent staff and the availability of translators, may begin to make it easier for Hispanic parents to access these services, even if they are not motivated by experiencing parenting stress. One intervention that has been directly applied to the pre-diagnostic period is Family Navigation (FN), which was the basis of the larger study from which this data was drawn. FN is a structured, manualized intervention consisting of core components (Broder-Fingert et al., 2020) designed to support families from the time of a positive ASD screen through engagement in recommended services. FN has been found to increase the likelihood of, and decrease time to, diagnostic ascertainment, particularly for Hispanic families (Feinberg et al., 2021).

In conclusion, Hispanic families seem to report lower stress, coping, and family impact throughout the diagnostic process, indicating significant protective and resilience factors. Although Hispanic families experience disparities in access to timely ASD diagnostic and treatment services (Christensen et al., 2016; Maenner et al., 2020; Magaña et al., 2013; Yingling et al., 2018), they also respond particularly well to culturally informed intervention efforts (Feinberg et al., 2021, Magaña, Lopez, & Machalicek, 2017), which may be due to their lower levels of stress and family impact. While there remains a need to test culturally informed interventions and better understand the unique needs of Hispanic families, our work suggests that differences in parent experiences may provide leverage and allow interventions and support to capitalize on the unique strengths of Hispanic families to reduce disparities.