Parental Cognitions, Treatment Engagement, and Child Outcomes of ADHD Behavioral Treatment among Asian American Families

Abstract

Asian American (AA) families remain critically underrepresented in clinical trials for ADHD interventions. Little is known about AA families’ engagement in and outcomes of behavioral treatment (BT). Comparing AA families to other minoritized (OM) families and White families, this study examined parental cognitions, treatment engagement, and child outcomes of BT for ADHD inattentive type (ADHD-I). Path analyses were conducted utilizing data from a randomized controlled trial of BT for ADHD-I (N = 199 children, ages 7–11). Racial/ethnic differences in pretreatment parental self-competence and treatment expectations were examined for AA (n = 29) compared to OM (n = 35) and White (n = 135) parents. Two additional path models were conducted to examine the relations among race/ethnicity, pretreatment parental cognitions, treatment engagement, and posttreatment child outcomes. Direct effects of race/ethnicity and parental cognitions on posttreatment child outcomes as well as their indirect effects via treatment engagement were estimated. At pretreatment, AA parents endorsed lower parental self-competence and treatment expectations compared to OM and White parents. At posttreatment, AA parents reported fewer improvements in ADHD symptoms than White parents and lower global psychosocial improvement than OM parents. For all parents, treatment expectations positively predicted parent- and observer-rated treatment engagement, which in turn predicted child global psychosocial improvement. Path analyses indicated that the relationship between treatment expectations and posttreatment child global improvement was fully mediated by treatment engagement. These findings suggest that treatment expectations impede AA parents’ engagement and success in BT. Implications for cultural adaptations of BT to improve AA families’ treatment experience are discussed.

The Asian American (AA) community has persistently faced mental health disparities in service use and research (Cook et al., 2017; Đoàn et al., 2019). This disparity extends to attention deficit/hyperactivity disorder (ADHD), now diagnosed in approximately one in 11 U.S. children (Bitsko et al., 2022). Current estimates indicate under-diagnosis of ADHD in AA youth, despite similar prevalence rates across North America and Asia (Polanczyk et al., 2014). Recent analyses of the 2018 National Survey of Children’s Health found that AA youth were 73% less likely than White youth to receive an ADHD diagnosis, followed by Latine (32%) and Black youth (22%; Zhao et al., 2023). AA children with ADHD were 46% less likely than their White peers to have received treatment for ADHD (Shi et al., 2021). Even within a single healthcare system – in which groups have relatively equal access to ADHD care, AA had lower rates of treatment utilization than White, Black, and Latine youth (Chung et al., 2019). Yet ADHD-related impairments may be especially detrimental to AA youth, who are already at heightened risk for low self-esteem, parent–child conflict, suicidal ideation, and academically related anxiety (Chen & Graham, 2018; Wyatt et al., 2015).

Treatment disparities for AAs is theorized to persist throughout the multi-stage process of ADHD service use (i.e., problem recognition, decision to seek help, service selection, and service utilization and engagement; Eraldi et al., 2006) due to intrapersonal and interpersonal factors (e.g., cognitive/affective processes, cultural and relational values) in addition to structural barriers (Kim & Lee, 2022). The literature on treatment barriers for AAs has largely focused on the impact of stigma and other relatively stable aspects of cultural identity (e.g., acculturation, interdependence, collectivistic values; Cook et al., 2017; Hall et al., 2019) on the early stages of service use (i.e., problem recognition, help-seeking attitudes; Kim & Lee, 2022). Little is known about the cognitive barriers proximal to the later stages—treatment engagement and outcomes – for AAs who use services given the challenges faced by AAs to even enter treatment (Interian et al., 2013). Exploration of cognitive barriers is essential because they are potentially modifiable factors that can improve the effectiveness of treatment for AAs. These gaps are especially apparent among AA families of children with ADHD.

Behavioral Treatment Trials among Asians and AA

Behavioral treatment (BT) is widely used for improving ADHD-related impairments (Evans et al., 2014). BT teaches parents to modify the contingencies in their child’s environment for behavior change. BT has shown to increase positive parenting, improve parent–child relationships, and reduce parenting stress (Fabiano et al., 2009). AA families, however, are critically underrepresented in BT research trials and little is known about their engagement and outcomes (e.g., Ortiz & Del Vecchio, 2013). Asian heritage parenting values and practices may conflict with skills taught in BT. Heritage values emphasizing parental control, family hierarchy, and child compliance (e.g., Chinese guan and Korean ga-jung-kyo-yuk; Chao, 1994; Choi et al., 2013) may be incongruent with BT skills of ignoring misbehavior and positive reinforcement strategies of reward and praise (Lau et al., 2010). Criticism, believed to motivate persistence, is instead often favored by Asian parents (Lau, 2006). Endorsement of Asian heritage values was associated with lower acceptability of BT (Ho et al., 2012).

A small body of literature examining culturally adapted BT in Asian countries has revealed preliminary efficacy (Malik et al., 2017; Shah et al., 2021; So et al., 2008). These studies delivered treatment by linguistically and ethnically matched therapists and modified BT protocols to reflect the local context and culture. Studies conducted in Hong Kong, India, and Pakistan included modifications to match parents’ high valuation of education (e.g., focus on classroom behaviors), address low acceptability of using rewards and expressing positive emotions (e.g., extended coaching of quality time, praise, and token economy system), and employ culturally-appropriate ways of coping (e.g., inclusion of spiritual coping techniques) (Malik et al., 2017; Shah et al., 2021; So et al., 2008). When cognitive restructuring and discussion of cultural barriers were added to a BT trial for Chinese American immigrant families in the U.S., results indicated parity in treatment engagement and outcomes (Lau et al., 2011).

Meta-analyses have shown that when BT was culturally adapted for specific ethnic groups, parenting behavior and child outcomes improved (van Mourik et al., 2017), and for psychosocial treatment outcomes more broadly, the average effect size of cultural adaptations for AAs was more than twice those for Black and Latine/x samples (Smith et al., 2011). As such, cultural fit of BT may be especially salient for AA engagement and outcomes. Adapted interventions, however, present significant challenges to fidelity and the feasibility of large-scale implementation in an increasingly multicultural and ethnically heterogeneous landscape (Bernal et al., 2009; Cabassa & Baumann, 2013). Empirical approaches are needed to clarify whether/when adaptations are warranted for AA families prior to taking on effortful and costly adaptation processes (Lau, 2006). Little research has examined whether AA parents have differential treatment engagement and outcomes relative to other parents when participating in non-culturally adapted BT for ADHD in an ethnically heterogeneous group. This knowledge, along with the identification of modifiable predictors of engagement and outcomes, would serve to inform the types of and extent to which adaptations could address AA disparities.

The Role of Parental Cognitions in ADHD Treatment Engagement and Outcomes

The current study applies the framework of the Health Belief Model (HBM; Rosenstock et al., 1988) to examine predictors of BT engagement and outcomes among AA families with children with ADHD. The HBM posits that key determinants of health behaviors comprise treatment expectations – perceived benefits relative to barriers to treatment – and self-efficacy – perceptions of one’s ability to influence outcomes – in addition to perceptions of the illness (Rosenstock et al., 1988). Parents’ perceptions of competence (a construct reflecting parenting efficacy and satisfaction; Ohan et al., 2000) and treatment expectations are viewed as especially salient predictors of ADHD treatment engagement and outcomes (Mah & Johnston, 2008). Given the cross-context challenges of ADHD, parents of children with ADHD endorsed lower parental self-competence than those of typically developing children (Primack et al., 2012). Lower parental self-competence was in turn linked to worse outcomes for ADHD (Chacko et al., 2017). Similar relations have been found among treatment expectations, treatment adherence, and outcomes (Johnston et al., 2010; Nock et al., 2007).

The HBM has been promoted as an apt explanatory framework for service use disparities in AAs given the influences that cultural norms and beliefs assert on HBM mechanisms. As illustrated by Kim and Zane (2016) with a large sample of ethnically heterogeneous young adults with high distress, AAs reported lower treatment expectations – fewer benefits and more barriers – and lower help-seeking intentions than their White peers. For AA parents of children with ADHD, parental self-competence and treatment expectations may be salient cognitive barriers as they navigate beliefs about ADHD and values tied to their children’s academic and behavioral functioning inherent to Asian heritage culture (Mah & Johnston, 2007; Wong et al., 2018) within the context of the model minority stereotype (MMS, Wong & Landes, 2022). While many structural barriers (e.g., racism, high costs, lack of culturally- and language-matched services) affect all racially/ethnically minoritized groups, AAs are impacted by unique structural barriers such as MMS (Kim & Lee, 2022). MMS is the externally imposed stereotype depicting AAs as independently high-achieving and free of psychological distress, underlying the belief that low academic performance and distress in AAs are a result of inferior traits or behaviors of the child or parent (Qin & Han, 2011). MMS is linked to the under-identification of ADHD in school settings (Wong & Landes, 2022), and when internalized, increases stigma and negative help-seeking attitudes in AA (Gupta et al., 2011; Kim & Lee, 2014).

Causes of ADHD and related impairment are often attributed to parental inadequacy among Asian parents, increasing shame and stigma (Lam & Ho, 2010; Shah et al., 2019). In a sample of Taiwanese parents, for example, inattention symptom severity (but not hyperactivity/ impulsivity) was found to be positively related to affiliate stigma (i.e., internalized stigma in caregivers of individuals with a disability), a finding the authors attributed to parental censures by relatives and teachers for inadequate monitoring of academic performance (Chang et al., 2020). Affiliate stigma was in turn linked to lower acceptability of BT. Engaging in BT for their child’s inattention and related academic impairments may be especially challenging for U.S. AA parents given the impact of MMS. Indeed, shame and guilt for their children’s behavior problems were negatively linked to treatment expectations and help-seeking among U.S. AA parents of children with attention/behavior challenges (Lau & Takeuchi, 2001; Mah & Johnston, 2007, 2008). Pretreatment assessment of cognitive barriers may thus be an efficient way to bolster engagement in treatment due to their readily modifiable nature (McCabe et al., 2020). To our knowledge, no study has examined whether AA parents have differential levels of parental self-competence and treatment expectations – and how these cognitions influence treatment engagement and outcomes of BT – compared to parents of other racial/ethnic backgrounds.

The Current Study

To better understand treatment experience among AA families and the need for adapted interventions, the current study compares relations among parental cognitions, treatment engagement, and outcomes of BT among AA, White, and other racially/ethnically minoritized (OM) parents with children with ADHD-inattentive type (ADHD-I). The comparison groups were derived considering the cognitive/affective treatment barriers related to: (a) structural racism faced by racially/ethnically minoritized families compared to White families, and (b) structural barriers unique to AA compared to OM groups (e.g., MMS, Kim & Lee, 2022).

Guided by the HBM, the aims of this study were twofold. (1) We examined the associations between race/ethnicity and parental cognitions at pretreatment, hypothesizing that compared to White and OM parents, AA parents would endorse lower parental self-competence and treatment expectations. (2) Based on the suggested effects of parental cognitions on engagement and outcomes (Mah & Johnston, 2008) and parent engagement on outcomes (Nock & Ferriter, 2005), we examined the associations among AA membership, parental cognitions, treatment engagement, and outcomes. We hypothesized that compared to White and OM parents, AA parents would demonstrate lower engagement and endorse less child improvement at posttreatment given: (a) the non-culturally adapted nature of our treatment, which may impact the treatment experience of AA parents more than OM parents (Smith et al., 2011), and (b) the suggested links between inattention symptoms and related academic impairment and BT acceptability in Asian parents (Chang et al., 2020). We hypothesized an indirect effect of parental cognitions on child outcomes, mediated by treatment engagement.

We operationalized parental cognitions as parental self-competence and treatment expectations. Treatment engagement was measured via observer and parent ratings of attendance, in-session participation, and adherence to in-between session practice (Nock & Ferriter, 2005). Treatment outcomes comprised parent-rated child ADHD severity and global psychosocial improvement. We elected to not include teacher-rated outcomes because previous research has shown racial/ethnic disparities in teacher identification of ADHD symptoms that more severely impact AA students. AA students were two to five times less likely than White and Black students to receive a teacher referral even after accounting for the effects of teacher-rated academic, social, behavioral, and emotional problems (Villodas et al., 2019), calling into question the validity of this measurement approach to address the study aims in the context of comparisons across racial/ethnic groups.

Method

Participants

The sample comprised 199 children and at least one parent who participated: (N = 199). Sample recruitment was conducted via mailings to schools and offices of pediatricians, child psychiatrists, and psychologists, postings on online parent networks or professional organizations, or through word-of-mouth. As a function of the data available from the original trial (Pfiffner et al., 2014), the inclusionary criteria comprised: (a) a primary Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) diagnosis of ADHD-I, as confirmed by the Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS-PL; Kaufman et al., 1997) clinical interview, (b) Full Scale IQ greater than 80, as confirmed by the Wechsler Intelligence Scale for Children, Fourth Edition (Wechsler, 2003), (c) age 7–11 years (grades 2–5), (d) attending school full time in a regular classroom, (e) living with at least one parent for the past year, (f) family schedule that permitted participation in scheduled groups, (g) school proximity within 45 min of either study site to allow for study personnel to conduct teacher consultation meetings, and (h) teacher consent to participate in a school-based treatment. Exclusionary criteria comprised: (a) children who were taking non-stimulant psychoactive medication due to the difficulty withholding medication to confirm ADHD-I symptoms, (b) families planning to initiate or change medication treatment (stimulant or otherwise) in the near term, and (c) children with significant developmental disorders (e.g., pervasive developmental disorder) or neurological illnesses.

Mean child age at randomization was 8.6 years (range 7 to 11), with 26% in 2nd grade, 31% in 3rd grade, 27% in 4th grade, and 17% in 5th grade. Boys comprised 58% of the sample. At randomization, 4.5% of youth were taking medication (all but one received stimulant medication) to address ADHD-related symptoms. Total household income was below $50,000 for 14%, $50,000–100,000 for 27%, $100,000–150,000 for 28%, and more than $150,000 for 31% of families; 13% of the participants were living in single-parent homes.

Most primary caregivers (i.e., those who completed all measures and attended treatment) were mothers (90.5%). Most primary caregivers had an advanced or professional degree (81, 40.7%) or were college graduates (79, 39.7%), 33 parents (16.6%) had some college/post high school education, and 4 (2.0%) were high school graduates. Their racial/ethnic background was as follows: 3 parents (1.5%) identified as American Indian/Alaska Native (AI/AN)/White, 1 (0.5%) as AI/AN/ Hispanic/Latine/White, 22 (11.1%) as Asian American (AA), 2 (1.0%) as AA/Native Hawaiian or Pacific Islander (NHPI), 5 (2.5%) as AA/White, 9 (4.5%) as Black, 1 (0.5%) as Black/White, 4 (2.0%) as Hispanic/Latine, 17 (8.5%) as Hispanic/Latine/White, and 135 (67.8%) as White. Parents were asked to select all racial/ethnic categories with which they identified. Related to our main study question regarding the unique treatment experiences for AAs, we combined the participants who endorsed AA as one of their racial identities (n = 29). The two comparison groups comprised: (1) all other Black, Indigenous, and parents of color who endorsed AI/AN, Black, or Hispanic/Latine as one of their racial identities (OM, n = 35), and (2) parents who endorsed White as their only racial identity (n = 135). Compared to the local demographics, our sample was over-represented with White and AI/AN individuals, slightly underrepresented with Black and Latine individuals, and markedly underrepresented with AA individuals (Bay Area Census, 2010). Education levels were slightly higher than the local area and income distribution was similar. Medication usage was markedly lower than the national average (approximated at 66% of children with ADHD; Visser et al., 2010).

Procedure

All study procedures received approval from the Committee on Human Research of the University of California, San Francisco. For full descriptions of participant screening, flow/attrition, diagnostic procedures, participant compensation, treatment fidelity, and therapist qualifications, see Pfiffner et al. (2014). All participating children met DSM-IV criteria for ADHD predominantly inattentive type (ADHD-I). Participating parents provided written consent and children provided written assent. Treatment was provided at no cost to families.

Design

Children were randomized to one of two BTs: (a) Child Life and Attention Skills Treatment (CLAS; n = 74; AA n = 11, OM n = 10, White n = 53), or (b) Parent Focused Treatment (PFT; n = 74; AA n = 10, OM n = 9, White n = 55), or treatment as usual (TAU; n = 51; AA n = 8, OM n = 16, White n = 27). Treatment occurred over a 10- to 13-week period. Following treatment, families attended laboratory visits and teachers received rating scales. All measures and treatment groups were administered in English; parents were eligible to participate upon endorsing an 8^th grade English reading level or above during the initial screening.

Treatment Conditions

See Pfiffner et al. (2014) for a full description of all treatment conditions including the development and rationale for the parent, child, and classroom components for CLAS. Parent group leaders and observing clinicians for the CLAS and PFT conditions comprised three licensed clinical psychologists and one clinical psychology postdoctoral fellow. Two group leaders co-led child groups for CLAS and comprised B.A. and M.A. level clinicians (N = 10).

Child Life and Attention Skills Treatment (CLAS; Pfiffner et al., 2014)

CLAS comprised three components: (a) ten 90-min parent group sessions with up to six 30-min family meetings (parent, child, and therapist), (b) ten 90-min child group sessions, and (c) one 30-min teacher orientation with up to five 30-min meetings family and teacher meetings and booster sessions as needed. All therapists identified as White females.

Parent Focused Treatment (PFT)

PFT comprised the parent training group identical to that of CLAS with two exceptions. PFT did not include training in working with teachers nor the child skills taught in CLAS. Child’s attendance was also excluded in individual family meetings. Teachers were mailed information about the study and ADHD-I as well as suggested classroom accommodations. All PFT therapists were female and identified as White except one, who identified as AA.

Treatment as Usual (TAU)

TAU families received a written diagnostic report of the pretreatment assessment and were given a list of community treatment providers without specific treatment recommendations. After the follow-up assessment, TAU families were offered the opportunity to participate in a two-session workshop on the strategies taught in the CLAS groups, with limited individual follow-up. TAU children received medication (14%), psychotherapy (33%), educational intervention (51%) and classroom accommodations (53%) between pre- and posttreatment.

Measures

Demographic Characteristics

Information about the family demographics (e.g., income, parental education level), family structure, and the child’s treatment history were collected during the screening interviews.

Children’s ADHD and Oppositional Behaviors Severity (Pre- and Posttreatment)

Parents completed the CSI-4 (Gadow & Sprafkin, 1994), which assessed for symptoms of ADHD and oppositional defiant disorder that correspond with DSM-IV criteria. Symptoms are rated on a 4-point scale (0 = never to 3 = very often); those rated as occurring often (2) or very often (3) are considered as present. The ADHD and ODD scales have shown internal consistency in samples of children with alphas ranging from 0.86 to 0.92 and scores have been correlated with ADHD diagnoses and externalizing problems (Gadow & Sprafkin, 1994; Sprafkin et al., 2002). Cronbach’s αs at baseline ranged from 0.80 to 0.90 for ADHD and 0.83 to 0.92 for ODD.

Children’s Global Psychosocial Impairment (Pretreatment) and Improvement (Posttreatment)

At pretreatment, parents completed the Clinical Global Impression Scale, Severity version (CGI-S; National Institute of Mental Health, 1985; 1 = no impairment to 7 = maximal impairment). The Clinical Global Impressions Scale, Improvement version (CGI-I) was administered at posttreatment (1 = much worse to 7 = much improved). The CGI is a medication-sensitive measure of treatment response used in clinical trials for ADHD (Solanto et al., 2009).

Parental Self-competence

Parents completed the Parenting Sense of Competence Scale (PSOC; Johnston & Mash, 1989), comprising 16 items rated on a 6-point scale (1 = strongly disagree to 6 = strongly agree). 8 items assess for Satisfaction (valuing and comfort with parenting role; e.g., “My talents and interests are in other areas, not in being a parent,” “Being a parent makes me tense and anxious”) and 8 items for Self-Efficacy (problem-solving ability and capability in parenting role; e.g., “Being a parent is manageable, and any problems are easily resolved,” “If anyone can find the answer to what is troubling my child, I am the one”). Higher scores indicate greater parental self-competence. PSOC has shown adequate internal consistency with alphas ranging from 0.77 to 0.85 in community samples in the U.S. and Asia and scores have been correlated with child behavior problems, parenting styles, self-esteem, and postnatal depression (Ngai et al., 2007; Ohan et al., 2000; Suwansujarid et al., 2013). Cronbach’s α was 0.81.

Pretreatment Treatment Expectations

Parents completed the Parent Treatment Expectations Scale (PTES; Kaiser et al., 2010), consisting of 20 items scored on a 1 (strongly agree) to 6 (strongly disagree) scale. Items assessed for parents’ motivation for treatment and expectations for positive treatment outcomes (e.g., “I’m not sure why I’m here; it’s my child who needs help,” “I expect that by participating in this program, my child’s problems at home and school will really improve.”). Items were recoded; a higher score indicated more positive treatment expectations. Cronbach’s α was 0.78.

Parents’ Treatment Engagement

Treatment engagement for CLAS and PFT was assessed via parent-report and clinician ratings. Parent completed 10 weekly ratings after each session with five items: “How interesting was today’s session for you?”, “How much did you participate in today’s session?”, “How well did you understand the material presented in today’s session?”, “How helpful was today’s session?” (four items rated 1 = not at all, 7 = a great deal) and homework completion (0 = did not complete, 1 = not well to 7 = very well). Mean scores for all items were calculated then aggregated across the weeks to also reflect attendance. Cronbach’s α was 0.78 for the five items.

Clinicians completed ratings during live observations of each session on four items indexing how much the parent appeared to engage in the session, understand the material presented, implement the homework assignment from the previous week, and adhere to the program overall (1 = not at all, 7 = a great deal). Ratings were based on verbal and non-verbal indications (e.g., asked a relevant question; looked at/completed handouts; took notes; brought in copies of child’s behavior chart). There was substantial agreement (κ = 0.71) between the clinician and a second observer who attended 12% of the sessions. Mean scores for the four items were calculated for all 10 sessions and summed into a total score to reflect attendance. Cronbach’s α was 0.86 for the four items. Missing scores for parent- and observer-rated engagement were imputed with the session mean. Observer ratings were missing 25 of 1480 scores (1.68%) and parent ratings were missing 41 of 1480 scores (2.77%).

Data Analytic Plan

All statistical analyses were conducted with RStudio version 4.2.1 (RStudio, PBC). Preliminary bivariate analyses were conducted such as Chi-squared tests, Fisher’s exact tests, Kruskal–Wallis tests, and post-hoc Dunn’s tests with Benjamin-Hochberg corrections to reduce false discovery rate (FDR) when appropriate. Three path models were conducted with the lavaan package in RStudio (Rosseel, 2012). Models were bootstrapped with 5,000 draws and fitted to proposed cutoff criteria (Hu & Bentler, 1999): comparative fit index (CFI ≥ 0.95), Tucker-Lewis Index (TLI ≥ 0.95), root mean square error of approximation (RMSEA < 0.06), and standardized root mean square residual (SRMR < 0.08). Race/ethnicity was dummy coded for AA/OM and AA/White comparisons (AA = 0, OM/White = 1) given the main study questions.

The first path model examining race/ethnicity differences in pretreatment parental cognitions used the full sample (N = 199) because parental cognitions were assessed prior to randomization. Race/ethnicity comparisons and covariates were treated as predictor variables and parental self-competence and treatment expectations as outcome variables. We conducted two other path models to assess observer-rated and parent-rated treatment engagement separately given our interest in parents’ perceptions of engagement compared to an objective third-party. We estimated indirect paths of race/ethnicity and pretreatment parental cognitions treatment engagement (assessed during treatment) posttreatment outcomes as well as the direct paths of pretreatment variables to posttreatment child outcomes. To capture the experiences of parents who received BT, these analyses used the subsample of parents who received CLAS or PFT (n = 148) and controlled for treatment type. Effects of these models were considered as significant at p < 0.025 to account for familywise error using the Bonferroni correction.

Results

Tables 1 and 2 present descriptive characteristics and Spearman correlations for the main study variables. Bivariate analyses showed that after Benjamin-Hochberg FDR corrections, parents in the CLAS condition reported higher engagement (p = 0.007) and lower posttreatment ADHD (p = 0.039) compared to parents in PFT. Parents endorsed higher posttreatment ADHD severity for male children than females (p = 0.026). Racial group differences in education (χ² = 18.28, p = 0.006) were such that more White parents had an advanced/graduate degree, followed by AA parents, and more White and AA parents were college graduates than OM parents (Fisher’s exact test ps range = 0.016 to 0.023). AA parents scored lower on observer-rated engagement than OM and White parents (respective ps = 0.016 and 0.016). AA parents scored marginally lower on parental self-competence than OM and White parents (respective ps = 0.065 and 0.065) and treatment expectations than OM parents (p = 0.090), and marginally higher on posttreatment ADHD severity than White parents (p = 0.058). OM and White parents did not differ on any variable. No differences in income were found.

Table 1.

Descriptive statistics of study variables by racial/ethnic identity and full sample

	AA n = 29		OM n = 35		White n = 135		Full sample N = 199
Variables	M	SD	M	SD	M	SD	M	SD	min	max	skew	kurtosis
PRE ADHD severity	25.60	5.60	26.69	6.85	24.82	7.12	25.31	6.91	11.00	44.00	0.36	−0.41
PRE ODD severity	6.08	3.94	5.94	5.17	7.21	4.45	6.67	4.36	0.00	24.00	0.87	0.97
PRE GPF impairment	4.08	0.77	4.75	0.93	4.45	0.83	4.46	0.82	2.50	7.00	0.39	−0.16
Parental self-competence	3.89a	0.63	4.17b	0.61	4.24b	0.66	4.18	0.66	2.50	5.56	−0.10	−0.62
Treatment expectations	4.45a	0.51	4.82b	0.26	4.82	0.40	4.74	0.43	3.10	5.65	−0.39	0.28
Treatment engagement (O)	52.10a	8.31	60.72b	8.56	56.57b	13.03	55.74	13.53	0.00	70.00	−2.14	5.33
Treatment engagement (P)	49.48	7.82	54.13	7.20	50.89	12.12	50.63	12.34	0.00	68.95	−2.03	5.23
POST ADHD severity	19.98a	5.70	17.44	5.05	16.96b	6.18	18.55	6.83	4.00	39.00	0.53	0.12
POST GPF improvement	5.60	0.68	6.50	0.52	5.89	0.82	5.70	0.94	3.00	7.00	−0.49	−0.13

Means not sharing subscripts differ at p < .10 after Benjamin-Hochberg FDR correction

PRE pretreatment, GPF global psychosocial functioning, P parent-rated, POST posttreatment, O observer-rated

Table 2.

Spearman correlations of study variables

	1	2	3	4	5	6	7	8
1. PRE ADHD severity	-
2. PRE ODD severity	0.35	-
3. PRE GPF impairment	0.20**	0.13+	-
4. Parental self-competence	−0.16*	−0.24**	−0.01	-
5. Treatment expectations	0.06	−0.03	0.17*	0.34***	-
6. O treatment engagement	0.03	−0.04	0.09	0.10	0.34***	-
7. P treatment engagement	0.05	0.03	0.24**	0.13	0.35***	0.70***	-
8. POST ADHD severity	0.55***	0.25***	0.15*	−0.16*	−0.14+	−0.14+	−0.16+	-
9. POST GPF improvement	−0.01	0.00	0.08	0.04	0.17*	0.33***	0.35***	−0.39***

GPF global psychosocial functioning, P parent-rated, PRE pretreatment, POST posttreatment, O observer-rated

⁺p < .10,

^*p < 0.05,

^**p < 0.01,

^***p < 0.001

Aim 1. AA Identification and Parental Cognitions at Pretreatment

The first path model estimated the effect of race/ethnicity on pretreatment parental self-competence and treatment expectations (see Fig. 1 for all estimated paths). Child’s clinical functioning at pretreatment (i.e., ADHD and ODD severity and global psychosocial impairment) and parent education were selected as covariates based on statistical and theoretical importance. To allow for positive degrees of freedom, paths estimated from the pretreatment clinical functioning variables to parental cognitions were limited to the relations that were significantly correlated: ADHD and ODD severity to parental self-competence (respective rs = −0.16 and −0.24, ps < 0.05) and global impairment to treatment expectations (r = 0.17, p = 0.014).

Fig. 1.

Path analyses from parent racial identification to parental cognitions at pretreatment

The model fit indices showed a good fit: χ² = 1.75, p = 0.625; CFI = 1.00, TLI = 1.10, RMSEA = 0.00, SRMR = 0.01. See Online Resource 1 for complete path statistics with standardized coefficients and bias-corrected (b-c) 95% confidence intervals. After controlling for pretreatment functioning and parent education, AA parents had lower parental self-competence compared to OM (B = 0.19, p = 0.030, b-c 95% CI [0.04, 0.63]) and White parents (B = 0.24, p = 0.005, b-c 95% CI [0.10, 0.58]). AA parents also had lower treatment expectations compared to OM (B = 0.23, p = 0.016, b-c 95% CI [0.05, 0.48]) and White parents (B = 0.21, p = 0.044, b-c 95% CI [0.02, 0.39]). ODD severity and parent education were negatively associated with parental self-competence and global impairment was positively associated with treatment expectations.

Aim 2. AA Identification, Parental Cognitions, Treatment Engagement, and Outcomes

Two path models estimated the effects of race/ethnicity and pretreatment parental cognitions on: (1) observer-rated engagement and (2) parent-rated engagement and posttreatment child outcomes (see Figs. 2 and 3 for estimated paths among the main study variables). Although observer- and parent-rated engagement were strongly correlated (r = 0.70, p < 0.001), we conducted separate analyses as planned given the study question and their differential bivariate relations to race/ethnicity (i.e., differences were found for observer- and not parent-rated engagement). Paths were estimated from: (a) race/ethnicity and pretreatment parental cognitions to engagement during groups, and (b) engagement to child functioning at posttreatment. The direct and indirect paths from parental cognitions to outcomes were limited to those correlated: (c) parental self-competence to ADHD severity, and (d) treatment expectations to global psychosocial improvement. Both models controlled for pretreatment ADHD severity (to assess pre-post improvement), ODD severity, treatment condition (CLAS vs. PFT), and child sex (female vs. male). The model for parent-rated engagement also controlled for pretreatment global psychosocial impairment because the two variables were correlated (r = 0.24, p = 0.003).

Fig. 2.

Path analyses from pretreatment variables to observer-rated treatment engagement to posttreatment outcomes

Fig. 3.

Path analyses from pretreatment variables to parent-rated treatment engagement to posttreatment outcomes

The model fit indices showed good fits for both models (observer-rated engagement: χ2 = 8.27, p = 0.408; CFI = 1.00, TLI = 0.99, RMSEA = 0.02, SRMR = 0.03; parent-rated engagement: χ2 = 7.82, p = 0.646; CFI = 1.00, TLI = 1.06, RMSEA = 0.00, SRMR = 0.03). See Online Resources 2 and 3 for all path statistics with b-c bootstrapped 95% confidence intervals. Similar significant effects were found in both models. At posttreatment for both models, AA parents reported less improvement in global psychosocial improvement than OM parents and less improvement in ADHD severity than White parents. Race/ethnicity comparisons were not significant for observer- or parent-rated engagement after accounting for parental cognitions and covariates. Treatment expectations positively predicted both observer-rated (B = 0.23, p = 0.003, b-c 95% CI [0.26, 1.17]) and parent-rated engagement (B = 0.22, p = 0.005, b-c 95% CI [0.21, 1.05]). Treatment engagement in turn positively predicted posttreatment global psychosocial improvement, but not posttreatment ADHD in both models. The indirect effect of treatment expectations treatment engagement posttreatment global improvement was significant for both observer- and parent-rated engagement (respective Bs = 0.09, ps < 0.05, b-c 95% CIs [0.03, 0.19] and [0.03, 0.20]). The significant total effects of treatment expectations -> treatment engagement -> posttreatment global improvement for both models (respectively for observer and parent-ratings, Bs = 0.33 and 0.32, ps < 0.05, b-c 95% CIs [0.05, 0.59] and [0.04, 0.58]) and nonsignificant direct effects of treatment expectations on global improvement indicated a full mediation effect. The direct effects of parental self-competence on treatment engagement (respectively for observer and parent-ratings, Bs = −0.03 and 0.01, ps > 0.05, b-c 95% CIs [−0.47, 0.28] and [−0.36, 0.33]) and on posttreatment ADHD (Bs = −0.01, ps > 0.05, b-c 95% CIs [−0.15, 0.12] and [−0.14, 0.13]) were not significant.

Discussion

This study utilized the HBM framework (Rosenstock et al., 1988) to examine the differential treatment mechanisms of functional improvement for children with ADHD between AA and OM/White parents. In support of our predictions, we found that AA parents endorsed lower parental self-competence and treatment expectations at pretreatment and reported attenuated improvement in ADHD symptoms and global clinical functioning posttreatment.

Our findings support the HBM (Rosenstock et al., 1988) as a framework for understanding mental health service utilization for AAs (Kim & Zane, 2016) and parents’ engagement with BT (Mah & Johnston, 2008). We found that treatment expectations was a significant predictor of treatment engagement for AA, OM, and White parents. In line with previous research (Nock & Ferriter, 2005), we also found that for all parents, higher engagement in turn predicted greater global psychosocial improvement in their children. Our study also adds to the literature on the importance of parents’ engagement for BT outcomes by demonstrating that treatment engagement mediated the relationship between treatment expectations and global psychosocial improvement. Our findings thus highlight the notable downstream effects of treatment expectations for AA parents, who began treatment with the lowest expectations.

While parental self-competence was not associated with treatment engagement or outcomes, our bivariate analyses showed a strong relationship between parental self-competence and treatment expectations. Mah and Johnston (2008) have posited that parents with low parental self-competence may perceive BT as too demanding. As such, parental self-competence may precede and influence treatment expectations. Given our finding that both parental self-competence and treatment expectations were associated with race/ethnicity, future research should assess parental self-competence as a mechanism by which AA parents develop lower treatment expectations. Our findings, nonetheless, are in line with previous work suggesting that AA parents of children with ADHD feel less efficacious and satisfied with parenting, perhaps due to cultural beliefs regarding parents’ responsibility and shame for their children’s behavior problems and academic failures (Lam & Ho, 2010). Our results reflect Kim and Zane’s (2016) finding that AA participants had lower treatment expectations than their White counterparts, indicating hesitance among AAs to self-disclose and seek help when facing psychosocial stressors. Despite having lower expectations, AAs endorsed similar levels of engagement. This result is in line with Nock and Kazdin’s (2001) work that showed that parents with the highest and lowest expectations had the highest participation in sessions, perhaps due to increases in expectations – and thus engagement – as a result of any improvement observed in their children.

We found that AA parents endorsed lower global psychosocial improvement than OM parents and lower ADHD symptom reduction than White parents. These findings – and those described above – should be interpreted within the context of the relatively high acculturation levels of the sample (given the English requirements for participation) as well as high treatment seeking behaviors (as evidenced by the TAU condition participants). We expect these relations to have larger effects among AA parents who are less acculturated to U.S. culture given the barriers faced to enter treatment. While many of these barriers (e.g., linguistic/cultural mismatch and help-seeking stigma; Hall & Huang, 2020) are shared with other racial/ethnic groups, our study found attenuated improvement reported by AA parents and lower treatment expectations when compared to OM parents, indicating the influence of barriers unique to the AA community. If certain skills do not fit cultural socialization practices, AA parents may not perceive the same benefits from BT as non-AA parents (Lau et al., 2011). Perhaps the AA parents in our study were slower to warm up to treatment or experienced a cultural mismatch, potentially made more salient when working with White therapists (Smith et al., 2011), which detracted from the full potential of the treatment. Future research should examine how cultural factors such as acculturation, heritage values, and immigration status impact AA treatment experience in BT as well as fluctuations in engagement across sessions. Engagement may be lower in sessions that teaches skills observed to be incongruent with the parent–child hierarchy and value of instilling humility (e.g., praise, rewards; Lau et al., 2010, 2011). Such data would inform how BT components could be adapted to align with AA cultural values and bolster outcomes.

These findings suggest that the efficacy of BT for AA families may be enhanced with the integration of several adaptations. Because treatment expectations underlie engagement, parents’ expectations can be assessed and addressed prior to treatment (McCabe et al., 2020). Clinicians can conduct an enhanced intake procedure to discuss and clarify misconceptions of BT (Chacko et al., 2012) and provide handouts with information about the effectiveness of BT, including results from clinical trials and potential improvements in child compliance (McCabe et al., 2020). Discussions of cultural barriers for using each skill can also be incorporated to assess for cultural congruence (Lau et al., 2011), and if present, content adaptations (e.g., language translations, culturally matched examples) and cognitive restructuring of thoughts inhibiting skill use (e.g., Wong et al., 2018) can be employed. Clinicians can also consider teaching culturally incongruent skills in later sessions after some child improvements have been observed and collaborate with parents on overcoming persisting barriers (Nock & Kazdin, 2001).

It is crucial, however, to bear in mind the structural barriers impacting service use when working with AA families. Given that this study was conducted in healthcare clinics, one interpretation of these findings is that AAs’ lower treatment expectations are an accurate reflection of the confluence of structural problems impacting their perceptions of treatment (e.g., MMS, racism, low access due to high costs, lack of culturally-matched therapists/treatments; Kim & Lee, 2022). Future research should measure the direct influence of structural barriers on parental cognitions and examine whether improving accessibility of BT in community settings (e.g., schools) bolsters treatment expectations and engagement. At minimum, any modifications addressing parental cognitions should be sensitively employed to avoid deepening stigma (e.g., not framing parents’ cognitions as deficient or placing blame on families’ cultural beliefs).

Several other important directions for future research are highlighted by key limitations of this study. First, race/ethnicities were used as a proxy for cultural identities. Future research is needed to examine the within-group heterogeneity of cultural and racial/ethnic identities (e.g., monoracial vs. multiracial) among AA (e.g., East vs. South Asians), Black, Latine/x, AI/AN, and White families. In addition, we were unable to address unique factors associated with NHPI given the lack of representation in our sample. Future research should be paid to the mono- and multiracial identities among other factors related to the NHPI community (e.g., NHPI only vs. AA/NHPI, Indigeneity, U.S. settler colonialism) through intentional recruitment (Sasa & Yellow Horse, 2022) and how these factors impact their perceptions of ADHD and ADHD treatment as well as mental health treatment broadly. Second, our AA sample was limited in size, ADHD-I presentation, and caregiver gender (mostly mothers). Only 14.6% of our sample were AA compared to the local population of 33.7% (Bay Area Census, 2010). While this reflects the low service use of AAs (Chung et al., 2019), future research with a larger sample of AAs is warranted to: (a) replicate the results of this study, (b) examine how the strength of the effects differ for AA/OM/White parents, and (c) examine how the effects generalize to children with ADHD hyperactive/impulsive and combined presentations and other caregivers. Fourth, while our sample was not affluent given the cost of living in the local area, it was highly educated. Future research should recruit samples with more varied parental education and its influence on parental cognitions and treatment. Nonetheless, the present study marks an important step towards increasing engagement in and improving BT treatment outcomes for AA parents of children with ADHD. The study identifies areas in which parental cognitions may be a target of adaptations to fit the AA community, shedding light on immediate initial steps towards the goal of achieving health equity in ADHD services across communities.