Improving Maternal Mental Health After a Child’s Diagnosis of Autism Spectrum Disorder: Results From a Randomized Clinical Trial

Importance  The prevalence of psychological distress among mothers of children with autism spectrum disorder (ASD) suggests a need for interventions that address parental mental health during the critical period after the child’s autism diagnosis when parents are learning to navigate the complex system of autism services.

Objective  To investigate whether a brief cognitive behavioral intervention, problem-solving education (PSE), decreases parenting stress and maternal depressive symptoms during the period immediately following a child’s diagnosis of ASD.

Design, Setting, and Participants  A randomized clinical trial compared 6 sessions of PSE with usual care. Settings included an autism clinic and 6 community-based early intervention programs that primarily serve low-income families. Participants were mothers of 122 young children (mean age, 34 months) who recently received a diagnosis of ASD. Among mothers assessed for eligibility, 17.0% declined participation. We report outcomes after 3 months of follow-up (immediate postdiagnosis period).

Interventions  Problem-solving education is a brief, cognitive intervention delivered in six 30-minute individualized sessions by existing staff (early intervention programs) or research staff without formal mental health training (autism clinic).

Main Outcomes and Measures  Primary outcomes were parental stress and maternal depressive symptoms.

Results  Fifty-nine mothers were randomized to receive PSE and 63 to receive usual care. The follow-up rate was 91.0%. Most intervention mothers (78.0%) received the full PSE course. At the 3-month follow-up assessment, PSE mothers were significantly less likely than those serving as controls to have clinically significant parental stress (3.8% vs 29.3%; adjusted relative risk [aRR], 0.17; 95% CI, 0.04 to 0.65). For depressive symptoms, the risk reduction in clinically significant symptoms did not reach statistical significance (5.7% vs 22.4%; aRR, 0.33; 95% CI, 0.10 to 1.08); however, the reduction in mean depressive symptoms was statistically significant (Quick Inventory of Depressive Symptomatology score, 4.6 with PSE vs 6.9 with usual care; adjusted mean difference, −1.67; 95% CI, −3.17 to −0.18).

Conclusions and Relevance  The positive effects of PSE in reducing parenting stress and depressive symptoms during the critical postdiagnosis period, when parents are asked to navigate a complex service delivery system, suggest that it may have a place in clinical practice. Further work will monitor these families for a total of 9 months to determine the trajectory of outcomes.

Trial Registration  clinicaltrials.gov Identifier: NCT01021384

Mothers of children with an autism spectrum disorder (ASD) consistently report high levels of psychological distress, including parental stress, depressive symptoms, and social isolation.^1-5 Almost 40% of such mothers report levels of clinically significant parenting stress,³ and between 33% and 59% report depressive symptoms warranting psychiatric evaluation.^1,2 The decreased energy, concentration, and lack of motivation associated with depressive symptoms adversely affect maternal functioning as well as a mother’s capacity to manage the health needs of her children.^6-9 The most effective interventions for young children with ASD are highly demanding, time intensive, and actively involve parents as “adjunct therapists.”^10,11 The prevalence of psychological distress among mothers of children with ASD suggests a need for interventions that specifically address parental mental health during this critical period.

To date, most intervention approaches for young children with ASD focus on training parents in behavioral techniques and methods to facilitate joint attention and communication skills.^12-17 These strategies, however, fail to deliberately address the types of psychological distress most often reported by parents.¹² Therefore, we conducted a randomized clinical trial of problem-solving education (PSE), a brief cognitive behavioral intervention, targeted to mothers of young children who recently received a diagnosis of ASD. We hypothesized that strengthening maternal problem-solving skills would serve as a buffer against the negative impact of life stressors and thereby reduce parental stress and attenuate depressive symptoms. This analysis reports outcomes after 3 months of follow-up in a newly diagnosed, largely low-income sample. We plan to monitor these families for 9 months but elected to present 3-month outcomes independently because they measure the intervention effect during the critical period following the child’s autism diagnosis. Because parents are learning to navigate the complex world of autism services during the immediate postdiagnosis period, intervention effects during this time have implications for clinical practice even if such findings are not sustained.

Between November 12, 2009, and August 16, 2012, we enrolled 122 English- or Spanish-speaking mothers with a child younger than 6 years who had recently received a diagnosis of an ASD. We aimed to enroll 140 participants, which was the sample size needed to detect a 25% change in the proportion of mothers reporting clinically significant parenting stress and a clinically relevant 3-point change in depressive symptoms. The ASD diagnosis was made by a qualified licensed health care provider, most often (92%) using the Autism Diagnostic Observation Schedule.¹⁸ Participants were recruited from an autism clinic that serves ethnically diverse, primarily low-income urban families and from 6 Boston area early intervention (EI) programs for children from birth to 3 years that provide services authorized under Part C of the Individuals with Disability Act.

Families of children with newly diagnosed ASD were introduced to the study by clinicians from the autism clinic or EI providers who had existing relationships with the families. Study personnel then contacted interested families by telephone to screen for eligibility and arrange face-to-face meetings to explain the study and obtain informed consent. Randomization procedures and requirement of participation in the treatment and control groups were explained during this face-to-face meeting. Participants’ understanding of study conditions and capacity to give informed consent were assessed using the MacArthur Competence Assessment Tool for Clinical Research.¹⁹ A total of 165 mothers were referred to the study and screened for eligibility; 122 were assigned to a study arm (Figure). Mothers were randomly allocated to intervention (PSE) or usual care after completion of baseline measures with a 1:1 allocation within strata based on high and low depressive symptoms. This random allocation sequence was created by the study statistician (H.C.) using a computer-generated random number sequence. Randomization occurred independently at each study site. Within strata, randomization was conducted within blocks of 4. The allocation sequence was concealed from all study personnel in sequentially numbered opaque, sealed envelopes.

The intervention group received PSE, a manualized cognitive behavioral intervention adapted from problem-solving treatment, itself an evidence-based depression treatment.²⁰ The adapted intervention was designed such that it could be delivered by the range of existing multidisciplinary staff who work with families of young children with ASD. Details of the adaptation process have been described previously.^21,22 During each workbook-based PSE session, mothers work one-on-one with a trained interventionist to identify a single, measurable problem and then proceed through a series of steps that involve goal setting, brainstorming, and evaluating solutions, choosing a solution, and action planning. For example, a mother might describe feeling lonely. In the PSE session, she would work with the interventionist to change this subjective feeling into an objective problem, such as not being able to find someone to watch her child so that she could go out with friends. Working through the defined steps, the mother sets an achievable goal, generates solutions to meet it, and plans specific steps to fulfill her solution. In this example, the mother might decide that she would ask her sister to watch her child after he goes to bed to avoid his challenging behaviors and plans to call her sister the next day. A full PSE course involved 6 individual sessions (30-45 minutes). It was delivered in the home or other location of the mother’s choosing. In EI, PSE was fully integrated into the existing service delivery system. For children who were not receiving EI, PSE was delivered as a supplement to school-based services.

Usual care group mothers received the services specified in the child’s Individualized Family Service Plan or Individualized Educational Plan. Typically, such plans include a package of services, including speech and language therapy, occupational therapy, and social skills training. Neither specifically includes parent-focused mental health services.

Ten EI staff representing the range of disciplines within EI programs (child development, nursing, speech and language, and social work) served as educators for mothers recruited from EI programs; 4 graduate students, pursuing degrees in social work and public health, served as educators for mothers who were identified from the autism clinic. Educator training entailed 1 day of didactics, followed by 3 to 5 practice sessions conducted under the supervision of investigators. All training sessions were audiotaped, and all educators received individualized feedback until they met predefined criteria for protocol adherence. Educators met in biweekly group sessions at each site for supervision with investigators (E.F., Z.F.C.S.). To assess intervention fidelity, we audiotaped one randomly selected PSE session from each intervention participant and used a standardized form developed in previous work^21,22 to assess whether all 7 components of the intervention were present and delivered according to protocol.

The trial was approved by the Boston University Medical Center institutional review board and has been registered on clinicaltrials.gov (NCT01021384). Formal protocols (available on request) to support women with severe depressive symptoms or suicidal ideation were developed and reviewed by the Boston University Medical Center institutional review board.

Parenting stress was assessed using the Parenting Stress Index Short Form (PSI).^23,24 A substantial body of published research links PSI scores to observed parent and child behaviors, specifically attachment style, social skills, and confidence. It is one of the only valid and reliable instruments that can assess a wide range of parenting behaviors in a single instrument, including the participant’s attachment to the child, social isolation, sense of competence in the parenting role, relationship with spouse/parenting partner, role restrictions, and parental mental health.²⁵ It has excellent psychometric properties and has been used with racially and ethnically diverse populations, including Spanish-speaking families, and families of children with ASD.^26-30 We analyzed parental stress as both a mean symptom score and relative to a prespecified clinical threshold, defined as a percentile score above the 90th percentile. A percentile score in this range represents clinically significant parenting stress.²⁵

Depressive symptoms were measured using the Quick Inventory of Depressive Symptomatology (QIDS).³¹ Because the QIDS has a wide scoring range, it can be used to detect depressive illness in populations with moderate- and low-level symptoms and is sensitive to change over time. It has excellent correlation with other depression scales,^32,33 has been validated with community samples, and demonstrated reliability in culturally diverse postpartum populations³⁴ and those with minor depression.³⁵ Similar to analysis of parenting stress, we analyzed QIDS depression scores as a mean symptom score and relative to a prespecified clinical threshold (≥11), corresponding to moderately severe symptoms occurring during a 1-week recall period. Based on previous work³⁴ among urban postpartum women, a QIDS score of 11 or more has greater than 88% specificity for predicting a major depressive episode.

Coping style was assessed using the Brief Coping Orientation to Problems, which measures 14 different adaptive and problematic coping reactions.³⁶ Based on previous research,^37,38 we looked at 3 subscales that were hypothesized to mediate the relationship between the intervention and parenting stress: problem-focused coping, avoidant coping, and social coping.

Child functioning was measured using the Parent/Primary Caregiver Form of the Communication Scale of the Adaptive Behavior Assessment System (ABAS II).³⁹ The ABAS II is one of the few caregiver-report measures validated for use in young children. For children from birth to 5 years, the ABAS II was normed based on 750 teacher/child care provider ratings and 1350 parent ratings of children with both typical and atypical development. Scaled scores are standardized by child age. Internal consistency, test-retest reliability, and convergent and divergent validity are all assessed to be in a good to very good range.⁴⁰

Data were collected through face-to-face interviews at baseline and 3 months after the first intervention session (PSE group) or 3 months after the baseline assessment (usual care). Assessments were conducted in English and Spanish. Baseline assessments were conducted by research staff before randomization. All outcome assessments were administered by research staff who were not involved in the intervention and who, along with investigators, were blinded to study allocation. Participants received a $25 gift card for completing the baseline and 3-month assessment. Mothers in the PSE group received a $100 gift card at completion of the full 6 sessions of the intervention. Data regarding the type of problems that mothers addressed during PSE sessions were obtained from individual session worksheets.

We evaluated the psychometric properties of key study measures using Cronbach α and assessed outcomes by intention-to-treat. For each outcome (parenting stress and depressive symptoms), we compared differences in mean levels of reported symptoms and the proportion of mothers in each group who experienced clinically significant symptoms. For coping style, we compared differences in mean scores on the 3 identified subscales. We estimated mean differences from linear regression models and relative risks from log binomial models. Models assessing parenting stress, depressive symptoms, and coping style were adjusted for the respective baseline measure because a strong correlation with the outcome was expected.⁴¹ All statistical analyses were performed using commercial software (SAS, version 9.3; SAS Institute).

Data regarding the problems that mothers addressed during PSE sessions were grouped a priori using existing established categories described in the copyrighted Problem Solving Therapy Manual.⁴² Problems were then recoded into categories relevant to the study’s population of mothers of children with ASD. For example, problems with relationships were divided into child-related issues and family and community relationships. All problems were coded independently by 2 research assistants. If consensus was not reached, another reviewer (E.F.) participated in the coding decision such that the group could render a final decision acceptable to all 3 reviewers.

Of the 122 randomized families, 111 mother-child pairs (91.0%; 53 PSE and 58 usual care) completed the 3-month assessment. Baseline maternal and child characteristics reflected the demographics of the participating sites (Table 1). All measures, with the exception of the avoidant subscale of the Brief Coping Orientation to Problems, performed acceptably in the study population, with Cronbach α values ranging from .75 to .92.⁴³

Fifty-five of 59 intervention participants (93.2%) received at least 1 session of PSE; 46 of the participants (78.0%) received the full 6-session course during a mean of 2 months. Of 328 PSE sessions that were delivered, mothers used the sessions to address child-related issues (38%), maternal personal issues (21%), and time and household management (15%) (Table 2). Almost all mothers (90%) reported that they continued to use the skills they learned through PSE to address problems almost weekly.

The study’s crisis management protocol was used for 5 mothers who reported suicidal thoughts and 11 mothers with severe depressive symptoms without suicidality (QIDS score, ≥16). All 16 of these individuals were assessed per study protocols and referred for more definitive mental health services if they were not already connected with a mental health provider. These participants remained in the study; there were no adverse advents. Of the 328 PSE sessions, 34 were randomly selected for an audiotaped fidelity assessment. All were judged to be meet fidelity standards based on a score that was higher than 90% on a standardized form.

Adjusting for baseline parenting stress, there was a significant difference by treatment arm in the risk of experiencing clinically significant parenting stress at the 3-month assessment. Mothers in the PSE group were significantly less likely to report clinically significant parenting stress compared with mothers receiving usual care (4% vs 29%; adjusted relative risk [aRR], 0.17; 95% CI, 0.04 to 0.65). The mean difference in parenting stress scores was significant in unadjusted models but did not remain so when adjusting for baseline parenting stress (PSI score: 85.5 with PSE vs 94.2 with usual care; adjusted mean difference, −5.61; 95% CI −12.79 to 1.56) (Table 3).

With adjustment for baseline depressive symptoms, the difference by treatment arm in the risk of experiencing clinically significant depressive symptoms approached statistical significance. Mothers receiving PSE were 67% less likely to report clinically significant depressive symptoms than were mothers receiving usual care (6% vs 22%; aRR, 0.33; 95% CI, 0.10 to 1.08). We found similar but statistically significant results favoring the intervention for reduction in mean depressive symptoms (QIDS score: 4.6 with PSE vs 6.9 with usual care; adjusted mean difference, −1.67; 95% CI, −3.17 to −0.18).

To assess potential effect modification, we examined the frequency of clinically significant depressive symptoms and parenting stress in analyses stratified by race/ethnicity, primary language, income, and marital status. Sparse data in multiple cells precluded conducting valid analyses.

Compared with mothers receiving usual care, at the 3-month assessment, PSE mothers had significantly higher use of social coping, including use of emotional support and instrumental support, which is defined as concrete, direct ways that people assist others in accessing services and other concrete resources (adjusted mean difference, 1.43; 95% CI, 0.40-2.46) (Table 4). We found no significant difference in the use of problem-focused or avoidant coping.

Our findings demonstrate evidence of PSE’s short-term efficacy to decrease parental stress and depressive symptoms in the months immediately following a child’s ASD diagnosis. Three-month outcomes suggest the intervention’s potential to reduce clinically significant psychological distress during this critical juncture—when parents are asked to navigate a complex service system on their child’s behalf.

Theoretically, the relational/problem-solving model of stress posits that problem-solving skills buffer the impact of daily-life stressors and improve coping, which in turn prevents episodes of clinically significant depressive symptoms.⁴⁴ In our study, receiving PSE was associated with an increase in social coping abilities and a decrease in clinically significant parenting stress. Although mean depressive symptoms decreased significantly—the risk reduction of clinically significant depressive symptoms—the more relevant outcome with respect to maternal functioning, was large but did not reach statistical significance. Mothers used their PSE sessions to address problems that were either directly related to their child or were made more difficult by having a child with ASD. Differences between PSE and usual care mothers in the use of instrumental and emotional support to address such problems may have contributed to their decreased risk for clinically significant parenting stress and depressive symptoms.

The American Academy of Pediatrics⁴⁵ recommends that all children be screened for ASD at age 18 or 24 months and that children with ASD diagnoses receive early and intensive treatment. Because very young children spend more time with their parents and less time in school, parental participation in the child’s treatment plan is essential. Recent rigorously evaluated interventions, such as the Early Start Denver Model,¹⁴ Preschool Autism Communication Trial,¹⁵ Parent Education and Behavior Management,¹³ and the study of a joint attention intervention by Kasari et al,¹⁶ all include parent-mediated interventions as either the primary component or a major component of intervention delivery. Researchers recognize that the quality and consistency of such participation, apart from treatment fidelity, affect child treatment outcomes. In the present study, we focused on assessing the feasibility and efficacy of an intervention whose primary goal is to support maternal mental health, based on the underlying premise that parental stress and depressive symptoms, which are highly prevalent among mothers of children with ASD, are associated generally with poor adherence to recommended health practices^6,7,46 and specifically with decreased effectiveness of autism teaching interventions.⁴⁷

The study findings described herein must be viewed cautiously. This small study is, to our knowledge, the first efficacy trial of this intervention. Despite a randomized design and participation data that minimize the potential for bias, findings should be replicated. Future analyses will examine the effect of intervention over a longer follow-up period and allow us to assess whether the intervention worked differently among subgroups of mothers, which is knowledge that could help us better target those most likely to benefit from the intervention. Despite the limited scope of the present analysis, the findings are encouraging. Evidence of successful intervention delivery and positive intervention effects support our ongoing efforts to more definitively test the efficacy of this brief, easily replicated intervention.

Corresponding Author: Emily Feinberg, CPNP, ScD, Department of Community Health Sciences, Boston University School of Public Health, Crosstown Center, Room 440, 801 Massachusetts Ave, Boston, Massachusetts 02118 (emfeinbe@bu.edu).

Accepted for Publication: July 7, 2013.

Published Online: November 11, 2013. doi:10.1001/jamapediatrics.2013.3445.

Author Contributions: Dr Feinberg had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Feinberg, Augustyn, Cabral, Beardslee, Silverstein.

Acquisition of data: Feinberg, Fitzgerald, Sandler, Ferreira-Cesar Suarez.

Analysis and interpretation of data: Feinberg, Augustyn, Chen, Cabral, Beardslee, Silverstein.

Drafting of the manuscript: Feinberg, Augustyn, Beardslee, Silverstein.

Critical revision of the manuscript for important intellectual content: Feinberg, Fitzgerald, Sandler, Ferreira-Cesar Suarez, Chen, Cabral, Beardslee, Silverstein.

Statistical analysis: Feinberg, Chen, Cabral, Silverstein.

Obtained funding: Feinberg.

Administrative, technical, and material support: Feinberg, Fitzgerald, Sandler, Ferreira-Cesar Suarez, Beardslee, Silverstein.

Study supervision: Feinberg, Augustyn, Fitzgerald, Ferreira-Cesar Suarez, Cabral, Beardslee.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grant R40MC15596 from the Maternal and Child Health Bureau (Dr Feinberg).

Role of the Sponsor: The Maternal and Child Health Bureau had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We acknowledge the contributions of the mothers who took part in this study and the support and effort of staff at the participating early intervention programs and Boston Medical Center’s Developmental and Behavioral Pediatric Division.