A community-based sleep educational intervention for children with autism spectrum disorder

Abstract

Background:

Sleep problems are common in children with autism spectrum disorder (ASD). Sleep education, effective in improving sleep in ASD, may be difficult to access. We determined if community-based pediatric therapists could successfully deliver sleep educational interventions to caregivers of children with ASD.

Methods:

A seven-week feasibility study was conducted consisting of 10 children and caregivers. This feasibility study informed the development of a 16-week preliminary effectiveness study, which consisted of 33 children and caregivers. Children, ages 2–12 years, with a clinical diagnosis of autism and caregiver-reported sleep onset delay of 30 min were included. Community therapists underwent comprehensive training in sleep education and then met with caregiver participants to provide sleep education to each family. Semi-structured qualitative interviews were conducted with all families who completed study procedures.

In the feasibility and preliminary effectiveness studies, child participants wore an actigraphy watch (at baseline and after sleep education) and caregivers completed the Child Sleep Habits Questionnaire and Family Inventory of Sleep Habits at baseline and after sleep education; the Child Behavior Checklist was also completed by caregivers in the preliminary effectiveness study.

Results:

Educator fidelity to the manualized curriculum was maintained. Caregivers showed appropriate understanding, comfort, and implementation of the curriculum. Qualitative and quantitative measures, including caregiver surveys and actigraphy, showed improvements in child sleep and behavior.

Conclusions:

Community-based therapists can successfully deliver sleep education to families of children with ASD, which has favorable implications for improving access to care in this population.

1. Introduction

Sleep problems, especially insomnia, are common in children with autism spectrum disorder (ASD) (Goldman, Richdale, Clemons, & Malow, 2012; Reynolds et al., 2019; Richdale & Schreck, 2009) and are associated with challenging daytime behaviors (Mazurek & Sohl, 2016; Veatch et al., 2017) and decreased health-related quality of life (Delahaye et al., 2014) and family health (Hodge, Hoffman, Sweeney, & Riggs, 2013). Parents of children with ASD report greater sleep problems (Lopez-Wagner, Hoffman, Sweeney, Hodge, & Gilliam, 2008), poorer sleep quality, and shorter sleep duration, as measured by actigraphy, than parents of typically developing children (Meltzer, 2008).

Sleep problems in ASD have multifactorial causes that include medical and psychiatric conditions, medications, and factors intrinsic to ASD (Maxwell-Horn & Malow, 2017). Behavioral and environmental causes also contribute to insomnia in ASD. Some of these causes are related to core features of ASD (e.g., difficulty with transitions from stimulating to calming activities when preparing for bed), sensory sensitivities (e.g., to noises, sleepwear) and restrictive interests/repetitive behavior (e.g., reciting dialogue from favorite videos at bedtime). Others causes encompass daytime and evening habits that impede sleep in any child, including screen time close to bedtime, limited physical exercise, or caffeine use.

Pharmacological treatment of insomnia in ASD includes the use of melatonin, alpha-agonists, gabapentin, anti-depressants, and antianxiety medications (Malow, Katz et al., 2016). These regimens are not always successful and are frequently associated with reports of adverse side effects (Reynolds & Malow, 2011).

Evidence supporting behavioral interventions for insomnia in young children of typical development is strong. Behavioral strategies have included unmodified extinction, graduated extinction and parent education/prevention (Meltzer & Mindell, 2014). Studies with younger typically developing children have shown that behavioral approaches are effective in decreasing sleep onset latency and night waking frequency and duration. This literature informed our study intervention. However, much less work has been done in this area for school-aged children, the adolescent population, or those with special needs such as children with ASD, mood disorders, and other chronic illnesses (Meltzer & Mindell, 2014).

Clinical guidelines advocate for behavioral sleep interventions in ASD as an initial approach to treatment (Buckley et al., 2020; Malow et al., 2012) while recognizing the need for more evidence-based research. Our previous work has focused on providing sleep education to parents of children and adolescents with ASD in the form of informational pamphlets and manuals to individual and group educational sessions (Adkins et al., 2012; Malow, MacDonald, Fawkes, Alder, & Katz, 2014, 2016). At our institution, psychologists and nurses have been trained to conduct sleep education with parents of children with ASD ages 2–10 years. Improvements in objective measures of sleep (actigraphy) and parent reports of sleep and daytime behavior were noted (Malow et al., 2014). These findings were replicated in an older age range that included 11 to 18-year olds (Loring, Johnston, Gray, Goldman, & Malow, 2016). Other studies have also shown behavioural approaches to be efficacious in improving sleep in children with ASD (McLay et al., 2020; Souders et al., 2017). These interventions, often provided to families at academic medical centers, are usually small scale with long waiting lists. Families who may benefit from such programs may not live near medical centers or may have difficulty managing the time and expense required to travel to and from a medical center campus.

The goal of these projects was to determine whether a community-based sleep education program, providing sleep education to families closer to home, was feasible and would show preliminary efficacy. To accomplish this goal, we initially carried out a seven-week feasibility study that informed the development of a subsequent 16-week preliminary effectiveness study. We collected both quantitative survey data and sleep actigraphy along with qualitative data through parent interviews to gain a further understanding of family dynamics and parent success with previous behavior change in the context of our intervention.

2. Methods

We conducted an initial seven-week feasibility study followed by a 16-week preliminary effectiveness study, both with a pre/post design and approved by the Vanderbilt Institutional Review Board (IRB). The two studies were similar in design except for participating practices, number of participants, and study length. Therefore, we describe procedures for the feasibility study in detail and then highlight the modifications that were made for the preliminary effectiveness study (see Tables 1a, 1b). All research staff completed good clinical practice training and were trained in consent procedures. At the time of consent, all participants’ caregivers were provided with a verbal and written explanation of the study and signed the IRB-approved consent form prior to their child’s participation in research activities. All participating children provided verbal assent and signed a consent form when able. This research was ethically conducted, and research staff adhered to the standards and principles of the revised Declaration of Helsinki regarding human experimentation.

Table 1a.

Procedures and Assessments for Study #1.

Study Phase	Baseline Phases Weeks 1–3	Week 4	Week 5	Week 6	Week 7
Procedures & Assessments
Consent/Assent	●
Record Review	●
Parent Interview	●
Parent education in actigraphy collection procedure	●
Actigraphy	●	●	●	●	●
Parent Questionnaires: Demographics, Hollingshead, NVS	●
Parent Questionnaires: CSHQ and FISH	●				●
Initial Sleep Education Session		●
Sleep Education Booster			●	●
Assessment of Parent Understanding, Comfort, and Implementation (Parent Absorption Scale)			●	●
Parent and Educator Qualitative Interviews					●

NVS = Newest Vital Sign; CSHQ = Children’s Sleep Habits Questionnaire; FISH = Family Inventory of Sleep Habits.

Table 1b.

Procedures and Assessments for Study #.

Study Phase	Baseline Phases Weeks 1–3	Week 4	Week 5	Week 6	Week 7	Week 10	Week 16
Procedures & Assessments
Consent/Assent	●
Record Review	●
Parent Interview	●
Parent education in actigraphy collection procedure	●
Actigraphy	●	●	●	●	●		●
Parent Questionnaires: Demographics, Hollingshead, NVS	●
Parent Questionnaires: CSHQ, FISH	●				●		●
Parent Questionnaires: CBCL							●
Initial Sleep Education Session		●
Sleep Education Booster			●	●		●
Assessment of Parent Understanding, Comfort, and Implementation (Parent Absorption Scale)			●	●		●
Parent and Educator Qualitative Interviews							●

NVS = Newest Vital Sign; CSHQ = Children’s Sleep Habits Questionnaire; FISH = Family Inventory of Sleep Habits; CBCL = Child Behavior Checklist.

2.1. Therapist training in sleep education

For both studies, the therapists first engaged in self-study, including reviewing the manualized curriculum (from our previously published sleep education program) and watching recorded mock sessions of a trained sleep educator and another staff member acting in a caregiver role. Investigators then met with the therapists individually to review their self-learning and answer questions. The physician sleep specialist (co-author BAM), focused on medical aspects of sleep related to the training program. For example, if a caregiver brought up medical concerns, such as snoring, therapists were trained to recommend to the caregiver that they reach out to the child’s physician or nurse practitioner. The psychologist/ behavioral sleep specialist (co-author WL) focused on the behavioral aspects of sleep and how to carry out the manualized curriculum.

Therapists then engaged in recorded practice sessions where the physician sleep specialist played the role of caregiver of a child with ASD and the therapist delivered the sleep education intervention. The psychologist/behavioral sleep specialist observed and rated each practice session and provided oral feedback immediately following each session and also written feedback which summarized this discussion. Therapists were given ratings of 0, 1, or 2 on each of the session elements (see Table 2) which was then calculated into an overall fidelity percentage score. There were three prearranged scenarios for the practice sessions, which became progressively more difficult in terms of both presenting sleep concern (e.g., night wakings) and caregiver style (e.g., a “chatty” caregiver, a “withdrawn” caregiver). Therapists were required to achieve an average of 80% fidelity on the practice sessions prior to delivering the education in the research study. If therapists did not meet the criteria of 80% fidelity across three practice sessions, they continued to conduct additional sessions until the fidelity criteria were met. All live education sessions with parent participants were video recorded for fidelity review by the psychologist. Therapists were also required to maintain an average of 80% fidelity on the first three live parent sessions. Based on our experience in the feasibility study, ongoing training, individualized to therapist needs, was provided throughout the preliminary effectiveness study via phone or conference calls, email, or in-person meetings.

Table 2.

Fidelity Checklist.

Session Integrity Goals	Rating
Introductions	0 1 2
Thanks parent for coming and for completing necessary paperwork	0 1 2
States, in general terms, what will happen during the session	0 1 2
Asks parent to state their primary sleep concern and what they hope to achieve from the program	0 1 2
Restates parent concerns	0 1 2
Introduces the Homework Sheet to parent and states this will be referred to throughout education session.	0 1 2
Targets specific areas related to components of successful sleep (daytime and evening habits) based on parent CSHQ and FISH	0 1 2
Reviews slide entitled, Bedtime Routines	0 1 2
Reviews Bedtime Routines Worksheet	0 1 2
Shows video CD	0 1 2
Completes Sleep Recommendations Form while parents view CD	0 1 2
Parent Assembles visual schedule for bedtime routine with support from facilitator	0 1 2
	0 1 2
Shows parent where to list bedtime routine on Homework sheet folder	0 1 2
Shows slides/strategies to address sleep resistance, night wakings, and early morning wakings (as applicable to child). Explains bedtime pass.	0 1 2
Shows parent where to enter bedtime, sleep resistance, minutes to fall asleep and night wakings on Homework sheet folder	0 1 2
Wrap-up
Restates parent’s major sleep concern	0 1 2
Summarizes key points listed on the Sleep Recommendations form and allows parent to edit if needed.	0 1 2
Educator asks if parent feels that the overall intervention plan is feasible for family and not too ambitious	0 1 2
Provides & reviews parent folder (Homework data sheet, Calendar pages for follow up visits).	0 1 2
General Adherence to Manual
Educator stays within allotted timeframe Total minutes=	0 1 2
Educator follows material (i.e., does not go outside information in manual)	0 1 2
Educator addresses child’s sleep difficulties as family problem	0 1 2
Educator tailors intervention to child’s needs	0 1 2
	N/A = Not Applicable
Personal Characteristics of Educator		Not Observed	Rarely	Sometimes	Usually	Always
Educator is sympathetic to family’s experience	N/A	1	2	3	4	5
Educator makes eye contact and is engaged with parent	N/A	1	2	3	4	5
Educator is non-judgmental	N/A	1	2	3	4	5
Educator provides positive reinforcement to parent
• Non-verbal cues	N/A	1	2	3	4	5
• Positive statements	N/A	1	2	3	4	5
• Consistently supportive throughout session	N/A	1	2	3	4	5
Educator’s Interaction With Parent
Educator’s interaction with parent is discussion-based rather than didactic	N/A	1	2	3	4	5
Educator consistently checks in throughout session to ensure that the parent is feeling comfortable with the material	N/A	1	2	3	4	5
Educator redirects parent back to sleep education as needed	N/A	1	2	3	4	5
Educator compromises when parent does not feel able to implement certain aspects of intervention	0 N/A	1	2	3	4	5
Educator asks after each intervention if the plan is feasible for the family and not too ambitious	N/A	1	2	3	4	5

Subject ID:________ Date of Session ________.

Site ________Therapist ________.

The following scale should be used to rate the degree to which session goals were attained.

0 = Goal was not introduced or covered by the clinician.

1 = Goal was partially achieved.

2 = Goal was fully achieved.

2.2. Study #1. Feasibility

2.2.1. Participants and recruitment

Children with ASD were recruited for participation through a pediatric practice in middle Tennessee that serves a variety of patient populations with racial, ethnic, and socioeconomic diversity. Families were enrolled between November 2015 and June 2016. Physicians, therapists, and nurses pre-screened children with ASD for sleep problems and asked families about their potential interest in participation. Interested families gave permission to be contacted directly by research staff from Vanderbilt University Medical Center. Research staff then called caregivers to confirm interest in participation and to screen for eligibility. Eligibility criteria included 1) age between 2 and 12 years; 2) a confirmed clinical diagnosis or educational classification of ASD from a specialized professional (e.g., developmental pediatrician, psychologist, psychiatrist, neurologist, or school psychologist) consistent with either the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) (American Psychiatric Association, 2014) or 4th edition (DSM-IV-TR (American Psychiatric Association, 2000); 3) nightly sleep onset latency of 30 min or more (on average) by caregiver report and 4) no untreated/unstable medical conditions or new medication(s). A waiver of consent was received by the IRB to conduct a short medical interview (by a research nurse) with the caregiver to screen for any medical conditions that might need to be addressed prior to participation and to document current medications.

2.3. Study #1 design

The study was designed to be completed over a period of 7 weeks; however, flexibility to this timeline was allowed to account for schedule changes and unforeseen family circumstances (e.g., family illness, new job, change in living situation). All study procedures took place at Mercy Community Healthcare or by telephone. The study included a baseline data collection period (weeks 1–3) and the sleep education intervention period (weeks 4–7). The licensed therapist provided the sleep educational intervention to the caregivers of participants. Children wore an actigraphy watch for two weeks prior to the sleep education intervention and for four weeks after their caregivers received their initial sleep education session.

2.3.1. Consent and actigraphy education visit

This was a 90-minute visit. After providing informed consent, the caregivers were given thorough training on the use and care of the actigraphy watch followed by a short review quiz to ensure all pertinent information was understood, as outlined in our previous work (Fawkes et al., 2015). Caregivers also completed the Newest Vital Sign (Weiss et al., 2005) (see Measures), an assessment of language fluency. Thorough instructions were given on the use of Research Electronic Data Capture (REDCap), a secure online website used for data collection. Secure emails sent through REDCap allowed caregivers to complete questionnaires and study-related homework at their convenience using their mobile phones or home computers. Hard copies of all study forms and questionnaires were available if parents preferred using paper versions to phones or home computers.

2.3.2. Baseline data collection

Caregivers completed a demographic form and, at baseline and week 7, two child sleep questionnaires (see Measures: Caregiver-Completed). Caregivers were asked to complete the questionnaires and homework during a baseline 14-day period when their child was wearing an actigraphy watch. For each night the actigraphy watch was worn, the caregivers were asked to complete “daily homework” related to their child’s sleep times. This homework included a paper time log for marking when the event marker on the watch was pushed at bedtime (lights out) and wake time, and answering questions (online through REDCap or on paper forms) about their child’s sleep (e.g., bedtime, waketime, naps, night wakings).

Another week was allotted for the family to return the actigraphy watch to the research coordinator via an express mail envelope and for the study staff to review the data. Upon review, if there was a sufficient amount of scorable actigraphy data (10–14 days) and the caregivers had completed the paper time log and baseline questionnaires, the caregivers were then scheduled to meet with the trained therapist at Mercy Community Healthcare.

2.3.3. Sleep education intervention

Based on our previously published curriculum (Malow et al., 2014), the intervention consisted of three weekly sleep education sessions provided by the therapist. The first session was a 90-minute in-person session which was structured by a slide presentation for organization. The therapist provided the caregiver with information about sleep-promoting daytime habits, evening habits, bedtime routines, visual schedules and other helpful strategies for helping children with ASD who may be resisting sleep and/or waking multiple times in the night. This session was informational as well as collaborative, with the caregivers providing information about their child’s specific sleep problems and the therapist tailoring the session to address those specific areas, including strategies that met the child’s and family’s individual needs rather than a “one-size fits all” approach. For example, for a child with minimal verbal skills, visual tools were emphasized. In a family with multiple children, bedtime routines were integrated with those of other siblings. With the caregivers’ input about what activities are preferred and relaxing for their children in the evening, the therapist created a bedtime routine with an accompanying visual schedule for the caregivers to use nightly as they put their children to bed. The caregivers were asked to immediately begin making any suggested changes (e.g., delay bedtime, limit screen time before bed) that were offered in the session and to begin implementing the bedtime routine with visual schedules that night.

At the end of this sleep education session, the caregiver was sent home with another actigraphy watch, which was programmed to collect child data during the next four weeks that followed this session. The caregivers again completed daily homework including the paper time log (bedtime and waketime).

Two shorter follow-up sessions were conducted weekly for two weeks after the initial session. These were conducted either in person or by phone depending on the caregivers’ preference. Most parents chose to meet by phone for convenience. The follow-up sessions allowed the therapist and caregiver to review the information presented at the initial sleep education session and the completed homework. Using a structured format which we have described in previous publications as the Parent Absorption Scale (Malow, MacDonald et al., 2016; Table 3), the educator asked open-ended questions (“Tell me how you are doing making changes to…”) to assess caregivers’ understanding, recall, and reported implementation over the previous week(s) in making changes to their child’s activities regarding sleep. The caregivers reported how they were doing in each area where they made changes (e.g., increasing exercise; limiting screen time) and the therapist reviewed with caregivers the rationale for why making those changes can help their child sleep better. If the caregiver did not spontaneously report on each area recommended during the intervention session, the therapist then followed up with more specific questions to allow report for all recommendation areas. The therapist also answered any additional questions regarding the intervention recommendations and offered trouble shooting and problem-solving support related to implementation concerns noted by the caregiver. Based on information provided by caregiver report, the therapist completed a short Parent Absorption Scale (PAS) which rated the caregivers’ understanding and comfort level with the material and how well they implemented the suggested changes and strategies (Table 3). After the second follow-up session, the caregivers were instructed to continue implementing the strategies which were offered in the session.

Table 3.

Parent Absorption Scale.

Part 1: Questions asked by therapists.

1. Ask the parent “During your education session, your therapist and you came up with a number of suggestions to help your child sleep. Which of the suggestions did you find most helpful?”

2. Say to the parent “I am now going to ask you some specific questions.”

If an area wasn’t discussed as not applicable, mark n/a.

a. Tell me about how you are doing making changes to ____’s daytime activities?

b. Tell me about how you are doing making changes to ____’s evening activities?

c. Tell me about how you are doing making changes to ____’s bedroom?

d. Tell me about how you are doing using the visual supports we made?

e. Tell me about how you are doing use the bedtime pass?

f. Tell me about how you are doing helping ____ learn to sleep without you/lights on/television on/ etc.?

g. Tell me about how you are doing with helping ____ when he wakes up during the night?

h. Tell me about how you are doing with helping __ when he wakes up early in the morning?

Parent Absorption Scale.

Part 2: Rating Scale.

Complete the following rating scales based on your impressions of parent responses to the above questions.

Parent understanding of concepts discussed

• 1=does not seem to understand the concept at all

• 2=understood one or two ideas, but needed much more information

• 3=understood some ideas, but not all

• 4=has good understanding of the concept, but needed a little additional information

• 5=has excellent understanding of the concept

NOTES:

Parent implementation of strategies discussed:

• 1 =did not implement any of the strategies discussed

• 2=implemented one of the strategies, but not any others

• 3=implemented some strategies, but not all

• 4=implemented most of the strategies discussed

• 5=implemented all the strategies discussed

NOTES:

Parent comfort level with implementation of strategies/techniques:

• 1=parent does not seem comfortable with ability to implement strategies

• 2=parent seems comfortable with one strategy, but not any others

• 3=parent seems comfortable with some strategies, but not all

• 4=parent seems comfortable with most of the strategies discussed

• 5=parent seems comfortable with all of the strategies discussed.

NOTES:

2.3.4. Post-intervention data collection

Four weeks after the initial sleep education session, caregivers were asked to complete post-intervention questionnaires (FISH and CSHQ) and return the actigraphy watch to research staff via a prepaid express mail envelope.

Wilcoxon signed-rank tests were used to compare baseline and week 7 data.

2.3.5. Qualitative interviews

When these study procedures were complete, a study coordinator contacted the caregiver by phone to conduct an end-of-study qualitative interview. This semi-structured interview was designed to capture a more complete picture of the participant’s family dynamics (e.g., number of family members, orderly vs. chaotic home life) and the parent’s history of previous health behavior changes. The interview questions (Table 7) provided parents with an opportunity to describe how activities and routines of the family unit played a role in sleep behaviors. They also related their personal history with being successful or unsuccessful when attempting a health behaviour change in the past.

Table 7.

Qualitative Themes, Subthemes, Examples, Quotes, and Questions.

Themes	Subthemes	Examples	Quotes
Theme 1. Factors related to disrupted sleep	1. Environmental factors 2. Caregiver factors 3. Child factors	1. Chaotic home environments; lack of routine 2 & 3. Aspects of parent and child behavior	If I’m not paying attention, if the girls get off their routine and schedule, if there is mess everywhere, no one gets fed or gets their medicine... If I’m not paying attention, they will go for days without a bath. My husband also has autism. I can’t lose track of anything right down to his medicine and his clothes. It’s been phenomenal.. . because we travel. . . more now. So, I’m able to take the visuals with me. I tried not using it one day and I was like not doing that again.
Theme 2. Program implementation	1. Facilitators 2. Barriers	1. Education tailored to family; visual sleep aids 2. Children removing actigraphy watch; completing sleep education homework
Theme 3. Intermediate outcomes	1. Family harmony 2. Family engagement	1. Committing to program for whole family not just child 2. Parent being held accountable through homework	Once we made sure to make a conscious effort. . . he got there, and he was happy and he was better for it. We were better for it and we all got to see the results of him being there because he got better.
Theme 4. Impacts		1. Parent responses concerning improved sleep for child and caregiver 2. Appreciation for receiving sleep education in their community	It was a process and that’s the whole point. Finding what works and what doesn’t But, oh yes. Because nighttime used to be a real nightmare. Now everything to us and to her is very predictable and we know what to expect. I appreciate y’all coming out this way. It was very helpful, not having to take a trip downtown, you know, every time.
Study 1: Qualitative Interview Questions.
• Do you consider your household to be more orderly or more chaotic?
• In what ways is it more orderly or more chaotic?
• Walk me through a typical day for your family.
• How many family members are in your home?
• How many caretakers?
• What are some of the positive effects that autism has had on your life or your family’s life?
• Can you give examples of health education you have received in the past that was helpful? (ex: workshops, seminars, pamphlets, websites).
• Can you describe why or how this education was helpful?
• Have you had success in the past when you set out to change a health behavior for yourself? For your child?
• Please provide examples for your answer.
• When it is challenging to change a health behavior for yourself or your child, how do you respond and what do you when you feel challenged?
• Do you think your child and family benefited from participation in this project?
• If yes, how?
Study 2: Qualitative Interview Questions.
Section A. Family Life.
• How many family members are in your home?
• How many caretakers?
• Walk me through a typical day for your family.
• In what ways is your household more orderly?
• In what ways is it more chaotic?
Section B. Success in Changing Behavior.
• Can you think of a time you tried to change something you were doing? (For example, some people try to lose weight or quit smoking).
• Tell me some things that were successful about that process.
• Tell me some things that were challenging about that process.
• How do you respond when you feel challenged?
Section C. Participation.
• Do you think that your child and family benefited from participation in this study?
• If yes, what did you like about it?

Each interview was audio-recorded, transcribed verbatim, de-identified, and formatted for data coding. A hierarchical coding system was developed and refined using an inductive-deductive approach based on the semi-structured interview guide questions and a preliminary review of transcripts. Two trained coders independently coded the interviews. They compared their work and reconciled any differences in coding to reach consensus, which enhanced interrater reliability of the coding process. Quotes were sorted by category, frequency distributions were examined, and then quotes were read in detail to identify higher order themes and relationships.

2.4. Study #2. Preliminary effectiveness

Minor modifications to the protocol were made for study #2 (See Table 1b) which was conducted between February 2016 and April 2018. We partnered with additional pediatric practices in the middle Tennessee area and trained additional licensed therapists (occupational therapists, a registered nurse with a bachelor’s degree, and a licensed professional counselor—none with doctorate degrees) to provide the sleep education intervention. Involving therapists without doctorate degrees in education was important because we wanted to show that we could generalize our results to most therapy practices. Partnering with additional practices increased recruitment, enrollment, and number of participants, allowing us to gain additional experience with the protocol. After reviewing the feasibility study procedures, we felt it was important to lengthen the study duration to determine if any improvements in sleep made during the intervention period were sustained over time. Therefore, a third follow-up session was added and conducted (by phone only) 3 weeks after the intervention period, and an additional week of actigraphy and daily homework were added approximately 9 weeks after the intervention. Also, the questions asked in the end-of-study qualitative interview varied slightly to be more relevant to the sleep education program. The CSHQ and FISH were completed at the end of the intervention period (week 7) as in Study #1. The CSHQ, FISH, and CBCL were completed again during the additional week of actigraphy collection.

2.4.1. Post-Intervention week 10 follow-up-call

Caregivers were called by the therapist approximately 3 weeks after the second follow-up session. These phone calls were the same in format to the weekly follow-up sessions which were conducted during the intervention period. The therapist asked open-ended questions assessing how the caregivers were doing with implementation, reiterated the rationale for the changes, and answered questions. The caregivers were instructed by the therapist to continue with implementation and to call if they had any concerns prior to the end of the study, which would be six weeks later. The PAS was completed at the end of the phone call.

2.4.2. Week 16 follow up/end of study

The final week of the study included another week of actigraphy data collection with daily homework and a repeat of all baseline survey measures (FISH, CSHQ, and CBCL). Actigraphy watches were mailed to the caregivers’ home with the paper time log and the online daily homework was sent via email. When these procedures were complete, a study coordinator contacted the family by phone to conduct the end-of-study qualitative interview.

Wilcoxon signed-rank tests were used to compare baseline and week 7 data, and baseline and week 16 data.

2.5. Measures: caregiver-completed

2.5.1. Hollingshead four factor index of socioeconomic status

This tool measures socioeconomic status based on four domains: marital status, employment status, education, and occupation (Hollingshead, 1975). SES is calculated using a coding scale for each domain (e.g., 7-point scale for education, 9-point scale for occupation). Each parent living in the home is rated in the four domains and given a composite score. This measure was used to characterize our sample.

2.5.2. Newest vital sign (NVS)

In this validated assessment of health literacy, patients are given a nutrition label to read (from a container of ice cream) and then asked to answer six questions demonstrating their ability to process and use the information.(Weiss et al., 2005). Four of the six questions involve using mathematical skills. This measure was used to confirm that participants would be able to understand and complete questionnaires appropriately.

2.5.3. Children's sleep habits questionnaire (CSHQ)

The CSHQ is a 33-item validated parent-completed questionnaire that has been used to examine sleep behavior in toddlers, preschool, and school-aged children with a variety of conditions, including ASD (Goodlin-Jones, Tang, Liu, & Anders, 2008; Owens, Spirito, & McGuinn, 2000). The eight subscales of the CSHQ address common parental sleep concerns including bedtime resistance, sleep onset delay, sleep duration, sleep anxiety, night wakings, parasomnias, sleep disordered breathing, and daytime sleepiness. A higher score indicates more sleep problems. We recently validated a shortened version of the CSHQ for children with ASD in the Autism Speaks Autism Treatment Network (AS-ATN) Registry (Katz et al., 2018), which includes a subset of questions from the original CSHQ and has a more relevant factor analysis for ASD. We analyzed the total CSHQ scores and associated subscales as well as the total score and subscales on the shortened CSHQ. Based on our previous work (Malow et al., 2014), we considered the CSHQ to be our primary outcome variable.

2.5.4. Family inventory of sleep habits (FISH)

The FISH is a validated 12-item scale which assesses family sleep habits, including bedtime routine, sleep environment, and parental interactions (Malow et al., 2009). There is also more comprehensive version which contains 22 items (Reed et al., 2009). A higher score indicates better sleep hygiene. The FISH has demonstrated improvement with a parental behavioral intervention in our prior work (Malow et al., 2009). The FISH was included as a measure of whether a child’s sleep habits changed with the intervention.

2.5.5. Child behavior checklist (CBCL)

The CBCL is comprised of 100 items (scored on a 3-point Likert scale from “not true” to “very true or often true”) that examine daytime behavior in children, including those with sleep problems. There are two versions, one for preschoolers ages 1 ½ to 5 and one for school children ages 6–18 (Achenbach & Rescorla, 2001a, 2001b). T-scores derived from scales common to both forms, which in our previous work have been shown to be affected by poor sleep, were selected for analysis. These included the syndrome scales of anxious/depressed, somatic complaints, attention problems, aggressive behavior, and the DSM-5 scale of affective problems. The CBCL was included as a measure of whether child behaviors would improve – we limited analysis to the preliminary effectiveness study as that allowed for more time (16 weeks) to see improvement.

2.6. Measures: actigraphy and associated materials

2.6.1. Actigraphy

Based on our previous work, (Malow et al., 2014), we considered actigraphy to be a secondary outcome variable, providing an objective measure of improvement in sleep patterns, with sleep onset delay and sleep duration selected for analysis. AW Spectrum Actiwatches® (Philips Respironics, Bend, OR) were used for actigraphy data collection. The devices were configured using a 1-min epoch with medium threshold. Sleep latency (SL) and sleep duration were calculated based on Philips Respironics algorithms. To determine SL, bedtime was recorded by caregivers using the time log and the watch event markers. Caregivers were given the option to use wrist or pocket watches. The pocket watches (without bands) are placed inside a pocket sewed onto the shoulder of a form fitting t-shirt. Younger children often tolerate the pocket watches better than the wrist watches (Adkins et al., 2012), and results support that data obtained is comparable to that obtained from wrist placement.

2.6.2. Time log

This is a one-page paper chart, created by research staff for this study, which is auto populated with the dates of scheduled actigraphy collection. This is used by caregivers for recording the times event markers are pressed on the actigraphy watch at bedtime (lights out) and the final awakening. Caregivers are asked to use the time on the actigraphy device (instead of a clock or phone) for recording times. This manual recording of the event markers assists the data analyst in scoring the actigraphy data. There is also a column for writing additional comments (e.g., my child was sick during the night and came to my room three times….)

2.6.3. Daily homework

Before the intervention, caregiver daily homework consisted of questions about their child’s sleep times (e.g. bedtime, wake time, night wakings, naps) and questions about pushing the event marker on the actigraphy device. After the intervention, the daily homework included the pre-intervention questions about sleep times and the watch and also asked additional questions about implementation of healthy daytime and sleep habits (e.g., exercise, caffeine, screen time, light), visual supports (see below) and the bedtime routine.

2.6.4. Visual supports

Laminated pictures of bedtime routine activities (e.g., putting on pajamas, brushing teeth, reading books), laminated visual supports that represented nighttime (e.g., a moon picture to place on the bedroom door) and laminated bedtime passes were provided to the families in the first sleep education session, along with instructions for establishing a bedtime routine and using the visual supports. A bedtime pass is a card that the child can use as a token (“ticket”) to receive parental interaction during the night (Friman et al., 1999), which can be “saved” and not used by the child. Our group has modified the use of the bedtime pass to allow the child to trade an “unused” pass for a reward in the morning. Additional supports were given to families based on the individualized needs of the children.

3. Results

3.1. Study #1 (feasibility)

3.1.1. Participants

Ten families of children [7 boys, 3 girls, with a mean age of 8.1 years (standard deviation; SD) of 2.5 years)] participated in Study #1. Caregivers were usually parents, although a grandparent, as well as an aunt and uncle guardian, participated in place of parents. Average Hollingshead score was 36.4 (SD = 11.4).

Seven children completed all study procedures. One participant withdrew prior to starting the intervention, one withdrew after the first follow-up session, and another completed all procedures except the end-of-study interview. This resulted in nine children completing both baseline and week 7 actigraphy, with their parents completing baseline and week 7 CSHQ and FISH instruments.

All families demonstrated appropriate fluency (score of at least 4 on the Newest Vital Sign) to receive sleep education materials.

3.1.2. Caregiver-completed surveys

3.1.2.1. CSHQ (n = 8 participants).

The standard and modified four-factor version of the CSHQ showed significant improvements in several of the insomnia-related domains at week 7 compared to baseline (Table 4). These included sleep onset delay, sleep duration, sleep anxiety, and the CSHQ total scale on the standard scales, and the four-factor sleep initiation/duration subscale. Subscales that were not expected to improve with treatment of insomnia, such as the sleep disordered breathing scale, did not change. These findings supported continued use of the CSHQ in the preliminary effectiveness study (Study #2).

Table 4.

Study 1. Children’s Sleep Habits Questionnaire (CSHQ) Results (Standard and 4-Factor).

	Baseline (n = 8) Mean/SD	WK 7 (n = 8) Mean/SD	p Value
CSHQ – Bedtime Resistance	10.38/2.45	8.00/2.0	p = .061
CSHQ – Sleep Onset Delay	2.00/0.76	1.25/0.46	p = .034
CSHQ – Sleep Duration	5.38/1.51	4.00/1.07	p = .026
CSHQ – Sleep Anxiety	7.38/2.33	6.25/2.19	p = .039
CSHQ – Night Wakings	4.88/1.73	4.25/1.17	p = .163
CSHQ – Parasomnias	9.75/1.7	9.00/1.51	p = .034
CSHQ – Sleep Disordered Breathing	3.13/0.35	3.63/1.06	p = .102
CSHQ Daytime Sleepiness	16.25/4.59	14.75/5.97	p = .305
CSHQ TOTAL	55.38/5.37	48.38/6.48	p = .017
CSHQ (4FAC) Sleep Initiation/Duration	10.50/1.60	7.88/1.64	p = .011
CSHQ (4FAC) Sleep Anxiety/Co-sleeping	8.63/2.45	6.5/1.41	p = .074
CSHQ (4FAC) Night Wakings/Parasomnias	9.25/1.67	8.63/1.60	p = .096
CSHQ (4FAC) Daytime Alertness	12.00/3.02	11.25/4.10	p = .522

3.1.2.2. FISH (n = 8 participants).

Sleep habits improved on the Family Inventory of Sleep Habits, increasing from a mean (standard deviation; SD) of 42.63 min (SD = 6.2) to 48.25 min (SD = 2.6); p < 0.028. These findings supported continued use of the FISH in the preliminary effectiveness study (Study #2).

3.1.3. Actigraphy

Actigraphy was well tolerated in all children. Based on our previous work, we focused our analysis on sleep onset latency (SOL) and total sleep time (TST). SOL improved from a mean of 45.5 min (standard deviation; SD = 12.9 min) to 35.8 min (SD = 6.3 min) and TST minimally changed from a mean of 492.1 min (SD = 76.3 min) to 487.9 min (SD = 65.3 min) with neither result reaching statistical significance (SOL: p = 0.091; TST: p = 0.409). These findings of tolerability supported continued use of actigraphy in the preliminary effectiveness study (Study #2).

3.1.4. Therapist fidelity

The therapist involved in Study 1 required additional practice sessions to achieve the 80% fidelity necessary to deliver the education in the research study. Additional support by the Vanderbilt psychologist/behavioral sleep specialist was provided to maintain fidelity during the study. As a result of Study #1, we modified our protocol to provide additional support to therapists in Study #2 via phone or conference calls, email, or in-person meetings.

3.1.5. Parent absorption scale

Seventy-five percent of caregivers achieved scores of 4 or 5 on each section of the Parent Absorption Scale (representing good to excellent levels of understanding, comfort, and implementation).

3.2. Study #2 (preliminary effectiveness)

Thirty-three families of children [25 boys, 8 girls, with a mean age of 6.2 years (standard deviation; SD = 2.7 years)] consented to participate in Study #2. All of the caregivers were parents in Study #2. Mean Hollingshead score was 40.0 (SD = 14.8). Twenty families completed all study procedures and were able to contribute to the end of study quantitative measurements and the qualitative interview. One additional family who withdrew prior to the intervention was willing to participate in the end of study interview. Eight families withdrew prior to the first sleep education session. The remaining five withdrew at various points after the intervention began. Twenty-three parents completed baseline and week 7 CSHQ and FISH instruments and 21 parents completed baseline and week 7 CSHQ and FISH instruments. Nineteen parents completed baseline and week 16 CBCL results. All families demonstrated appropriate fluency (score of at least 4 on the Newest Vital Sign) to receive sleep education materials.

3.3. CSHQ

The standard and modified four-factor version of the CSHQ showed significant improvements in several of the insomnia-related domains at week 7 compared to baseline (Table 5), including bedtime resistance, sleep onset delay, sleep duration, sleep anxiety, and night wakings (standard subscales) and the four-factor sleep initiation/duration, sleep anxiety/co-sleeping, and night wakings/ parasomnias subscales. The CSHQ total score also improved. Most of these scales/subscales showed continued improvements at 16 weeks, with significance noted when comparing baseline with week 7, with maintenance of improvement at week 16.

Table 5.

Study 2. Children’s Sleep Habits Questionnaire (CSHQ) Results (Standard and 4-Factor).

	n = 23 Baseline Mean (SD)	n = 23 Week 7 Mean (SD)	n = 21 Week 16 Mean (SD)	p-value comparing baseline & Wk 7	p-value comparing baseline & Wk 16
CSHQ – Bedtime Resistance	11.04 (3.27)	8.70 (2.53)	9.05 (3.41)	P = 0.001	p = 0.005
CSHQ – Sleep Onset Delay	2.30 (0.64)	1.39 (0.66)	1.33 (0.48)	P = 0.000	p = 0.000
CSHQ – Sleep Duration	5.7 (1.36)	4.70 (1.36)	4.29 (1.45)	p = 0.007	p = 0.003
CSHQ – Sleep Anxiety	7.91 (2.52)	6.61 (2.17)	7.05 (2.46)	p = 0.009	p = 0.115
CSHQ – Night Wakings	5.83 (1.78)	4.70 (1.64)	4.57 (1.78)	p = 0.002	p = 0.027
CSHQ – Parasomnias	11.43 (2.78)	10.78 (2.58)	10.43 (1.96)	p = 0.191	p = 0.094
CSHQ – Sleep Disordered Breathing	3.65 (1.23)	3.43 (1.20)	3.48 (0.75)	p = 0.096	p = 0.564
CSHQ Daytime Sleepiness	14.96 (3.46)	13.9 (3.62)	14.05 (3.26)	p = 0.047	p = 0.236
CSHQ TOTAL	58.78 (7.47)	50.87 (7.02)	50.67 (8.78)	p = 0.000	p = 0.001
CSHQ (4FAC) Sleep Initiation/Duration	11.39 (2.19)	8.65 (2.145)	8.05 (2.13)	p = 0.000	p = 0.000
CSHQ (4FAC) Sleep Anxiety/Co-sleeping	9.13 (3.38)	7.57 (2.52)	8.10 (3.39)	p = 0.003	p = 0.306
CSHQ (4FAC) Night Wakings/Parasomnias	11.48 (2.61)	9.83 (2.21)	9.76 (2.41)	p = 0.002	p = 0.044
CSHQ (4FAC) Daytime Alertness	11.13 (3.53)	10.35 (3.55)	10.43 (3.68)	p = 0.164	P = 0.360

3.4. FISH

Sleep habits improved on the Family Inventory of Sleep Habits, increasing from a mean (standard deviation; SD) of 47.4 (SD = 5.0) to 50.7 (SD = 3.9); p < 0.001 at week 7 (n = 23 participants) and 50.7 (SD = 4.2); p = 0.002 at week 16 (n = 21 participants).

3.5. CBCL

The CBCL subscales of aggressive behavior and affective problems improved from baseline to week 16 (Table 6). Improvements were not seen in the subscales of anxious depressed, somatic complaints, and attention problems.

Table 6.

Study 2. Child Behavior Checklist (CBCL) Results.

CBCL Subscale	N = 19 Baseline Mean/SD	N = 19 Week 16 Mean/SD	p-value
Anxious Depressed	59.14/9.409	58.38/9.856	p = .856
Somatic Complaints	62.81/8.370	60.76/7.822	p = .277
Attention Problems	65.10/8.179	64.29/9.461	p = .522
Aggressive Behavior	62.86/9.961	59.10/7.905	p = .016
Affective Problems	66.71/7.988	63.76/8.694	p = .035

3.6. Actigraphy

Actigraphy was tolerated well in all participants except one. Twenty-four participants completed actigraphy at week 7 and 19 at week 16. Sleep-onset delay (SOL) and total sleep time (TST) improved at weeks 7 and 16 compared to baseline although statistical significance was not achieved (p > 0.1). Mean SOL (standard deviation; SD) was 43.1 min (SD = 35.7) at baseline, 32.5 min (SD = 22.2) at week 7, and 28.2 min (SD = 26.6) at week 16. Mean TST was 491.8 min (SD = 48.0) at baseline, 499.7 min (SD = 41.0) at week 7, and 493 min (SD = 52.3) at week 16. Of note, 12 participants had a mean SOL that was greater than 30 min at baseline. Their SOL improved from 70.1 min (SD = 32.4) at baseline to 48.8 at week 7, although significance as not achieved (p = 0.1).

3.7. Therapist fidelity

All therapists achieved an average of 80% or higher fidelity after conducting practice sessions and continued to exhibit fidelity throughout the study.

3.8. Parent absorption scale

Ninety-five percent of caregivers reported good to excellent levels of understanding, 100% of caregivers reported good to excellent levels of comfort, and 90% of caregivers reported good to excellent levels of implementation.

3.9. Qualitative interviews

A total of 7 caregivers participated in the qualitative interview in Study #1 and 21 caregivers participated in Study #2. Four major themes and multiple subthemes emerged. Major themes identified included: 1) factors related to disrupted sleep; 2) program implementation; 3) intermediate outcomes; and 4) impacts. Within theme 1, subthemes included environmental, parent and child factors related to disrupted sleep. Parents identified chaotic home environments, lack of routine, as well as aspects of parent and child behavior that contributed to disrupted sleep. Within theme 2, parent responses fell into subthemes of facilitators and barriers to participating in and implementing the sleep education program with their children. One barrier identified was children taking off the actigraphy watch without parent knowledge. Another barrier was keeping up with homework that went along with the sleep education. Conversely, parents identified the education tailored to their family needs and visual sleep aids as facilitators to program participation and implementation. Theme 3, intermediate outcomes, included subthemes of family harmony and family engagement. Several parents discussed benefits of holding themselves accountable through homework and committing to the program for the family rather than just the child with ASD. The fourth theme, impacts, included parent responses concerning improved sleep for child and parent. Parents observed that their understanding of sleep and the process of preparing for sleep had changed as a result of the sleep education. Themes and subthemes with example quotes from caregivers are outlined in Table 7.

4. Discussion

In this report, community therapists, trained by academic sleep professionals, provided sleep education to caregivers of children with ASD. Our findings demonstrate the feasibility and preliminary efficacy of this training program. Therapists showed fidelity in carrying out the sleep training, both with the practice interviews and in the actual study with caregivers. They exhibited both completeness in carrying out the goals of the training and also in being empathetic to families. In turn, families showed a high level of absorption, comfort, and implementation of the training program based on therapist rating. The pilot study helped us recognize the need for additional therapist support in the preliminary effectiveness study, and we saw improved therapist fidelity in the preliminary effectiveness study. Scores on the Parent Absorption Scale also improved in the preliminary effectiveness study, potentially in keeping with our refined training procedures of therapists and caregivers.

Modest improvements in child sleep and some aspects of child behavior, by parent-completed measures, were documented. Of note, caregivers did not report improvements globally, but in areas that would be expected to improve with the intervention. For example, the CSHQ subscale related to sleep disordered breathing did not show improvements with the intervention, while the subscales related to insomnia did show improvement. Differences between the four-factor CSHQ (Katz et al., 2018) and the standard CSHQ (Owens et al., 2000) were also noted in the preliminary effectiveness study. The four-factor CSHQ subscale combining night wakings and parasomnias, for example, showed improvement with treatment while the standard CSHQ parasomnia subscale did not. This may reflect challenges in distinguishing parasomnias and night wakings from each other in children with ASD, who may have limited verbal skills or ability to recall night wakings or to be fully conversational during them. Additional versions of the CSHQ have been developed for ASD (Johnson et al., 2016; Zaidman-Zait et al., 2020). These versions, along with the standard and four-factor CSHQ, should be examined in large, interventional studies. It should be noted that actigraphy, an objective measure of a child’s sleep, did not show improvements that reached statistical significance.

The qualitative end of study interviews supported and helped explain sleep improvements demonstrated in the quantitative results. These interviews also emphasize the clinical relevance of broadening the reach of sleep interventions for children with ASD into community settings. Families of children who have ASD and sleep problems were able to identify sources of sleep problems, articulate effects of child disrupted sleep, and appreciate benefits of community-based sleep education programs that offer research-based strategies for improving child sleep in a more accessible format. The interviews also acknowledged that families who have children with ASD may benefit from support aimed at structuring family routines. Families also commented favorably on the ability to gain access to sleep education in their communities without having to travel long distances to academic medical centers. They were appreciative of receiving specialized education from a therapist who was familiar to them. While therapists did not participate in the qualitative interviews, they also commented positively on being able to work directly with families on a skill rather than having to refer them outside of their practices. In addition, the therapists who received extensive training in sleep education for their clients reported a sense of satisfaction at being able to provide this service within their practice rather than sending families to another specialist.

Our study has several limitations. First, our study numbers were relatively small and enrollment of families was challenging. As the families verbalized, being able to access the study sessions in their community reduced time and financial resources compared to accessing the sessions in an academic medical center. However, for many families, being enrolled still resulted in an additional time spent attending an additional appointment and completing survey measures and study calls. Second, while all therapists met fidelity, training of therapists was intensive and time-consuming. They continued to have questions while working with families and required more individualized levels of support than anticipated given our experience in previous studies utilizing this curriculum. Third, given the primary focus of feasibility for this study, we did not include a control group of parents who were not given sleep education. Fourth, while actigraphy, an objective measure of sleep patterns, was included, many children did not have sleep latencies of greater than 30 min as documented by actigraphy, even though parents reported this finding, which limited analysis of actigraphy data. Confirmation of prolonged sleep latency by actigraphy, as carried out in our previous multisite study conducted in academic health centers (Malow et al., 2014) would have strengthened our findings. However, it is important to note that, even though not all families met this threshold regarding this objective measure of sleep, these were families who identified that their child’s sleep difficulties were concerning and were affecting their child and family. Given the focus on the community health setting in this study, the social validity of receiving this intervention despite not meeting that threshold is noteworthy.

Next steps for our work include using this preliminary information as a basis for recruiting larger sample sizes of families, with attention to a broad range of racial, ethnic, socioeconomic, and geographic disparities. For example, limited work has been done examining how to make sleep education accessible in rural and underserved populations/areas. Our development and evaluation of such a model was a very important first step in best identifying structures for providing larger scale training on these methods to the general population of providers who serve children with ASD but may not have this specialized knowledge. Our findings can be used to formalize training structures and assess the types of professionals, settings, and roles that may be best able to adopt this model.

Inclusion of a control group will be critical to demonstrate efficacy. Ways to continue to support therapists while working with families is important. Use of collaborative telehealth models in which therapists can present cases to experts may be one solution. Such models include Project ECHO, Extension for Community Healthcare Outcomes (Arora et al., 2007), which has been expanded to ASD (Mazurek, Brown, Curran, & Sohl, 2017). Identifying how the provision of these services can be integrated into billable models and determining which types of professionals may have more of an ability to do so will also be critical for sustainability. It was qualitatively noted that professionals with certain types of experiences and training seemed to more quickly feel comfortable with the curriculum, but having a larger number of professional types in a larger scale study would allow for a systematic analysis of which types of professionals and settings may be best equipped to provide this information at a high level of fidelity.

Finally, additional ways to train parents in sleep education are much needed. This highlights the importance of the emerging field of implementation science that promotes the systematic uptake of research findings into routine practice to improve the quality and effectiveness of health services (Bauer, Damschroder, Hagedorn, Smith, & Kilbourne, 2015). In prior work, we provided parents with access to a pamphlet, as well as the text of a book on sleep written for parents, with mixed success. Parents spoke to the need to be accountable to a professional, and to also have access to a professional to answer their questions. Programs that include on-line modules, including the incorporation of videos and tip sheets that parents can access and review on their own (before and between formal therapy sessions), are worthy of future development and evaluation.