Research in Developmental Disabilities

Transcription

1 Research in Developmental Disabilities 34 (2013) Contents lists available at SciVerse ScienceDirect Research in Developmental Disabilities Evaluation of a behavioral treatment package to reduce sleep problems in children with Angelman Syndrome Keith D. Allen a, *, Brett R. Kuhn a, Kristi A. DeHaai b, Dustin P. Wallace c a Department of Psychology, Munroe-Meyer Institute for Genetics and Rehabilitation, Nebraska Medical Center, Omaha, NE , USA b Human Genetics Laboratory, Munroe-Meyer Institute for Genetics and Rehabilitation, Nebraska Medical Center, Omaha, NE , USA c Department of Pediatrics, University of Missouri Kansas City, Children s Mercy Hospitals and Clinics, Integrative Pain Management, 2401 Gillham Road, Kansas City, MO 64108, USA ARTICLE INFO ABSTRACT Article history: Received 10 August 2012 Received in revised form 2 October 2012 Accepted 2 October 2012 Available online Keywords: Angelman Syndrome Sleep problems Bed-time problems Behavioral treatment Disruptive behavior The purpose of this investigation was to evaluate the effectiveness of a behavioral treatment package to reduce chronic sleep problems in children with Angelman Syndrome. Participants were five children, 2 11 years-of-age. Parents maintained sleep diaries to record sleep and disruptive nighttime behaviors. Actigraphy was added to provide independent evaluations of sleep wake activity. The treatment package targeted the sleep environment, the sleep wake schedule, and parent child interactions during sleep times. Treatment was introduced sequentially, across families, and evaluated in an interrupted time series, multiple baseline design. Data show that prior to treatment, baseline rates of nighttime disruptive behavior were stable or increasing and none of the participants were falling to sleep independently. With the introduction of treatment, all participants quickly learned to initiate sleep independently. Gradual reductions were reported in disruptive behaviors and these improvements were sustained over time. Results were replicated with two participants when treatment was withdrawn and reinstated. Changes in disruptive bedtime behaviors and in sleep onset were found to be statistically significant. Parents indicated high satisfaction with the treatment. A behavioral treatment package was found to be effective with five children with long histories of significant sleep-related behavior problems. These results suggest that behavioral treatment may be a reasonable way to address sleep problems in some children with Angelman Syndrome. ß 2012 Elsevier Ltd. All rights reserved. 1. Introduction Angelman Syndrome (AS) is characterized by developmental delay, happy demeanor, hypermotor behavior, stereotypies, and reduced adaptive skills. Behavior problems are frequently present in males and females of all ages, and include language deficits, excessive laughter, hyperactivity, short attention span, and problems sleeping (Summers, Allison, Lynch, & Sandler, 1995). Researchers have highlighted the importance of examining environmental influences on common behavior problems in genetic syndromes like Angelman and evidence suggests that some of these common problem behaviors in children with AS can be managed by systematically manipulating various aspects of the environment (Horsler & Oliver, 2006). Indeed, management of problematic behaviors has been based primarily on behavioral approaches that emphasize changing the environment, though psychoactive medication is sometimes required as well (Pelc, Cheron, Boyd, & Dan, 2008). * Corresponding author at: Munroe-Meyer Institute, Nebraska Medical Center, Omaha, NE , USA. Tel.: ; fax: addresses: kdallen@unmc.edu (K.D. Allen), brkuhn@unmc.edu (B.R. Kuhn), kdehaai@unmc.edu (K.A. DeHaai) /$ see front matter ß 2012 Elsevier Ltd. All rights reserved.

2 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Sleep problems are one of the most common behavior problems associated with AS. In one study, severe sleep problems were reported in as many as 35% of individuals with AS (Didden, Korzilius, Smits, & Curfs, 2004). Another study found marked problems with independently initiating and maintaining sleep as well as frequent night awakenings and reduced total sleep (Bruni et al., 2004). In many cases, heightened sensitivity to the sleep environment has also been identified as a common problem (Conant, Thibert, & Thiele, 2009; Walz, Beebe, & Byars, 2005). Perhaps not surprisingly, sleep problems have been found to be more severe in younger children, peaking between 2 and 6 years-of-age (Clayton Smith, 1993). Indeed, some have found bedtime problems and night awakenings in as many as 90% of young children with AS (Summers et al., 1995), and children with AS have been identified by sleep experts to be among the most challenging populations to treat (Mindell, Emslie, et al., 2006; Mindell, Kuhn, Lewin, Meltzer, & Sadeh, 2006b). Not surprisingly, sleep problems also create detrimental effects for families. Most parents report that dealing with a child s ongoing sleep problems can result in fatigue, irritability, distress, and impaired social functioning (Didden & Sigafoos, 2001). In addition, sleep problems in children with AS can create anger, anxiety and feelings of helplessness in family members who, because of their caregiving role for a child with AS, experience their own lack of sleep. Parents of children with AS report higher levels of psychological distress compared to parents of other children with rare genetic syndromes (Griffith et al., 2011). Behavior therapy is currently the consensus first-line treatment for pediatric sleep disturbance, with numerous studies showing that this approach produces durable changes in more than 80% of young children (Mindell et al., 2006b; Morgenthaler et al., 2006). Parents serve as the active change-agent (Owens, 2006), and are coached in how to implement these principles to shape healthier sleep habits in their children. The basic elements of a quality sleep intervention include teaching parents how to (1) create a quality sleep-compatible environment, (2) adjust and regulate the sleep wake schedule to consolidate sleep and (3) manage parent child interactions to reinforce appropriate bedtime behaviors and to promote independent sleep initiation (Kuhn, 2007; Mindell & Owens, 2003). These behavioral methods have empirical support and wide acceptance in the sleep literature and have been recommended in the management of sleep problems in AS (Pelc et al., 2008). Yet, few parents of children with AS report ever receiving this type of advice in treatment of sleep problems (Didden et al., 2004). This may be due, at least in part, to concerns about whether behavioral treatment of sleep problems is appropriate for children with AS. That is, some providers as well as parents may question whether behavioral interventions can be expected to solve sleep problems that are thought to reflect structural or chemical abnormalities (Walz et al., 2005), or the dysregulation of cortical interactions (Pelc et al., 2008). Additionally, providers may be reluctant to recommend behavior treatments without clear empirical support. Indeed, practitioners and researchers alike have acknowledged the need for more research of behavioral methods for managing common sleep problems in children with AS (e.g., Didden et al., 2004; Pelc et al., 2008). To date, much of the empirical evidence to support the efficacy of behavioral sleep interventions has been derived primarily from research with more typically developing populations (e.g., Mindell, Emslie, et al., 2006; Mindell, Kuhn, et al., 2006). Numerous other studies have investigated behavioral treatments for children with other types of neurodevelopmental disabilities. For example, Piazza et al. have evaluated a procedure called faded bedtime with response cost (e.g., Piazza & Fisher, 1991; Piazza, Fisher, & Sherer, 1997). In the procedure, if a child does not fall asleep within 15 min of being placed in bed, the bedtime is pushed 30 min later until they do. This procedure has been found effective in children with intellectual disabilities and severe behavior disorders, but has not been evaluated for children with AS. Only a single case study has been published evaluating a behavioral treatment in children with AS (Summers et al., 1992). Summers et al. treated a nine year old boy with AS who was sleeping less than two hours a night. The behavioral components included restriction of daytime sleep, installing a consistent sleep schedule, and reducing nighttime adult child interactions. However, the treatment included administering the medication diphenhydramine HCI and it was conducted primarily in an inpatient behavioral treatment unit by staff. Treatment in the home by parents was only introduced after 55-days of inpatient treatment. Moreover, there was no actual experimental design. So, although the treatment worked well and the child increased nighttime sleep to over eight hours a night, the independent effects of the behavioral intervention or the parent s ability to implement it was never assessed. Thus, although behavioral sleep interventions are often recommended for children with AS (Pelc et al., 2008; Walz et al., 2005), there are no well controlled, empirical investigations of behavioral treatment of sleep problems in children with AS. The primary purpose of this investigation was to evaluate a behavioral treatment package to address severe and chronic sleep disturbance in young children with AS. 2. Methods 2.1. Participants Five children between 2 and 11 years-of-age with a diagnosis of AS (confirmed by genetic testing) served as participants. All were recruited from a general registry maintained by The Human Genetics Laboratory at the Munroe-Meyer Institute for Genetics and Rehabilitation at the University of Nebraska Medical Center and from the AS Foundation ListServ. Inclusion criteria were adapted from Mindell and Durand (1993) and included a minimum 4 week history of sleep disturbances as defined by bedtime resistance, delayed sleep onset and nighttime awakenings. More specifically, sleep disturbances were operationally defined as

3 678 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) resistance to going to bed for a minimum of three nights per week (as specified by refusal to go to bed, persistent calling out or getting out of bed, more than 20 min between bedtime and onset of sleep); or 2. falling asleep in a location other than his/her bed or crib or requiring parental intervention (e.g., rocking, holding, feeding) or parental presence until he/she is asleep at bedtime for a minimum of three nights per week; or 3. spontaneously awakenings on an average of at least one time per night for a minimum of four nights per week that require parental intervention (e.g., rocking, holding, feeding, and parental presence) for return to sleep. No restrictions were placed on where participants resided, resulting in participants from across the U.S. Participant 1, Annie, was a 2-year-old female. Her parents reported that she was resistant to going to bed, never fell asleep without parental presence, and awakened repeatedly throughout the night. Participant 2, Bobby, was a 4-year-old male. His parents reported that he exhibited little bedtime resistance as long as they lay next to him at bedtime. However, he awakened repeatedly throughout the night, and each time sought parental presence to reinitiate sleep. Participant 3, Cindy, was an 11-year-old female. Cindy s parents always had to lie next to her each night to get her to fall asleep at bedtime and after any night awakening. Participant 4, Darcy, was a 5-year-old Caucasian female. Darcy s parents reported significant disruptive behavior at bedtime and throughout the night. They stated that they could not get her to sleep without lying down with her. Participant 5, Eddie, was a 6-year-old male. His parents reported that Eddie exhibited frequent bedtime resistance and required parental presence to fall asleep. They had resorted to sleeping on the floor next to his bed. All five participants had seizure activity and were on anti-seizure medications. Anti-epileptics as a drug class tend to be associated with increased sedation and reduced sleep latency, so medications were closely monitored and parents were asked to avoid making changes in medication status to ensure that the introduction of the sleep intervention was not confounded Setting The treatment team was located at a genetics and rehabilitation center at a large Midwestern University and Medical Center. Because participants came from across the United States (California, Iowa, Nebraska, North Dakota, and Texas), initial evaluations and subsequent delivery of treatment recommendations were conducted either via video telehealth connections in health clinics proximal to the participants when possible or via phone contact when video telehealth was not an option Measures Pediatric sleep diary The primary outcome measure consisted of a daily sleep wake diary completed by parents. Diaries have high internal consistency and good agreement with other measures of children s sleep such as videotapes and actigraph (Corkum, Tannock, Moldofsky, Hogg-Johnson, & Humphries, 2001; Mindell & Durand, 1993). Parents recorded sleep latency (i.e., time from lights out to asleep), total sleep time, frequency and duration of night-time awakenings, and sleep efficiency (i.e., total sleep time divided by total time in bed). Parents also recorded (1) the number of disruptive behaviors during the bedtime routine, (2) the number of disruptive behaviors after the lights out, but before the child fell asleep, and (3) the number of disruptive behaviors throughout the night. These numbers were totaled each night, summed across the week and divided by 7 to produce a mean frequency per night or the disruptive behavior composite (DBC). Independent sleep onset (ISO) was calculated by counting the number of nights each week in which the child fell asleep without a parent or caregiver physically present in the room. This number was divided by 7 and multiplied by 100 for the weekly percentage of nights with independent sleep onset. Study personnel maintained regular phone contact with parents throughout the study to prompt completion of the diary Actigraphy For the current study, a MicroMini Motionlogger actigraph (Ambulatory Monitoring, Inc.) was worn continuously by each child for the duration of the study, except during bathing or swimming. The actigraph was attached to the child s ankle with a plastic hospital band to prevent removal. Actigraphy is a valid and reliable addition to the sleep diary when assessing sleep wake patterns in the child s natural setting (Sadeh, 2011). Actigraphy objectively quantifies motor movement via accelerometer. These data are entered into an algorithm index of sleep wake patterns. Action ME software (Act Millennium version ) was used to interface between the actigraph and the computer and Action W software (ver 2.6), was used to score the data using the Sadeh scoring algorithm for young participants (Sadeh, Sharkey, & Carskadon, 1994) The Developmental Behavior Checklist parent report form This is a 96 item instrument used for the assessment of behavioral and emotional problems in children with developmental and intellectual disabilities and it has high test/retest reliability and excellent concurrent validity (Einfeld & Tonge, 2002). Parents rate each item on a 0 2 point scale (0 = not true, 1 = sometimes true, 2 = very true or often true) while thinking about their child s behavior in the past six months. This assessment yields a total behavior problem score and five subscale scores: disruptive/antisocial, self-absorbed, communication disturbance, anxiety, and social rating. Scores are compared to those of a national sample of children of the same age and gender. There is no clinically significant limit for this assessment.

4 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) The Abbreviated Children s Sleep Habits Questionnaire (CSHQ) This is a parent-completed sleep screening instrument designed for children ages 4 10 years with good test/retest reliability as well as good sensitivity and specificity (Owens, Spirito, & McGuinn, 2000). The abbreviated version consists of 33-items related to common sleep behaviors in children. Each sleep behavior is rated on a three-point Likert scale (rarely = 0 1 night per week; sometimes = 2 4 nights per week; usually = 5 7 nights per week). The CSHQ items are grouped into three domains: Dysomnias (difficulty getting to sleep or staying asleep), Parasomnias (sleepwalking/ talking, night terrors, bedwetting, restless leg syndrome, etc.), and Sleep-Disordered Breathing. A score of 41 or higher has been identified as the cut-off score above which referral to a sleep specialist should be considered. Normative data on the CSHQ were recently made available for younger children, 2 5 years-of-age (Goodlin-Jones, Sitnick, Tang, Liu, & Anders, 2008) Customer satisfaction questionnaire Finally, at the end of treatment, parents provided judgments regarding their overall satisfaction with the treatment protocol using the Abbreviated Acceptability Rating Profile (AARP) that was modified to specifically reflect the acceptability of the behavioral treatment package. Research has found that the AARP possesses acceptable internal consistency, reliability, and validity (Tarnowski & Simonian, 1992). The modified AARP included eight items regarding treatment acceptability rated on a Likert-type scale (1 = strongly disagree to 6 = strongly agree). Parents are asked whether they agree that the intervention was acceptable and effective, whether they would keep using it and recommend it to others, and whether they felt it was a good way to solve their child s sleep problems. Quantitatively, with a range of scores from 8 to 48, a score above the midpoint (28+) indexes minimal treatment acceptability Research design and data analysis The research employed a multiple baseline (MBL) design across participants (Barlow, Nock, & Hersen, 2009), in which each participant spent varying time in baseline before treatment implementation. Once treatment was implemented, the results were evaluated using visual analysis of observed changes in level, trend, and variability. This time-lagged control design demonstrates experimental control when the sequential introduction of the treatment produces changes in sleep activity for the targeted participant, while sleep activity in concurrent or untreated individuals remains unchanged. To supplement the visual analysis, data were also subjected to statistical evaluation Procedures During subject recruitment, participants were interviewed by one of the research team to determine if the prospective subject met the inclusion criteria. Those that met the inclusion criteria were mailed the consent forms as well as the behavior rating scales. Those that consented and completed the forms were then contacted to begin baseline recording Baseline Those enrolled in the study were contacted via phone or proximal telehealth connections in order to describe how to complete the sleep diary and use the actigraph. Caregivers then began baseline recording of sleep activity using the sleep diary and the actigraph. Parents were asked to record sleep variables and behavior disruptions on the sleep diary. To determine sleep onset, parents were asked to monitor their child every 10 min from bedtime until the time they found their child asleep. They were also asked to report the times at night when they heard or observed that their child was awake. Completed sleep diaries and actigraphs were returned via express mail on a weekly basis to the investigators, who calculated and plotted the disruptive behavior composite (DBC) and the percentage of independent sleep onset (ISO) for each child in order to guide decision-making about when to begin treatment. In deciding how long to continue in baseline, the first criterion was whether stability had been demonstrated in the data. Data were considered stable if there were no trends toward improvement in both the DBC and ISO data. The other criterion was to insure that there were varying numbers of weeks in baseline (ranging from two-six weeks) before treatment was introduced so that it could be determined whether changes in DBC and ISO began when and only when treatment was introduced (a critical element of establishing experimental control and demonstrating internal validity) Behavioral treatment package (BTP) The BTP included 3 principle components; parents served as the active agents of change and they were taught to (1) create a quality sleep-compatible environment, (2) adjust and regulate the child s sleep wake schedule, and (3) manage parent child interactions to reinforce appropriate child behavior and promote independent sleep initiation Component 1: sleep environment. Experts in sleep ecology have long recognized that sleep quality and sleep duration are maximized in an environment that is dark, quiet, non-stimulating, and perceived as safe (e.g., Siegel, 1995). This component involved asking parents to create a quality sleep-compatible environment. They were asked to insure no light or low light (a nightlight was permitted), to eliminate visual and auditory stimulation (i.e., turn off stereo or television/videos), and to adjust ambient temperature if necessary.

5 680 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Component 2: sleep schedule. Some children experience difficulty falling to sleep simply because they are not always tired when their parents put them to bed. This component of the treatment plan involved asking parents to maintain a consistent sleep wake schedule across the week, putting the child to bed at the same time each night and getting them up at the same time every day to help each participant maintain a consistent sleep wake schedule (Spielman, Saskin, & Thorpy, 1987). For children with long sleep onset times, parents were asked to temporarily delay the child s bedtime to promote quick sleep onset (Piazza & Fisher, 1991). The new bedtime was determined by calculating the average sleep onset time during baseline and then adding 30 min (e.g., if the average time of sleep onset was 9:30 p.m. during baseline, then the initial bedtime was set at 10:00 p.m.). The child was not allowed to go to bed or fall asleep prior to this time or sleep past the scheduled wake time. Once the child was falling asleep quickly (e.g., min), the bedtime was gradually moved earlier in 30 min increments (as long as the child was falling asleep quickly) until reaching a parent determined bedtime goal. Parents were told that by not allowing the child to make up for lost sleep (i.e., allowing the child to sleep in or allow them to go to bed earlier), parents would increase the likelihood that their child would fall to sleep more quickly Component 3: parent child interactions. This treatment component relied on empirically supported options that focus on reducing parental responses to disruptive bedtime behaviors (Mindell, Emslie, et al., 2006; Mindell, Kuhn, et al., 2006). Parents were asked to avoid responding (i.e., ignore) disruptive bedtime behaviors including crying, tantrums, calling out, or coming out of the bedroom. Parents who had low tolerance for ignoring were taught to use the Excuse-Me Drill (Kuhn, 2011). In this procedure, parents were permitted to check in on the child periodically, but only when the child was being calm and quiet. That is, parents used their attention and physical presence to reinforce the child for displaying sleepcompatible behaviors, such as remaining calm, quiet, and in bed. This was repeated during night awakenings. Together, these recommendations about parent child interactions are seen as necessary to promote self-soothing skills which in turn allow young children to initiate sleep independently at bedtime, and to reinitiate sleep in their own bed following normal nighttime awakenings that are part of the normal sleep cycle. The fact that the treatment involved multiple components raised obvious concerns about treatment adherence. As a result, the project coordinator maintained weekly contact with families implementing treatment via both phone and communications. In addition, parents could phone or at any time with questions or concerns. Treatment was expected to last about six eight weeks, but was not considered to be completed until the DBC and ISO data were observed to be stable and/or improving. Thus, the actual number of weeks in treatment was somewhat flexible. In two cases treatment continued on longer because there were unplanned treatment withdrawals. Parents/ caregivers were then asked for post-treatment ratings of behavior using the CSHQ and the Developmental Behavior Checklist as well as the modified AARP. Follow-up evaluations were attempted at 1 and 3 month post treatment. At that time, parents were asked to record their child s sleep behavior for one week using the sleep diary and actigraphy. 3. Results 3.1. Visual analysis of primary dependent measures The primary results of the investigation are depicted in Fig. 1. The disruptive behavior composite (DBC) is depicted on the primary ordinate while independent sleep onset (ISO) is depicted on the secondary ordinate. For each of the five participants, DBC was stable or increasing (i.e., getting worse) during the baseline phase. Rates of DBC ranged from only 1 to 2 per night (Cindy) to as many as 200 per night (Darcy). In addition, none of the five participants were falling asleep independently, with the ISO at 0% across baseline. With the introduction of the behavioral treatment package, the DBCs for all 5 participants showed a reversal in trend, decreasing over the next month. After four weeks in treatment, the DBCs reached near zero levels for Annie, Cindy, Darcy, and Eddie. In addition, the changes began when and only when the treatment package was initiated, providing repeated demonstrations of experimental control. Particularly noteworthy is that Darcy showed an unusually high DBC rate in baseline that was nearly five times greater than the DBC of the other participants. These rates were showing an increasing trend that was reversed and reached near zero levels within 1 month. Bobbie and Darcy also had unplanned treatment withdrawals during which the parents stopped the treatment package. Bobbie s parents had job-related changes that interfered with implementation while and Darcy s parents began allowing her more time-in-bed. For both participants, there were marked increases in DBCs that were then reduced again when treatment was reinstated as planned. These changes provide additional demonstrations of a functional relationship between the introduction of treatment and the reductions in DBC. Fig. 1 also shows that the introduction of treatment produced dramatic changes in independent sleep onset for all 5 participants. All participants went from requiring their parents to be present to fall asleep during baseline, to falling asleep independently within the first week of treatment. The only time this changed at all was when Darcy s parents stopped treatment and increased her time-in-bed. This eventually resulted in her beginning to have problems falling to sleep independently. Independent sleep onset was recovered when treatment was reinitiated. The changes in ISO were not only immediate for all of the participants but they were stable and sustained across treatment and follow-up.

6 [(Fig._1)TD$FIG] K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Fig. 1. Disruptive behavior composite (DBC) and independent sleep onset (ISO) for each of the 5 participants. DBC is mean frequency per week (primary ordinate) and ISO is percentage of days per week (secondary ordinate). Follow ups were conducted at 1 and 3 months Statistical analyses of primary dependent measures To determine appropriate statistical methods for each of the two primary dependent measures, preliminary visual and statistical analyses were conducted to assess the presence and degree of autocorrelation as well as the pattern of differences between baseline and treatment conditions. These analyses found significant autocorrelation within DBC. More specifically,

7 682 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Annie, Bobby and Eddie had no statistically significant autocorrelations (statistical calculations based on the Ljung Box test), however, Cindy showed significant autocorrelation at the lag-1 level (.48, p <.05), and Darcy showed autocorrelation at lag-1 (.55, p <.01), lag-2 (.36, p <.01), lag-3 (.23, p <.01), lag-4 (.09, p <.05), lag-5 (.17, p <.05), and lag-6 (.01, p <.05). In addition, the average disruptive behavior composite (DBC) showed gradual rather than immediate effects including apparent changes in slope and overlapping data points across baseline and treatment conditions. In contrast, the ISO showed no autocorrelation (all p s > 0.36), changes in level but not slope, and marked and immediate treatment effects with no overlapping data points (except in reversal conditions). As a result, different analysis strategies were chosen Independent sleep onset (ISO) For the dependent variable ISO, a binomial sign test was selected. This type of test is appropriate when there is no significant autocorrelation, baseline trend is fairly level, and the primary outcome of interest is a change in level (not slope) from baseline to intervention (Campbell & Herzinger, 2010). The test uses probabilities to determine whether a set of observations is likely to occur by chance. Statistical significance was determined by using a binomial probability table with assumed probability of success (observation above all baseline points) at Calculated separately for each participant, this revealed statistically significant effects in each case; more specifically, Annie had 6/6 observations above all baseline values (p <.01), Bobby had 8/8 (p <.01), Cindy had 9/9 (p <.01), Darcy had 12/12 (p <.001), and Eddie had 6/6 (p <.01). Across the four participants with follow-up data, 8/8 data points fell above the predicted lines (p <.01) Disruptive behavior composite (DBC) The characteristics of the dependent variable DBC required a statistical approach that could account for autocorrelation. In addition, the statistical approach needed to model changes in both the level of behavior as well as the slope of the behavior. Hierarchical linear modeling (HLM) was selected because it allows for this type of analysis (DeLucia & Pitts, 2006; Raudenbush & Bryk, 2002). This HLM statistical approach evaluated: (1) whether there was a stable baseline before intervention began, (2) whether the introduction of the behavioral treatment package (BTP) resulted in a change in the slope of DBC observed, (3) whether the introduction of BTP resulted in a change in the level of DBC observed, and (4 and 5) whether the follow-up data points also demonstrated a change in slope and/or level from baseline. The analysis was conducted with restricted maximum-likelihood estimation, which is less biased in small samples (Singer & Willett, 2003). After starting with a basic (intercept only) model, we introduced each predictor sequentially, and removed predictors as they lost interpretability. This process is presented in Table 1. Initial models include few predictors, and the final model (Model J) presents a parsimonious and statistically supportable model of the observed data. That is, interpretation of Model J allows a statistical answer to each of the 5 research questions above. First, Model J has a significant intercept. This specifically supports the presence of a stable baseline prior to intervention. Second, the inclusion of BTP-slope and the noninclusion of BTP-level suggest that behavior significantly decreases during intervention, and is more accurately modeled as a gradual rather than immediate change at the introduction of the intervention. Finally, the inclusion of FU-level and noninclusion of FU-slope support the conclusion that at follow-up, participants demonstrated significantly fewer disruptive behaviors than at baseline Analyses of secondary measures Actigraphy Results from actigraphs worn by the participants are presented in Table 2. Data from the last two weeks of baseline are presented alongside the last two weeks of treatment for each of the main sleep variables calculated by the actigraphy. The results from total sleep time (TST) show that, on average, the participants were sleeping about 30 min more each 24 h period at the conclusion of treatment. Even after treatment, however, this sample of children with AS were sleeping, on average, about h less than 2 6-year old children from comparison groups (Goodlin-Jones et al., 2008), presented in Table 2. We removed the data from our 11-year-old participant (Cindy) to allow more direct comparisons between the Goodlin-Jones data and our 2 6-year-old participants, but the relative differences remained unchanged so Cindy s data are included in the results. Actigraphy results also show that sleep efficiency was only slightly improved. In addition, the actigraphy latency data show that the participants were falling asleep about 5 min faster each night. Finally, WASO actigraph data show that the participants were waking about the same after treatment Behavioral assessments Fig. 2 shows the results of the parent reported behavioral assessments. These results show that in baseline, the participants were described by their parents on the Developmental Behavior Checklist (DBC-P) as exhibiting marked nonsleep behavior problems that were consistent with individuals who have clinically significant behavior problems (e.g., tantrums, repetitive behaviors, self-injury, sensitivity to changes in routine). After treatment, the scores on the DBC-P showed a reduction in the incidence of non-sleep behavior problems. Fig. 2 also shows results from the Abbreviated Children Sleep Habits Questionnaire. Parents reported that prior to treatment, their children showed sleep habits consistent with other children who are considered problem sleepers or

8 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Table 1 Comparison of alternative HLM models of the disruptive behavior composite. Model Intercept b Time BTP a slope BTP a -level FU-slope FU-level Deviance A ** B ** * C ** ** *** *** D ** ** E ** *** y F ** * y y G ** * y y H ** * y I ** * *** J ** *** *** Time represents the change in DBC over time, not accounting for treatment. BTP-slope represents the change in slope during intervention phases. BTP-level represents the change in level of disruptive behavior during intervention phases. FU-slope represents the change in slope during follow-up. FU-level represents the change in level of disruptive behavior during follow-up observations. a BTP = behavioral treatment package. b Intercept represents the level of disruptive behavior at time zero, set at first week of baseline. * p <.05. ** p <.01. *** p <.001. y p <.10. Table 2 Pre-post actigraphy measures of total sleep time (TST), sleep efficiency, sleep latency, and time awake after sleep onset (WASO). TST (h) Efficiency (%) Latency (min) WASO (min) Angelman Baseline 7: Angelman PostTx 8: Autism sample a 10: DD sample a 11: TD sample a 11: a From Goodlin-Jones et al. (2008) (DD: developmentally delayed without autism and TD: typically developing). children with developmental disabilities. However, after treatment, parents reported that the participants more closely resembled individuals who are considered nonproblem sleepers. Finally, results from the AARP show that overall, the parents found the treatment to be highly acceptable. Each of the parents endorsed strong agreement with each of the questions on the AARP, resulting in an average score of 47.5 (possible 48 total), indicating that they liked the treatment, that they found it effective, that they intended to keep using it, and that they would recommend it to others. In addition, the parents often wrote anecdotal endorsements of treatment such as, Wow, it s so amazing to sleep all night and She has slept through the night for probably 6 10 weeks now. First time in 12½ years Discussion The results of this investigation provide strong support for the notion that the sleep problems common to AS can be successfully treated with behavioral intervention. By enhancing the sleep environment, modifying the sleep schedule, and altering parent child interactions during bedtime and throughout the night, the behavioral treatment package produced immediate improvements in children s ability to fall asleep independently. This is a clinically significant outcome because establishing independent sleep initiation is the cornerstone of all empirically supported treatments for sleep disturbance in young children (Mindell, Emslie, et al., 2006; Mindell, Kuhn, et al., 2006). The intervention also produced gradual but marked reductions in children s disruptive bedtime behavior and increased their sleep duration on average, by 30 min a night. Indeed, by the conclusion of treatment, parents described their children as nonproblem sleepers and even noted improvements in nonsleep-related daytime behavior. Equally important, parents reported that they felt the treatment was acceptable and effective and worth recommending to others. These results are important because they confirm that sleep habits, even in children with a neurodevelopmental genetic condition such as AS, are highly influenced by the environment, just as with typically developing children and children with other developmental disabilities. This suggests that behavioral interventions can be an effective alternative to medication and should provide the empirical evidence that some providers and parents have been seeking regarding the efficacy of behavioral interventions for treating sleep problems in children with AS.

9 [(Fig._2)TD$FIG] 684 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Problem sleeper (55) 50 Clinically significant cut off (46) Children with DD (50.5)* Non-Problem sleeper (41) Total Score 30 pre post DBC-P CSHQ * see Goodlin-Jones et al., 2008 Fig. 2. Mean scores from Developmental Behavior Checklist (DBC-P) and Abbreviated Children Sleep Habits Questionnaire (CSHQ). Confidence in these conclusions is strengthened in a number of ways. First, changes in disruptive bedtime behaviors and independent sleep onset were stable and/or deteriorating in baseline and subsequently began to change when and only when treatment was introduced. This is a compelling demonstration of experimental control in a small n multiple baseline design and increases confidence that the behavioral treatment package produced the observed changes. Second, improvements were lost during two unplanned withdrawals of treatment and recovered when treatment was reinstated. These two replications of the treatment effect are also important demonstrations of experimental control. Finally, additional analyses confirmed that the observed changes were statistically significant. Although these results are highly encouraging, it is difficult to say how well the outcomes might generalize to other children with AS. For example, each of the five participants presented with a similar type of sleep disturbance typically classified as Behavioral Insomnia of Childhood Sleep Onset Association Type (International Classification of Sleep Disorders, 2nd edition). Although the behavioral insomnias are predominant in young children with neurodevelopmental disabilities (Souders et al., 2009), it is unlikely that this type of insomnia is representative of all children with AS. Note, however, that the participants in this study were selected from a larger sample of research recruits because these five all demonstrated disruptive bedtime behaviors, not because they showed sleep onset insomnia. Finally, although there were only five participants in this study, AS is a rare genetic condition, making large group studies difficult to implement. In addition, small n research designs are not intended to provide strong external validity (i.e., generalization). Instead, they are intended to emphasize interval validity; that is, a clear demonstration of a functional relationship between the treatment and the outcomes. The small n design accomplished that by demonstrating that, in the presence of stable baseline responding, the DBC and ISO changed when and only when treatment was introduced, thereby enhancing one s confidence that the treatment produced the observed changes (i.e., internal validity). One may also question the generalizability of the results because the parents of these participants were highly motivated. Evidence of this strong motivation appears in their willingness to keep daily sleep diaries and make their child wear actigraphs daily for an average of weeks. Further evidence appears in their willingness to try ignoring or at least reducing their responsiveness to disruptive behaviors, something that is often extremely difficult to do, especially in the middle of the night. Finally, parents who were asked to reduce a child s allowable time-in-bed were committing to increased supervision requirements each evening. Yet these parents may be very representative of parents of children with AS, many of whom have been struggling with sleep-related problems for many years and, as a result, are highly motivated as well.

10 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) In spite of what were perceived as high levels of motivation, parents did find sticking with treatment recommendations to be challenging. Two of the five families stopped adhering to treatment, even after marked changes had been observed in child sleep behavior, and all families required weekly contact with the investigators to maintain good treatment integrity. This should not be surprising since parents were often being asked to alter well-established (even if ineffective) strategies for managing bedtime struggles. In addition, the intervention had multiple components, making the treatment more complex, a known risk factor in treatment nonadherence (e.g., Rapoff, 1999). Thus, a behavioral treatment package like this may be effective but challenging to implement. It is important to note that the behavioral treatment package was designed to be flexible and individualized. This was done to allow more or less emphasis on each of the three treatment components to match the needs of each child and the parents involved and to reduce unnecessary complexity. For example, the sleep schedule was not altered if it already matched the child s sleep need and if it was appropriately timed. However, the three treatment components were reviewed with all parents to insure they understood the importance of each and to insure that while changes were being made in some areas, parents were not inadvertently changing well-functioning aspects of their child s routine or sleep schedule. The next steps for this line of research should focus on generalizing these promising results. First, the treatment team included a behavioral sleep specialist; therefore it is not clear to what extent non-experts could implement the behavioral treatment package and achieve similar results. In addition, this current investigation was not designed to answer questions about external validity. A randomized controlled trial would help answer questions about generalizability. Finally, this study focused on younger children, whose parents still have considerable control over the sleep environment, sleep schedule, and nighttime interactions. Future research will need to explore whether a behavioral sleep intervention can be successful with older children, adolescents, or young adults with AS. 5. Conclusion Sleep problems, which are one of the most common behavior problems associated with AS, can lead to fatigue, irritability, distress, and adverse effects on social functioning of the entire family. Although behavioral interventions are well established as a means of treating sleep problems in typically developing children, many clinicians have been reluctant to attempt these methods to treat sleep problems in those with Angelman s Syndrome because of fears that structural or chemical abnormalities would interfere. The results from this study show that those fears can now be put to rest. Results show that the behavioral treatment package produced immediate improvements in children s ability to fall asleep independently as well as marked reductions in children s disruptive bedtime behavior. Furthermore, parents felt the treatment was acceptable and even noted improvements in nonsleep-related daytime behavior. Clinicians should now feel more comfortable recommending behavioral sleep interventions for children with AS. Acknowledgements This research was supported in large part by the Angelman Syndrome Foundation. Support was also provided by Project #8188 from the Maternal and Child Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services and in part by grant 90DD0533 from the Administration on Developmental Disabilities (ADD), Administration for Children and Families, Department of Health and Human Services. References Barlow, D. H., Nock, M. K., & Hersen, M. (2009). Single case experimental designs: Strategies for studying behavior change (3rd ed.). Boston: Allyn & Bacon. Bruni, O., Ferri, R., D Agostino, G., Miano, S., Roccella, M., & Elia, M. (2004). Sleep disturbances in AS: A questionnaire study. Brain & Development, 26, Campbell, J. M., & Herzinger, C. V. (2010). Statistics and single subject methodology. In D. L. Gast (Ed.), Single subject methodology in behavioral sciences. New York: Routledge. Clayton Smith, J. (1993). Clinical research on AS in the United Kingdom: Observations on 82 affected individuals. American Journal of Medical Genetics, 46, Conant, K. D., Thibert, R. L., & Thiele, E. A. (2009). Epilepsy and the sleep wake patterns found in AS. Epilepsia, 50, Corkum, P., Tannock, R., Moldofsky, H., Hogg-Johnson, S., & Humphries, T. (2001). Actigraphy and parental ratings of sleep in children with attention-deficit/ hyperactivity disorder (ADHD). Sleep, 24, DeLucia, C., & Pitts, S. C. (2006). Applications of individual growth curve modeling for pediatric psychology research. Journal of Pediatric Psychology, 31, Didden, R., Korzilius, H., Smits, M. G., & Curfs, L. M. (2004). Sleep problems in individuals with AS. American Journal of Mental Retardation, 109, Didden, R., & Sigafoos, J. (2001). A review of the nature and treatment of sleep disorders in individuals with developmental disabilities. Research in Developmental Disabilities, 22, Einfeld, S. L., & Tonge, B. J. (2002). Manual for the developmental behaviour checklist (2nd ed.). Authors: University of New South Wales & Monash University. Goodlin-Jones, B. L., Sitnick, S. L., Tang, K., Liu, J., & Anders, T. F. (2008). The children s sleep habits questionnaire in toddlers and preschool children. Journal of Developmental and Behavioral Pediatrics, 29, Griffith, G. M., Hastings, R. P., Oliver, C., Howlin, P., Moss, J., Petty, J., et al. (2011). Psychological well-being in parents of children with Angelman, Cornelia de Lange and Cri du Chat syndromes. Journal of Intellectual Disability Research, 55, Horsler, K., & Oliver, C. (2006). The behavioural phenotype of AS. Journal of Intellectual Disability Research, 50, Kuhn, B. R. (2007). Sleep disorders. In M. Hersen & J. C. Thomas (Eds.), Handbook of clinical interviewing with children (pp ). New York: Sage Publications. Kuhn, B. R. (2011). The excuse-me drill for bedtime problems: A behavioral protocol to promote independent sleep initiation skills and reduce bedtime problems in young children. In M. Perlis, M. Aloia, & B. R. Kuhn (Eds.), Behavioral treatments for sleep disorders: A comprehensive primer of behavioral sleep medicine treatment protocols (pp ). Boston: Elsevier/Academic Press.

11 686 K.D. Allen et al. / Research in Developmental Disabilities 34 (2013) Mindell, J. A., & Durand, V. M. (1993). Treatment of childhood sleep disorders: Generalization across disorders and effects on family members. Special issue: Interventions in pediatric psychology. Journal of Pediatric Psychology, 18, Mindell, J. A., Emslie, G., Blumer, J., Genel, M., Glaze, D., Ivanenko, A., et al. (2006). Pharmacologic management of insomnia in children and adolescents: Consensus statement. Pediatrics, 117, e1223 e1232. Mindell, J. A., Kuhn, B., Lewin, D. S., Meltzer, L. J., & Sadeh, A. (2006). Behavioral treatment of bedtime problems and night wakings in infants and young children. Sleep, 29, Mindell, J. A., & Owens, J. A. (2003). A clinical guide to pediatric sleep. Philadelphia: Lippincott, Williams & Wilkins. Morgenthaler, T. I., Owens, J., Alessi, C., Boehlecke, B., Brown, T. M., Coleman, J, Jr. et al. (2006). Practice parameters for behavioral treatment of bedtime problems and night wakings in infants and young children. Sleep, 29, Owens, J. A. (2006). When child can t sleep, start by treating the parents. Current Psychiatry5 pp , 27 30, Owens, J. A., Spirito, A., & McGuinn, M. (2000). The children s sleep habits questionnaire (CSHQ): Psychometric properties of a survey instrument for school-aged children. Sleep, 23, Pelc, K., Cheron, G., Boyd, S. G., & Dan, B. (2008). Are there distinctive sleep problems in AS? Sleep Medicine, 9, Piazza, C. C., & Fisher, W. W. (1991). Bedtime fading in the treatment of pediatric insomnia. Journal of Behavior Therapy and Experimental Psychiatry, 22, Piazza, C. C., Fisher, W. W., & Sherer, M. (1997). Treatment of multiple sleep problems in children with developmental disabilities: Faded bedtime with response cost versus bedtime scheduling. Developmental Medicine & Child Neurology, 39, Rapoff, M. A. (1999). Adherence to pediatric medical regimens. New York: Kluwer Academic/Plenum Publishers. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications. Sadeh, A. (2011). The role and validity of actigraphy in sleep medicine: An update. Sleep Medicine Reviews, 15, Sadeh, A., Sharkey, K. M., & Carskadon, M. A. (1994). Activity-based sleep wake identification: An empirical test of methodological issues. Sleep, 17, Siegel, J. M. (1995). Phylogeny and the function of REM sleep. Behavior and Brain Research, 69(1 2), Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York: Oxford University Press. Souders, M. C., Mason, T. B., Valladares, O., Bucan, M., Levy, S. E., Mandell, D. S., et al. (2009). Sleep behaviors and sleep quality in children with autism spectrum disorders. Sleep, 32, Spielman, A. J., Saskin, P., & Thorpy, M. J. (1987). Treatment of chronic insomnia by restriction of time in bed. Sleep, 10, Summers, J. A., Allison, D. B., Lynch, P. S., & Sandler, L. (1995). Behaviour problems in AS. Journal of Intellectual Disability Research, 39, Summers, J. A., Lynch, P. S., Harris, J. C., Burke, J. C., Allison, D. B., & Sandler, L. (1992). A combined behavioral/pharmacological treatment of sleep wake schedule disorder in AS. Journal of Developmental and Behavioral Pediatrics, 13, Tarnowski, K. J., & Simonian, S. J. (1992). Assessing treatment acceptance: The abbreviated acceptability rating profile. Journal of Behavior Therapy and Experimental Psychiatry, 23, Walz, N. C., Beebe, D., & Byars, K. (2005). Sleep in individuals with AS: Parent perceptions of patterns and problems. American Journal of Mental Retardation, 110,