Extending Caregiver Training Via Telecommunication for Rural Icelandic Children With Autism

Abstract

This study describes the development and evaluation of a behavioral parent training protocol via telecommunication for three parents of preschool children with autism, with limited access to behavioral expertise. A single-subject, multiple baseline experimental design across child behaviors, replicated across parents, was used to evaluate the effects of the training protocol. Dependent measures were collected via telecommunication for most assessments and included parent and child responses during naturalistic play. During intervention, the parents were taught methods to increase their child’s sociocommunicative behavior. All targeted skills increased during intervention. The increase was maintained with two families and some generalization to other settings occurred. The results indicate that training via telecommunication is a viable approach for rural families with low speed Internet connection. Continued development and refinement of telehealth training methods is discussed in context of technological challenges and procedures that fit telecoaching for rural families.

Keywords

parent training telehealth preschool children autism rural families

Families of children with autism are more likely to encounter difficulties in accessing expert services including evidence-based early behavioral intervention (National Autism Center, 2015), than families of children with other disabilities. Rural families experience even more barriers in accessing the expertise required for successful treatment (Chiri & Warfield, 2012; Hoogsteen & Woodgate, 2013; Murphy & Ruble, 2012; Pickard & Ingersoll, 2016; World Health Organization [WHO], 2007). Furthermore, transportation to and from the limited pool of experts is difficult (Birkin, Anderson, Seymour, & Moore, 2008; Elder, Brasher, & Alexander, 2016). For example, winter travel conditions make it sometimes impossible for rural families to travel from their local community to access necessary services. Similarly, mileage or other travel costs and travel time for experts, which would otherwise be spent on providing services, are also a barrier for service agencies and institutions. However, families of children with autism who received family-centered and coordinated care are more likely to experience increased access to the necessary expertise (Cheak-Zamora & Farmer, 2016).

The importance of parents in their children’s development and implementation of intervention has been well documented in the behavioral parent training research literature (Strauss, Mancini, The SPC Group, & Fava, 2013; Suppo & Floyd, 2012). This includes studies documenting the parents’ role as main behavior change agents in comprehensive programs or enhancing generalization and maintenance of skills (see, for example, Strauss et al., 2013). Also included are studies using single-subject experimental design in looking at the effectiveness of directly training parents as main behavior change agents, teaching their children social communication skills, play, adaptive skills, and managing challenging behavior (see, for example, Fettig & Barton, 2014; Gena, Galanis, Tsirempolou, Michalopulou, & Sarafidou, 2016; Patterson, Smith, & Mirenda, 2012; Rakap & Rakap, 2014). Without skilled parent involvement, children’s outcomes are significantly worse than if parents are involved and skilled in promoting the development of their child with autism (Strauss et al., 2013). Furthermore, the research literature also indicates that parent participation in treatment reduces parent stress and has a positive effect on the quality of interactions between parent and child (Brookman-Frazee, 2004; Estes et al., 2014). However, parent training research studies that include rural families of children with autism are scarce.

To enable rural families increased access to information and intervention based on the parent training research literature, experts and researchers have used different strategies (Boisvert, Lang, Andrianopoulos, & Boscardin, 2010; Farmer & Reupert, 2013; Reese et al., 2013), with the majority of experimental evaluations involving telehealth procedures. In telehealth (synonymous with telemedicine), information and communications technology (ICT) is used to provide health care to people who reside in a different physical location than the health care provider (WHO, 2010b). Although a promising innovation in health care, there is still limited research evidence in behavioral interventions and very few of the studies have a targeted focus on families living in rural areas (Hamad, Serna, Morrison, & Fleming, 2010; Jang et al., 2012; Kobak et al., 2011; Simacek, Dimian, & McComas, 2017; Suppo & Mayton, 2014; Wainer & Ingersoll, 2015). It is important, however, to include rural families as participants as there may be unique variables likely to have an effect on progress, such as limited access to regular feedback from a trainer and the technical challenges possibly influencing the consulting process via telecommunication.

The existing telehealth behavioral parent training studies have approached the training in two ways: with web-based self-directed learning modules and via teleconferencing that has included coaching with guided feedback. Some studies have also combined these two methods. Only two studies to date have taught parents via video coaching combined with minimal in situ training (e.g., on site) at the start of the intervention (Guðmundsdóttir, Sigurðardóttir, & Ala’i-Rosales, 2017; McDuffie et al., 2013).

Most of the telehealth evaluations have focused on teaching parents to assess and decrease challenging behavior (see, for example, Lindgren et al., 2016; Machalicek et al., 2016; Simacek et al., 2017) and to increase social communication skills with naturalistic strategies (see, for example, Guðmundsdóttir et al., 2017; Ingersoll & Berger, 2015; Ingersoll, Wainer, Berger, Pickard, & Bonter, 2016; Meadan et al., 2016; Simacek et al., 2017; Vismara, McCormick, Young, Nadhan, & Monlux, 2013; Wainer & Ingersoll, 2015). Other telehealth studies have taught parents about discrete trial training and general behavior change strategies (Hamad et al., 2010; Heitzman-Powell, Buzhardt, Rusinko, & Miller, 2014; Jang et al., 2012; Kobak et al., 2011; Suppo & Mayton, 2014).

Numerous parent training studies have demonstrated the effectiveness of in situ coaching (instructions, modeling, and feedback; Barton & Fettig, 2013; Kaiser, Hancock, & Trent, 2007; Schreibman et al., 2015). Modeling is especially important if the skills are new to the parent, the parent does not have a strong history of rule-following behavior, the child engages in challenging behavior, or skills to be learned are complex. However, as the child is not physically with the trainer, it is difficult for the trainer to model correct responses during telecoaching Many of the telehealth parent training studies have addressed this issue by including video models of the procedures to be viewed asynchronously during self-directed modules (McDuffie et al., 2013; Meadan et al., 2016; Suppo & Mayton, 2014; Vismara et al., 2013; Wainer & Ingersoll, 2015). Only two studies included the modeling during video coaching (Machalicek et al., 2016; Meadan et al., 2016). However, both studies included modeling with people other than the parent and the child. Another approach using video modeling that appears to be effective is videotaped self-observation where the parent observes her own behavior interacting with the child, from a video recording (Fox & Westling, 1991; Kaiser, Hancock, & Nietfeld, 2000; Reamer, Brady, & Hawkins, 1998; Townley-Cochran, 2014). This approach was included as part of the training package in the current study during the telecoaching sessions. It involved each parent watching a video clip of her interaction with the child during the current training session. At least one study suggests this is an effective alternative during telecoaching (Guðmundsdóttir et al., 2017).

The current study was the second project in a series of experiments conducted with families living in rural areas with access to low-speed Internet connection. As in the previous research (Guðmundsdóttir et al., 2017), this study was based on the procedures from a community-based parent training program, Sunny Starts, at the University of North Texas (Ala’i-Rosales, Cermak, & Guðmundsdóttir, 2013). This is one of several such programs based on previous research. It is informed by the evidence base of successful responsivity-based parent training approaches. The program uses naturalistic teaching methods to teach meaningful child behaviors (such as social communication skills and play) by using frequent preference assessments, environmental arrangements, and responsive parental consequences. Furthermore, the program includes monitoring of family enjoyment as well as frequent evaluation of family progress. The methods taught to the parent in Sunny Starts and in this study are represented by the acronym and mnemonic Teaching “DANCE,” which describes teaching strategies based on naturalistic behavioral intervention procedures. The Teaching “DANCE” consists of five components: (D—decide) deciding whether it is a good time to teach, (A—arrange) identifying the child’s preferences and arranging these events to allow teaching, and (N—now) responding to the child’s target skill in an immediate and responsive manner. In addition, as part of the Teaching “DANCE,” the parent is taught to (C—count and contemplate) monitor progress and pursue interactions that were (E—enjoy) effective and enjoyable to both the parent and child.

The purpose of this study was to further develop and evaluate a telehealth training protocol for rural families with limited access to behavioral expertise. Furthermore, the aim was to evaluate the effects of a behavioral parent training package conducted via telecommunication on the frequency of parent’s teaching skills and the child’s sociocommunicative behavior (e.g., social attending and requesting).

Method

To extend the previous research, the current study examined the effects of a behavioral parent training package (Ala’i-Rosales et al., 2013; Guðmundsdóttir et al., 2017) on both adult and child skills during an initial in situ training with follow-up and continued telecommunication coaching. A videotaped self-observation was included in the training package to enhance parent learning of the “Teaching DANCE.” Unlike the previous study, all of the training sessions, following the initial in situ training, were conducted via telecommunication with all of the families. To analyze the effects of training, a valid single-subject experimental design was used with precise behavioral measures. As in the previous experiment, the relationship between the parent skills and the children’s social communication skills on the same dimension of behavior was analyzed.

Participants

Three families participated in this experimental study. The families were selected in collaboration with the State Diagnostic and Counseling Center according to predetermined selection criteria regarding age (3–6 years), diagnosis, and geographic location of the family. Parent criteria included interest and availability, no simultaneous participation in another behaviorally based program, and access to the technical resources necessary for participation. Criteria also included living in a rural location at a distance of 35 to 165 km and 259 to 311 km from the trainer’s location. The parents did not receive reimbursement for participating in the study and carried all costs for Internet use during the course of their assessment and training. They gave informed consent before initiation of the study and understood the purpose of the study was to develop and evaluate a procedure for effective telehealth with rural families in their country.

All the families were typical middle-class families of Icelandic origin and spoke Icelandic. They all lived in rural villages or towns in Iceland and had limited access to the necessary expertise for evidence-based intervention for their children with autism. Each of the families included two to three preschool- and elementary-age children. Two of the families had more than one child diagnosed with autism. Most of the parents had completed vocational training as well as the first 2 years of upper secondary education according to the European education system (equivalent to 11th and 12th grades of high schools in the United States). The parents were all working full- or part-time in their field of study. One parent was pursuing a graduate degree.

All the children were diagnosed with autistic disorder (code F84.0) according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, 2010 (ICD-10) at the State Diagnostic and Counseling Center. Before the beginning of each family’s participation in the study, each child was assessed with the revised version of Wechsler Preschool and Primary Scale of Intelligence–Revised (WPPSI-R; Wechsler, 1989) and with Vineland Adaptive Behavior Scales, Second Edition (VABS-II; Sparrow, Cicchetti, & Balla, 2005). The assessment results are described in Table 1.

Table 1.

Assessment Information for the Children Participants.

Assessment tool	Ari	Gunnar	Daníel
WPPSI-R^is
Performance scale	106	69	65^a
Verbal scale	95	61	50^a
Full-scale IQ	95	60	52^a
VABS-II
Communication domain	69	69	61
Socialization domain	63	77	72

Note. WPPSI-R = Wechsler Preschool and Primary Scale of Intelligence–Revised; ^is = Icelandic translation and standardization of the assessment; VABS-II = Vineland Adaptive Behavior Scales, Second Edition.

Test scores inconclusive due to challenges with cooperation and comprehension. Not all subtests completed.

All of the children attended preschool daily and received various eclectic special education services at their preschool for a few hours every week. These services included speech therapy, individual instruction of academic and daily living skills, and shadowing during group activities. Furthermore, the services included visual support throughout the child’s school day. None of the parents or children had participated in a behavioral intervention. The majority of the children’s intervention and family support came from the children’s preschools where there was no expertise in behavioral interventions for children with autism. Two of the children received eclectic services at school from trained social education workers and paraprofessionals. The third child received services from paraprofessionals only.

All of the families had busy schedules that included school, work, and extracurricular activities. Two of the families had daily transport between their rural home, school, and work and the children’s extracurricular activities. For two of the families, work, school, and other activities were either all or partly located in another town, in about 30 min driving distance. During the wintertime, when the study was conducted, this transport often became difficult because of weather and road conditions. The families all had close ties with their extended family and received some support from them, mostly from grandmothers. Two of the families also received additional and regular respite care for the child with autism. Mothers were the primary change agents and fathers participated in the initial assessment and goal setting but not in the training. Siblings were often present but did not participate in training.

Trainer

The first author served as the parent trainer in the study. She is an instructor in behavior analysis at a university, holds a master’s degree in behavior analysis, is a board certified behavior analyst (BCBA) and had over 10 years’ experience in early and intensive behavior intervention for children with autism, including an emphasis on naturalistic teaching methods. During the study, she received supervision from the second and third authors, both doctoral-level behavior analysts with more than 60 years combined experience in interventions with families.

Observers

A graduate student and the trainer collected the behavioral data in the study. The graduate student was working toward a master’s degree in psychology at the university where the study was conducted. She had completed an undergraduate course on basic behavior principles and was trained by the first author in observing and scoring behaviors.

Family 1

Saga and her son Ari (pseudonyms) were the first family to participate in this study. Saga was 33 years old at the beginning of the study. She was pursuing a graduate degree in engineering and working part-time. Ari was 3 years and 11 months at the beginning of the study. He had also received a diagnosis of unspecified disorder of psychological development (code F89, according to ICD-10; WHO, 2010a) 7 months before the beginning of the study. At the beginning of the study, Ari initiated communication in three- to four-word utterances and had limited social attending.

Family 2

Katla and her son Gunnar (pseudonyms) were the second family to participate in the study. Katla was 31 years old at the beginning of the study and had completed 1 year of upper secondary education (equivalent to 11th grade of high school in the United States) as well as vocational training. During the time of the study, she stayed at home but accepted freelance projects in her area of vocational training. Katla had attended a daylong informational workshop on behavior intervention methods in autism intervention 3 years before her participation in this study. Gunnar was 4 years and 7 months at the beginning of the study. He had also received a diagnosis of unspecified disorder of psychological development 10 months before the beginning of the study and a diagnosis of expressive language disorder (code F80.1 according to ICD-10; WHO, 2010a). Gunnar was verbal at the beginning of the study and communicated in four- to five-word utterances. However, he did not frequently initiate communication with his mother, but mainly responded to her initiations. He also did not socially attend to his mother.

Family 3

Thora and her son Daníel (pseudonyms) were the third family to start participation in the study. Thora was 29 years old at the beginning of the study and had completed 2 years of upper secondary education (equivalent to 12th grade of high school in the United States). During the time of the study, she was working full-time. Daníel was 4 years and 1 month at the beginning of the study. He had also received a diagnosis of unspecified intellectual disabilities (code F79, according to ICD-10; WHO, 2010a). At the beginning of the study, Daníel was able to communicate with his family in what appeared to be one- to three-word understandable utterances, part of which were echoic; however, most of his utterances were vocal or incomprehensible. Some of his utterances included phrases in English. However, Daníel did not frequently initiate communication with his mother and he had limited social attending.

Setting, Schedule, and Materials

Sessions took place in the families’ living rooms and in Daníel’s case also in his playroom/bedroom. Baseline and training sessions with all families took place twice per week on average, mostly on weekdays and afternoons, with some exceptions (one third of the sessions with Saga and Ari took place on weekend mornings and the first 13 sessions with Thora and Daníel took place in early mornings). The trainer was located in the same room as the parent and child, during all in situ sessions, and was video recording and collecting data on target behaviors. During all telecommunication sessions, the trainer was located in her university office. The children’s preferred toys and activities were available during all sessions (e.g., blocks, figurines and dolls, puzzles and beads, books, vehicles, garages and train tracks, colors, games). During in situ data collection, the parent trainer used paper data sheets and a pencil.

Technical Tools

Equipment and technical setup

All families used their personal laptop during telecommunication sessions and a Microsoft LifeCam Cinema web camera. The web camera was attached to the computer screen, located on a dining table, chair, or a small table and in Daníel’s case on a windowsill in his bedroom. The laptops were placed so that the dyads were located 1- to 1.5 m from the front of the camera. Thora and Daníel were sometimes located up to 4 m from the camera. An external Universal Serial Bus (USB) MXL®-UCHAT (AC-406) boundary microphone, placed on the floor or a table close to each dyad, was used by all families. A video camera and a compact camera were used for recording in situ assessments in the families’ homes. Sennheiser HD202 headphones were used for data collection from video samples of assessments.

Software

Skype™ videoconferencing software was used for the telecommunication sessions and Windows Media Player® video playing software was used for data collection from video samples of assessments and showing each parent samples from video assessments during training. Telecommunication sessions were recorded with VODburner recording Software developed by Netralia Pty Ltd. During in situ sessions, a timer on a Samsung GTI8160 smartphone was used to keep track of the assessment time. A Free Desktop timer™ developed by Informer Technologies Inc., was used for the same purpose during the telecommunication sessions.

Closed file system and website

All families had access to their information via a closed file system on an encrypted website only accessible to the family and the trainer with a personal username and a logon password. The website was designed for this study by the first author and two employees at Stefna Software. The website also included instructions on technical matters, videos with general information about the study and its procedures, and written information and website links about behavior analysis and autism intervention.

Internet connection

Ari’s family used a 3G wireless Internet connection via a mobile WiFi device until Session 24 and the same connection via a wireless box router from Session 24. Gunnar’s and Daníel’s families used an asymmetric digital subscriber line (ADSL) wireless Internet connection available in their homes. Despite potential low bandwidth of the families’ Internet connection, all fulfilled the minimum entry requirements and were able to use the Skype™ software for all video calls. The trainer used a high-speed Internet connection via fiber optic cable.

Measures

Behavioral measures were based on the Sunny Starts observation code (Cermak, 2011; Townley-Cochran, 2014), other measures from the naturalistic teaching and parent training literature (Ingersoll & Dvortcsak, 2010), and a handbook on grammar (Höskuldur Þráinsson, 2006). Child and parent behaviors were scored for the first 5 min during all baseline and training sessions. Generalization probes were conducted for Saga’s and Ari’s skills and Katla’s and Gunnar’s skills in other settings in the home (during family mealtimes and baking in the kitchen). The behaviors were scored by paper and pencil from a video recording by the first author who served as the primary observer for communication measures with all families and social attending measures for Daníel’s family. A graduate student served as the primary observer for social attending measures with Ari’s and Gunnar’s family. The first author served as the secondary observer for social attending measures with Ari’s and Gunnar’s family. The graduate student served as a secondary observer for communication measures with all families. Below is a summary of parent and child measures. The complete observation code is available from the first author.

Parent behaviors

In each training session, the parent and the teacher were taught methods to evoke initiations from the child and respond to the child’s initiations. They were taught the Teaching “DANCE” (previously described) as a general strategy to increase children’s sociocommunicative behaviors. For the purpose of this study, the correct implementation of two components of the Teaching “DANCE,” Arrange and Now, were observed and scored as one unit (i.e., as teaching episodes). Each teaching episode included three behaviors: identifying high preference events, environmental arrangement, and responsive event delivery. Identifying high preference events was scored when the parent selected an event to which the child showed an interest. For example, the parent presented a choice of a bubble gun and a doll to the child and the child reached for the bubble gun. Environmental arrangement was scored when access to a preferred event was regulated or arranged so that the child could not contact the event at that current time, needed assistance with the event, or came into contact with an unexpected situation regarding the event. For example, the parent held the bubble gun but did not activate the blower until the child engaged in the target behavior. Responsive event delivery was scored when the parent gave the child the apparent preferred event within 10 s after the end of the child’s target response. For example, the parent blew a bubble when the child said “bubble” and began to pop bubbles with the child. Expanding language was scored when the parent immediately added, while providing a responsive event, a single word or a sentence to the child’s target response.

Child behaviors

Social attending was scored for all children during all interactions with their mothers. Social attending included any orientation of the child’s face or eyes to the eyes or face of the parent. The behavior ended when child’s face or eyes oriented away from the parent’s eyes or face. The observer also scored, but differently, when it was not clear whether social attending had occurred due to lack of video quality or obscured view of the child’s face. Requesting was scored for all children and included spoken words, word approximations, phrases or sentences that were directed to the parent, asking for an item or information (e.g., questions), specifying an action or an activity to be completed, or requesting information, permission, or attention. A new request was scored if it occurred 3 s or later after the end of the last verbalization of the child, when the child asked a different question with a different content and when the child made a request following the parent’s nonverbal action. Requesting was a target behavior for all three children. In addition to requesting, modifiers in requests were scored for Ari. Modifiers are words (e.g., an adjective, adverb, demonstrative pronoun, or quantity) that accompany another word (noun or verb) and give that word a more elaborate meaning (Höskuldur Þráinsson, 2006). For example, the yellow car, this house, driving fast, I want two cars. Number of words in each request were also counted for Ari and Daníel. One word included all understandable words or approximations of a word in the child’s request that were different grammatical units, such as noun, verb, adjective, and pronoun. Exceptions included short versions of phrases, common in the families’ native language that refer to location of an object. These versions were scored as one word. When the scorer could not hear clearly or understand the child’s verbal behavior, he scored it as an unintelligent verbalization.

Research Design

A multiple baseline experimental design across child’s skills, replicated with two families, was used to evaluate the effects of the training. Each parent taught her child two skills where (a) the parents’ and children’s behaviors were observed repeatedly during baseline and then (b) each of them received training, first in situ, and then continued via telecommunication.

Interobserver agreement (IOA)

IOA calculations were based on the occurrence of the parent’s Teaching Episodes (a free operant) during each 5-min assessment. An episode-by-episode IOA of scored occurrences and nonoccurrences of the participants’ target behaviors was calculated using the following formula: Number of episodes agreed ÷ Number of episodes agreed + Number of episodes disagreed × 100. IOA was calculated for approximately 30% of assessments for each participating dyad across each target behavior and experimental phase. A trained observer scored the assessments. The average IOA for Saga’s episodes for teaching social attending and for teaching longer phrases was 86.7% (range = 50–100) and 96.7% (range = 71.4–100), respectively. The average IOA for Katla’s episodes for teaching social attending and for teaching requesting was 86.8% (range = 58.3–100) and 82.6% (range = 0–100), respectively. The average IOA for Thora’s episodes for teaching social attending and teaching longer phrases was 85.35% (range = 42.8–100) and 100%, respectively. The average IOA for Ari’s social attending and requesting with longer phrases was 86.9% (range = 0–100) and 89.89% (range = 88.8–100), respectively. The average IOA for Gunnar’s social attending and requesting was 89% (range = 75–100) and 82.4% (range = 0–100), respectively. The average IOA for Daníel’s social attending and requesting with longer phrases was 82.49% (range = 50–100) and 85% (range = 0–100), respectively. In summary, low IOA scores were few, most of them were high, and hence the high averages in all cases.

Procedures

Intake process and rapport building

An initial meeting with each child’s parents was held via telecommunication to explain the study. Written informed consent was obtained, further information was provided to each family as necessary, and the first home visit was scheduled. At the same time, the parent trainer developed rapport with the children and their parents. For example, the trainer conversed about issues important to the parent and played with the child and assessed the child’s preference for events. Rapport building continued during the following sessions and was maintained throughout the intervention (Ala’i-Rosales et al., 2013). These meetings lasted from 40 min to 1 hr.

Preintervention assessments and goal setting

Intervention goals for the children were selected in collaboration with the parents. Generally, the goals were focused on increasing social communication skills of the child, such as eye contact and verbal initiations, as these behaviors address the core deficits of autism (American Psychiatric Association, 2013). Each family provided information about their child’s social communication by answering questionnaires based on modified and translated Sunny Starts tools (Ala’i-Rosales et al., 2013; Ingersoll & Dvortcsak, 2010). The parents also participated in setting specific goals for their child through discussion and review of assessments. Goal selection was also guided by a planning guide for children’s skills, translated into the parents’ native language and adapted for this study from Ala’i-Rosales et al. (2013). Ari’s family chose improved attending and longer phrases, Gunnar’s family chose attending and requesting, and Daníel’s family chose attending and longer phrases. Electronic versions of the questionnaires were made available for the families on the study’s website. Preassessments took approximately 4 weeks.

Baseline

Baseline assessments were conducted in situ and via telecommunication with all families. They were conducted during a preferred play activity with all families. The adults selected play materials for each assessment. They were asked to interact as they usually did with the child but received no further information or instructions regarding teaching procedures.

Training

Training consisted of a package of behavior intervention procedures based on the Sunny Starts program. The parents were taught in their native language, using a translated and adapted version of the Sunny Starts Teaching “DANCE” as previously described in this article. All training sessions lasted between 1.75 and 2 hr. Each started with several minutes of rapport building. This was followed by a 5-min assessment of the dyad’s skills conducted under the same conditions as baseline assessments. Meanwhile, the trainer recorded data on the child’s target behavior and use of targeted components of the Teaching “DANCE.”

During the first in situ training session, the teaching goals for the child were reviewed and the basis of the Teaching “DANCE” was explained and modeled by the parent trainer. Although the purpose of using telehealth was to provide an alternative to overcome distance barriers, it was decided to run one training session in situ to provide initial modeling and in person contact. During all telecommunication sessions, technical issues such as interruption of the Internet connection, bad sound, or the position of the web camera were resolved prior to the assessment and during the session as needed.

In addition to the instruction regarding teaching interactions, part of the training included a review of progress over time via review of graphs of parent and child behavior and review of video clips of the parent–child interactions during the current session. This addressed the C of the Teaching “DANCE” and focused on the parent contemplating and considering the changes in behavior they observed while engaged with the child, through review of the graphs, and by watching videos of themselves. During this time, there was also discussion of the E of the Teaching “DANCE,” that is, both the parents’ and child’s enjoyment of the teaching interactions.

During all training sessions, an Excel® graph was shown to the parent with the parent trainer sharing her screen via the telecommunication software. Meanwhile, the child engaged in a preferred activity in the same room. Then, during all telecommunication sessions, each parent continued interacting with the child, that is, she practiced the “DANCE” with guidance from the trainer. Throughout practice, the trainer taught the parent to use the Decide, Arrange, and Now components of the Teaching “DANCE” by giving positive and corrective feedback on their “DANCE” skills, along with discussion related to the parent’s and child’s target skills.

Once the parents were making progress, the video reviews were introduced. This involved a self-observation component with guided practice with the parent. Each parent watched selected episodes of her “DANCING” with her child. This was done with the parent trainer playing the video on her computer and sharing the screen via the telecommunication software.

The order of all of the training components, during each telecommunication session described above, was changed with Gunnar after Assessment 21 and Daníel after Assessment 13. The guided practice was completed right after each session’s assessment before progress was reviewed with the parent and the videotaped self-observation took place. This was done because both children had difficulty engaging in an independent activity while the parent and the trainer were viewing and discussing the parent’s progress.

At the end of all training sessions, the trainer provided the parent with an electronic written summary with specific guidelines for the “DANCE” components they had agreed should be practiced between sessions. The parent was instructed to practice the “DANCE” with the child during preferred activities, time, and place. At the start of each session, the trainer asked the parent whether they had practiced or not and noted in her session notes what the parent said.

Exit meeting and follow-up assessment

Parents were provided with a verbal and written review of the intervention procedures and outcomes upon completion of the study. Follow-up assessments were conducted with Saga and Ari and Katla and Gunnar 1 month and 3 months following the last training sessions. No direct coaching was provided but the data regarding the progress were reviewed and discussed. No follow-up assessments were conducted with Thora and Daníel due to logistical circumstances. The total training time, minus follow-up, over the course of the study was 29 hr for Saga and Ari, 14 hr for Katla and Gunnar, and 18 hr for Thora and Daníel. This took place over a time span of several months with each family: Saga and Ari spanned 1 year, Katla and Gunnar 7 months and 2 weeks, and Thora and Daníel 8 months and 3 weeks.

Results

Saga and Ari

The figures are displayed by family and contain both parent responding and child responding during each condition. Each figure displays the effects of the “DANCE” training package on frequency of parent teaching episodes and the frequency of child’s responding. Also indicated on each graph are telehealth sessions. A summary of both responding and relationships between parent and child responding is presented here.

Saga teaching social attending

Figure 1 displays the effects of the “DANCE” training package on Saga and Ari.

Figure 1.

Teaching episodes for Saga and the frequency of Ari’s social attending and longer phrases of requesting.

The frequency of Saga’s teaching episodes in the baseline phase when she was teaching Ari social attending (top panel), stayed at zero levels throughout the baseline phase (Assessments 1–8). When the “DANCE” training package was implemented, there was a slight but variable increase in the frequency of her teaching episodes for the first four assessments (Assessments 9–12). Following that, her frequency of responding increased sharply during Assessment 13. Saga’s responding maintained at high but variable levels until a break that was taken due to her illness (Assessment 23). Three months later, when the “DANCE” training package was continued (Assessment 24), her frequency of teaching episodes was maintained at similar levels as before the break and continued at fairly stable levels until the “DANCE” training package was implemented for longer phrases (Assessments 24–27). After the “DANCE” training package for longer phrases had been implemented (Assessment 28), Saga’s frequency of teaching episodes for social attending decreased sharply but increased again and continued at fairly high, but variable, levels (Assessments 30–35) although that stabilized for the last three assessments of the training phase (Assessments 36–38). The percentage of nonoverlapping data (PND) between the baseline phase and training phase was 97.3%, which indicates the training was very effective (Scruggs & Mastropieri, 2001). One month after completion of training, Saga’s frequency of teaching episodes for social attending was maintained at similar levels as during the end of the training phase. Three months after completion of training, her frequency of teaching episodes for social attending was low, just above baseline levels.

Saga teaching longer phrases

Saga’s frequency of teaching episodes stayed at zero levels (lower panel) during the baseline phase before she taught Ari to request in longer phrases during requesting. When the “DANCE” training package for longer phrases was implemented, her frequency of teaching episodes did not increase until Assessment 30. Saga’s pattern of responding was variable throughout the training phase but stayed above baseline levels with exception of Assessment 35. The PND between the baseline phase and training phase was 72.7%, which indicates the training was effective (Scruggs & Mastropieri, 2001). One month after completion of training, her frequency of teaching episodes for longer phrases was maintained at fairly high levels (Assessments 39–41). Three months after completion of training, her frequency of teaching episodes was also maintained (Assessment 42) but at slightly lower levels than at 1-month follow-up.

Ari’s social attending

During baseline, when Saga was teaching Ari social attending, the frequency of his responding (top panel) was at low stable levels throughout the phase (Assessments 1–8) with the exception of Assessment 6. When the “DANCE” training package was implemented with his mother, Ari’s frequency of social attending increased to high levels, although not immediately (Assessment 10), and stayed at high but variable levels for majority of the training phase. The PND between the baseline phase and training phase was 97.3%, which indicates the training was very effective (Scruggs & Mastropieri, 2001). One and 3 months after completion of training, Ari’s frequency of social attending was maintained at similar levels as during the training phase.

Ari’s longer phrases

Ari’s frequency of responding during the baseline phase, before Saga taught him to request in longer phrases (lower panel), was at low but variable levels (Assessments 1–27). After the “DANCE” training package had been implemented for longer phrases, Ari’s frequency of requesting in longer phrases increased immediately (Assessment 28) but stayed at variable levels throughout the training phase. The PND between the baseline phase and training phase was 45.4%, which indicates the training was not effective (Scruggs & Mastropieri, 2001). One month after completion of training, his frequency of requesting in longer phrases was maintained at similar levels (Assessments 39–41) as during the latter part of the training phase.

Comparison of Saga’s and Ari’s responses

Comparison of Saga’s and Ari’s data patterns, during training of both skills, shows correspondence. When frequency of Saga’s teaching episodes changed as training was implemented, similar changes occurred in Ari’s responding.

Generalization

Figure 1 also displays results from the generalization probes during the family’s mealtimes. Results from the probes during training of social attending (top panel) show that during the implementation of the “DANCE” training package, generalization of Saga’s teaching episodes did not occur during the mealtime (Assessment 20) nor did generalization occur 1 month after completion of training (Assessment 41). In comparison, probes during baseline show zero frequency of Saga’s teaching episodes. Generalization of Ari’s social attending occurred during the implementation of the “DANCE” training package (Assessment 20) where he displayed very high frequency of responding. Generalization also occurred 1 month after completion of training with high frequency of responding, similar to the training situation. In comparison, probes during baseline show high frequency of social attending. However, results from the generalization probes 1 month after completion of training (bottom panel) after the implementation of the “DANCE” training package for longer phrases (Assessment 41) show no generalization of Saga’s teaching episodes or Ari’s requesting in longer phrases. In comparison, probes during baseline show zero frequency of Saga’s teaching episodes and low levels of Ari’s frequency of requesting in longer phrases.

Katla and Gunnar

Katla teaching social attending

Figure 2 displays the effects of the “DANCE” training package on Katla and Gunnar.

Figure 2.

Teaching episodes for Katla and the frequency of Gunnar’s social attending and requesting.

During the baseline phase, when she was teaching Gunnar social attending (top panel), responding stayed at zero levels throughout the phase (Assessments 1–9). When the “DANCE” training package was implemented, there was an immediate increase in the frequency of her teaching episodes, which continued at variable, but above baseline levels, until training for requesting was implemented (Assessment 19). The frequency of Katla’s teaching episodes for social attending decreased when the “DANCE” training package for requesting was implemented but stayed fairly stable for most of the remainder of the phase (Assessments 20–23), with the exception of an increasing trend for the last three sessions of the phase (Assessments 23–25). The PND between the baseline phase and training phase was 100%, which indicates the training was very effective (Scruggs & Mastropieri, 2001). One and 3 months after completion of training, Katla’s frequency of teaching episodes for social attending was maintained at similar levels as during the end of the training phase.

Katla teaching requesting

Katla’s frequency of teaching episodes, when teaching Gunnar requesting (lower panel), stayed at stable low levels until after the “DANCE” training package for social attending was implemented (Assessment 10), when it increased immediately. Katla’s responding decreased again for the next three assessments (Assessments 11 and 12) and remained fairly stable at low levels for the remainder of the baseline phase. When the “DANCE” training package for requesting was implemented, there was an immediate increase in the frequency of teaching episodes (Assessment 19), which stayed slightly variable but well above baseline levels throughout the phase. The PND between the baseline phase and training phase was 42.8%, which indicates the training was not effective (Scruggs & Mastropieri, 2001). However, these results are affected by an outlier in the baseline phase. If the next highest data point in baseline is used as a reference, the PND is 85.7%, which indicates the training was very effective. One and 3 months after completion of training, Katla’s frequency of teaching episodes for requesting was maintained at similar levels (Assessments 26–29) as during the end of the training phase.

Gunnar’s social attending

During baseline, when Katla was teaching Gunnar social attending, the frequency of his responding (top panel) was at fairly low levels but with a slightly increasing trend throughout the phase (Assessments 1–9). When the “DANCE” training package was implemented with his mother, there was an immediate increase in his responding, which decreased sharply during the next assessment (11) but continued with an increasing trend at high levels until training for requesting was implemented (Assessments 11–17), with an exception of the training session itself (Assessment 18). When the “DANCE” training package for requesting was implemented, Gunnar’s frequency of social attending increased immediately to very high levels but continued with a decreasing trend until Assessment 23, when it increased again for the remainder of the phase. The PND between the baseline phase and training phase was 93.7%, which indicates the training was very effective (Scruggs & Mastropieri, 2001). One month after completion of training, Gunnar’s frequency of social attending was maintained at similar levels for the first two assessments (Assessments 26 and 27) but decreased sharply for the third assessment. Three months after completion of training, his frequency of social attending was also maintained but at slightly higher levels. During both of those assessments, his responding was above baseline levels.

Gunnar’s requesting

During baseline, before Katla was taught to teach Gunnar requesting (lower panel), his frequency of requesting was at low levels until the “DANCE” training package for social attending was implemented (Assessments 1–9). After the “DANCE” training package had been implemented for social attending, Gunnar’s frequency of requesting increased immediately (Assessment 12) but decreased again and stayed at low levels and was fairly stable throughout the remainder of the baseline phase, with the exception of Assessment 16. After the “DANCE” training package had been implemented for requesting, there was an immediate increase in Gunnar’s frequency of requests, which stayed fairly stable for the remainder of the phase. The PND between the baseline phase and training phase was 28.5%, which indicates the training was not effective (Scruggs & Mastropieri, 2001). However, these results are affected by an outlier in the baseline phase. If the next highest data point in baseline is used as a reference, the PND is 85.7%, which indicates the training was very effective. One and 3 months after completion of training, Gunnar’s frequency of requesting was maintained at similar levels (Assessments 26–29) as during the end of the training phase.

Comparison of Katla’s and Gunnar’s responses

Comparison of Katla’s and Gunnar’s data patterns, during training of both skills, shows correspondence. When frequency of Katla’s teaching episodes changed as training was implemented, similar changes occurred in Gunnar’s responding.

Generalization

Figure 2 also displays results from the generalization probes during baking in the kitchen. Results from the probes during training of social attending (top panel) show that generalization of both Katla’s teaching episodes and Gunnar’s social attending occurred (Assessment 17) and also 1 and 3 months after completion of training (Assessments 26–29). In comparison, generalization probes during baseline (Assessment 6) show zero frequency of Katla’s teaching episodes and low levels of Gunnar’s social attending. Similarly, generalization of Katla’s teaching episodes and Gunnar’s requesting occurred 1 and 3 months after completion of training (Assessments 26–29). In comparison, probes during baseline (Assessment 6) show low frequency of Katla’s teaching episodes and Gunnar’s requesting. During the implementation of the “DANCE” training package for social attending, Gunnar’s requesting generalized to baking, occurring at high levels.

Thora and Daníel

Thora teaching social attending

Figure 3 displays the effects of the “DANCE” training package on Thora and Daníel.

Figure 3.

Teaching episodes for Thora and the frequency of Daníel’s social attending and longer phrases of requesting.

The frequency of Thora’s teaching episodes in the baseline phase, before she was taught to teach Daníel social attending (top panel), stayed at zero levels throughout the phase (Assessments 1–8). When the “DANCE” training package was implemented, there was a slight increase from baseline levels in the frequency of her teaching episodes (Assessments 9 and 10), followed by increased but variable pattern of responding until a break at Assessment 20. After the breaks (Assessments 20 and 22), her frequency of responding continued to stay variable, but well above baseline, until the “DANCE” training package for longer phrases was implemented.

After the “DANCE” training package for longer phrases had been implemented (Assessment 27), Thora’s frequency of teaching episodes for social attending decreased until the end of the training phase. The PND between the baseline phase and training phase was 95.4%, which indicates the training was very effective (Scruggs & Mastropieri, 2001).

Thora teaching longer phrases

During the baseline phase, before Thora taught Daníel to request with longer phrases (lower panel), her frequency of teaching episodes stayed at zero levels (Assessments 1–26). When the “DANCE” training package for longer phrases was implemented, there was slight but variable increase in Thora’s frequency of teaching episodes, which continued to the end of the training phase (Assessments 27–30). The PND between the baseline phase and training phase was 100%, which indicates the training was very effective (Scruggs & Mastropieri, 2001).

Daníel’s social attending

During baseline, before Thora was taught to teach Daníel social attending, the frequency of his responding (top panel) was variable during the majority of the baseline phase (Assessments 1–8) but decreased toward the end of the phase. When the “DANCE” training package was implemented with his mother, Daníel’s frequency of social attending increased to high levels (Assessment 11) and continued at high but variable levels. When the “DANCE” training package for longer phrases was implemented (Assessment 27) Daníel’s frequency of social attending decreased sharply and continued to decrease to baseline levels at the end of the training phase (Assessment 29). The PND between the baseline phase and training phase was 81.8%, which indicates the training was effective (Scruggs & Mastropieri, 2001).

Daníel’s longer phrases

Daníel’s frequency of requesting in longer phrases during the baseline phase, before Thora taught him to request in longer phrases (lower panel), was at low but slightly variable levels until after the second break (Assessments 1–21). Following the second break (Assessment 22), there was an increase in Daníel’s frequency of responding that continued with an increasing trend until the “DANCE” training package for longer phrases was implemented. After the “DANCE” training package had been implemented for longer phrases (Assessment 27), Daníel’s frequency of requesting in longer phrases decreased from the levels before the training but continued to be stable and above baseline levels throughout the training phase (Assessments 27–30). The PND between the baseline phase and training phase was 0%, which indicates the training was ineffective (Scruggs & Mastropieri, 2001).

Comparison of Thora’s and Daníel’s responses

Comparison of Thora’s and Daníel’s data patterns, during training of both skills, shows correspondence. When frequency of Thora’s teaching episodes changed as training was implemented, similar changes occurred in Daníel’s responding. This occurred throughout all sessions.

Social Validity

The complete social validity and qualitative study and further analysis of the mothers’ experience with the training and telehealth methods will be described in another article. A preliminary summary of the qualitative data is presented here. In general, the mothers in the study expressed gratitude for having the opportunity to receive the “DANCE” training and thought it was useful in helping their child. For example, they described the training enabled them to get into better contact with the child and that it was easier to obtain the child’s attention and maintain it. Saga told us that the training with Ari had strengthened their relationship bond and their “DANCING” resulted in a very enjoyable quality time. Saga, Katla, and Thora all reported satisfaction with the goals that they and the trainer created at the outset of the training and Katla and Thora discussed that they liked being able to take part in the goal setting process with the trainer. The mothers all described changes in their children’s behaviors and skills that were consistent with the study’s outcomes. All of them said they had learned to teach their children appropriate social skills and engage their child in appropriate play. Saga also said the training enabled her to utilize the Teaching “DANCE” to help Ari’s younger sibling who has a disability. They all thought their children benefited from the training and that they had made much progress. For example, they reported that social attending increased, and they were able to engage in more frequent and improved contact with family members and friends. Thora reported that Daníel’s expressive communication had improved considerably and was pleased that even his peers were starting to understand him. Saga told us, for example, that Ari was more socially engaged and got better along with his younger sibling. Similarly, Katla described Gunnar being more attentive, more in “contact with life,” and his social and communication skills had improved.

Saga, Katla, and Thora all felt the training via Skype was useful and they described how accessible it was, as well a convenient way to obtain training. Saga and Thora told us they generally preferred training via Skype compared with in situ, although Saga said a regular visit from the trainer would also be good. Katla, however, preferred that the training occurs more frequently at home, in addition to having the regular Skype sessions. She said she liked watching the trainer model the procedures instead of just giving verbal instructions. They all gave favorable reports about the guided practice via Skype and thought the videotaped self-observation was useful. The most challenging aspect of training in their view included technical problems with the Internet connection, computers not working, and the children’s challenging behaviors. In Katla’s and Thora’s cases, the behaviors were related to delays caused by these technical problems. In general, the more technically proficient the mothers were, the more comfortable they were in problem solving through technical issues. Furthermore, Saga, Katla, and Thora did not think that their interaction with their children was different during the Skype session compared with the in situ training sessions or daily routine. However, they were all a little anxious about the child’s challenging behavior having an effect on the session flow. They also said they had to accommodate the child in a certain spot in the room, for example, a play mat, to stay within view for the trainer who was observing and video recording. This was noticed by the trainer and it was clear they were all very aware of this during the session and regularly glanced at the screen to make sure they were within view. Thora said she found the training periods very stressful, as Daníel periodically engaged in very challenging behavior. Despite this, she said she was more comfortable with receiving training and support via Skype than having the trainer located beside her in the home. All of the mothers discussed the importance of engaging in the Teaching “DANCE” with their children in other settings within the home and community and described various situations where they “DANCE-d” with their child.

Discussion

The purpose of this study was to further develop and evaluate a telehealth training protocol for rural Icelandic families with limited access to behavioral expertise. The results from the three families show that teaching parents via brief in situ training and ongoing telecommunication training increased their skills and had a positive effect on the children’s’ skills, especially social attending. These results extend previous research on the effectiveness of teaching naturalistic strategies to caregivers via telecommunication (Guðmundsdóttir et al., 2017; Heitzman-Powell et al., 2014; Ingersoll et al., 2016; McDuffie et al., 2013; Meadan et al., 2016; Simacek et al., 2017; Vismara et al., 2013; Wainer & Ingersoll, 2015). They also extend and confirm the findings of previous research in which training began in situ and continued via telecommunication (Guðmundsdóttir et al., 2017; McDuffie et al., 2013).

The effects in this study were maintained 1 month after completion of training for two families and 3 months after completion of training for one family. This is in accordance with earlier studies (Meadan et al., 2016; Vismara et al., 2013; Vismara, Young, & Rogers, 2012; Wainer & Ingersoll, 2015). However, frequency of teaching episodes for social attending decreased to almost baseline levels for one of those families 3 months after completion of training. No follow-up assessments were conducted for the third family. Generalization to other settings occurred for one parent and two children, and to a sibling’s social attending with one parent. Generalization probes were not conducted for the third family. These results were similar to Meadan et al. (2016), who demonstrated that parents’ naturalistic strategy use generalized to an “uncoached setting.” They were also similar to Suess, Wacker, Schwartz, Lustig, and Detrick (2016), who conducted functional communication training with parents and produced inconsistent generalization of parent and child skills.

Limitations

The mean IOA in this study was over 80% for all participants and behaviors; however, very low instances of IOA occurred occasionally as in the previous study (Guðmundsdóttir et al., 2017). Similar to the previous study, most of these low instances can be explained by low frequency of responding and video quality not being optimal because the room where the training took place was either too dark or very bright due to sunlight in the Icelandic summer. As a result, the behavior to be scored could be very ambiguous to the observer despite systematic training and clear operational definitions. Furthermore, low IOA scores are very sensitive to few instances as the percentage of agreement drops very quickly with each instance of disagreement. Low upload speed of the family’s Internet connection may also have contributed to low video quality. Also, according to an error analysis of low IOA results, the children’s complex play behavior contributed to low instances of IOA, but only to a small proportion of them. Thus, many of the factors that contributed to low instances of IOA were beyond the researchers’ control, such as the context in the family’s setting and perhaps upload speed as mentioned above. Although the range of IOA scores displays very low IOA scores, this does not necessarily have an impact on the interpretation of the degree of experimental control in the study. The method used for assessing IOA (episode by episode) is a stringent method and reflects the accuracy of the observers’ scoring. However, results from calculations of Total IOA show even higher scores, and comparison of the scores of the primary observer and the IOA observer shows very little difference in scoring. Thus, it does not have an effect on the interpretation of the degree of experimental control for all dependent variables with each of the families. Issues with sound quality were seldom related to low instances of IOA except for a few that can be explained by scoring difficulties due to unclear utterances of the child.

In the few existing studies, difficulties with technology such as Internet connection have been noted but have not been systematically explored (Boisvert et al., 2010; Ekeland, Bowes, & Flottorp, 2010; Guðmundsdóttir et al., 2017). Furthermore, very few of the existing telehealth parent training studies that used a video coaching component using an Internet-based telecommunications software (as opposed to teleconferencing equipment) reported difficulties with Internet connection, equipment, or audio or video quality of the samples. Of those studies that did (Machalicek et al., 2016; Suess et al., 2016; Vismara et al., 2012), only Suess et al. (2016) described negative effects on data collection. Although disruption of Internet connection and problems with the families’ computers occurred quite frequently with some of the families in the present study, error analysis of IOA scores did not reveal that these issues were related to low instances of IOA or observed decreases in parent responding. Factors that were beyond the researchers’ control, like few occurrences of responding and video quality due to factors in the family’s setting, contributed most to the occurrences of low IOA.

Part of the current study was to explore the use of “live” consulting via telecommunication. Skype was selected because it was accessible to the families, and at the time the study started, Skype was the software that most Icelandic families were familiar with using. The effects of video quality on the data collection was not anticipated at the start of the study as the video call quality was considered acceptable during the first Skype contact with the families.

When looking at Thora’s pattern of teaching episodes for social attending following the booster session (see Thora’s Assessment 13), it may have been better for experimental and practical reasons related to the family’s summer schedule, to implement this sooner. One reason the training was not implemented at that time point was an artifact of the research protocols and development. There were difficulties in observing and scoring Daníel’s social attending because his face could not be seen clearly from the screen. This affected the trainer’s decision making and the related feedback may in turn have affected Thora’s responding. The trainer and Thora worked on Daníel’s orientation for the next training sessions. By that time, however, Thora reported more stress because of the family’s summer schedule. This prolonged training period combined with development of the coding procedures may have caused more variable pattern of responding.

Implications for Research and Practice

Related to the schedule issues described above, we recommend that both clinicians and researchers assess family schedules and demands that occur seasonally. This is especially important for rural families that may have agricultural responsibilities. Such information may change how rapidly technical problems or training issues are addressed.

Telehealth training models also need to be flexible. Depending on the family’s needs during training, the trainer needs to decide when and where to visit the family and when to provide face-to-face training. Along these lines, Daníel’s behavior became more challenging during the summer break, which may have affected Thora’s responding. McDuffie et al. (2013), also encountered a similar issue in which the child’s challenging behavior was reported to have affected the parents’ teaching skills. Daníel’s challenging behaviors could have been related to the family’s scheduling issues or the challenges related to the scoring described above. The trainer worked on solving this issue with Thora via telecommunication. It is not clear if it would have been easier and perhaps more effective to visit the family in their home as the trainer could have guided Thora with modeling rather than having to rely solely on verbal instructions and videotaped self-observation via the telecommunication. Related to this, researchers of other telehealth parent training studies have also discussed the issue of helping with the child’s challenging behavior during telecommunication sessions when the parent has difficulty learning the teaching techniques, needs more support, or if safety issues are a concern (Machalicek et al., 2016; Vismara et al., 2013; Wainer & Ingersoll, 2013).

Although it is important that the parent is able to teach the child under various conditions during the family’s daily routine, such as during mealtimes and bedtime, there are particular barriers to conducting training via telecommunication in these conditions. For example, it can be challenging to position and reposition a camera, insuring Internet connection and clear visibility with several moving family members, while a parent is preparing and serving a meal. If regular telecoaching is not possible under such conditions, it is important, however, to explore ways to conduct generalization probes of parent and child skills in such settings. Furthermore, it is necessary to incorporate evidence-based methods into training to increase the likelihood of generalization (Allen & Warzak, 2000; Stokes & Baer, 1977).

As an example from the current study, the results from the generalization assessments of Saga (see Saga’s Assessments 20 and 41) show lack of her “DANCING” during mealtime. This could be due to the mealtime being a very different stimulus condition than the playtime setting in which the training was conducted. During mealtime, Saga also needed to care for Ari’s younger sibling and, as a result, the mealtime setting did not have sufficient stimulus control over Saga’s “DANCING,” in competition with other discriminative stimuli. It should be noted, however, that Ari’s frequency of social attending during mealtime was fairly high during baseline and very high during training. Furthermore, his mother was indeed reinforcing his attending even though she did not manage to arrange opportunities for the skill to occur. Perhaps, identifying a high preference event and arranging that event was too complex for Saga in that particular situation.

We have learned through the study that when observing and recording behavioral data, the video quality and speed of the Internet connection may contribute to the reliability of data collection. Because the quality of Internet connection in many rural areas of the world, such as some rural parts in Iceland, may not be good, it is important to explore factors that may affect this and thus the quality of training and evaluation methods. To avoid sound problems in the present study, compared with the previous study (Guðmundsdóttir et al., 2017), the researchers provided the participants with an external boundary microphone, which improved the quality of data collection considerably. However, sound quality still warrants careful consideration when conducting parent training and evaluation via telecommunication to assure reliable data collection and consulting methods. As described in our previous study (Guðmundsdóttir et al., 2017), the reason for poor video quality in that study can possibly be traced to poor quality of Internet connection, including low upload speed, and weather (Póst- og fjarskiptastofnun, 2014). It may also be necessary to adapt and simplify the observation protocol and scoring methods to these different technical challenges to improve reliability of data collection and enhance the trainer’s data-based decision making.

Conclusion

To conclude, the results from this experiment show that parent training via telecommunication is indeed a viable approach for rural families with low speed Internet connection. The effects of the behavioral parent training protocol conducted in situ and during telecoaching had positive effects on both the parent and child skills. Generalization to other settings occurred for some of the families and the increase in skills was maintained for two of the families as well. The research design in the study did not allow for a separate evaluation of the effects of the package training components, such as graph reviews and videotaped self-observation. Such procedures have been supported in past research (Fox & Westling, 1991; Kaiser et al., 2000; Reamer et al., 1998) and the parents in the current study reported all the package components as highly useful. It might be important to develop and evaluate experimentally the effects of the components of the training such as videotaped self-observation.

Although some technical problems occurred, it did not seem to affect the parents’ responding nor reliability of data collection. Hybrid training methods, such as described here, should continue to be explored so as to better understand the conditions necessary for individualization and success for families that have more challenging needs. Continued development of telehealth methods that refine training methods that fit telecoaching and that fit the specific needs of rural life, should help rural children with disabilities make progress despite geographic and sociological barriers. This study extended training procedures and measures by replicating previous findings with additional participants and behaviors, by adding a videotaped self-observation for the parents, and by assessing generalization in other settings in the families’ homes. Furthermore, technical aspects of the telecommunication were improved, by adding an external microphone during training sessions, to improve reliability of data collection. These extensions were a valuable addition, yet further refinements are important, for example, fine-tuning the videotaped self-observation and generalization procedures as well as revising the observation protocol to further enhance reliable data collection and thus the trainer’s data-based decision making. Finally, development and evaluation of hybrid training methods that include telecommunication is important. There are many other activities and settings to examine (mealtimes, community outings, self-care activities, etc.), many other behaviors to increase (conversation skills, complex play actions, dressing, expanded eating, etc.), many other participants to include (fathers, siblings, grandparents, etc.), and many other technical advances that could be incorporated (virtual reality, moving cameras, etc.). With each of these areas to explore, there will be specific technical and clinical considerations for families of children with disabilities in rural settings. The authors hope that continued research in this area will be undertaken in the future.

Footnotes

Authors’ Note

This research was conducted in partial fulfillment of the requirements for the doctoral degree in psychology at the University of Iceland under the supervision of the second and third authors.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported in part by grants from the Icelandic Research Fund, University of Iceland Research and Equipment Purchasing Funds, University of Akureyri Research Fund, KEA University Fund, and Oddur Ólafsson Fund.

ORCID iD

Kristín Guðmundsdóttir

References

Ala’i-Rosales

Cermak

Guðmundsdóttir

(2013). Sunny starts: DANCE instruction for parents and toddlers with ASD. In Bondy

Weiss

M. J.

(Eds.), Teaching social skills to people with autism: Best practices in individualizing interventions (pp. 26–46). Bethesda, MD: Woodbine House.

Allen

K. D.

Warzak

W. J.

(2000). The problem of parental nonadherence in clinical behavior analysis: Effective treatment is not enough. Journal of Applied Behavior Analysis, 33, 373–391. doi:10.1901/jaba.2000.33-37

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: Author.

Barton

E. E.

Fettig

(2013). Parent-implemented interventions for young children with disabilities: A review of fidelity features. Journal of Early Intervention, 35, 194–219. doi:10.1177/1053815113504625

Birkin

Anderson

Seymour

Moore

D. W.

(2008). A parent-focused early intervention program for autism: Who gets access? Journal of Intellectual & Developmental Disability, 33, 108–116. doi:10.1080/13668250802036746

Boisvert

Lang

Andrianopoulos

Boscardin

M. L.

(2010). Telepractice in the assessment and treatment of individuals with autism spectrum disorders: A systematic review. Developmental Neurorehabilitation, 13, 423–432. doi:10.3109/17518423.2010.499889

Brookman-Frazee

(2004). Using parent/clinician partnerships in parent education programs for children with autism. Journal of Positive Behavior Interventions, 6, 195–213. doi:10.1177/10983007040060040201

Cermak

S. M.

(2011). Parent-toddler training: The merits of further analysis (Unpublished master’s thesis). University of North Texas, Denton.

Cheak-Zamora

N. C.

Farmer

J. E.

(2016). The impact of the medical home on access to care for children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 45, 636–644. doi:10.1007/10803-014-2218-3

10.

Chiri

Warfield

M. E.

(2012). Unmet need and problems accessing core health care services for children with autism spectrum disorder. Maternal and Child Health Journal, 16, 1081–1091. doi:10.1007/s10995-011-0833-6

11.

Ekeland

A. G.

Bowes

Flottorp

(2010). Methodologies for assessing telemedicine: A systematic review of reviews. International Journal of Medical Informatics, 81, 1–11. doi:10.1016/j.ijmedinf.2011.10.009

12.

Elder

J. H.

Brasher

Alexander

(2016). Identifying the barriers to early diagnosis and treatment in underserved individuals with autism spectrum disorders (ASD) and their families: A qualitative study. Issues in Mental Health Nursing, 37, 412–420.

13.

Estes

Vismara

Mercado

Fitzpatrick

Elder

Greenson

. . . Rogers

(2014). The impact of parent-delivered intervention on parents of very young children with Autism. Journal of Autism and Developmental Disorders, 44, 353–365. doi:10.1007/s10803-013-1874-z

14.

Farmer

Reupert

(2013). Understanding autism and understanding my child with autism: An evaluation of a group parent education program in rural Australia. Australian Journal of Rural Health, 21, 20–27. doi:10.1111/ajr.12004

15.

Fettig

Barton

E. E.

(2014). Parent implementation of function-based intervention to reduce children’s challenging behavior: A literature review. Topics in Early Childhood Special Education, 34, 49–61. doi:10.1177/0271121413513037

16.

Fox

Westling

(1991). A preliminary evaluation of training parents to use facilitative strategies with their children with profound disabilities. Research and Practice for Persons With Severe Disabilities, 16, 168–176. doi:10.1177/154079699101600305

17.

Gena

Galanis

Tsirempolou

Michalopulou

Sarafidou

(2016). Parent training for families with a child with ASD: A naturalistic systemic behavior analytic model. The European Journal of Counselling Psychology, 4, 4–31. doi:10.5964/ejcop.v4i1.72

18.

Guðmundsdóttir

Sigurðardóttir

Z. G.

Ala’i-Rosales

(2017). Evaluation of caregiver training via telecommunication for rural Icelandic children with autism. Behavioral Development Bulletin, 22, 215–229.

19.

Hamad

C. D.

Serna

R. W.

Morrison

Fleming

(2010). Extending the reach of early intervention training for practitioners. A preliminary investigation of an online curriculum for teaching behavioral intervention knowledge in autism to families and service providers. Infants & Young Children, 23, 195–208. doi:10.1097/IYC.0b013e3181e32d5e

20.

Heitzman-Powell

Buzhardt

Rusinko

L. C.

Miller

T. M.

(2014). Formative evaluation of an ABA outreach training program for parents of children with autism in remote areas. Focus on Autism and Other Developmental Disabilities, 29, 23–38. doi:10.1177/1088357613504992

21.

Hoogsteen

Woodgate

R. L.

(2013). Embracing autism in Canadian rural communities. Australian Journal of Rural Health, 21, 178–182. doi:10.1111/ajr.12030

22.

Höskuldur Þráinsson. (2006). Handbók um málfræði [Handbook on grammar]. Reykjavík, Iceland: Námsgagnastofnun.

23.

Ingersoll

Berger

N. I.

(2015). Parent engagement with a telehealth-based parent-mediated intervention program for children with autism spectrum disorders: Predictors of program use and parent outcomes. Journal of Medical Internet Research, 17(10), e227. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4642401/

24.

Ingersoll

Dvortcsak

(2010). Teaching social communication to children with autism: A practitioner’s guide to parent training and a manual for parents. New York, NY: Guilford Press.

25.

Ingersoll

Wainer

A. L.

Berger

N. I.

Pickard

K. E.

Bonter

(2016). Comparison of a self-directed and therapist-assisted telehealth parent-mediated intervention for children with ASD: A pilot RCT. Journal of Autism and Developmental Disorders, 46, 2275–2284. doi:10.1007/s10803-016-2755-z

26.

Jang

Dixon

D. R.

Tarbox

Granpeesheh

Kornack

de Nocker

(2012). Randomized trial of an eLearning program for training family members of children with autism in the principles and procedures of applied behavior analysis. Research in Autism Spectrum Disorders, 6, 852–856. doi:10.1016/j.rasd.2011.11.004

27.

Kaiser

A. P.

Hancock

T. B.

Nietfeld

J. P.

(2000). The effects of parent-implemented enhanced milieu teaching on the social communication of children who have autism. Early Education and Development, 11, 423–466. doi:10.1207/s15566935eed1104_4

28.

Kaiser

A. P.

Hancock

T. B.

Trent

J. A.

(2007). Teaching parents communication strategies. Early Childhood Services: An Interdisciplinary Journal of Effectiveness, 1, 107–136.

29.

Kobak

K. A.

Stone

W. L.

Wallace

Warren

Swanson

Robson

(2011). A web-based tutorial for parents of young children with autism: Results from a pilot study. Telemedicine and E-Health, 17, 804–808. doi:10.1089/tmj.2011.0060

30.

Lindgren

Wacker

Suess

Schieltz

Pelzel

Kopelman

. . . Waldron

(2016). Telehealth and autism: Treating challenging behavior at lower cost. Pediatrics, 137(Suppl. 2), S167–S175. doi:10.1542/peds.2015-28510

31.

Machalicek

Lequia

Pinkelman

Knowles

Aulston

Davis

Alresheed

(2016). Behavioral telehealth consultation with families of children with autism spectrum disorder. Behavioral Interventions, 31, 223–250. doi:10.1002/bin.1450

32.

McDuffie

Machalicek

Oakes

Haebig

Weismer

S. E.

Abbeduto

(2013). Distance video-teleconferencing in early intervention: Pilot study of a naturalistic parent-implemented language intervention. Topics in Early Childhood Special Education, 33, 172–185. doi:10.1177/0271121413476348

33.

Meadan

H. M.

Snodgrass

M. R.

Meyer

L. E.

Fisher

K. W.

Chung

M. Y.

Halle

J. W.

(2016). Internet-based parent-implemented intervention for young children with autism. Journal of Early Intervention, 38, 3–23. doi:10.1177/1053815116630327

34.

Murphy

M. A.

Ruble

L. A.

(2012). A comparative study of rurality and urbanicity on access to and satisfaction with services for children with autism spectrum disorders. Rural Special Education Quarterly, 31(3), 3–11. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=aph&AN=83166725&site=ehost-live

35.

National Autism Center. (2015). Findings and conclusions: National standards project, phase 2—Addressing the need for evidence-based practice guidelines for autism spectrum disorder. Randolph, MA: Author.

36.

Patterson

S. Y.

Smith

Mirenda

(2012). A systematic review of training programs for parents of children with autism spectrum disorders: Single subject contributions. Autism, 16, 498–522. doi:10.1177/1362361311413398

37.

Pickard

K. E.

Ingersoll

B. R.

(2016). Quality versus quantity: The role of socioeconomic status on parent-reported service knowledge, service use, unmet service needs, and barriers to service use. Autism, 20, 106–115. doi:10.1177/1362361315569745

38.

Póst- og fjarskiptastofnun. (2014). Leiðbeiningar um uppbyggingu ljósleiðaraneta og ríkisaðstoðarreglur EES. Ljósleiðarauppbygging sveitarfélaga og annarra opinberra aðila [Guidelines for constructing fiber optic cable networks and the EEA rules for governmental support]. Retrieved from http://www.pfs.is/library/Skrar/Ljosleidarauppbygging/Leidbeiningar_um_uppbyggingu_ljosleidaraneta_og_rikisadstodarreglur_EES.pdf

39.

Rakap

(2014). Parent-implemented naturalistic language interventions for young children with disabilities: A systematic review of single-subject experimental research studies. Education Research Review, 13, 35–51. doi:10.1016/j.edurev.2014.09.001

40.

Reamer

R. B.

Brady

M. P.

Hawkins

(1998). The effects of video self-modeling on parents’ interactions with children with developmental disabilities. Education and Training in Mental Retardation and Developmental Disabilities, 33, 131–143.

41.

Reese

R. M.

Jamison

Wendland

Fleming

Braun

M. J.

Schuttler

J. O.

Turek

(2013). Evaluating interactive videoconferencing for assessing symptoms of autism. Telemedicine & E-Health, 19, 671–677. doi:10.1089/tmj.2012.0312

42.

Schreibman

Dawson

Stahmer

A. C.

Landa

Rogers

S. J.

McGee

G. G.

. . . Halladay

(2015). Naturalistic developmental behavioral interventions: Empirically validated treatments for autism spectrum disorder. Journal of Autism and Developmental Disorders, 45, 2411–2428. doi:10.1007/s10803-015-2407-8

43.

Scruggs

T. E.

Mastropieri

M. A.

(2001). How to summarize single-participant research: Ideas and applications. Exceptionality, 9, 227–244.

44.

Simacek

Dimian

A. F.

McComas

J. J.

(2017). Communication intervention for young children with severe neurodevelopmental disabilities via telehealth. Journal of Autism and Developmental Disorders, 47, 744–767. doi:10.1007/s10803-016-3006-z

45.

Sparrow

S. S.

Cicchetti

D. M.

Balla

D. A.

(2005). Vineland Adaptive Behavior Scales, Second Edition (Vineland™–II). Circle Pines, MN: American Guidance Service.

46.

Stokes

Baer

D. M.

(1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10, 349–367. doi:10.1901/jaba.1977.10-349

47.

Strauss

Mancini

The SPC Group Fava

(2013). Parent inclusion in early intensive behavior interventions for young children with ASD: A synthesis of meta-analyses from 2009 to 2011. Research in Developmental Disabilities, 34, 2967–2985. doi:10.1016/j.ridd.2013.06.007

48.

Suess

A. N.

Wacker

D. P.

Schwartz

J. E.

Lustig

Detrick

(2016). Preliminary evidence on the use of telehealth in an outpatient behavior clinic. Journal of Applied Behavior Analysis, 49(3), 1–7. doi:10.1002/jaba.305

49.

Suppo

Floyd

(2012). Parent training for families who have children with autism: A review of the literature. Rural Special Education Quarterly, 31(2), 12–26. Retrieved from http://web.b.ebscohost.com/ehost/detail/detail?vid=2&sid=f1ff8fe5-336d-44c6-8a4d-e5d13c9a3fad%40sessionmgr112&hid=124&bdata=JnNpdGU9ZWhvc3QtbGl2ZQ%3d%3d#db=aph&AN=77757819

50.

Suppo

J. L.

Mayton

M. R.

(2014). Expanding training opportunities for parents of children with autism. Rural Special Education Quarterly, 33(3), 19–28. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=aph&AN=99234977&site=ehost-live

51.

Townley-Cochran

(2014). DANCE: A training package utilizing videotaped self-observation to teach parents to enhance social interactions with children at-risk for a developmental delay (Unpublished master’s thesis). University of North Texas, Denton.

52.

Vismara

L. A.

McCormick

Young

G. S.

Nadhan

Monlux

(2013). Preliminary findings of a telehealth approach to parent training in autism. Journal of Autism and Developmental Disorders, 43, 2953–2969. doi:10.1007/s10803-013-1841-8

53.

Vismara

L. A.

Young

G. S.

Rogers

S. J.

(2012). Telehealth for expanding the reach of early autism training to parents. Autism Research and Treatment, 2012, 1–12. doi:10.1155/2012/121878

54.

Wainer

A. L.

Ingersoll

B. R.

(2013). Disseminating ASD interventions: A pilot study of a distance learning program for parents and professionals. Journal of Autism and Developmental Disorders, 43, 11–24. doi:10.1007/s10803-012-1538-4

55.

Wainer

A. L.

Ingersoll

B. R.

(2015). Increasing access to an ASD imitation intervention via a telehealth parent training program. Journal of Autism and Developmental Disorders, 45, 3877–3890. doi:10.1007/s10803-014-2186-7

56.

Wechsler

(1989). Manual for the Wechsler Preschool and Primary Scale of Intelligence—Revised Edition. San Antonio, TX: The Psychological Corporation.

57.

World Health Organization. (2007). Atlas: Global resources for persons with intellectual disabilities. Retrieved from http://www.who.int/mental_health/evidence/atlas_id_2007.pdf

58.

World Health Organization. (2010a). The international statistical classification of diseases and related health problems (10th revision, Volume 2, Instruction manual). Retrieved from http://apps.who.int/classifications/icd10/browse/Content/statichtml/ICD10Volume2_en_2010.pdf

59.

World Health Organization. (2010b). Telemedicine: Opportunities and developments in member states: Report on the second global survey on eHealth (Global Observatory for eHealth series, Volume 2). Retrieved from http://whqlibdoc.who.int/publications/2010/9789241564144_eng.pdf