An Evaluation of “Balance”: a Home-Based, Parent-Implemented Program Addressing Emerging Problem Behavior

Abstract

Programs that prevent the development of severe problem behavior in young children with autism spectrum disorder (ASD) are critically needed. We describe a program designed to do this, and we report on a preliminary evaluation of its effects with four 3- and 4-year-old children with ASD. Parents served as the primary implementers, with twice-weekly coaching from a Board Certified Behavior Analyst. Direct measures and Aberrant Behavior Checklist scores reflected decreases in emerging problem behavior. Direct measures also reflected increases in child communication, social, and cooperation skills, and parents rated the process as highly acceptable. A randomized controlled trial will be required to evaluate the extent to which the program prevents the development of problem behavior in young children with ASD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40617-020-00490-3.

The average age at an autism spectrum disorder (ASD) diagnosis in the United States is 4 years 4 months (Centers for Disease Control and Prevention, 2019), and research focuses on ever-earlier identification (e.g., English, Tenenbaum, Levine, Lester, & Sheinkopf, 2019; Franchini et al., 2018). The American Academy of Pediatrics recommends early intervention services as soon as ASD is suspected (American Academy of Pediatrics, 2014; Myers & Plauché Johnson, 2007). A critical function of these services would seem to be preventing the development of severe problem behavior, given that it is more common in children with ASD than in typically developing children and children with other disabilities (American Psychiatric Association, 2013; Bodfish, Symons, Parker, & Lewis, 2000; Farmer & Aman, 2011; Mayes et al., 2012). However, most early intensive behavioral intervention outcome studies report little or no data on problem behavior (Smith & Iadarola, 2015). To our knowledge, no programs from any discipline have empirically demonstrated the prevention of severe problem behavior in children with ASD.

However, there is research suggesting the potential utility of certain procedures. The Preschool Life Skills Program (Hanley, Heal, Tiger, & Ingvarsson, 2007) focuses on systematically and explicitly teaching children how to respond to challenging situations they are likely to encounter in daily life. Target skills include responding to name calls, functional communication, responding calmly to delays and denials, following instructions, and interacting with peers. It has been demonstrated to prevent the development of problem behavior in typically developing children (Luczynski & Hanley, 2013). The Preschool Life Skills Program also reduces existing problem behavior and strengthens social skills in typically developing preschoolers and children with developmental disabilities (Fahmie & Luczynski, 2018; Gunning, Holloway, & Healy, 2019; Gunning, Holloway, & Grealish, 2020; Robison, Mann, & Ingvarsson, 2020). No published studies have evaluated its preventive effects for children with developmental disabilities. However, given the program’s preventive effects for typically developing children and its intervention effects for children with developmental disabilities, this seems a logical next step for researchers.

Recent research outside of that investigating the Preschool Life Skills Program suggests some additional components that might enhance its effects for children with ASD. Comparative research suggests that synthesized reinforcers (e.g., escape from work to tangibles, attention, and sensory reinforcement) produce stronger treatment effects than isolated reinforcers (e.g., escape only or praise only; Slaton & Hanley, 2018; Slaton, Hanley, & Raftery, 2017).1 Indeed, research has shown that synthesized reinforcers delivered continuously at first, then intermittently and unpredictably, establish social skills to the exclusion of problem behavior in individuals with ASD (e.g., Beaulieu, Van Nostrand, Williams, & Herscovitch, 2018; Falcomata, Roane, Muething, Stephenson, & Ing, 2012; Falcomata, Muething, Gainey, Hoffman, & Fragale, 2013; Greer, Fisher, Saini, Owen, & Jones, 2016; Hanley, Jin, Vanselow, & Hanratty, 2014; Herman, Healy, & Lydon, 2018; Jessel, Ingvarsson, Metras, Kirk, & Whipple, 2018; Rose & Beaulieu, 2019; Santiago, Hanley, Moore, & Jin, 2016; Strand & Eldevik, 2017; Taylor, Phillips, & Gertzog, 2018). Therefore, the effects of the Preschool Life Skills Program might be enhanced by incorporating synthesized, personalized reinforcers to shape skills.

This synthesized, personalized reinforcement interval could be thought of as a child-led playtime (see also Luckett, Bundy, & Roberts, 2007). A prevention program might begin by targeting adult–child interaction during child-led play to create the most reinforcing context possible for the child. Access to this context could then be delivered contingent on the demonstration of target skills during the remainder of the program.

The Preschool Life Skills Program might also be enhanced by the addition of an emphasis on developing a balance between child- and adult-led activities. The National Association for the Education of Young Children (NAEYC) recommends both child- and adult-led experiences (NAEYC, 2009), and yet access to both types of experiences might be limited for children with ASD. Research suggests that adults avoid adult-led activities with children with a history of uncooperative behavior (Carr, Taylor, & Robinson, 1991). At the same time, the quality of child-led experiences might be routinely compromised due to common treatment recommendations. For example, caregivers are often encouraged to embed teaching into play activities for children with ASD, but research suggests that children may not prefer this, and that embedded teaching can even punish some children’s play (Heal & Hanley, 2011; Heal, Hanley, & Layer, 2009). Therefore, separating child- and adult-led activities and striking a balance between the two might improve the quality of both and ensure that children access important learning opportunities.

This general framework of child-led time and adult-led time is used in other programs, such as parent–child interaction therapy (Masse, McNeil, Wagner, & Chorney, 2007). Parent–child interaction therapy is a treatment program typically used with children with significant, existing behavior problems. Although researchers have begun to experimentally evaluate its use with children with ASD (e.g., Ginn, Clionsky, Eyberg, Warner-Metzger, & Abner, 2017; Scudder et al., 2019; Solomon, Ono, Timmer, & Goodlin-Jones, 2008), no prevention studies have been published. In fact, we suspect that there may be components important to preventing the development of problem behavior in children with ASD that are not included in parent–child interaction therapy, such as the explicit teaching of functional communication and tolerance for request denials (Ala’i-Rosales et al., 2019; Fahmie, Iwata, & Mead, 2016). We also suspect that some components that are included may not be optimal for children with ASD, such as the recommendation that parents model, expand, describe, and imitate during child-led time (Ginn et al., 2017; Masse et al., 2007; Scudder et al., 2019). These could present as demands to children with ASD, decreasing the quality of the child-led time and potentially evoking problem behavior.

Therefore, we developed a program that is the subject of the current study—a home-based, parent-implemented program specifically designed to prevent the development of severe problem behavior in children with ASD. It is based on the Preschool Life Skills Program, with modifications based on recent research (i.e., synthesized reinforcement, a focus on balance between child- and adult-led activity). Given this emphasis, we call it the “Balance Program.” It also contains modifications to make it accessible to children not receiving intensive early intervention services: The primary interventionists are parents, with 2 hr of Board Certified Behavior Analyst (BCBA) support per week, and teaching occurs in the home.

In this preliminary investigation, we evaluated the effects of the Balance Program using a single-subject experimental design, and we report on direct and indirect measures of emerging problem behavior and direct measures of communication, social, and cooperation skills. We also report on procedural integrity, the amount and distribution of teaching and support required to achieve the observed outcomes, and the social validity of the goals, procedures, and outcomes. Evaluating the preventive effects of the Balance Program will necessitate a randomized controlled trial, and the present study is a precursor to such an effort.

Method

Participants

Participants were four young children and their primary caregivers (see Tables 1 and 2). Child inclusion criteria were are follows: (a) between 3 and 6 years old, (b) diagnosed with ASD by an outside evaluator within the last year, (c) living at home, and (d) reported by parents to engage in frequent uncooperative behavior or nondangerous problem behavior that appeared to be socially mediated.

Table 1.

Child characteristics

Name	Age (yr:mo)	Sex	Diagnosis	Race/ethnicity	Language levela	EPB	Tasks
Walt	4:3	M	ASD	White	3	Whining, yelling, hitting	Cleaning up, putting on outerwear, sharing with brother, listening to story
Jaden	4:7	M	ASD	White	4	Yelling, throwing,b kicking air, stomping, squeezing parent’s face	Cleaning up, handwriting, putting on outerwear, making bed, brushing teeth
Max	4:1	M	ASD	White	4	Whining, yelling, throwing,b flopping, verbal attacks	Cleaning up, putting on particular clothes, playing parent-selected game, playing alone
Kelly	3:3	F	ASD	Black/African American	1	Whining, kicking air, pushing items/people away, flopping	Cleaning up, listening to story, early VB tasks, playing alone

Tasks include those used in baseline and Steps 6 and thereafter; yr, years; mo, months; ASD, autism spectrum disorder; EPB, emerging problem behavior; VB, verbal behavior

^a1 = nonverbal; 2 = one-word utterances; 3 = short disfluent sentences; 4 = full fluency

^bThis included small/lightweight items, such as socks or game pieces, not in the direction of another person

Table 2.

Parent characteristics

Child name	Participating caregivers	Race/ethnicity	Primary language	Highest education	Employment	Other children	Training/professional experience—early childhood
Walt	Mother, mother	White	English	College degree, postgraduate degree	Full time outside home	1	No
Jaden	Mother, father	White	English	College degree, high school degree	Full time outside home	0, 1	Yes (general), no
Max	Mother	White	English	Some college	Not employed outside home	1	No
Kelly	Mother	Black/African American	English	College degree	Full time outside home	2	Yes (general)

For Walt and Jaden, when a column contains one response, it applies to both parents. When a column contains two responses separated by a comma, one response pertains to each parent

We recruited participants by distributing flyers to local service providers, with the request that they offer them to families of young children recently diagnosed with ASD. The first author screened respondents using the inclusion criteria. For the fourth criterion, the researcher asked the parent whether the child exhibited any uncooperative or problematic behaviors, and if so, to describe the topographies, frequency, intensity, and situations in which they tended to occur. Children were included when parent report suggested “emerging” socially mediated behavior problems. An “emerging” behavior problem was defined as the parent-reported daily occurrence of uncooperative behavior or nondangerous problem behavior. If parents reported the occurrence of severe problem behavior (i.e., self-injury, pica, elopement, aggression resulting in injury to others, or behavior that resulted in observable damage to property),2 or if parent report suggested automatic reinforcement, we referred the child to other programs within our clinic.

Although our participants were not referred for dangerous problem behavior, their parents nevertheless reported that the emerging problem behavior had adverse effects on the family. Parents had difficulty running errands, completing household tasks, and caring for siblings, and some reported persistent family conflict and/or mental health concerns related to parenting issues.

Three master’s-level BCBAs participated, all of whom were enrolled in a behavior-analytic doctoral program at the time of the study and were supervised by the second author, a doctoral-level behavior analyst. One BCBA completed all visits for a family.

Overview

The Balance Program was designed to fit within existing service systems for children with ASD. Practice guidelines for the behavior-analytic treatment of ASD include at least 2 hr per week of case supervision by a BCBA (The Council of Autism Service Providers, 2020). Therefore, the Balance Program involves two 1-hr visits by a BCBA per week, with the recommendation that parents practice at least five trials with their child twice daily between visits. To promote child skill acquisition and provide parents with immediate reinforcement for their teaching efforts, we asked parents to teach in short, circumscribed practice sessions with short intertrial intervals (Francisco & Hanley, 2012) early in the program. We specifically suggested that they not attempt to teach all day or embed the trials in typical routines until Step 10. To minimize demands on parents, we did not ask them to collect data on practice sessions. We anticipated that it would be apparent from the child’s responding at the start of the next visit whether high-quality practice had occurred between visits, and we planned to continue practicing with the parent and child at the current level until performance criteria were met (see Appendix D).

We conducted assessments during the first and second visits. During subsequent visits, the BCBA taught the parent to teach the child a skill (see Tables 3 and 4). Following mastery of all skills, there was a break of at least 1 month, after which follow-up was conducted.

Table 3.

Objectives by teaching steps

		Parent objective	Child objective	Selected references
Teaching process	1	Present no EOs for EPB.	Play in the absence of EPB.	Ghaemmaghami, Hanley, Jin, & Vanselow, 2016; Hanley et al., 2014 & repl.; Heal et al., 2009; Heal & Hanley, 2011; McLaughlin & Carr, 2005; NAEYC, 2009
	2	Shape responding to name.	Stop, look, say “yes” to name call.	Beaulieu, Hanley, & Roberson, 2012, 2013; Beaulieu & Hanley, 2014; Everett et al., 2005; Hamlet, Axelrod, & Kuerschner, 1984; Hanley et al., 2007 & repl.; Kraus, Hanley, Cesana, Eisenberg, & Jarvie, 2012; Miller et al., 2017
	3	Shape functional communication.	Emit part of request.	Carr & Durand, 1985; Ghaemmaghami, Hanley, Jin, & Vanselow, 2016; Hanley et al., 2007 & repl.; Hanley et al., 2014 & repl.; Horner & Day, 1991; Reeve & Carr, 2000
	4	Teach complete request.	Emit complete request.	Ghaemmaghami, Hanley, Jessel, & Landa, 2018; Hanley et al., 2007 & repl.; Hanley et al., 2014 & repl.; Hernandez, Hanley, Ingvarsson, & Tiger, 2007
	5	Shape tolerance.	Respond to denials calmly.	Carr & Carlson, 1993; Fagen & Hill, 1987; Hanley et al., 2007 & repl.; Hanley et al., 2014 & repl.; Skinner, 1953
	6	Shape cooperation.	Follow 1–3 instructions.	Ghaemmaghami, Hanley, & Jessel, 2016; Hanley et al., 2007 & repl.; Hanley et al., 2014 & repl.; Lalli & Casey, 1996; Piazza et al., 1996; Tarbox et al., 2007
Post	7	Shape persistence.	Complete all steps of an activity.	Hanley et al., 2007 & repl.; Hanley et al., 2014 & repl.
Generality	8	Practice with new instructions.	Complete new activities.	Stokes & Baer, 1977
	9	Practice without bins.	Demonstrate all skills without bins.
	10	Practice during typical routines around house.	Demonstrate all skills during typical routines around house.

EO, establishing operation; EPB, emerging problem behavior; repl., replications

Table 4.

Teaching procedures

	Content	Process	Dosage
BCBA teaches parent	How to: Initiate trials Prompt desired child responses & fade prompts Differentially reinforce child responding	Written, vocal description and rationale Modeling Prompting, feedback as parent practices with child	Two 1-hr visits/week
Parent teaches child	How to: Respond to name calls Emit FCR Emit TR Cooperate with adult-led activities of increasing length, diversity	Adult-initiated trials Prompting:            Three step or wait out Differential reinforcement:            Synthesized reinforcers            Intermittent, unpredictable schedule	Two 1-hr visits/week with BCBA and Two 5-trial sessions/day independently

BCBA, Board Certified Behavior Analyst; FCR, functional communication response; TR, tolerance response. See Appendix A for BCBA and parent implementation integrity checklists

Setting and Materials

BCBA visits were conducted in one room in the participants’ homes until Step 10, at which point skills were practiced throughout the home. All materials originated in the participants’ homes, except for video equipment, program-related written materials, and empty bins, which the BCBA provided. Items related to child- and adult-led activities were identified via interview and sorted into two bins of the same size (approximately 30 cm³ or 60 cm × 40 cm × 30 cm). Parents were not asked to obtain any new items for the program.

Data Collection and Response Definitions

Data Collection

The primary dependent variables were emerging problem behavior, child-led play, responding to name, functional communication, tolerance responding, and cooperation. Responses per minute were used to analyze emerging problem behavior; percentage of opportunities correct was used for responding to name, functional communication, and tolerance responding; and percentage of 20-s whole intervals was used to analyze play and cooperation. Data were also collected on parent integrity of the teaching procedures (see Appendix A), which are reported as a percentage of 20-s intervals in which no errors were made. A checklist was used to measure each BCBA’s fidelity to the support procedures (see Appendix A). In baseline, all direct measures were collected in 5-min sessions. In the teaching process, data were collected from one 5-min or 5-trial session at the start of each visit. In follow-up, data were collected from one or two 5-trial sessions at the start of the visit. All sessions were video-recorded and scored later.

Indirect measures included a semistructured parent interview called the Individualized Features of Reinforcement Meeting (InFORM; see Appendix B); the Aberrant Behavior Checklist, Second Edition (ABC-2; Aman & Singh, 2017); and social validity assessments (see Appendix C). The purpose of the InFORM, administered during the first visit, was to identify the materials and activities to be used during teaching. The ABC-2 is a 58-item rating scale completed by parents/caregivers and often used in evaluations of interventions for children with ASD outside of behavior analysis (Kaat, Lecavalier, & Aman, 2014). The ABC-2 was completed by each participant’s primary caregiver at baseline and again after Step 10. Parents completed versions of the social validity assessment at baseline, after Step 10, and at follow-up.

Response Definitions

The definition of emerging problem behavior was individualized per child (see Table 1). Hitting was rarely observed and of low intensity—that is, distinct from “severe problem behavior.”

Child-led play was defined as the child manipulating toy(s) and/or engaging in activity-related conversation with the caregiver in the absence of problem behavior. Responding to name definitions were individualized per participant but generally involved stopping, looking toward the parent, and/or saying “yes.” Functional communication was defined as any instance of a request for access to items or to play the child’s way, to escape or avoid items or parental instruction, or any combination thereof, emitted in the absence of emerging problem behavior. The topography of the functional communication response (FCR) taught was individualized for each child based on the child’s language level. A tolerance response was defined as any instance of an explicit and calm response to a delay or denial imposed by the parent (e.g., “OK”). Cooperation was defined as behaving in accordance with the adult instruction or expectation in place at the time. For example, if the adult asked the child to color and left the area, the expectation that the child color would be considered in place until the adult returned. See Table 1 for participant-specific tasks.

Interobserver Agreement

Interobserver agreement was evaluated by having a second independent observer review the video recordings and collect data on all target behaviors at a separate time from the primary data collector. Each session was divided into 20-s intervals. Interobserver agreement for emerging problem behavior was calculated by dividing the smaller total number of occurrences by the larger total number of occurrences in each interval, adding up the proportions, dividing by the number of intervals, and multiplying by 100. Interobserver agreement for all other measures was calculated by dividing the number of agreements (intervals marked with the same symbol by both data collectors) by the number of agreements plus disagreements, and then multiplying by 100. Interobserver agreement data were collected for at least 20% of sessions across all conditions and participants and averaged at least 86% (range 67%–100%) for all participants and all measures.

Design

Functional control by the independent variable over child emerging problem behavior and over parent teaching behavior was demonstrated using a nonconcurrent multiple-baseline across-participants design. Functional control over child skills (responding to name, FCR, tolerance responding, and cooperation) was demonstrated via repeated measures and replication of the effects of the independent variable on multiple dependent variables across points in time. The effects of the package on the balance between child- and adult-led time were inferred rather than demonstrated.

Procedures

Natural Baseline

The BCBA distributed the materials identified in the parent interview and directed the parent to engage with the child as they normally would. Two or three 5-min sessions were conducted. These data were used to assess the balance between time spent on child-led and adult-led activities, prior to any instruction to the parent in this regard.

Prompted Baseline

The BCBA instructed the parent to allow the child to play for about half of the time and try to have the child complete tasks for the other half of the time. No further prompts were given. Three to five 5-min sessions were conducted. These data were used to evaluate the parent’s ability to create balance between child-led and parent-led activities when specifically asked to do so, as well as to evaluate child behavior under these conditions.

Teaching (Steps 1–10)

Following baseline, participants moved into the teaching process. Beginning with Step 1, the BCBA taught the parent to teach each skill outlined in Table 3. At least one visit was dedicated to each step, with progression to the next step based on parent and child performance (see Appendices for details).

Teaching Procedures

The procedures by which BCBAs taught parents and parents taught children are outlined in Table 4. Before teaching began, the BCBA described two prompting approaches—three-step prompting (verbal, model, physical prompt; e.g., Tarbox, Wallace, Penrod, & Tarbox, 2007) and a “wait-out” (parent delivers a verbal instruction, removes tangibles unrelated to the instruction, and waits for the child to follow the instruction; Piazza, Moes, & Fisher, 1996). Parents were asked to select the approach they preferred. Jaden’s, Max’s, and Kelly’s parents opted to use three-step prompting. Walt’s parents opted to use a wait-out.

Each teaching step built on the previous ones such that the child continued to practice and, intermittently, contact immediate reinforcement for all previously learned skills. The order of trial types varied across sessions such that the response that would result in immediate reinforcement on each trial was unpredictable (see Appendix D).

Posttest

The data from Step 7 were considered the posttest. At this time, all the skills had been taught and the same contextual features as in the baseline were present.

Generality

Steps 8, 9, and 10 were considered generality steps, as they extended the process beyond that which was assessed in baseline. Step 8 involved introducing new activities, Step 9 involved items kept in their regular locations (rather than bins), and Step 10 involved practice in new areas of the home and during natural routines.

Performance Criteria

The teaching process was considered complete following two consecutive sessions at Step 10 in which the child displayed (a) near-zero rates of emerging problem behavior, (b) independent skills (responding to name, functional communication, and tolerance responding) in most opportunities, (c) cooperation and play in most applicable intervals, and when parent- and child-led times each composed roughly half of the session.

Booster and Follow-Up

The BCBA conducted a booster session between 4 and 10 weeks after the last Step 10 visit, depending on the family’s availability. The BCBA observed the parent and child practice Step 10 and provided feedback as needed. The BCBA then returned later the same week to record a five-trial session of Step 10. If the child’s performance met the criteria noted previously, a second session was recorded during that visit. If not, the BCBA provided feedback, asked the parent to practice with the child, and returned to record another session when the family’s schedule allowed.

Modifications

Based on Max’s and Kelly’s performance and on feedback from their parents, we made modifications for these participants. These are detailed in Table 5. The reasons for the modifications were different for the two children, leading us to make different changes for each.

Table 5.

Modifications for Max and Kelly

Child	Modifications	Rationale	Timing
Max	(1) Moved to clinic-based practice: (i) BCBA practiced Steps 8, 9 with child     Parent collected treatment integrity data     Parent & BCBA discussed scores (ii) Parent practiced Step 9 with child & BCBA, then child alone	- Low child skill emission suggests need for more practice - Low parent-reported confidence suggests need for:     More modeling by BCBA     Practice under more controlled conditions	After first follow-up phase
	(2) Returned to home-based practice:    (i) BCBA practiced Steps 9, 10 with child before parent did
	(3) BCBA conducted parent education modules	-Parent questions suggest need for additional education
Kelly	(1) Added BCBA practice with child in home	- Low child skill emission suggests need for more practice	After Step 9 visits
	(2) Eliminated response to name and tolerance response trials	- Decrease in parent time for practice for family/personal reasons

Note. BCBA, Board Certified Behavior Analyst. All BCBA-run sessions were conducted using the same procedures parents had used. Parent education modules consisted of didactic instruction and discussion only, without the child present, and are detailed in Appendix E

For Kelly, we attempted a set of modifications (marked as “Modif.” phase on Figure 5) that were unsuccessful, prior to the successful modifications detailed in Table 5 and marked as “Express” on the figure. The unsuccessful modifications consisted of adding BCBA practice with Kelly in the home, returning to Step 2, and reinforcing only independent responses. When these modifications did not yield the desired rates of independent FCRs over six visits, we consulted with Kelly’s mother to determine which skills she most wanted Kelly to learn. She indicated the FCR and cooperation. Therefore, we developed a shortened version of the program called “Balance Express,” which omitted responding to a name call and tolerance responding.

Figure 5.

Individual Data for Kelly. Note. Numbers in brackets refer to teaching steps. At parent request, Step 4 (gain attention) was terminated prior to mastery. Nat. = natural baseline; Pr. = prompted baseline; Teach. Integ. = teaching integrity; BCBA = Board Certified Behavior Analyst; Modif. = modification

Results

Summary

Figure 1 summarizes data for all participants. Black symbols represent performance given the standard program, and gray symbols represent Max’s and Kelly’s performance with their parents, in the home, following modifications. For all participants, emerging problem behavior (top-left panel) decreased from the prompted baseline to postprogram (posttest, generality, and follow-up), even given balance between child- and adult-led times (top-right panel) and given that social skills and cooperation were required (third and fourth panels on the left). All participants played in the absence of emerging problem behavior for most child-led intervals in all conditions (second panel on the left), although an increase in play was observed with Jaden after teaching. Parental integrity to the teaching procedures improved from the prompted baseline to postprogram for all participants, with more substantial improvements observed for Max’s mother following modifications to the program (bottom-right panel).

Figure 1.

Summary Data for All Participants. Note. The baseline mean was calculated from all sessions of the prompted baseline for all measures except for “Balance,” which was calculated from all sessions of the natural baseline. The posttest mean was calculated from all sessions of Step 7. The generality test mean was calculated from all sessions of Steps 8, 9, and 10. The follow-up mean was calculated from all sessions of follow-up. Gray symbols (Max, Kelly) represent performance in the home, with the parent as implementer, following modifications to the package

Emerging Problem Behavior

Individual participant data are displayed in Figures 2, 3, 4 and 5. Figure 6 displays emerging problem behavior data from the prompted baseline and the first three steps of teaching for all participants, to highlight the multiple-baseline across-participants design.

Figure 2.

Individual Data for Walt. Note. Numbers in brackets refer to teaching steps. Nat. = natural baseline; Prompt. = prompted baseline; Teach. Integ. = teaching integrity

Figure 3.

Individual Data for Jaden. Note. Numbers in brackets refer to teaching steps. Nat. = natural baseline; Prompt. = prompted baseline; Teach. Integ. = teaching integrity

Figure 4.

Individual Data for Max. Note. Numbers in brackets refer to teaching steps. At parent request, Step 4 (gain attention) was terminated prior to mastery. Nat. = natural baseline; Prompt. = prompted baseline; Teach. Integ. = teaching integrity; BCBA = Board Certified Behavior Analyst

Figure 6.

Emerging problem behavior multiple baseline across participants

During the natural baseline, all participants except Walt emitted some emerging problem behavior. Jaden and Max emitted it following certain parent actions during play (e.g., when parents did not follow their instructions), and Kelly emitted it during her mother’s attempts to implement adult-led time. During the prompted baseline, all parents imposed some adult-led activities, and all participants engaged in increasing (Walt, Jaden, Max) or stable (Kelly) rates of emerging problem behavior. With the introduction of Teaching Step 1, rates of emerging problem behavior decreased to zero or near zero and remained below the baseline average throughout the teaching process and follow-up for all participants, except for a few isolated spikes for Max and Kelly.

Play

Walt and Kelly engaged in play in the absence of emerging problem behavior for 100% of the child-led time in baseline (Figures 2 and 5, second panels). Jaden and Max engaged in play for the entire child-led time, but their play was sometimes accompanied by emerging problem behavior (Figures 3 and 4, second panels; black portions of the bars reflect play in the absence of emerging problem behavior; gray portions reflect the time in which play was permitted, but the child either did not play or played while emitting emerging problem behavior). In Step 1, Jaden’s and Max’s play in the absence of emerging problem behavior increased to 100%. All participants engaged in play for all or most of the allotted time across all teaching sessions and follow-up. Because there was little change in the quality of play from baseline to teaching, we characterize our participants as having had initially strong child-led play skills.

Responding to Name

Walt’s and Max’s parents did not call their names during baseline, and Jaden’s and Kelly’s did so infrequently. Neither Jaden nor Kelly responded to these name calls. Responding emerged only following the introduction of Step 2. Across subsequent teaching steps and follow-up, responding generally remained high for Walt, Jaden, and Max. Kelly’s responding was more variable. Following the first set of modifications to Kelly’s program, responding immediately increased and became more stable; however, it became variable again following the reintroduction of Step 3.

Functional Communication

All caregivers attempted to restrict access to tangibles and/or place demands during the prompted baseline, creating opportunities for children to emit requests to retain the tangibles and escape the demands. However, only Max did so in the absence of problem behavior, and he only did so on one out of four opportunities. Responding emerged for all participants in Steps 3 and 4. Walt and Jaden emitted FCRs in most or all opportunities across subsequent steps. Max’s and Kelly’s responding was variable until modifications were made, at which time responding for both became high and stable and remained so across subsequent phases.

Tolerance Responses

Because participants emitted few or no requests using words in baseline, there were few explicit request denials by parents and therefore few opportunities to emit a tolerance response. Tolerance responding emerged following the introduction of Step 5 for Walt, Jaden, and Max and in the third visit at Step 5 for Kelly. Responding maintained in most opportunities across subsequent steps for Walt and Jaden. Max’s responding was variable until modifications were made, at which time it became high and stable and remained so across subsequent phases. Kelly’s responding decreased following Step 7, and the tolerance response was not included in her modified program.

Cooperation

All participants engaged in some cooperation during baseline; however, it averaged less than 50% of requested intervals for all participants. Cooperation was targeted beginning in Step 6, and it increased for Walt, Jaden, and Kelly. Max’s cooperation was variable but increased to consistently high levels following modifications.

At follow-up, Walt and Max emitted cooperation in all requested intervals, and Kelly emitted cooperation in 95% of requested intervals. Jaden emitted cooperation in 100% of requested intervals with his father but only in 29% of intervals with his mother.

Balance

During the natural baseline, most of the time was spent in child-led activity for all participants. During most prompted baseline sessions, time was either allocated disproportionately to child- or adult-led activity. During Step 1, time was again allocated exclusively to child-led activity by design, but this decreased gradually across teaching steps for all participants until follow-up, at which point child- and adult-led time were allocated more equally.

Parent Integrity to Teaching Procedures

Parent integrity to the teaching procedures increased upon the introduction of the teaching process for all participants. It remained high for Walt’s, Jaden’s, and Kelly’s parents across teaching and follow-up. For Max’s mother, it decreased upon the introduction of Step 2 and remained variable across teaching, although it was generally higher than in baseline. Following modifications to the program, the level of teaching integrity increased and variability decreased.

BCBA Integrity to Support Procedures

BCBA integrity to the support procedures was scored for an average of 17% of sessions (range 14%–22%) across all three conditions (baseline, teaching process, and follow-up) for all participants and averaged at least 93% (range 86%–100%).

Time Requirement

The entire process, from the InFORM through follow-up, required 15 one-hour BCBA visits for Walt, 20 for Jaden, 59 for Max, and 56 for Kelly.

Aberrant Behavior Checklist–2

The results for each subsection of the ABC-2 are summarized in Table 6, with raw scores appearing first and percentile ranks3 in parentheses. According to the ABC-2’s authors, a score in the 75th percentile or above suggests the need for intervention (Aman & Singh, 2017). At baseline, both of Walt’s parents rated his behaviors at or above the 75th percentile for Hyperactivity/Noncompliance and Inappropriate Speech. Max’s mother rated his behavior at the 75th percentile for Irritability. All participants’ parents rated behaviors in all the other categories below the 75th percentile, supporting our designation of the participants’ behaviors as “emerging” (rather than “severe”) problems.

Table 6.

Aberrant behavior checklist (ABC-2) scores (with percentiles)

Name	Irritability		Social Withdrawal		Stereotypic Behavior		Hyperactivity/ Noncompliance		Inappropriate Speech
	Max. 45 Clin. Sig. 20		Max. 48 Clin. Sig. 15		Max. 21 Clin. Sig. 8		Max. 48 Clin. Sig. 29		Max. 12 Clin. Sig. 6
	Pre	Post	Pre	Post	Pre	Post	Pre	Post	Pre	Post
Walt
Parent A	9 (37th)	1 (<16th)	4 (25th)	0 (<16th)	0 (16th)	1 (25th)	35 (84th)	9 (16th)	6 (75th)	0 (16th)
Parent B	18 (70th)	3 (<16th)	13 (70th)	4 (25th)	1 (25th)	1 (25th)	32 (75th)	13 (32nd)	6 (75th)	1 (32nd)
Jaden	6 (25th)	6 (25th)	9 (50th)	6 (37th)	1 (25th)	1 (25th)	19 (50th)	21.4 (50th)	2 (37th)	2 (37th)
Max	21 (75th)	12 (50th)	1 (<16th)	0 (<16th)	4 (50th)	0 (16th)	28 (70th)	19 (50th)	3 (50th)	4 (63rd)
Kelly	7 (32nd)	2 (<16th)	7 (37th)	3 (16th)	1 (25th)	2 (25th)	12 (32nd)	12 (32nd)	1 (45th)	1 (45th)

Clin. Sig. = clinically significant score per ABC-2 authors. Boldface denotes a clinically significant reduction according to criteria provided in the ABC-2 manual. Post measures reported for Max were collected following Session 58. Mean scores for a sample of children with autism spectrum disorder, as rated by their parents, are as follows: Irritability = 12.8 (50th percentile); Social Withdrawal = 10 (50th percentile); Stereotypic Behavior = 5 (50th percentile); Hyperactivity/Noncompliance = 18.7 (50th percentile); and Inappropriate Speech = 3.7 (63rd percentile; Aman & Singh, 2017)

The ABC-2’s authors state that generally, a reduction of one standard deviation or more might be considered clinically significant (Aman & Singh, 2017). Walt’s Parent A’s ratings met this criterion in the areas of Hyperactivity/Noncompliance and Inappropriate Speech, and his Parent B’s ratings met this criterion in all areas except Stereotypic Behavior. Max’s mother’s ratings met this criterion in the areas of Irritability and Stereotypic Behavior.4 By contrast, Jaden and Kelly did not meet this criterion in any area. However, their baseline ratings were lower (at or below the 50th percentile), making this outcome unsurprising. All participants’ raw scores improved by at least three points in at least one category, and none worsened by the same amount.

Social Validity

All parents rated the goals of the program as very important (M = 6.9 on a 7-point Likert-type scale), the procedures as highly acceptable (M = 7), and their comfort implementing the procedures as high (M = 6.7). At Step 10 and follow-up, parent ratings of their satisfaction with the effects outside of BCBA visits varied across participants (M = 5.6, range 3–7). Generally, Walt’s and Jaden’s parents rated themselves as more satisfied than Max’s and Kelly’s. Walt’s, Jaden’s, and Kelly’s parents provided only positive responses to open-ended questions, and Max’s mother did not respond to any open-ended questions. (See Appendix C for participant-specific scores and comments.)

Discussion

Direct measures suggest that the Balance Program reduced emerging problem behavior, improved social and cooperation skills, and was associated with the development of balance between child- and adult-led activities for all participants. These outcomes were demonstrated in the home, with parents providing all or most of the teaching. Parents socially validated the goals and procedures, and an indirect, standardized measure of problem behavior detected general improvements. In other words, the Balance Program produced important child behavior outcomes relevant to children with an ASD diagnosis.

Our evaluation also revealed some areas for improvement and important directions for future research. First, two out of four children acquired all the targeted skills given the “standard” package. Modifications to the program were associated with more robust performance for the other two children, and some of these modifications might be incorporated from the beginning for future participants. In addition, on the social validity assessment, some parents provided equivocal ratings for the outcomes of the program outside of BCBA visits; therefore, steps might be taken to improve outcome generality. Once these areas have been addressed, future researchers might consider a randomized controlled trial to evaluate the program’s preventive effects on problem behavior in children with ASD.

All children in the present study acquired at least a simple FCR, and three out of four acquired responding to their name and tolerance to denial responses. It is likely that initial language skills are a moderator of outcomes in this program. The three participants who acquired responding to their name and tolerance response communicated in short, disfluent sentences or with full vocal-verbal fluency at the start of the program, whereas the participant who did not acquire these responses (Kelly) was nonverbal at the start of the program. Kelly’s individual data suggest that these responses were attainable for her in isolation; in several sessions, she emitted high levels of responding to her name, but she never emitted criterion levels of responding to her name and the FCR in the same session, suggesting that the desired conditional discrimination was not established. Future researchers should evaluate the program with additional participants who have little verbal behavior in baseline. If similar outcomes are observed, the program might prescribe a modified approach (e.g., Balance Express) for children with little initial language and/or parents who foresee having limited time to dedicate to practice.

Max’s modifications also suggest revisions to the program. Max’s performance improved substantially in sessions implemented by the BCBA in the clinic with higher teaching integrity than his mother had exhibited. When his mother returned to teaching, she demonstrated improved teaching integrity. Although the current data do not allow for conclusions about which changes were responsible for either Max’s or his mother’s improvements, his mother may have benefited from the higher dose of modeling from the BCBA and from the experience of scoring teaching integrity until she did so with no errors. In future in-home applications, the BCBA could model implementation or show video models and teach the parent to accurately evaluate teaching integrity.

The in-home, parent-implemented Balance Program could also be considered a Tier 1 intervention in a response-to-intervention model. That is, most or all children admitted to early intervention services could receive the Balance Program, implemented by parents in the home. If the Balance Program does not produce the desired outcomes for a child, a Tier 2 intervention would be indicated and considered. A Tier 2 intervention might consist of professionally implemented sessions a few times per week, either in the home or in a clinic. A Tier 3 intervention, should it be required, might consist of intensive outpatient or inpatient services. Therefore, under this conceptualization, given that the Tier 1 intervention did not yield a general-enough change in Max’s behavior, Max experienced a Tier 2 version of the program.

Walt’s, Jaden’s, and Max’s performances suggest a possible relation between teaching integrity averaging at least 80% and robust child outcomes—child performance reached criteria in phases in which teaching integrity averaged 80% or higher and did not otherwise. Future implementers might consider adding a role-play component to the program and coaching parents until they demonstrate at least 80% correct implementation before beginning to teach their children a given step. In other words, in future iterations, the parent-training procedures might include all the steps of behavioral skills training (BST; e.g., Drifke, Tiger, & Wierzba, 2017; Sawyer, Crosland, Miltenberger, & Rone, 2015). We did not initially incorporate role-play because of expected time constraints in practice and because we hoped the program might be useful in telehealth settings, in which role-play is more difficult. However, future researchers might consider adding opportunities for parents to demonstrate robust teaching skills prior to practicing with their children.

Parents’ ratings of the goals and procedures of the program were overwhelmingly positive, and parents reported that the BCBA support was enough to make them feel comfortable teaching the skills. However, parents’ ratings of the outcomes of the program outside of BCBA visits were mixed, with Walt’s and Jaden’s parents rating the outcomes more highly than Max’s and Kelly’s. Walt and Jaden progressed through the program more quickly than Max and Kelly, with no modifications. Some variables that might have influenced both the child’s progress and parents’ satisfaction with outcomes outside of visits include the amount and/or procedural integrity of practice outside of visits throughout the program. In the future, researchers might collect data on between-visit practice and evaluate the effects of procedural modifications on the frequency and quality of practice, as well as the generality of outcomes. Interestingly, Kelly’s mother’s open-ended comments on the effects of the program seem to contrast with her rankings, suggesting that future researchers might find multiple types of social validity measures informative.

Future researchers implementing the program using a single-subject design might also consider design modifications to enhance the demonstration of functional control over all child skills by the program. In the present study, there were no baseline opportunities for children to demonstrate some of the skills. Future researchers could ask parents to provide opportunities for specific skills during the prompted baseline or afterward, as its own baseline condition, if no opportunities are observed in the prompted baseline.

However, even when parents did not present opportunities in baseline, we have several reasons to infer that children lacked these skills and acquired them as a function of the program. First, it is unlikely that a child has had opportunities to acquire a skill if, over repeated observations, the parent does not provide opportunities for the child to emit that skill. Skill acquisition requires practice; Luczynski and Hanley (2013) demonstrated that, in the absence of explicit teaching, typically developing preschoolers did not develop similar skills. The fact that the parents did not provide baseline opportunities for commonplace parent–child interactions suggests that parental attempts had been extinguished by persistent child noncooperation or punished by emerging problem behavior, as suggested by Carr et al. (1991). Finally, all parents rated the goals of the program as very important. It seems unlikely they would have done so if children had a history of reliably emitting these skills in relevant opportunities.

We inferred, rather than demonstrated, control by the program over the balance between adult- and child-led activities. In early teaching steps, child-led play was delivered contingent on relatively short responses, and a great deal of the session was spent in child-led play by design; balance was achieved gradually, emerging around Step 6 or later. We deemed it undesirable to extend participants’ baselines for the time that would be required to demonstrate stability until after balance emerged for earlier participants. However, given the demographic and geographic differences among the families, and the varying amounts of time between baseline and demonstration of balance, it seems implausible that balance would have emerged as a function of some factor other than the program.

There is precedent for teaching each of the skills in the Balance Program (e.g., Durand & Moskowitz, 2015; Everett, Olmi, Edwards, & Tingstrom, 2005; Hanley et al., 2014; Luczynski & Hanley, 2013); nevertheless, whittling down the program to the behaviors that are essential for preventing the development of severe problem behavior is an important task for researchers. Such an evaluation might be most productive after a demonstration of the preventive effects of the program via a randomized controlled trial; the question will then be which components of the program are necessary to produce those preventive effects. It is possible, for example, that differentially reinforcing various durations of cooperation with personalized and synthesized reinforcers may be enough.

Future researchers, before or after a randomized controlled trial, might also investigate the moderating role of initial play skills. All children engaged with the materials their parents had nominated as preferred for most of the child-led time in baseline, and no parents prompted their children to do so. Despite these robust child performances in baseline, all BCBAs conducted instruction in Step 1 to address parent behavior—such as interfering with child toy manipulation or attempting to teach—that might have made access to the play materials a less potent reinforcing context. It is unclear what the effects of the program would have been had the participants not demonstrated robust child-led play in baseline. It seems probable that baseline play behavior moderates the effects of the program; whether this is the case and how to address it are important research questions.

Despite the many remaining questions, the promising effects of the Balance Program shown in the current study—and the studies showing positive effects of similar skill-based teaching programs in different service models (e.g., Hanley et al., 2007; Hanley et al., 2014; Jessel et al., 2018; Robison et al., 2020)—suggest that Balance may be a useful program for professionals supporting parents of young children with ASD in the home.

Compliance with Ethical Standards

Conflict of interest

During the peer review and publication process for this manuscript, Kelsey Ruppel and Gregory Hanley received salaries from FTF Behavioral Consulting, an organization that offers training on the Balance Program among other services. Neither were employed by FTF Behavioral Consulting before or during the conduct of the study. Robin Landa and Adithyan Rajaraman declare no conflict of interest.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional review board of the university and with the 1964 Helsinki declaration and its later amendments.

Informed consent

Informed consent was obtained from all individual participants included in the study.