Introduction

Selective eating is common in individuals with autism spectrum disorder (ASD) across the lifespan and wide range of intellectual ability (Ahearn et al. 2001; Bandini et al. 2010; Cermak et al. 2010; Kuschner et al. 2015; Ledford and Gast 2006; Schreck et al. 2004; Sharp et al. 2013; Williams et al. 2005; Zimmer et al. 2011). Persistence over time and risk for negative outcomes highlight the need for effective and available treatment. Behavioral interventions are widely used to treat selective eating; however, most of these programs are time intensive, have not been evaluated for use in outpatient settings, and do not typically include older youth. Despite the functional impact and risk for negative outcomes associated with selective eating across ages, there are no empirically supported treatments available for older children, adolescents, or adults, either with or without ASD.

This paper will review the development of BUFFET: the Building Up Food Flexibility and Exposure Treatment program. Using stakeholder input from an advisory board and pilot trial participants, BUFFET was developed as a multi-family group cognitive behavioral treatment (CBT) for selective eating in children (8–12 years) with ASD. The intervention aims to (1) capitalize on the growing evidence for effective CBT use with individuals with ASD, (2) pair these approaches with exposure-based methods commonly used in feeding treatments, (3) maintain a framework of autonomy and self-determination, and (4) develop a manualized, outpatient, multi-family group treatment program that could eventually be evaluated via a randomized controlled trial (RCT). This paper will present the development and pilot testing of this novel treatment.

Background

Selective Eating in ASD

Definition and Prevalence

Historically, selective eating was so common in children with ASD that it was considered a criterion for diagnosis (Ritvo and Freeman 1978). Generally, selective eating refers to having limited diet variety (i.e., narrow range of foods one is willing to eat). Selective eating is commonly considered one of the behaviors that can fall under the umbrella of a feeding problem or disorder, if there are also negative health or psychosocial sequelae (Bryant-Waugh et al. 2010; Kedesdy and Budd 1998). For the purposes of this review, we will describe selective eating as a set of behaviors that limits flexible food intake. Food repertoire is limited, with selection and refusal of foods based on features like texture, food group, color, or brand. Idiosyncratic requests and rules also drive selectivity, including how foods are prepared, the dishware and utensils used to serve them, and whether foods are mixed or touching when presented.

Prevalence of selective eating in children, with or without ASD, is challenging to estimate because the field has not yet established a consistent way to define and operationalize selective eating. However, current estimates in ASD are higher than the ~15–20% seen in typical development (Mascola et al. 2010; Zucker et al. 2015). Ledford and Gast (2006) reviewed a number of studies and estimated that problem feeding behaviors were present in 46–89% of children with ASD. In these studies, selective eating generally referred to frequent food refusals, limited repertoire of foods, excessive intake of a few foods, and selective intake of certain food categories.

Developmental Course

Children without ASD are commonly selective eaters between 2 and 6 years (Birch 1999); however, research suggests a more pervasive problem in individuals with ASD. The majority of studies on selective eating in ASD have focused on early childhood and elementary school-age children, though several studies have highlighted that at least a subgroup of children with ASD show persistent selectivity into the adolescent years and early adulthood (Bandini et al. 2016; Fodstad and Matson 2008; Kuschner et al. 2015). Thus, selective eating is a pervasive concern across both the developmental and heterogeneous course of ASD.

Etiology

The etiology of selective eating in ASD is complex and likely multifactorial. Sensory processing differences are commonly linked to selective eating in ASD (Ben-Sasson et al. 2008; Lane et al. 2009). Sensory sensitivity to textures, tastes, and smells has been hypothesized to contribute to food preferences, aversions, and refusal (Cermak et al. 2010; Johnson et al. 2015a; Lane et al. 2014). Gastrointestinal problems, including abdominal pain, constipation, and diarrhea, are also common in individuals with ASD (Kang et al. 2014; McElhanon et al. 2014) and may be associated with selective eating (e.g., a low-fiber diet contributing to constipation). Oral motor impairments (Dowell et al. 2009; Dziuk et al. 2007) and fine motor impairments (Barron-Linnankoski et al. 2014; Green et al. 2009a), frequently present in individuals with ASD, may impact swallowing and chewing as well as utensil use and further contribute to feeding challenges.

Eating and mealtime also require skills that are already challenging for children with ASD (reviewed in Twachtman-Reilly et al. 2008), priming children with ASD for feeding difficulties. Several core challenges experienced by children with ASD present themselves in the context of mealtime. For example, behavioral inflexibility and need for sameness, commonly seen in ASD (D’Cruz et al. 2013; Yerys et al. 2009), are in direct contrast to the variability of eating experiences; most people regularly change mealtime menus, utensils, dishes, and eating environments within and across days. Moreover, mealtime is typically social, and demands for interaction and conversation are high. High rates of co-occurring anxiety seen in ASD (Ung et al. 2015; van Steensel et al. 2011) increase the likelihood that children with ASD will develop food neophobia, a fear of eating new foods. The development of food preferences and diet variety requires repeated exposure to novel and perhaps anxiety-provoking foods (Birch and Marlin 1982; Wardle et al. 2003). Processing the similarities and differences of items within a food group or category (e.g., cheese varies widely in color, flavor, and form) may be even more challenging for individuals with ASD given known difficulties with cognitive skills related to managing novelty (Maes et al. 2010), prototype formation (Klinger and Dawson 2001), and generalization (Gastgeb et al. 2011).

Impact and Risk for Negative Outcomes

The functional impact of selective eating and risk for negative outcomes are both broad and multifaceted. As a daily, recurring challenge, families are impacted in numerous ways, including increased parent–child conflict and parent–spouse stress, disrupted family mealtimes, and anticipatory anxiety when facing encounters with food (Ausderau and Juarez 2013; Curtin et al. 2015; Marquenie et al. 2011; Rogers et al. 2011). Parents and caregivers are often concerned about both possible and actual nutritional consequences of a child’s selectivity, including issues related to nutrient deficiencies, gastrointestinal problems, and under- or overweight status (Ma et al. 2016; Sharp et al. 2013; Shmaya et al. 2015; Vissoker et al. 2015). As noted earlier, eating is also very often paired with social interactions, thus compounding already stressful settings for a child with ASD. Unfortunately, this stress may increase as the social implications of selective eating expand with age: Imagine the child who has been invited to a birthday party and is then faced with pizza from somewhere other than his or her preferred pizza delivery choice. Increased insight into these social pressures may increase stress, but may also offer motivation for change and treatment.

Treatment of Selective Eating in ASD

Despite the high frequency and negative impact of selective eating in children with ASD, few empirically supported interventions are available. Treatments generally use behavioral principles, sensory-based strategies, or a combination of the two, and have usually targeted children under the age of eight years. Sensory-based treatments, such as the sequential oral sensory (SOS) approach (Toomey and Ross 2011), use systematic desensitization across a hierarchy of textures in a play-based or social context; although commonly used and thought to be effective, these treatments have not undergone rigorous empirical evaluation. Behavioral interventions are widely used in clinical feeding programs and have been empirically tested to show improvements in chronic feeding problems (Lukens and Silverman 2014; Sharp et al. 2010, 2016). Behavioral pediatric feeding treatments are typically administered in the context of a time-intensive inpatient or day-treatment environment, depending on the extent of medical risk for the patient as well as other family and system factors. Supported by a multidisciplinary team (e.g., psychologist, dietitian, speech language pathologist, occupational therapist, social worker, nurse, gastroenterologist), treatment involves therapeutic, structured meals with repeated exposure to novel or non-preferred foods. The exposure allows for systematic desensitization and habituation to the foods. Treatment plans generally include withholding food outside of these mealtimes (appetite manipulation) and establishing clear rewards for bites taken (contingency management), while also incorporating escape extinction and graduated plans for adjusting and increasing food bites based on texture and quantity. Parents are often not initially involved in the treatment meals, but are trained to take over the meals toward the end of the program as discharge approaches.

Although effective, this inpatient/day-treatment approach generally treats one patient at a time and limits access for many affected youth, particularly those who are medically stable and may not need urgent and time-intensive treatment. Two manualized, outpatient selective eating interventions have been specifically designed and empirically evaluated for children with ASD in outpatient settings. The Autism MEAL Plan (Sharp et al. 2014) program is a group-based intervention for parents/caregivers of children with ASD and selective eating (3–8 years old; no IQ information included). This program includes eight, one-hour didactic training sessions with parents covering a breadth of behavior modification strategies (6 weeks), as well as more specific support for introducing new foods (2 weeks). Results from a pilot study (n = 19) indicated high social validity of Autism MEAL Plan, as well as a significant reduction in parent-reported stress following intervention relative to waitlist control. However, the authors reported no significant reduction in challenging behaviors during mealtime and no improvement in diet variety.

The Behavioral Parent Training Program for Feeding Problems (Johnson et al. 2015b) extended an efficacious parent training program for children with ASD and challenging behaviors (i.e., non-compliant behaviors, poor adaptive behavior skills, or sleep problems; Johnson et al. 2007, 2013) to the treatment of feeding problems. In the pilot trial, parents of 14 children with ASD and selective eating [2–7 years old, mean IQ = 77 (SD = 28)] participated in up to nine individually administered sessions (60–90 min) with a therapist over 16 weeks. Session content focused first on basic behavior modification principles, nutritional counseling, and strategies for mealtime routines. Grounded in understanding the antecedent–behavior–consequence model for addressing feeding problems, sessions focused on prevention and reinforcement strategies, compliance procedures, and teaching strategies for parents to advance the child’s feeding skills. Findings from the open trial demonstrated high treatment integrity (therapists) and adherence (parents) to program implementation. Additionally, parent ratings of challenging mealtime behavior, as well as irritability and hyperactivity more broadly, decreased significantly from baseline to endpoint. There was no significant change in nutritional status following treatment, as measured by three-day food records. Diet variety was not directly evaluated in this trial.

Remaining Treatment Gap

Autism MEAL Plan and the Behavioral Parent Training Program for Feeding Problems highlight the field’s move toward greater use of outpatient therapy models to treat selective eating, along with the possible use of a group model and more involvement of parents throughout all stages of treatment. Promising findings from these initial studies suggest feasibility of intervention for selective eating in children with ASD outside the typically structured medical settings. However, neither of these empirically evaluated treatments included children older than eight years; BUFFET was created as a start to fill this gap.

Development of BUFFET

Theoretical Foundation for Treatment

With a dearth of available treatment for this group of older children with ASD and selective eating, it is important to determine what treatment model would be most appropriate. Behavioral feeding principles and treatment strategies could be applied, but these elements may not be as developmentally appropriate for older children with ASD with average intellectual ability. These strategies would focus on altering the child’s behavior (i.e., selective eating), but may not address the distorted or maladaptive thoughts that can trigger or perpetuate rigid behaviors. These strictly behavioral approaches may also fail to build upon emerging cognitive skills, autonomy, and self-determination, which are important in development during later childhood. Ideally, a treatment for this older group of children would leverage emerging cognitive skills, empower youth to access and increase awareness of their own cognitions as part of treatment, and give them a “seat at the table” for addressing selective eating.

A growing body of literature suggests that CBT can successfully treat core and co-occurring symptoms in youth with ASD and is particularly effective for reducing anxiety symptoms. With appropriate modifications, research has indicated that CBT is an effective treatment in school-aged children and adolescents with ASD (without an intellectual disability) for anxiety (Reaven et al. 2012; Storch et al. 2013, 2015; Wood et al. 2009a), social impairment (White et al. 2013; Wood et al. 2009b), and inflexibility (Kenworthy et al. 2014). In general, CBT helps an individual recognize the thought–feeling–behavior connection. Effective treatment increases the individual’s awareness of emotional experiences as well as their ability to adjust distorted thoughts to reduce maladaptive behavior. In the case of CBT for anxiety, once the individual has these skills in place and has developed a set of coping skills to use when faced with anxiety-provoking situations, they complete exposure practices. Exposure is a crucial treatment component to allow individuals to gradually experience, and adjust to, an unpleasant stimulus (Kendall et al. 2005). Taking all of this evidence together, a CBT approach could be a helpful framework for addressing selective eating in older, verbally able children with ASD.

Considering the hypothesized etiology of selective eating in ASD, as well as the interplay between cognitions and behavior, the model for BUFFET (see Fig. 1) was developed for older youth with ASD. As described above, there are many possible etiologies for an individual’s selective eating; in some cases, the underlying problem can be treated directly (e.g., medical treatment of GI problems; occupational and physical therapy to address sensory sensitivities, motor impairments or swallowing difficulties). However, in many cases, the underlying cause of the individual’s selective eating may be more psychological in nature (or at the very least a combination of biological and psychological factors). In fact, even when more biologically based problems are addressed, resolving these physical concerns may not fully resolve a child’s selective eating. Regardless of why a particular food was initially avoided (e.g., difficulty with chewing and swallowing, a sensory property of the food, the brand name, the way it looks, or a previous negative experience like GI distress), selective eaters are often “stuck” and unable to flexibly eat new or non-preferred foods. Moreover, the prospect of trying these foods creates anxiety and stress. Based on our clinical experience, this anxiety and stress then perpetuates negative thoughts about (and further avoidance of) the food, thus suggesting the utility of a cognitive behavioral approach. Even if anxiety and inflexibility are not the etiologies driving selective eating, reducing anxiety and supporting flexibility are crucial parts of the treatment in this underserved and understudied group of children.

Fig. 1
figure 1

BUFFET conceptual model

Full size image

BUFFET leverages CBT strategies to help children develop skills to cope with anxiety, and to think and act flexibly with new or non-preferred foods. This skill set can then be applied as children tackle the unique challenges of any individual food (e.g., sensory issue, brand preference), with the goal of reducing the avoidance of new or non-preferred foods and expanding diet variety. In other words, addressing each unique challenge may be necessary but not sufficient for long-term reduction of selective eating; instead, flexibility and anxiety-management skills will be the key to providing an overarching skill set for maintenance of treatment gains.

Fortuitously, there is a growing body of literature and evidence base for treatments targeting flexibility and anxiety in school-age children with ASD. BUFFET incorporates and emphasizes components of two evidence-based treatments for ASD: (1) Facing Your Fears (Reaven et al. 2011), an outpatient multi-family group CBT program for anxiety in ASD, and (2) Unstuck and On Target (Cannon et al. 2011), a school-based curriculum for improving flexibility in ASD. Both interventions target and have been effective for the age range and verbal ability of the children with ASD that BUFFET aims to support. In addition, eating in the social environment of a group can positively contribute to food intake, suggesting that a group treatment model may maximize treatment effects (Herman 2015). While some youth with co-occurring social anxiety may find this challenging initially, this social facilitation approach may actually provide multiple exposure foci and reduce anxiety about both food and social environments. Given that mealtimes are generally social experiences, a group intervention model could, ultimately, be a preferred treatment context. Facing Your Fears and Unstuck and On Target are both successfully conducted in a group setting, and there is a well-established evidence base for the effectiveness of other interventions for ASD that occur in group settings (e.g., Laugeson et al. 2012). The combination of elements from Facing Your Fears, Unstuck and On Target, and the commonly used exposure strategies from behavioral feeding treatments (described above) set the stage for the BUFFET treatment model.

Target Population for BUFFET

BUFFET was developed for children with ASD who have not been previously served by the treatments developed for selective eating (i.e., chronologically eight years or older). In general, CBT is thought to be appropriate for a child who has a verbal ability of at least a seven-year-old (Kendall and Hedtke 2006). Language impairments are common in ASD and verbal ability is variable, even in the context of age-appropriate intellectual ability (Tager-Flusberg et al. 2011). For individuals with ASD, research with CBT programs has generally targeted a verbal IQ above 70 (Reaven et al. 2012; White et al. 2009; Wood et al. 2009a) or a verbal mental age of at least eight years (Kenworthy et al. 2014). Taken together, for this initial trial of BUFFET, treatment was conservatively geared toward youth with a verbal IQ ≥ 80 to focus on refining treatment content without the possible confound of more limited verbal abilities. Preliminary efficacy data will be evaluated to identify the most appropriate inclusion criteria to maximize treatment benefits.

In its current format, BUFFET targets children who are selective eaters but are medically stable. The goal is to build skills for trying new foods, which will hopefully then expand dietary variety and have a positive functional impact. This impact could be better nutrition to benefit overall health, or could be providing children with additional foods they can flexibly eat with peers or family members. BUFFET is not currently designed to directly target nutrition inadequacies. Therefore, if a child were medically unstable (e.g., severe nutritional instability, underweight, undernourished, at risk for placement of a G tube), in its current version, BUFFET would not be appropriate. Once BUFFET is established as an efficacious treatment, future trials can examine the inclusion of children with co-occurring medical complications and conditions, paired with interdisciplinary collaboration and oversight from pertinent medical professionals. It is important to note, however, that while BUFFET is designed without the requirement of an interdisciplinary team or medical professional, consultation from these clinician specialists (e.g., physician, nurse, occupational therapist) will certainly be valuable and could help reduce risks that would be more challenging in an outpatient mental health environment, such as gagging, vomiting, or choking on novel foods.

Stakeholder Advisory Board

A BUFFET Stakeholder Advisory Board was assembled to engage a participatory process for developing BUFFET. Advisory board members included 11 professionals across pertinent disciplines (psychology, speech/language pathology, occupational therapy, nursing, nutrition and dietetics, social work) as well as parents and self-advocate adults with ASD. Stakeholders were invited to participate in the advisory board if they had experience with selective eating and an interest in discussing the development of this type of intervention. The advisory board met three times across treatment development and pilot testing. An initial meeting was conducted as a focus group. The meeting addressed questions regarding the stakeholders’ previous experiences with treatment for selective eating in ASD and stakeholders offered feedback on the initial skeletal outline of the treatment manual. The group discussion was transcribed and reviewed for key themes; coding was completed by hand. A Delphi approach (Keeney et al. 2011; Linstone and Turoff 1975) was also used to generate consensus regarding primary treatment components. The advisory board viewed all proposed treatment components (i.e., anxiety, flexibility, sensory dimensions/taste training, and ASD-focused intervention features) as “crucial” and necessary for treatment of selective eating in ASD. Parent involvement in treatment was highlighted as vital to treatment success. The child’s self-determination and autonomy in treatment was also repeated as an important theme for this age group throughout the discussion. Illustrative quotations from stakeholders are shared below:

“I feel strongly that kids have to feel that they have a sense of control over what’s going on. Whether it’s they’re picking the foods or they’re picking the approach level that they’re using, or they know exactly what the minimum is that they need to do in order to gain—whether it’s a reinforcer or something else but there’s that perceived feeling of control and the second that gets jeopardized, you’re in murky waters. And sometimes you have to fight to give them that, at least in the older kids.”

“I have very young children that might be just touching broccoli or making broccoli trees, but not for this age group because you want to respect their intellect too and their abilities.”

When stakeholders were asked what does not work well in treatment for selective eating in this age group:

“Force feeding”

“Putting food to their lips or mouths, sneaking”

“Telling them what they can’t get until they eat”

“It’s healthy or it’s good for them”

At a second meeting, the advisory board reviewed the drafted treatment manual. Feedback was conceptual and overarching (e.g., change the order of presentation of treatment constructs) as well as detail-focused (e.g., the visual support used in a particular treatment activity was too “busy”). A final advisory board meeting was held following completion of the pilot study. Pilot data were presented to the stakeholders, and discussion focused on the successes and potential limitations of the program.

Additional stakeholder input was collected via feedback following each treatment session in the pilot study. Child and parent participants, as well as clinicians, provided ratings on a number of variables related to the acceptability of and satisfaction with treatment; these data are provided below in the Treatment Feasibility section.

Structure and Content of BUFFET

Overview

As depicted in Fig. 2, BUFFET is a 14-week multi-family group program based on CBT principles. Treatment sessions are 90-minutes each, with a combination of large group sections (all parents, children, and clinicians together), child-alone and parent-alone sections, and child–parent dyad sections. Each treatment session is led by at least two clinicians, so the child- and parent-alone sections can occur simultaneously.

Fig. 2
figure 2

Global structure of BUFFET across 14 weeks of treatment

Full size image

The first six sessions of the program emphasize psychoeducation and provide training in cognitive behavioral strategies for increasing flexibility and decreasing acute anxiety. Exposure happens in a group context (Snack Time, all sessions) as well as individually (BUFFET Building, sessions 7–13). The final treatment sessions involve a culminating activity of creating a Food Friend Commercial to showcase how participants have learned to eat new foods [similar to video strategies used in Facing Your Fears (Reaven et al. 2011) and Coping Cat (Kendall and Hedtke 2006)]. Homework assignments provide practice and generalization opportunities outside of the treatment clinic setting. See Table 1 for an overview of treatment sessions.

Table 1 BUFFET session outline
Full size table

Parents are actively involved throughout all treatment sessions, participating in full group activities (with all parents and children present) and supporting exposure and strategy use in child–parent dyad work. However, parents and children are separated for portions of many treatment sessions. This allows parents to gain skills for helping their children face new foods, while children learn the same skills in parallel, at a developmentally appropriate level. Including parents in this way aims to improve generalization of skills taught during group to the home environment. When parents and children are separated during treatment sessions, parents also hone specific skills pertinent to parents of selective eaters (e.g., how to re-approach a longstanding family challenge from a new treatment perspective; when to encourage children to go outside of their comfort zone with food versus when to stay with familiar and preferred food options).

The intervention also incorporates treatment strategies generally believed to be effective and helpful for children with ASD. Abstract concepts, particularly those related to cognitions and emotions, are presented visually and with repetition, response choices are provided for worksheets, scripts and routines are employed whenever possible, and video modeling is used to depict and demonstrate the novel experience of BUFFET Building. A positive reinforcement and reward system is incorporated with BUFFET Bucks “punch cards” (Fig. 3). This approach is similar to the reward plan incorporated in Facing Your Fears (Reaven et al. 2011). Children receive “punches” in their card for implementing the skills of flexibility and bravery in the face of novel or non-preferred foods (e.g., completing a BUFFET Building exposure practice step, tasting a novel Snack Time food). Punches can then be traded in for a prize previously identified during child–parent discussion.

Fig. 3
figure 3

BUFFET Bucks punch card

Full size image

Each component of the intervention program is described in more detail below.

Psychoeducation

The first six weeks of the intervention focus on building skills for managing anxiety and increasing flexibility. The anxiety topics presented are similar to concepts included in Facing Your Fears, and encourage bravery around new and non-preferred foods. The treatment content includes establishing the thought–feeling–behavior connection (i.e., mind–body regulation) inherent to CBT as well as introducing the concept of restructuring negative, maladaptive thoughts (i.e., “Food Foe Thoughts”) into positive, constructive ones (i.e., “Food Friend Thoughts”). Cognitive restructuring is used to help children solve the problem identified in the Food Foe Thought. For example, if the Food Foe Thought about a peanut butter and jelly sandwich is that “The sandwich is so gooey and messy,” the Food Friend Thought could address the problem by encouraging alternative self-statements and strategies (“No big deal. I will scrape off some of the jelly so it isn’t as messy.”). In preparation for exposure practice, children also build a “toolbox” of coping skills for managing their anxiety (e.g., deep breathing, counting to 10, thinking about a fun post-exposure activity). Flexibility topics build on Unstuck and On Target activities to help children understand cognitive flexibility. Children learn why it might be useful to be flexible (e.g., you get more of what you want in a situation) and are oriented to how the self-regulatory flexibility scripts presented in Unstuck and On Target can help them have food flexibility. For example, the “Plan A/Plan B” script could help children prepare for going out to eat with their family when they do not know what will be on the menu; “Plan A will be to order chicken fingers, but if they don’t have chicken fingers, my Plan B is to order pasta.” In addition to skills for managing anxiety and increasing flexibility, psychoeducation related to food dimensions is incorporated with the Food Dictionary activity (see Group-Based Exposure section below).

Across each of these sets of topics, there is a theme of self-advocacy and empowerment. As children learn skills, they are also learning why they may be struggling with food (e.g., Food Foe Thoughts are focused on the problems with food: they are stuck with only one idea for a restaurant dinner meal option), and then learning how they can improve things (e.g., generate a Food Friend Thought to identify a solution for the problem with the food: come up with a Plan B restaurant dinner meal choice). The children can then have a seat at the table to drive their own treatment.

Exposure

As a key component of any CBT for anxiety and any behavioral feeding treatment, exposure is heavily incorporated into BUFFET. Exposure activities are presented across the duration of treatment at the group level, and are more targeted toward each child’s specific treatment goals via individualized exposure during the second half of the intervention.

Group-Based Exposure: Snack Time

Didactic sessions are paired with ongoing exposure opportunities. First, at the group level, each session begins with Snack Time. A preferred food and a non-preferred food are presented for all children, parents, and therapists. Importantly, children are not required to eat the snack, but the snack must stay on their plate. This approach removes the initial pressure of being asked to try a food, but provides (at minimum), visual, proximity, and olfactory exposure practice. If children try a taste or eat a portion of the novel or non-preferred snack, they earn a BUFFET Buck punch (i.e., a reinforcement token).

During Snack Time, clinicians lead children and parents in the completion of a Food Dictionary “entry” for each Snack Time food (see Fig. 4 for a sample entry completed during one of the pilot groups). The Food Dictionary helps children develop a vocabulary for the dimensionality of food, guiding them to characterize the look, smell, flavor, texture, and food category of a food. They discuss the different ways someone could eat the food (e.g., utensils or not, in a sandwich, dipped in a sauce), practice restructuring negative thoughts about the food, and give the food a review (i.e., a five fork review if they love it down to a one fork review if they hate it). The Food Dictionary activity also targets prototype formation and generalization around foods. Over time, children are exposed to varying examples of the same food category (e.g., Swiss cheese slices, cheese dip, string cheese) and begin to draw parallels to generalize preferences (e.g., liking hard, mild cheese rather than soft cheese with strong flavors). These generalizations across foods can also be used to support children in advocating for their own preferences as a way to compromise and be flexible with a parent’s request.

Fig. 4
figure 4

Food Dictionary sample

Full size image
Individualized Exposure: BUFFET Building

Individual exposure practice is called BUFFET Building, with the goal of helping children expand the “buffet” of foods they are willing to try to eat. There is significant flexibility in the target foods chosen for BUFFET Building. The child–parent dyad works together to identify a target food, which should be something meaningful for both (e.g., a food the child often sees available at school events, a food the family frequently eats at mealtimes). Ideally, the child chooses the target food; the more investment the child has in the utility of the target, the more “buy-in” they will have for the challenging process of exposure (and more feelings of self-determination and empowerment across treatment). Targets could include trying a new food, increasing the quantity eaten of a particular food, mixing two foods together, or practice relative to challenges associated with mealtime (e.g., the sound of silverware clinking against dishes, introducing a new plate into the rotation of dishware used at mealtime).

Target food goals are individualized, but BUFFET Building practice sessions are highly structured for predictability. As is typical of fear hierarchies and exposure plans in CBT for anxiety, children and parents work together and are supported by clinicians to plan the exposure steps (see Fig. 5 for a sample BUFFET Building Planner worksheet). The BUFFET Building plan should maximize initial success in early steps, but still evoke enough anxiety and stress to make the exposure effective. Hierarchy steps and coping thoughts/strategies should be designed to address the problem identified in the Food Foe Thoughts related to the food. In other words, two children may both have the same target food, but may need different steps in their BUFFET Building plans if one dislikes the texture of the food while the other will only eat a particular brand of the food. This is where the hypothesis regarding the etiology of the selectivity should be factored in; if texture is driving the selectivity, then the hierarchy steps can gradually expose the child to the aversive texture. Perhaps different from a typical exposure plan for anxiety, BUFFET Building hierarchies are standardized to six steps and may not encapsulate the entirety of a particular goal. For instance, if a CBT hierarchy is designed to address a fear of dogs, the goal is that when the final step is reached, the patient is no longer afraid of dogs (or is at least able to manage the anxiety and stress associated with being in the presence of dogs). In BUFFET, the goal is not necessarily to “conquer” the food; the goal is to have enough exposure to the food to either integrate it into the child’s regular diet or to confirm that it is a non-preferred food (i.e., food preferences are still allowed!). Hierarchies are tailored to the child’s progress with respect to the target goal, and there may be multiple six-step hierarchies designed and completed prior to a child possibly integrating a food into their regular diet. For example, an early six-step sequence may end with the child taking one small bite of a target food, whereas a subsequent and more advanced six-step sequence may involve the child eating an entire portion of the food. The quantity of six steps was chosen as a manageable number that maintains enough “distance” from the start point to challenge the child, but is not overwhelming (since another hierarchy may quickly be designed and implemented). A set, rather than variable, number of steps was also employed to be a predictable routine/script children could use when facing a new food. Scripts and routines are generally appealing for individuals with ASD and are useful for self-regulation. When each hierarchy is complete, children and parents reflect upon and summarize the experience, decide whether the food has moved onto the child’s “buffet,” and then make a plan for their next step with the target food. Clinicians and parents work together to ensure that the child has had enough exposures to the food within their BUFFET Building plans to establish a preference for or against the food (Fig. 5).

Fig. 5
figure 5

BUFFET Building planner sample

Full size image

Parent Involvement

Parents are engaged alongside children throughout the entire course of BUFFET. In parent-alone sections, parents are oriented to the psychoeducation topics presented to the children. They learn in parallel about coping strategies for anxiety and stress, the concepts of Food Foe and Food Friend Thoughts, as well as the utility of and strategies for flexibility. Parents can then support their children through weekly homework (BUFFET Takeout) assignments, which promote generalization of skills and guide exposure practice. Parents practice using new concepts (e.g., restructuring their own Food Foe Thoughts into Food Friend Thoughts), so they can be effective with modeling and role-plays.

Psychoeducation for parents addresses topics pertinent to their own role in their child’s selective eating. Early in treatment, there is discussion about the cycle of anxiety, and how continued avoidance of a feared situation can perpetuate and worsen anxiety. This is linked to the parent’s role in the anxiety cycle, and the risk of parents inadvertently maintaining the anxiety, such as doing something to help the child feel better quickly or avoiding situations that are known anxiety triggers. Discussion focuses on how parents can gradually support their child to be brave and build skills for facing fears (adaptive protection) without limiting their opportunities for exposure to anxiety (excessive protection) (Reaven and Hepburn 2006). As treatment progresses, parent discussions focus on debriefing about BUFFET Building practice and provide parents with a forum to share their successes and struggles. Parents are supported to consider how to help children harness their skills and coping strategies to complete steps in BUFFET Building hierarchies without forcing them to eat. In the final sessions, discussion turns to how parents will continue to support their children after BUFFET ends. Clinicians highlight different approaches for maintaining or furthering treatment gains (e.g., scheduling regular BUFFET Building exposure practice sessions, transitioning to more naturalistic and family-based strategies like “flexible Friday” dinners to incorporate new foods), keeping the momentum of progress, and possible hurdles to anticipate. Parents are encouraged to brainstorm and troubleshoot their plans with clinicians prior to ending treatment in order to maximize continued progress.

Homework: BUFFET Takeout

Weekly homework assignments (“BUFFET Takeout”) are provided to support repetition and rehearsal of treatment concepts as well as generalization beyond the clinic. In the first half of the intervention, BUFFET Takeout is focused on practicing the skills learned in psychoeducation sessions. During the second half, BUFFET Takeout is the planned BUFFET Building exposure practice to happen at home.

Treatment Manual

The BUFFET Treatment Manual Bundle resulting from the development process and pilot study includes (1) BUFFET Therapist Manual: Each of the 14 sessions includes delineation of the week’s schedule with recommended activity timing, materials lists, session purpose and goals, therapist guidelines and discussion points, weekly BUFFET Takeout assignment information, session take home points, “Helpful Hints” and “What ifs…” sections for possible session challenges, and key “Optional Modifications,” (2) BUFFET Parent Guide: A set of handouts and accompanying worksheets to support parent involvement in the intervention, (3) BUFFET Child Guide: A combination of handouts and worksheets to visually present treatment information and provide opportunities for practice and application of treatment concepts, and (4) BUFFET Treatment Materials: Visual supports related to treatment activities, including posters for the treatment room and hands-on activity materials.

Treatment Feasibility

Following development of the treatment manual guided by the BUFFET Stakeholder Advisory Board, an open pilot trial was conducted to evaluate the feasibility and initial efficacy of BUFFET. Feasibility data, evaluated by acceptability of and satisfaction with treatment, are presented below following a description of the pilot study. Initial efficacy data will be presented in a separate manuscript; results are promising, suggesting improvements in children’s willingness to eat new foods and a reduction in the daily impact of the child’s selective eating.

Pilot Trial Participants

Youth 8–12 years with verbal IQ ≥ 80 were recruited for this study. The recruited sample yielded three BUFFET groups (n = 4, 3, 4 families) with a total sample of 11 maleFootnote 1 participants (ages 8–11 yearsFootnote 2) and their parent(s). Participants all met criteria for ASD, confirmed with scores above cutoffs on the Autism Diagnostic Observation Schedule, Second Edition (Lord et al. 2013), a brief diagnostic interview with parents informed by Social Communication Questionnaire (Rutter et al. 2003) responses, and a DSM-5 checklist. All participants had a verbal IQ ≥ 80 (M = 108, SD = 16; range 85–138), as measured by the Differential Abilities Scales-II (Elliott 2007) or documented results of other recent (<12 months) cognitive testing.

Selective eating was confirmed via initial screening and baseline testing documenting avoidance of foods based on sensory characteristics (e.g., selectivity based on food group, texture, brand, or color) and parent report of impact on the child’s and family’s daily functioning. For this initial feasibility and efficacy trial, participants were excluded if there was a known medical condition or restricted diet that could account for the child’s selective eating (e.g., Celiac disease, gastrointestinal disorders, parent-chosen gluten/casein free diet).Footnote 3 General exclusion criteria included co-occurring diagnoses of major depressive disorder, bipolar disorder, or schizophrenia. Diagnoses of anxiety, ADHD, mild depression, and other mild psychopathology were not exclusionary, provided that these symptoms were not in a crisis state and did not prevent the participant from engaging in the majority of treatment activities. Additional exclusion criteria included gestational age below 32 weeks, perinatal injury, traumatic brain injury, known medical or neurological abnormality, and English as secondary language. Participants were excluded at either the screening or eligibility phase for the following reasons: ASD diagnosis not confirmed at eligibility visit (n = 2), selective eating was too mild (n = 1), medical condition that could account for selective eating (n = 1), gestational age below 32 weeks (n = 1), or verbal abilities below inclusion criteria (n = 12). Four additional families were deemed eligible for the study but were unable to participate due to logistical/scheduling conflicts.

Pilot Trial Procedures

Once enrolled following the eligibility visit, families were seen for a total of 16 weeks. Children and their parent(s) completed baseline and outcome testing at Week 1 and Week 16, respectively, with the 14 intervention sessions completed at Weeks 2–15. Families were also asked to complete follow-up assessments at 4 and 12 weeks following treatment completion.

In the pilot trial, 90-minute treatment sessions were held weekly during after school (or after summer camp) hours in a medical center research clinic environment. Sessions were led by two doctoral- or masters-level clinicians; all clinicians had extensive experience working with children with ASD, but varied substantially in their familiarity with feeding interventions. Sessions were additionally supported by 2–4 trained research assistants who had experience with children with ASD.

Results: Treatment Acceptability

In this pilot study, acceptability of BUFFET was evaluated via session attendance, individual session ratings (by children, parents, and clinicians), and global parent satisfaction ratings.

Session Attendance

There was a high attendance rate (91%) with no attrition. Participants missed an average of 1.45 sessions each (n = 7 missed one session, n = 3 missed two sessions, n = 1 missed three sessions). Clinicians offered to meet individually with families following missed sessions to review material, and several families took advantage of this opportunity (n = 4).

Individual Session Ratings

Following each treatment session, parents and children completed a rating form with questions about (1) the clarity of the material covered, (2) the quantity of material covered, (3) how helpful they perceived the session to be, and (4) their enjoyment of the session. They were further provided with free response space for additional comments. Clinicians rated each session based on their perception of (1) the quantity of material reviewed, (2) how effective the session was, and (3) how comfortable they were as a clinician in the session. Clinicians were also asked to provide open-ended responses for what they really liked about the session and what they would change.

Parent and Child Ratings

Overall, parents and children (see Table 2) reported that they understood what was discussed in treatment sessions and believed that sessions generally had the right amount of material. Children rated a small portion (13%) of sessions as having too much material; data suggest that these ratings were distributed across the course of treatment and did not consistently indicate one particularly overloaded session. Of note, ratings were anonymous and compiled at the end of each session; therefore, it was not possible to identify whether one specific participant found the sessions to have too much material. Parents and children showed variability in their perceptions of how helpful each session was in treating their selective eating. A closer look at the data suggested that the latter half of sessions in the treatment (~sessions 7–13; those that primarily focus on individualized exposure) were perceived as more helpful than the initial half. Parents and children generally rated sessions as enjoyable; the sessions rated by children to be less enjoyable were also in the first half of the treatment. Interestingly, this highlights differences between the content and activities of the initial half of the intervention, which focuses more on psychoeducation, and the latter half of the intervention, which focuses on individualized food exposures.

Table 2 Individual session satisfaction ratings—parents and children
Full size table
Clinician Ratings

Consistent with child participant report, clinicians (see Table 3) also indicated that a subset of sessions (16%) contained too much material. Closer analysis highlighted three sessions that were most likely to be rated by both children and clinicians as having too much material: Session 1 (Introduction and Overview), Session 7 (BUFFET Building Introduction), and Session 11 (BUFFET Building and Food Friend Commercial Planning). Clinicians reported that treatment sessions were generally effective and they felt comfortable leading or supporting the intervention process. Importantly, several instances were highlighted as making clinicians feel “a little uncomfortable” during sessions. Examination of the free responses for those particular ratings suggested that clinicians were concerned that session content may have been too abstract or complex for the participants.

Table 3 Individual session satisfaction ratings—clinicians
Full size table

Results: Treatment Satisfaction

Parent report of global satisfaction was evaluated via the Client Satisfaction Questionnaire (CSQ-8; Attkisson and Zwick 1982), an 8-item parent report questionnaire designed to elicit the client’s perspective on the value of services received (internal consistency α = .93). The CSQ-8 yields a total score (8–32), with higher scores representing greater satisfaction.

High total scores on the CSQ-8 (Mean = 30.5, range 27–32) reflected general satisfaction with BUFFET. All (100%) of the reporting families (n = 8)Footnote 4 indicated that they were very satisfied with BUFFET, believed they received excellent service quality, and would return to the program if seeking help again. Looking more specifically at individual items on the CSQ-8, parents indicated that they:

  • Definitely (75%) or generally (25%) got the service they wanted with BUFFET

  • Would definitely (88%) or probably (12%) recommend BUFFET to a friend

  • Believed BUFFET helped a great deal (63%) or helped (37%) to deal more effectively with their child’s selective eating

  • Felt very (88%) or mostly (12%) satisfied with the amount of help received in BUFFET

Compared to the responses above, there was more variability in the extent to which families felt their treatment needs were met by BUFFET (i.e., “To what extent has our program met your needs?”). Half of the reporting families indicated that almost all of their needs were met (50%), while others indicated that most (38%) or only a few (12%) of their needs were met.

Implications and Next Steps

Selective eating is a common challenge across a wide range of individuals with ASD. Despite a lifespan trajectory of selective eating and daily, functional impact, evidence-supported treatments for older children (≥8 years) with ASD are quite limited. BUFFET was developed to fill this treatment gap. Acceptability data suggest that BUFFET was successfully administered to 11 families who found the treatment to be generally acceptable and satisfactory. Data highlighted particular sessions that could be streamlined and suggest that children preferred the hands-on, individualized exposure activities to the psychoeducational activities.

Pilot Trial Limitations

Limitations from the pilot study offer considerations for future trials. As an open trial with no control group and small sample size, results are limited. It is also possible the three families who did not complete global satisfaction ratings felt less positively about the intervention and their experiences. It is important to note that treatment fidelity was not evaluated in this small pilot trial because the treatment manual was being iteratively refined throughout the study; therefore, the extent to which session content was delivered in a similar manner by clinicians across groups is unknown. However, consistency and appropriateness of curriculum implementation were maintained by the PI (EK) via her administration (or supervision of administration) of intervention material at each group session.

Treatment Manual Revisions

The treatment manual was revised following feedback from the first pilot treatment group; the second and third groups were started only a few weeks apart, so the same manual version was used for both groups. Following the first group, content across sessions was streamlined for sessions perceived as having too much material. In addition, some of the concepts and activities clinicians perceived as too abstract were revised and made more concrete.

Additional treatment manual revisions are being considered in preparation for a randomized controlled trial. The individual session acceptability data related to perceived helpfulness and enjoyment of treatment sessions point toward careful consideration of the balance of psychoeducation versus direct exposure practice. These data align with the anecdotal feedback from child participants in group discussions and feedback questionnaires at the end of treatment. At minimum, the earlier psychoeducation-focused sessions will be revised to be more engaging, hands-on, and enjoyable for participants. Elimination of these components altogether (or significant consolidation) could yield a leaner treatment, or one that would allow time for more direct exposure practice. Alternatively, spending several sessions on these concepts early in treatment may, in fact, be a critical component of the intervention by functioning as a “proxy” exposure that helps prepare participants for jumping into the more challenging task of trying new foods later in treatment. Once the treatment is considered efficacious, a dismantling study with a components analysis approach could be helpful to evaluate whether or not the early weeks of psychoeducation are key ingredients for efficacious results.

Fast-Tracking to Community Use

At this early stage in treatment development, efficacy, and effectiveness testing, the 17-year research-to-practice gap (Dingfelder and Mandell 2011; Green et al. 2009b) can seem daunting. Thus, it is anticipated that future treatment trials will maintain an eye toward using hybrid research designs (Curran et al. 2012; Glasgow and Lichtenstein 2003; Wells 1999) to fast-track dissemination for community use. The next study will need to include a larger sample with a randomized design to evaluate BUFFET relative to a comparison group to truly examine its efficacy. However, this could be done in the context of a hybrid efficacy–effectiveness design (Glasgow and Lichtenstein 2003; Wells 1999) to begin to examine external validity.

BUFFET is certainly several steps away from community-based implementation; however, “lessons learned” from this trial may lay the groundwork for successful effectiveness evaluation and implementation (Bowen et al. 2009; Proctor et al. 2011). For example, the involvement of food in BUFFET sessions adds a treatment expense to implementation cost, which would need to be addressed for lower resourced providers and families. Demand and need for an intervention like BUFFET appear to be strong and suggest motivation for adoption of the treatment. An initial high recruitment response rate (nearly 100 families responding to a single initial recruitment email) and ongoing inquiries into the study highlight the need and perceived demand for this type of intervention in the community. The results of this study suggest BUFFET is a feasible and acceptable intervention for families of youth struggling with selective eating.

Conclusion

The development of BUFFET shows promise for filling an important selective eating treatment gap. BUFFET provides an outpatient treatment option for older youth with ASD and selective eating who have, thus far, been underserved. The treatment design is grounded in a theoretical model appropriate for this older age group and supports self-advocacy and empowerment. Once effectiveness has been established, BUFFET has the potential to address an important clinical issue for children and their families in an accessible and easy to implement treatment.